INDUSTRIAL RESTRUCTURING AND IMPLICATIONS FOR LABOUR MARKETS IN THE NEW EU MEMBER STATES

Transcription

1 INDUSTRIAL RESTRUCTURING AND IMPLICATIONS FOR LABOUR MARKETS IN THE NEW EU MEMBER STATES COMPILATION OF STUDIES PRESENTED AT A SEMINAR ORGANISED IN BRUSSELS ON 24/09/ "Structural change, productivity and employment in the new EU Member States" by Peter Havlik - "Foreign direct investment and restructuring in the automotive industry in Central and Eastern Europe" by Slavo Radoševic and Andrew Rozeik - "Foreign direct investment in the new Central and Eastern European member countries" by Ingo Geishecker - "The skill-bias of foreign direct investment in Central and Eastern Europe" by Ingo Geishecker - "Employment, education and occupation structures: a framework for forecasting" by Robert Stehrer - "FDI and the skill composition of the workforce: the case of the electronics industry in Hungary" by Kushal Kataria and Harald Trabold - "Impact of FDI inflows on labour market differences in Hungary stylized facts and policy implications" by Károly Fazekas and Éva Ozsvald - "Agricultural regions and regional policy in Poland" by Eugeniusz Kwiatkowski, Pawel Gajewski and Tomasz Tokarski - "Inter-industry labour mobility in Poland" by Eugeniusz Kwiatkowski, Pawel Kubiak and Leszek Kucharski - "Active labour market policy (ALMP) in Bulgaria" by Iskra Beleva - "Development of SMEs in Bulgaria" by Iskra Beleva - "Developments in education and training in the Czech Republic" directed by Filip Zeman, EuroProfis - "Industrial policy in the Czech Republic" by Filip Zeman

2 Economic Restructuring and Labour Markets in the Accession Countries Research Project commissioned by EU DG Employment, Social Affairs and Equal Opportunities, Contract No. VC/2003/0367 Coordinated by in cooperation with Alphametrics and The Vienna Institute for International Economic Studies Structural change, productivity and employment in the new EU Member States by Peter Havlik* December 2004 *) The Vienna Institute for International Economic Studies (wiiw)

3 Contents Executive summary 1 Development of GDP, employment and macro-productivity 2 Changes in broad sectoral structures 3 Structural change and productivity growth 4 Patterns of productivity catching-up in manufacturing 5 Productivity and employment growth dilemmas References Appendix List of Tables and Figures Table 1 Table 2 Table 3 Table 4 Table 5 Long-term growth and productivity catching-up of NMS Decomposition of aggregate and manufacturing productivity growth in NMS (shift-share analysis), Size of European manufacturing industry after enlargement to EU-25 Labour productivity catching-up in manufacturing: NMS vis-à-vis the EU-15, Regression estimates of NMS employment elasticity to GDP growth, Table A1 Labour productivity levels in MNS manufacturing industry, 2002 Table A2 Relative productivity gains in NMS manufacturing, (average annual change in % for total manufacturing (D) and relative gains DA to DN, in percentage points) Figure 1 GDP, employment and productivity in EU-15, NMS and Poland, 1995 = 100 Figure 2 GDP, employment and macro-productivity in the NMS and EU-15, 1995 = 100 Figure 3 Figure 4 Levels of macro-productivity and of GDP per capita in the NMS, EU-15 and EU-25, year 2003 Comparison of NMS and EU-15 gross value added structures in 1990, 1995 and 2002, % of GVA

4 Figure 5 Figure 6a Comparison of NMS and EU-15 employment structures in 1990, 1995 and 2003, % of total Productivity growth in NMS economic sectors, (annual averages, gross value added per employed person) Figure 6b Productivity levels in NMS economic sectors, 2002, EU-15 = 100 Figure 7 Figure 8 Manufacturing employment concentration ratios (CR3) in NMS Deviations of NMS and EU-15 manufacturing employment structures, 2002 and 1995 Figure 9 Manufacturing production and employment growth in NMS and EU-15, 2002 (1995 = 100) Figure 10 Employment elasticity of GDP growth in selected NMS, Figure A1 Manufacturing labour productivity in selected NMS (UVR-based), 1996 and 2002 (Austria = 100)

5 Executive summary This paper provides an overview of longer-term structural developments in the new EU Member States from Central and Eastern Europe (NMS). It analyses structural changes in NMS economies and patterns of productivity catching-up at both macro level and within the individual industries. With the transformational recession of early 1990s left behind, the majority of NMS embarked on a path of rapid economic growth during the past decade. They have experienced an impressive productivity catching-up, at both macroeconomic level and in manufacturing industry in particular. Yet the growth of labour productivity went in most NMS hand in hand with declining employment, and even with considerable job losses in manufacturing industry. Structural changes observed during the past decade brought the NMS economies nearer to the economic structure observed in the EU-15, but the shifts of labour among individual sectors or industries themselves did not have any marked impact on aggregate productivity growth. Similar to EU-15, the recent productivity catching-up observed in the NMS resulted overwhelmingly from across-theboard productivity improvements in individual sectors of the economy while employment shifts among sectors had only a negligible effect on aggregate productivity growth. Notwithstanding fast productivity catching-up, the estimated productivity levels indicate that NMS are in this respect still lagging behind EU-15 economies considerably, implying a huge catching-up potential. Estimated elasticity of employment to production growth is low in all NMS; the recently observed and expected rates of economic growth will in all likelihood not be sufficient for the creation of additional jobs. The required further productivity convergence with EU-15 may thus be in conflict with urgently needed employment growth in the NMS; the net job creation occurred in just a few services sectors and could not offset job losses in agriculture and industry. Keywords: Structural change, economic growth, productivity, employment, EU enlargement. JEL classification: E24, F43, J21, J60, O11, P52 i

6 Peter Havlik Structural change, productivity and employment in the new EU Member States 1 Development of GDP, employment and macro-productivity In the first half of the 1990s, the Central and East European countries which became members of the EU on 1 st May 2004 the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, the Slovak Republic and Slovenia, in the following termed NMS-8 went through the dramatic phase of the 'transitional recession' in which their GDP and employment recorded considerable declines (Figure 1), due to supply as well as demand shocks caused by the loss of traditional export markets, the disruption of existing supply chains and decision-making structures, sudden trade liberalisation and restrictive macroeconomic policies. During , the NMS-8 experienced a cumulated decline of real GDP by 4.7%. This translated into a substantial negative growth differential ( falling behind ) for the NMS-8 vis-à-vis the EU-15 (Table 1). Figure GDP, employment and productivity in EU-15, NMS and Poland 1995 = 100 GDP NMS Employment NMS Employment PL Productivity NMS Employment EU-15 Productivity PL Source: wiiw Database incorporating national statistics and AMECO, wiiw estimates (weighted averages). From 1993/94 onwards (in Poland already in 1992), economic recovery gained momentum in the NMS-8 and their average growth began to exceed that of the EU-15. However, a closer look reveals that most of these countries experienced further at times sharp 1

7 interruptions in their growth processes due to delayed/failed corporate restructuring and occasional financial crises (often called 'secondary transformational recessions') and/or macroeconomic imbalances, sometimes caused by unsustainable current account or fiscal deficits. Also, the growth process became more differentiated across the region, with the two candidate countries, Romania and Bulgaria, lagging behind significantly. For the period , the average annual growth rate of GDP was 3.7% for the NMS-8. 1 GDP growth accelerated moderately after 1995 in the EU-15 as well, with an average annual growth rate of 2.3% over the period The growth differentials thus turned in favour of the NMS and reached almost 16 percentage points in cumulative terms and 1.3 percentage points per annum for the NMS-8. Taking into consideration the whole period , there has been no difference in cumulative GDP growth for the NMS-8 relative to the EU-15 and therefore no catching-up (Table 1). Employment in the NMS-8 declined even more strongly than GDP in the first years of transition (-13% between 1990 and 1995) and did not fully recover even afterwards (Table 1). For the whole period , the cumulated employment decline in the NMS-8 reached nearly 17% (nearly 6 million jobs were lost) again with notable differences across the region. In the more recent period for which comparable data are available (after 1995), declining employment in Poland has been the main contributor for the dismal labour market performance of NMS as a group (Figure 1 and Landesmann et al, 2004). In the EU-15, overall employment declined in the first half of the 1990s as well, but to a much lesser extent than in the NMS. In the second half of the 1990s, EU-15 employment has been moderately growing (1.1% annually), resulting in a cumulated increase of employment throughout the whole period by 7.3%. Turning now to aggregate developments of income and productivity, macro-productivity in the NMS-8 rose on average at a similar pace as in the EU-15 in the period (Table 1). 2 But productivity gains in the NMS-8 during that period resulted mainly from massive labour shedding which overcompensated the fall in output. Thus, productivity gains reflected at that time the painful adjustment process going on in these countries rather than a successful restructuring and modernisation of their economies. GDP per capita, as a common measure for living standards, declined substantially in particular in the first years of transition (see Figure 2) % if Bulgaria and Romania, which recorded average annual growth rates of 1.1%, were included. Macro-productivity is defined as GDP per employed person employees and self-employed. 2

8 Table 1 Long-term growth and productivity catching-up of NMS Country groups growth rate growth differential growth rate growth differential growth rate growth differential in % against EU-15 in pp in % against EU-15 in pp in % against EU-15 in pp cumu- annual cumu- annual cumu- annual cumu- annual cumu- annual cumu- annual lated average lated average lated average lated average lated average lated average NMS-8 1) GDP Employment Macro-productivity Cyprus GDP Employment Macro-productivity Malta GDP ) 5.3 2) ) 5.3 2) ) 4.1 3) ) 4.1 3) Employment Macro productivity ) 3.8 2) ) 3.8 2) ) 3.2 3) ) 3.2 3) NMS-8+BG and RO GDP Employment Macro productivity EU-15 GDP Employment Macro productivity Notes: 1) NMS-8: Central and east European new EU member states, comprising the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, the Slovak Republic and Slovenia.- 2) ) Sources: wiiw Database incorporating national statistics, WIIW calculations using AMECO 3

9 4 Figure 2 GDP, employment and macro-productivity in the NMS and EU-15 (1995 = 100) 135 GDP NMS-8 EU Employment* NMS-8 EU Macro-productivity (GDP per persons employed) NMS-8 EU GDP per capita NMS-8 EU * employees and self-employed Source: wiiw Database incorporating national statistics, wiiw calculations using AMECO.

10 In the second half of the 1990s, the rise of macro-productivity strongly accelerated in the NMS-8 and this time productivity growth was supported by fast rising GDP at relatively constant employment levels in most NMS (Poland was the main exception). During , productivity growth was significantly higher in the NMS-8 than in the EU-15 (4.3% per annum as compared to 1% in the EU-15). The process of impressive 'productivity catching-up' of the NMS after 1995 is clearly demonstrated in Figure 2. The cumulated 'productivity gain' of the NMS-8 vis-à-vis the EU-15 over the whole period reached nearly 34 percentage points, almost all of which was achieved after 1995 (Table 1). Figure 3 Levels of macro-productivity and of GDP per capita in the NMS, EU-15 and EU-25, year 2003 Macro-productivity (GDP per person employed*) EU-15 = 100 EU-15 NMS EU at exchange rates at PPPs EU-25 = 100 EU-15 NMS EU at exchange rates at PPPs *) employees and self-employed; PPPs = purchasing power parities. GDP per capita EU-15 = 100 EU-15 NMS EU at exchange rates at PPPs EU-25 = 100 EU-15 NMS EU at exchange rates at PPPs Note: NMS include Cyprus and Malta. Source: wiiw calculations using national statistics and AMECO Database. Despite remarkable productivity catching-up in the recent period, the level of macroproductivity in the NMS is still very low compared to the EU-15 average, leaving ample space for further productivity growth and catching-up. In the year 2003, the average level 5

11 of macro-productivity (compared at current exchange rates) for all ten first-round NMS (including Cyprus and Malta) taken together was only 26% of the average EU-15 level. Measured at purchasing power parities (PPPs), which correct for undervalued currencies still prevailing in many NMS, the average level of macro-productivity reached just 52% of the EU-15 average (56% if compared to the enlarged EU-25; see Figure 3). 3 2 Changes in broad sectoral structures Economic developments in the NMS during the transition period were characterized by large shifts in the sectoral composition of GDP and employment, indicating a clear tendency of adjustment towards the broad economic structures in the EU-15. The NMS started off in 1990 with a larger agricultural and industrial sector on the one hand and a smaller services sector than the EU-15 countries on the other (see Figures 4 and 5). 4 The broad shifts occurring after 1990 in the NMS can thus be summarized under the headings of de-agrarianization, de-industrialization and tertiarization. However, there are a few interesting cases of 're-agrarianization' and 're-industrialization' as well. But while the former are considered to be of a transitory nature, the latter may become a more common phenomenon in the future. An overall tendency for de-agrarianization, de-industrialization and tertiarization can be observed in the EU-15 throughout this period as well, but here it has been much less pronounced than in the NMS. There has been also one example of re-industrialization within the EU-15, namely that of Ireland, where the share of industrial value added in GDP increased from 32% in 1990 to 37% in 2001 yet employment shares remained constant (European Commission, 2003). De- and re-agrarianization In all NMS-8, the shares of agriculture in GDP and employment fell dramatically during 1990s ('de-agrarianization'). 5 Employment declined significantly in absolute terms as well However, for the more advanced NMS such as Cyprus, Malta and Slovenia, macro-productivity measured at exchange rates has already reached between 50 % and 60 % of the EU-15 level and between 70% and 80%, if PPPs were used for conversion. Due to lower employment rates in some of the Central and East European countries GDP per capita for the NMS reached only 24% (at exchange rates) and 47% (at PPPs) of the EU-15 level. Under the previous regime, industry was emphasized at the expense of services and, furthermore, service activities were often supplied within big industrial combines, which meant that they were classified under 'industry' and to some extent 'agriculture' as well. Most services were considered 'unproductive labour' and their contribution to the efficient functioning of the economy was neglected (Stare and Zupancic, 2000). Also, many modern services that play an important role in market economies (such as marketing, financial services, real estate and other business services) were simply not needed under socialism. (Soubbotina and Sheram, 2000). Sector shares in this section are defined as gross value added (GVA) of agriculture (industry, services) in gross domestic product (GDP). Because of the so-called 'Financial intermediation services indirectly measured' (FISIM), which are included in GDP but not in gross value added, the so defined shares of the three sectors will not add up exactly to 100 %. 6

12 Figure 4 Comparison of NMS and EU-15 gross value added structures in 1990, 1995 and 2002 % of GVA 30 Agriculture and fishing CZ EE HU LV LT PL SK SI NMS-8 BG RO EU-15 Industry and construction CZ EE HU LV LT PL SK SI NMS-8 BG RO EU-15 Services CZ EE HU LV LT PL SK SI NMS-8 BG RO EU-15 Note: GVA = gross value added. Sources: wiiw Database incorporating national statistics; wiiw calculations using AMECO. Despite massive de-agrarianization in the NMS-8, the shares of agriculture in GDP and employment of these countries is on average still higher than in the EU. 6 In the more advanced NMS such as the Czech Republic, Hungary and Slovenia, the difference to the EU-15 was minimal in the share of gross value added (GVA), though not in terms of employment shares. In general, the differences between GVA shares and employment shares in agriculture are larger in the NMS than in the EU-15, due to the relatively low productivity in NMS' agriculture as compared to the other sectors of the economy. With 6 In Bulgaria and Romania the share of employment in agriculture has been very high (25% and almost 40%, respectively). This results from the severe employment crises in both countries due to the dramatic decline in industrial employment and the so far limited absorption capacity of the services sectors. 7

13 competitive pressures rising and modernization in agriculture accelerating after accession, we may thus expect agricultural employment in the new EU member countries to fall. This is particularly relevant for Poland, some of the Baltic countries and for the candidate countries Bulgaria and Romania, where the differences between GVA shares and employment shares in agriculture are huge (compare Figures 4 and 5), and productivity levels particularly low (Figure 6b). Figure 5 Comparison of NMS and EU-15 employment structures in 1990, 1995 and 2003 % of total Agriculture and fishing CZ EE HU LV LT PL SK SI NMS-8 BG RO EU-15* Industry and construction CZ EE HU LV LT PL SK SI NMS-8 BG RO EU-15* Services CZ EE HU LV LT PL SK SI NMS-8 BG RO EU-15* *) Year 2002 Sources: wiiw Database incorporating national statistics; wiiw calculations using AMECO. 8

14 De- and reindustrialization The share of industry (comprising manufacturing, mining, water & electricity supply and construction) declined strongly in terms of both GVA and employment in most NMS. This decline was more pronounced in the first years of transition and levelled off after Yet industrial employment dropped sharply in absolute terms even after 1995 (by nearly 1.3 million persons, over half of them in Poland). However, by around 1998/1999, labour shedding in industry bottomed out and employment started to rise slightly in some NMS (e.g. in Hungary, in the Czech and Slovak Republics; Poland is again an exception). On average, the shares of industry and construction in both GVA and employment in NMS still tend to be somewhat higher than in the EU-15 on average (30% and 27%), with some countries having particularly high employment shares of industry (e.g. Czech Republic, Slovakia, Slovenia Figure 5). 7 As illustrated by the recent example of Hungary and the Czech Republic, there is a possibility for a few additional NMS (e.g. Slovakia) to experience some kind of reindustrialization in the future. Low labour costs and the pool of skilled labour make the NMS an attractive location for FDI in export-oriented manufacturing productions and, as demonstrated by many south-east Asian economies, strong export orientation might well lead to a higher share of industry in both GDP and employment than would be typical for a certain stage of economic development. However, whether this process will lead to the creation of a substantial number of additional jobs is not sure (see below). 8 Tertiarization The share of services, in both GVA and employment, has increased significantly in most NMS during the past decade, indicating a clear 'catching-up' of this sector. However, at the beginning of transition, the rise of GVA and employment shares was mainly of a 'passive nature', reflecting a less pronounced decline of employment in services than in industry and agriculture. Only when growth of the overall economy gained momentum, employment in services started to rise in absolute terms as well: between nearly 1 million of services jobs were created in NMS-8. Despite rapid expansion, the shares of services in GVA and especially in employment in the NMS are still distinctly lower than in the EU Moreover, in all NMS the gap vis-à-vis the EU-15 is largest in the field of financial and other business services (marketing, consulting, auditing etc.). Within the services sector, employment gains were due to job creation in the market services segment (especially in trade, tourism and real estate see Landesmann et al, 2004). The services sector thus Figure 4 uses GVA data at current prices. The available evidence from selected NMS suggests that changes in relative prices did not affect the respective GVA shares to a large extent. Urban (2001), Landesmann et al (2004) and Stehrer (2004) for more details and development scenarios. Services shares are particularly low in the second-round accession countries, Bulgaria and Romania. 9

15 may become the major provider of new employment. But again, whether this process will lead to the creation of additional jobs is not sure. Parts of the service sector (especially financial services and retail trade) currently experience a restructuring process (as witnessed by industry earlier) which is associated with considerable efficiency improvements and layoffs of redundant workers Structural change and productivity growth In this section we will examine the effects of recent structural changes on NMS labour productivity growth. The traditional assumption of the growth accounting literature is that structural change is an important source of growth and overall productivity improvements. The standard hypothesis assumes a surplus of labour in some (less productive) parts of the economy (such as agriculture), thus shifts towards higher productivity sectors (industry) are beneficial for aggregate productivity growth. Even within industry shifts towards more productive branches should boost aggregate productivity. On the other hand, structural change may have a negative impact on the aggregate productivity growth if labour shifts to industries with slower productivity growth. The structural bonus and burden hypotheses were examined on example of Asian economies by Timmer and Szirmai (2000), on a large sample of OECD and developing countries (Fagerberg, 2000), and more recently by Peneder and DG Employment for USA, Japan and EU member states (Peneder, 2002, European Commission, 2003b). None of these studies has covered transition economies from Central and Eastern Europe. The overall developments regarding output, employment and productivity described above mask substantial structural changes within NMS economy and its individual sectors. Structural changes reflect inter alia different speeds of restructuring and resulting efficiency gains or losses at branch level. The impact of structural change on NMS aggregate productivity growth will be evaluated by a frequently applied shift-share analysis in analogy with Timmer and Szirmai (2000), Fagerberg (2000), Peneder (2002) and others. Shiftshare analysis provides a convenient tool for investigating how aggregate growth is linked to differential growth of labour productivity at sectoral level and to the reallocation of labour between industries. It is particularly useful for the analysis of productivity developments in NMS where data limitations prevent us to use more sophisticated econometric approaches (see Box 1) The evidence for productivity gains in NMS services sectors has been mixed so far. Moreover, a proper assessment is plagued by numerous conceptual and statistical problems (Wölfl, 2004). Rough estimates of labour productivity growth in services is provided in Section 4 below. Even this kind of analysis encounters a number of serious statistical problems. The majority of NMS does not publish sectoral value added data at constant prices. Owing to the lack of sector-specific price indexes we have applied GDP price deflators to calculate series at constant prices. Moreover, the measurement of output in certain services sectors is especially problematic (Wölfl, 2004). 10

16 BOX 1 Decomposition of aggregate labour productivity growth Using the same notation as presented in Peneder (2002), we decompose the aggregate growth of labour productivity into three separate effects: growth( LP ) T LP n n i, by i, fy i, by i, fy i, by i, fy i, by i, fy i, by i, by T, fy T, by i= l i= l i= l = = (1) LP LP T, by I: static shift effect II: dynamicshift effect III: within growtheffect LP ( S S ) + ( LP LP LP )( S T, by S ) + n ( LP LP where LP=labour productivity; by=base year, fy=final year; T=S over industries i; S i=share of industry i in total employment. ) S First, the structural component is calculated as the sum of relative changes in the allocation of labour across industries between the final year and the base year, weighted by the value of sector s labour productivity in the base year. This component is called the static shift effect. It is positive/negative if industries with high levels of productivity (and usually also high capital intensity) attract more/less labour resources and hence increase/decrease their share of total employment. The standard structural bonus hypothesis of industrial growth postulates a positive relationship between structural change and economic growth as economies upgrade from low to higher productivity industries. The structural bonus hypothesis thus corresponds to an expected positive contribution of the static shift effect to aggregate growth of labour productivity: The structural bonus hypothesis: n i, by ( S i, fy S i, by i = l LP ) > 0 (2) Second, dynamic shift effects are captured by the sum of interactions of changes in employment shares and changes in labour productivity of individual sectors/industries. If industries increase both labour productivity and their share of total employment, the combined effect is a positive contribution to overall productivity growth. In other words, the interaction term becomes larger, the more labour resources move toward industries with fast productivity growth. The interaction effect is however negative, if industries with fast growing labour productivity cannot maintain their shares in total employment. Thus, the interaction term can be used to evaluate Baumol's hypothesis of a structural burden of labour reallocation which predicts that employment shares shift away from progressive industries towards those with lower growth of labour productivity (Baumol, 1967). We would expect to confirm the validity of structural burden hypothesis in the NMS due to the above sketched shifts from industry to services (with lower productivity levels) at the macro level, respectively due to shifts from heavy (and capital-intensive) to light industries within manufacturing. The structural burden hypothesis: n ( LPi, fy LPi, by)( Si, fy Si, by) < i= l 0 (3) Thirdly, the within growth effect corresponds to growth in aggregate labour productivity under the assumption that no structural shifts in labour have ever taken place and each industry (sector) has maintained the same share in total employment as in the base year. We must, however, recall that the frequently observed near equivalence of within growth effect and aggregate productivity growth cannot be used as evidence against differential growth between industries. Even in the case that all positive and negative structural effects net out, much variation in productivity growth can be present at the more detailed level of activities As productivity has a robust tendency to grow, the within growth effect is practically a summation over positive contributions only. Conversely, for each industry the sign of the contribution to both shift effects depends on whether 11

17 Table 2 shows a decomposition of productivity growth in NMS (as well as in Bulgaria and Romania) at both macro level (total gross value added) and in manufacturing industry for the period As far as the economy as a whole is concerned, structural bonus hypothesis is mostly confirmed, though the contribution of labour shifts from low to high productivity growth sectors to aggregate productivity growth was in most cases rather small, in Romania and Latvia even negative. Having in mind the above mentioned data caveats regarding productivity measurement in the services sector, a detailed inspection of sectoral productivity performance gives a widely heterogeneous picture. 13 In most NMS, agriculture, construction, hotels and restaurants, as well as health and social work sectors recorded below average labour productivity growth (Figure 6a). On the other hand, data would suggest positive contributions of trade, real estate and other (community and social services) activities to aggregate productivity growth. Dynamic shift effects play an even smaller role as far as the contribution to aggregate productivity growth is concerned; structural burden (a small negative dynamic shift effect) was detected only in Slovenia and Romania. The overwhelming part (more than 80%) of aggregate productivity growth in NMS during the period can be attributed to productivity growth within individual economic sectors. This is broadly in line with productivity developments observed in advanced market economies, 14 but still somewhat surprising given the major restructuring that had occurred in NMS in that period. Obviously, aggregate productivity growth in NMS has mostly resulted from productivity improvements within individual sectors and their across the board productivity catching-up. In this respect, NMS economies display similarities with the more advanced EU-15 member states (Peneder, 2002, European Commission, 2003b) yet their overall productivity growth has been impressive. Data presented in the second part of Table 2 reveal that structural features of productivity growth in manufacturing industry were only marginally different. The evidence for individual NMS is mixed again, but a structural bonus (positive static shift effect) was detected for all NMS except the Czech Republic, Hungary, Estonia, Bulgaria and Romania. The negative static shift effect present in these countries means that labour moved away from (initially) high productivity manufacturing branches. As a rule, this effect resulted largely from labour shifts away from high labour productivity level industries (which are usually more capital labour shares have increased or decreased. The shift effects therefore capture only that comparatively small increment to aggregate growth which is generated by the net difference in productivity performance of the shifting share of the labour resources. Even that increment can either be positive (structural bonus) or negative (structural burden). In short, offsetting effects of shifts in employment shares of industries with high and low levels of labour productivity, as well as high and low productivity increases, explain why shift share analyses regularly fail to reveal substantial direct contributions of structural change to aggregate growth The measurement of output (and productivity) in services sector especially in trade, real estate and financial intermediation poses serious problems see O Mahony and van Ark (2003), Wölfl (2004). Peneder (2002) and European Commission (2003b) have found similar results for EU-15 countries and the USA in the period

18 intensive and use more intermediate inputs) like coke and refined petroleum, chemicals and basic metals branches. 15 Structural burden hypothesis a negative dynamic shift effect could be confirmed for nearly all NMS. The only exception is Hungary (and to a lesser degree also Poland and Slovakia) where dynamic shifts were dominated by simultaneous productivity improvements and growing employment shares in just a few branches (usually in electrical, optical equipment and transport equipment). Nevertheless, the aggregate productivity growth in NMS manufacturing was again dominated by productivity improvements within individual manufacturing branches. 16 Havlik (2003a), Hunya (2002), as well as the case studies undertaken in the framework of this project, provide some evidence for the key role played by foreign direct investments in productivity improvements and restructuring of NMS manufacturing. Van Ark and Piatkowski (2004) show that the main contribution to productivity growth in selected NMS (the Czech Republic, Hungary, Poland and Slovakia) during came from ICT-using manufacturing and non-ict manufacturing. Contrary to EU-15 and USA, the contribution of ICT-producing branches to aggregate productivity growth was much lower (with the exception of Hungary). Decomposition of manufacturing industry productivity growth in NMS thus again shows similar characteristics to those observed for EU-15 countries. For these countries, Peneder (2002) found only a weak evidence for the reallocation of labour towards high productivity branches (at 3-digit NACE level) and could not confirm the structural bonus hypothesis even for a longer time period ( ). Similar findings were obtained earlier by Timmer and Szirmai (2000) for a small sample of Asian economies, as well as by Faberberg (2000) for a number of OECD and developing countries. In this respect, we may conclude that the recent industrial restructuring in NMS did not differ too much from the earlier experience of other countries since shifts of labour among individual (2 digit NACE) industries apparently did not play a major role in total productivity improvements. There is some evidence of a structural burden effect in NMS manufacturing since employment shifts towards slower productivity growth industries had, on average, slightly negative impact on aggregate productivity growth in manufacturing. The overwhelming part of Note that due to limited data availability we use gross production as a measure of output. The negative static shift effect was particularly large in Bulgaria and Romania. Exemptions from a general tendency of productivity growth were in most cases only food, beverages, textiles and leather branches see Table 7 below. 13

19 Figure 1.6a Productivity growth in NMS and EU-15 by economic sectors, (annual averages, gross value added per employed person) EU15 NMS4 NMS7 NMS8 PL BG RO A-B CDE F GH I JK LQ Total Source: wiiw calculations based on wiiw Database and OECD STAN Database. Figure 1.6b Productivity levels in NMS economic sectors, 2002 (gross value added per employed person, at PPP, EU-15 =100) NMS4 NMS7 NMS8 PL BG RO AB CDE F GH I JK LQ Total NACE sectors: AB: Agriculture, forestry and fishing; CDE: Mining, quarrying, manufacturing, electricity, gas and water supply; F: Construction; GH: Wholesale, retail trade; Hotels and restaurants; I: Transport, storage and communications; JK: Financial intermediation; Real estate, renting and business activities; L: Public administration and defence; Education; Health and social work; Other activities. Source: wiiw calculations based on wiiw Database and OECD STAN Database. 14

20 Table 2 Decomposition of aggregate and manufacturing productivity growth in NMS (shift-share analysis), Percentage of total labour productivity growth explained by: static shift effect dynamic shift effect within growth effect Total productivity LPby*(Sfy-Sby)/LPby (LPfy-LPby)*(Sfy- Sby)/LPby (LPfy-LPby)*Sby/LPby effect growth in % p.a. Bulgaria, gross value added (without FISIM) Bulgaria, manufacturing output Czech Republic, gross value added (without FISIM) Czech Republic, manufacturing output Hungary, gross value added (without FISIM) Hungary, manufacturing output Poland, gross value added (without FISIM) Poland, manufacturing output Slovak Republic, gross value added (without FISIM) Slovak Republic, manufacturing output Slovenia, gross value added (without FISIM) Slovenia, manufacturing output Romania, gross value added (without FISIM) Romania, manufacturing output Estonia, gross value added (without FISIM) Estonia, manufacturing output Latvia, gross value added (without FISIM) Latvia, manufacturing output Lithuania, gross value added (without FISIM) Lithuania, manufacturing output Notes: Aggregate productivity based on gross value added at constant prices (without FISIM) and employment according to LFS statistics: Bulgaria: 12 NACE 1-digit sectors ( ), Czech Republic: 8 sectors ( ), Hungary and Poland: 12 sectors ( , resp. 2000), Slovak Republic:12 sectors ( ), Slovenia: 12 sectors ( ), Romania: 12 sectors ( ), Estonia: 12 sectors ( ), Latvia: 12 sectors ( ), Lithuania: 12 sectors ( ). FISIM: Financial intermediation services indirectly measured. Manufacturing labour productivity based on gross output at constant prices and employment for 14 NACE 2-digit manufacturing sectors. Sources: Countries in Transition wiiw Handbook of Statistics, wiiw, Vienna, 2003; wiiw Industrial Database. 15

21 overall manufacturing productivity growth in NMS can be attributed to productivity improvements taking place in nearly all manufacturing industry branches (albeit at widely different rates see Section 4) a process stimulated particularly by effects of FDI. In several NMS (especially in Hungary, Poland, Slovakia and Estonia), manufacturing labour productivity has recently expanded even faster than it did in the Asian Tigers countries during their rapid catching-up period. 4 Patterns of productivity catching-up in manufacturing This section looks in more detail at patterns of structural convergence of NMS manufacturing industry and evaluates the impact of structural changes on manufacturing industry labour productivity growth. Manufacturing industry provides 21% of all jobs in NMS-8 slightly more than in EU-15 (19.4%). However, the output of the sector, compared to aggregate production in the EU-15, is relatively small. Taken together, manufacturing production of all NMS-8 made up less than 5% of the total manufacturing output in the enlarged EU-25 in However, in view of the still grossly undervalued currencies, the 'real' shares of NMS' manufacturing are higher around 9% of the total EU- 25 manufacturing, and in some industries such as wood products, non-metallic minerals, rubber and plastics, food & beverages and manufacturing n.e.c. (mainly furniture) even more than that see Table 3. As far as employment is concerned, NMS-8 account for nearly 15% of EU-25 manufacturing jobs, with particularly high employment shares in the food and beverages, textiles, wood, coke and refinery industries. Large differences between production and employment shares point at substantial productivity gaps between the NMS and EU-15 member states. On average, NMS manufacturing labour productivity was below 30% of EU-15 level in 2002, respectively about 55% of that level if output values were converted with PPPs (with huge differences among individual NMS see Table A1 in Appendix). A crucial issue in the context of EU cohesion and NMS future productivity catching-up is whether (and in what manner) these gaps will be narrowed in future. Will NMS production shares in an enlarged EU-25 increase or, rather, will their employment shares decline? What will be the speed of these adjustments and how they will differ across individual countries and industries? 17 These and other questions will be addressed below. 17 The closure of NMS productivity gap in ten years (i.e. the equalization of their production and employment shares in EU-25) would require annual output growth differential of about 7 percentage points above EU-15 (in 15 years about 5pp). 16

22 Table 3 Size of European manufacturing industry after enlargement to EU-25 EU-15 NMS-8 NMS-10 NMS-8 NMS-10 NMS-8 NMS-10 NMS-8 NMS-10 Production (gross output) mn euro mn euro mn euro share in share in mn euro mn euro share in share in ER ER ER EU-15+NMS-8 EU-15+NMS-10 PPP PPP EU-15+NMS-8 EU-15+NMS-10 in % in % in % in % DA Food products; beverages and tobacco DB Textiles and textile products DC Leather and leather products DD Wood and wood products DE Pulp, paper & paper products; publishing & printing DF Coke, refined petroleum products & nuclear fuel DG Chemicals, chemical products and man-made fibres DH Rubber and plastic products DI Other non-metallic mineral products DJ Basic metals and fabricated metal products DK Machinery and equipment n.e.c DL Electrical and optical equipment DM Transport equipment DN Manufacturing n.e.c D Total manufacturing (Table 3 contd.) 17

23 Table 3 (contd.) EU-15 NMS-8 NMS-10 NMS-8 NMS-10 Employment Employment share in EU-25 share in EU-25 Persons in % in % DA Food products; beverages and tobacco DB Textiles and textile products DC Leather and leather products DD Wood and wood products DE Pulp, paper & paper products; publishing & printing DF Coke, refined petroleum products & nuclear fuel DG Chemicals, chemical products and man-made fibres DH Rubber and plastic products DI Other non-metallic mineral products DJ Basic metals and fabricated metal products DK Machinery and equipment n.e.c DL Electrical and optical equipment DM Transport equipment DN Manufacturing n.e.c D Total manufacturing Note: Production values in the year 2002 converted with current exchange rates (ER), resp. purchasing power parities (PPP) for NMS-10 comprise NMS-8 plus Bulgaria and Romania. Source: wiiw estimates based on national statistics and Eurostat New Cronos. 18

24 Before turning out to issues of productivity catching-up let us recall a few additional stylised facts regarding NMS manufacturing. Generally, manufacturing industry production in the NMS is more specialised than in the EU-15 and thus potentially more vulnerable to various shocks (European Commission, 2003). In terms of employment, the NMS specialisation of manufacturing industry is somewhat less pronounced, though still rather high. Employment specialisation measured by concentration ratios (CR3) 18 did not change much during the last decade (except in Bulgaria and Latvia where specialisation increased see Figure 7). The three biggest industries account for 40% (Czech Republic) to 60% (Latvia and Lithuania) of manufacturing employment, compared to fairly constant 44% over the last decade in EU-15 on average. 19 Figure 7 Manufacturing employment concentration ratios (CR3) in NMS LV LT BG CZ HU EE PO SI RO SK Note: CR3 is the share of 3 biggest NACE 2-digit industries in total manufacturing employment. Source: Own calculations based on wiiw Industrial Database. In terms of employment, the most important manufacturing branches in NMS are food and beverages, textiles, wood and wood products see Table 3. The majority of Central and East European NMS have nowadays an industrial structure which is very close to that of EU-15. Manufacturing employment structure in the Czech Republic, Hungary, Slovakia and Slovenia came very close to that observed in EU-North by On the other hand, Concentration ratios are here defined as the share of 3 largest manufacturing branches in the total CR3. However, in some old EU member states is the employment concentration also rather high (e.g. 49% in Greece and even 55% in Ireland). 19

25 Figure 8 Deviations of NMS and EU-15 manufacturing employment structures, years 2002 and BG EU-15 (2002) EU-North EU-South HU 9 RO CZ PL LT SK LV SI EE EU-15 (1995) EU-North EU-South HU RO BG CZ PL LT SK LV SI EE Note: Structural deviations (S) are calculated from 2-digit NACE rev. 1 data for manufacturing employment. A lower value of S indicates more structural similarity. For a definition see the following formula: t t t S* = ( sh sh ) 2 ( sh / 100) k k k k k = individual industry sh k = share of industry k in total employment (in %) t i = country index, where i = 1,2; 1 denoting the EU. Source: Own calculations based on wiiw Industrial Database and Eurostat. 20 EU-South is defined as an average of Greece, Portugal and Spain, EU-North as an average of Germany, France and the United Kingdom. 20

26 employment structures in the Baltic states tend to be more distinct in particular compared to EU-North. It is also interesting to note that in Bulgaria and Romania industrial employment structures increasingly deviate from that of EU-15 (and especially from EU- North), largely as a result of the collapse of machinery, electrical and transport equipment industries and rising shares of food, beverages and textiles (Figure 8). Given the lack of comparable data for manufacturing employment in some NMS for earlier periods, this analysis will again focus on the period after Moreover, since detailed data on value added are not available for most NMS we use gross production instead. Between 1995 and 2002, manufacturing production in the NMS-8 rose on average much faster (6.4% per annum) than in the EU-15 (2.1% per annum see Table 4). This translates into a growth differential in favour of the NMS of 4.3 percentage points per year, substantially higher than the growth differential for GDP during the same period (compare Section 1). On the other hand, manufacturing employment in the NMS declined strongly (-2.1% per annum) while it stayed more or less constant in the EU-15, resulting in a negative growth differential for the NMS-8 vis-à-vis the EU of -2.1 percentage points per annum, again significantly higher than for total employment. As a result, NMS' productivity catching-up, already impressive at the GDP level, was even more pronounced in manufacturing: between 1995 and 2002, the cumulated productivity gain in NMS manufacturing industry amounted to 79%, compared to 16.4% for the EU-15 (Table 4). The annual growth differential was 6.5 percentage points, by far exceeding the growth differential in terms of macro-productivity. Maintaining this speed of catching-up would help to eliminate NMS productivity gap in about ten years. Table 4 Labour productivity catching-up in manufacturing: NMS vis-à-vis the EU-15, NMS-8 1) Growth rate NMS growth differential Growth rate in % against EU-15 in pp in % cumu- annual cumu- annual cumu- annual lated average lated average lated average EU-15 Production Production Employment Employment Productivity Productivity Notes: Gross production and productivity in real terms.-1) Central and East European New Member States, weighted average. Sources: wiiw Database, incorporating national statistics, WIFO and wiiw calculations using AMECO. Figure 9 shows indexes of production and employment for individual countries in the period which indicate an impressive productivity recovery in most NMS. Hungary even managed to slightly increase the number of manufacturing jobs, in the remaining NMS productivity gains were associated with further lay-offs of workers. Hungary s outstanding 21

27 productivity performance in recent years thus resembles that of Ireland. Estonia, Poland and Slovakia outperformed Austria, Denmark and Finland, which have been the best performers in terms of labour productivity growth among the EU-15 (European Commission, 2003a). In several NMS and, as will be shown below, in a few manufacturing branches, there has been a spectacular productivity catching-up. But in contrast to the EU- 15 where manufacturing employment has been stagnating, productivity catching-up in most NMS has been associated with considerable job losses. The new EU member states will require specific growth and employment strategies (training, support of SMEs, regional policies for attracting FDI, etc.) to stabilize employment levels in manufacturing (and to create new employment opportunities in other sectors especially services) while simultaneously maintaining the recent pace of productivity improvements. 21 Figure 9 Manufacturing production and employment growth in NMS and EU-15, 2002 (1995=100) EU-15 production CZ employment EE RO HU BG 0 LV SI LT SK PL Source: Own calculations based on wiiw Industrial Database and AMECO. The NMS productivity gaps for the whole economy discussed in Section 1 above are similar to those in the manufacturing industry although their proper assessment poses considerable problems (see Appendix). On average, NMS manufacturing labour productivity was only 30%-55% of that in EU-15 in the year 2002 (see also Figure 1.6b above). Table A1 provides several alternative estimates of manufacturing labour productivity levels (compared to EU-15 average) and their sectoral variation. Hungary s productivity leadership in NMS manufacturing (roughly half of the average productivity 21 See also European Commission (2004), Celin (2003) for a more detailed discussion of employment strategies in the NMS. 22

28 level in EU-15) is confirmed, Slovenia s productivity (about the same as in the Czech Republic) is surprisingly low given its relatively high per capita income. There are large productivity gaps among individual NMS and also the sectoral variation of labour productivity is relatively high, especially in Hungary, Slovakia and Slovenia (such comparisons are of course affected by varying capital intensity of individual industries and the use of intermediate inputs). Contrary to broader sectoral developments shown above (Figure 6), a comparison of productivity changes across individual manufacturing branches displays a quite clear pattern: The most obvious productivity winner in the period was the electrical & optical equipment industry, performing much above average in all NMS, followed by the transport equipment industry and manufacturing n.e.c. (mainly furniture see Table A2). Note that all these branches were attractive targets for FDI. In the Baltic states, nonmetallic mineral products and basic metals are clear productivity winners as well. Typical productivity losers are the food & beverages industry, textiles & textile products, leather & leather products, paper & printing, chemicals and rubber and plastics. The poor productivity performance of food industry is both surprising and disturbing: this industry received large amounts of FDI, it also ranks among biggest employers in most NMS. In general, we find certain evidence that technologically more sophisticated industries have strongly improved their productivity performance, while traditional sectors using standard techniques and low skilled labour have been falling behind Productivity and employment growth dilemmas Productivity growth recorded in most NMS in the period after 1995 has been associated with only meagre increases of employment (in manufacturing industry even with considerable job losses here with the exception of Hungary). In the context of EU Lisbon Strategy which aims at both improved competitiveness and high employment growth, the NMS thus face an even greater challenge than the EU-15 Member States. Focusing on both targets simultaneously (i.e. fast productivity growth and employment growth) may be conflicting. 23 Taking into account that NMS are confronted with a situation of low productivity levels (about half of the EU-15 average see above) and, at the same time of high unemployment (on average nearly twice the EU-15 level), they need to foster both productivity and employment growth simultaneously. Realistically, the main accent of economic policies in these countries should focus on at least keeping existing jobs while simultaneously maintaining the recent pace of productivity catching-up Using a different classification, van Ark and Piatkowski (2004) found that the largest contribution to aggregate labour productivity growth in selected NMS (the Czech Republic, Hungary, Poland and Slovakia) during the period originated not from ICT-producing manufacturing, but rather from ICT-using and non-ict manufacturing branches. Policies aiming at higher employment may have negative consequences for labour productivity growth at least in the short run see O Mahony and van Ark et al.,

29 This is a formidable task: the relation between employment and production growth (employment elasticity to output growth see Employment in Europe, 2002) in the NMS has been rather disappointing since even in the recent period of relatively robust economic growth (that is after 1995) there has been little effect on the job creation; the employment elasticity to GDP growth has been much below unity. This is illustrated in Figure 10 where indexes of GDP and employment growth (and the respective trend lines for the period ) are plotted for three NMS. There are differences between individual countries: a constant employment would require GDP growth of at least 3% in Hungary, yet more than 4% in the Czech Republic and about 6% in Poland (the latter two countries could enjoy such rates of GDP growth only twice during the last decade). Regression estimates covering a sample of all NMS-8 (that is without Bulgaria and Romania) for the time period show that the average critical rate of GDP growth which would prevent further employment decline in the NMS has been nearly 6% per year in the period , which is again much more than the GDP growth actually achieved during that period (Table 5, see also Table 1 above). 24 For the manufacturing industry, the same estimation method yields even more disturbing results: the critical rate of production growth is here more than 10% per year, 25 nearly twice as high as the average manufacturing growth rate actually achieved during the (high growth) period of (Table 4). Seen from this angle, and taking into account the expected rates of economic growth and NMS evolving economic structures, the prospects for rising employment outside of services are not very encouraging. Without a substantial acceleration of their economic growth and/or a significant job creation in the services sector, the NMS seem to be condemned either to remain substantially less productive than EU-15 Member States, or to face the challenge of an even higher unemployment in the future This compares with a critical GDP growth rate of just 0.5% estimated for the same period for the EU, USA and Japan, resp. 1.3% GDP growth estimated for these countries for the period As shown in Figure 10, there are differences in estimated critical growth rates among individual NMS. However, regression estimates with countryspecific dummies did not yield statistically significant parameters. Compared to 3.2% production growth estimated for the EU, USA and Japan for the same period. During the last couple of years, the only sectors where additional jobs were created in the NMS are trade, hotels and restaurants, real estate, public administration and other activities see Landesmann et al (2004) for more details. 24

30 Figure 10 Employment elasticity of GDP growth in selected NMS, CZ HU PO CZ trend HU trend PO trend Employment growth index GDP growth index Source: wiiw calculations from wiiw Database based on national statistics. 25

31 Table 5 Regression estimates of NMS employment elasticity to GDP growth, Part A: Employment (yemp) and GDP growth (xgdp) Source SS df MS Number of obs = F( 1, 70) = 7.46 Model Prob > F = Residual R-squared = Adj R-squared = Total Root MSE = yemp Coef. Std. Err. t P> t [95% Conf. Interval] xgdp _cons Note: The estimated regression equation was: where: yemp = const + b*xgdp yemp: index of employment growth, xgdp: index of GDP growth. Min. estimated GDP growth index (critical growth rate)needed for employment staying at least constant (yemp = 1) is thus: ((1-cons)/xGDP) = Part B: Manufacturing employment (yemp) and output growth (xout) Source SS df MS Number of obs = F( 1, 70) = Model Prob > F = Residual R-squared = Adj R-squared = Total Root MSE = yemp Coef. Std. Err. t P> t [95% Conf. Interval] yout _cons Min. estimated manufacturing output growth index (critical growth rate) needed for manufacturing employment staying at least constant (yemp = 1) is thus: ((1-cons)/xOUT) = Source: Own calculations, wiiw Database. 26

32 References Aiginger, K. and M. Landesmann (2002), Competitive Economic Performance: USA versus EU, wiiw Research Reports, No. 291, The Vienna Institute for International Economic Studies (wiiw), Vienna, November. Barrel, R. and D. T. te Velde (2000), 'Catching-up of East German Labour Productivity in the 1990s', German Economic Review, Vol. 1, No. 3, August, pp Baumol, W. J. (1967), Macroeconomics of Unbalanced Growth: The Anatomy of Urban Crisis, The American Economic Review, Volume 57, pp Celin, M. (2003), 'European Employment Strategy: The Right Answer for the Candidate Countries?', Die Union, No. 1, European Commission, Representation in Austria, Vienna, pp Commission of the European Communities (2003), 'Impact of Enlargement on Industry', Commission Staff Working Paper SEC(2003) 234, February. Damijan, J., M. Knell, D. Majcen and M. Rojec (2003), Technology Transfer through FDI in Top-10 Transition Countries: How Important are Direct Effects, Horizontal and Vertical Spillovers?, Institute for Economic Research, Working Paper, No. 17, Ljubljana. EBRD (2000), Transition Report Employment, skills and transition, London. European Commission (1999), 'The Competitiveness of European Industry Report', Working document of the services of the European Commission, Luxembourg. European Commission (2000), European Competitiveness Report 2000, European Commission, (See European Commission (2001), Employment in Europe Recent Trends and Prospects, European Commission, DG Employment and Social Affairs. European Commission (2002), Employment in Europe Recent Trends and Prospects, European Commission, DG Employment and Social Affairs. European Commission (2003a), 'European competitiveness report 2003', Commission staff working document, Luxembourg. European Commission (2003b), Employment in Europe Recent Trends and Prospects, European Commission, DG Employment and Social Affairs. European Commission (2004), Strengthening the implementation of the European Employment Strategy. COM(2004) 239 final, Brussels, April. Eurostat (2001), The European Comparison Programme: Aggregate-level results for 1999, Luxembourg, September. Eurostat (2003a), 'Structural Indicators' (see Fagerberg, J. (2000), 'Technological progress, structural change and productivity growth: a comparative study', Structural Change and Economic Dynamics, Vol. 11, No. 4, pp Havlik, P. (2003a), 'Restructuring of Manufacturing Industry in the Central and East European Countries', Prague Economic Papers, No. 1, pp Havlik, P. (2003b), 'CEE Industry in an Enlarged EU: Restructuring, Specialization and Competitiveness', in S. Richter (ed.), The Accession Treaty and Consequences for New EU Members, wiiw Current Analyses and Country Profiles, No. 18, The Vienna Institute for International Economic Studies (wiiw), April, pp

33 Hunya, G. (2002), 'Recent Impacts of Foreign Direct Investment on Growth and Restructuring in Central European Transition Countries', wiiw Research Reports, No. 284, The Vienna Institute for International Economic Studies (wiiw), Vienna. Landesmann, M. (2000), 'Structural change in the transition economies, ', Economic Survey of Europe, United Nations Economic Commission for Europe, No. 2/3 pp Landesman, M., Vidovic, H., and Ward, T. (2004) CHAPTER 2 for DG EMPL Monnikhof, E. and B. van Ark (2002), 'New estimates of labour productivity in the manufacturing sectors of the Czech Republic, Hungary and Poland, 1996', Groningen Growth and Development Centre, University of Groningen & The Conference Board, January. O Mahony. M., van Ark, B. (2003), EU productivity and competitiveness: An industry perspective. Can Europe resume the catching-up process?, European Communities, Luxembourg. Palme, G. (1999), Impacts of an EU Eastern enlargement on Austria s manufacturing, Austrian Economic Quarterly, No. 1, pp Peneder, M. (2001), Entrepreneurial Competition and Industrial Location, Edward Elgar, Cheltenham, UK. Peneder, M. (2002), 'Structural Change and Aggregate Growth', WIFO Working Papers, No. 182, Vienna. Soubbotina, T. and K. Sheram (2000), 'Beyond economic growth: meeting the challenges of global development', World Bank, Washington DC, October. Stare, M. and S. Zupancic (2000), 'Liberalisation of Trade in Services: Slovenia's Experience', Round Table on 'Ten Years of Trade Liberalisation in Transition Economies', OECD, document CCNM/TD (2000)52. Stehrer, R. (2004) CHAPTER 6 for DG EMPL Timmer, M. P. and A. Szirmai (2000), 'Productivity growth in Asian manufacturing: the structural bonus hypothesis examined', Structural Change and Economic Dynamics, Vol. 11, No. 4, pp Urban, W. (2000), 'Patterns of Structural Change in CEEC Manufacturing', in M. Landesmann (ed.), Structural Developments in Central and Eastern Europe. wiiw Structural Report 2000, The Vienna Institute for International Economic Studies (wiiw), Vienna, pp Urban, W. (2001), 'Industry and growth in a global economy theoretical considerations and empirical results', Project Report (Jubiläumsfondsprojekt Nr. 7615), The Vienna Institute for International Economic Studies (wiiw), March van Ark, B. and Piatkowski, M. (2004), 'Productivity, Innovation and ICT in Old and New Europe', Research Memorandum GD-69, Groningen Growth and Development Centre, University of Groningen & The Conference Board. Vidovic, H. (2002), 'The Services Sectors in Central and Eastern Europe', wiiw Research Reports, No. 289, The Vienna Institute for International Economic Studies (wiiw), Vienna. wiiw (2001), 'Competitiveness of Industry in CEE Candidate Countries', Report to the European Commission, DG Enterprise, Final Report, July 2001; available on the EU DG Enterprise Website and at wiiw (2004), Countries in Transition. wiiw Handbook of Statistics, The Vienna Institute for International Economic Studies (wiiw), Vienna. Wölfl, A. (2004), Productivity Growth in Services Industries: Is There a Role for Measurement?, International Productivity Monitor, No. 8, Spring, pp Zysman, J. and A. Schwartz (eds.) (1998), Enlarging Europe: The industrial foundations of a new political reality, University of California International and Area Studies Digital Collection, Research Series, No

34 APPENDIX Labour productivity in international comparison International productivity level comparisons are hampered mainly by the difficult conversion of the national output data to a common currency unit (in the services sector even by the proper measurement of national output see Wölfl, 2004). The use of market exchange rates is not appropriate for the conversion to common currency units (especially for NMS, mainly due to their still grossly undervalued currencies and fluctuating exchange rates). Alternative proxy converters are either purchasing power parities (PPPs), or much better branch-specific unit value ratios (UVR) which compare prices of representative industrial products. UVR estimates for the manufacturing industry (for the year 1996) are available only for 3 NMS: the Czech Republic, Hungary and Poland relative to Germany from a research project jointly conducted by the wiiw and the University of Groningen. 27 The estimated Hungarian manufacturing industry labour productivity was slightly less than 40% of the German level in 1996, the respective Czech-German productivity relation was 35%, the Polish-German productivity relation was 25%, all with fairly large sectoral differences. Figure A1 shows productivity comparisons of these 3 NMS with Austria; the year 2002 was obtained after extrapolation from the above quoted 1996 UVR-based benchmarks with country and branch-specific rates of productivity growth. The results show that Hungarian manufacturing productivity reached close to half of Austrian level by the year 2002; there was a closure of productivity gap by nearly 10 percentage points since In Poland, the closure of the gap was even faster, whereas the productivity gap of the Czech manufacturing relative to Austria declined by less than 2 percentage points. A closer look at the performance of individual branches shows that relatively smaller productivity gaps (and impressive productivity catching-up) were observed especially in manufacturing of rubber and plastics, electrical, optical equipment and transport equipment, but virtually no catching-up occurred in other branches. Hungary's labour productivity in transport equipment industry, Polish productivity in rubber and plastics were apparently higher than in Austria. On the other hand, productivity gaps in food & beverages, leather and wood products have even widened since For a cross-country comparison, data in national currencies were converted with both exchange rates (ER) and purchasing power parities (PPPs). PPPs were adopted from the ECP 1999 see Eurostat (2001). The first data set presented in Table A1 (PPP for GDP) results from national productivity figures converted with purchasing power parities for the whole GDP. This conversion leads to higher productivity estimates for the NMS. The second data set uses as a conversion factor partial PPPs for gross fixed capital formation (PPPCAP) where the price levels in the NMS are relatively high (presumably due to imports of machinery and equipment). This conversion thus leads to lower productivity estimates. Given the close correspondence of the latter productivity estimates to the theoretically superior UVR-based productivity data for the Czech Republic, Hungary and Poland (UVRs are not available for other NMS), and assuming that a similar correspondence between UVR and PPPCAP exists for other NMS as well, one can assume that productivity levels expressed at PPPCAP in Table A1 are probably closer to reality at least for manufacturing industry as a whole. 27 See Monnikhof and van Ark (2002). 29

35 Table A1 Labour productivity levels in MNS manufacturing industry, year 2002 Czech Slovak Republic Estonia Hungary Latvia Lithuania Poland Republic Slovenia Bulgaria Romania Manufacturing total, productivity in EUR (at PPP for GDP) EU(15) = Manufacturing total, productivity in EUR (at PPPCAP) EU(15) = Manufacturing total, productivity in EUR (at ER) EU(15) = Manufacturing total = 100 DA Food products; beverages and tobacco DB Textiles and textile products DC Leather and leather products DD Wood and wood products DE Pulp, paper & paper products; publishing & printing DF Coke, refined petroleum products & nuclear fuel DG Chemicals, chemical products and man-made fibres DH Rubber and plastic products DI Other non-metallic mineral products DJ Basic metals and fabricated metal products DK Machinery and equipment n.e.c DL Electrical and optical equipment DM Manufacture of transport equipment DN Manufacturing n.e.c Others Standard deviation Standard deviation (without DF) Sources: wiiw estimates based on national statistics, OECD, EUROSTAT and UNIDO. 30

36 Table A2 Relative productivity gains in NMS manufacturing, (average annual change in % for total manufacturing (D) and relative gains DA to DN, in percentage points) 1) Czech Slovak Republic Estonia 2) Hungary Latvia 2) Lithuania 2) Poland Republic Slovenia Bulgaria Romania D Manufacturing total DA Food products; beverages and tobacco DB Textiles and textile products DC Leather and leather products DD Wood and wood products DE Pulp, paper & paper products; publishing & printing DF Coke, refined petroleum products & nuclear fuel DG Chemicals, chemical products and man-made fibres DH Rubber and plastic products DI Other non-metallic mineral products DJ Basic metals and fabricated metal products DK Machinery and equipment n.e.c DL Electrical and optical equipment DM Transport equipment DN Manufacturing n.e.c Notes: 1) Calculation of relative gains: DA ( ) minus D ( ) = relative gain DA. Positive values indicate higher, negative values lower than average productivity growth relative to total manufacturing (D). - 2) Sources: wiiw estimates based on national statistics; wiiw Industrial Database. 31

37 Figure A1 Manufacturing labour productivity in selected NMS (UVR-based), years 1996 and 2002 (Austria = 100) HU1996 HU2002 PL1996 PL2002 CZ1996 CZ Food Textile Leather Wood Paper Coke Chem. Rubber Mineral Metals Machin. Electr. Transport Others Total Source: wiiw Industrial Database, own estimates based on Monnikhof and van Ark (2002). 32

38 Economic Restructuring and Labour Markets in the Accession Countries Research Project commissioned by EU DG Employment, Social Affairs and Equal Opportunities, Contract No. VC/2003/0367 Coordinated by in cooperation with Alphametrics and The Vienna Institute for International Economic Studies Foreign direct investment and restructuring in the automotive industry in Central and Eastern Europe by Slavo Radošević* and Andrew Rozeik* June 2004 *) School of Slavonic and East European Studies (SSEES), University College London

39 Contents 1 Introduction 2 Globalization of the automotive industry and the emergence of Central and Eastern Europe as a global production location 2.1 CEE in the world automotive industry 2.2 Value creation potential of CEE for auto MNC 3 Restructuring of the automotive industry in CEE: key features 3.1 CEE as an automotive market 3.2 CEE as an automotive producer Socialist heritage in automotive industry The role of the automotive industry in CEE manufacturing Employment: OEMs and suppliers Productivity Trade Foreign Direct Investments (FDI) 4 Micro view: OEMs in CEE key drivers of restructuring? 4.1 Market and production dynamics at firm level 4.2 OEMs and building of a local supply base 5 Policies influencing automotive industry restructuring in CEE 5.1 Tariffs 5.2 FDI and privatization policies 5 3 Clusters policy 6 Conclusions Bibliography Tables and Figures Table 1 Old and new models of competition in automotive industry Table 2 Share of production of cars in West and East Europe, Table 3 Production of cars in global regions, (in thousand units) Table 4 The fate of major indigenous socialist manufacturers after 1989 Table 5 Changes in employment in the CEE automotive industry Table 6 Škoda Superb suppliers, 2002 Table 7 Estimated production plant capacity in CEECs, 2000

40 Table 8 Table 9 Trade in road vehicles and parts (imports + exports) EU-15 candidate countries, EUR million Shares in automotive exports and imports to/from CEE and Turkey of three major partner countries (= 100%), 1999 Table 10 FDI stock and share of automotive industry, 2000 Table 11 Top automotive MNCs in CEE, Table 12 Sales of top MNCs in the CEE automotive industry by home country, 1999/2000 (USD million) Table 13 Distribution of value-added at Magyar Suzuki (%) Table 14 Škoda Auto suppliers and supply volume, Table 15 Number of Fiat Auto suppliers, Table 16 Import duties on EU imports, Table 17 Figure 1 Ownership changes in CEE manufacturers Relationship between initial level of cars per 1000 pop in 1989 and index of increase 1998/1989 Figure 2 Registered cars in Central and Eastern Europe, Figure 3 Car sales in Figure 4 Relationship between GDP per capita and number of cars per 1000 population, 2001 Figure 5 Relationship between household consumption and price level of passenger cars, EU15 = 100, 2002 Figure 6 Socialist car production in CEECs, Figure 7 Share of automotive industry in manufacturing, 2001 (Czech Republic, 2000) Figure 8 Car production in CEECs, Figure 9 Number of employees in the automotive industry in CEE, 2001 Figure 10 Share of foreign affiliates in employment in car industry Figure 11 Change in employment in car industry and Figure 12 Figure 13 Productivity, unit personnel costs and unit labour costs in manufacturing of motor vehicles compared to manufacturing (= 100), 2001 Productivity, unit personnel costs and unit labour costs of assemblers, car parts suppliers = 100, 2001 Figure 14 Manufacturer productivity, Figure 15 EU-15 exports and imports to/from CEE and Turkey, 1999, in EUR million Figure 16a Market shares of key players (market share >4%) in the CEE markets Figure 16b Market shares of key players in the CEE markets Figure 17 Production of CEE OEMs,

41 Abstract This study reviews and explores the major effects of FDI on industry restructuring of the CEE automotive industry. In particular, we are interested if automotive companies have exploited the value creation potential of CEE? Which factors explain the scale and depth of automotive industry restructuring in CEE? What are the economic effects of restructuring the automotive industry in terms of employment, trade and technology? What has been the role of national and EU policies in shaping FDI and restructuring in the CEE automotive industry? Our results suggest that the value creation potential of CEE as a global automotive location has not yet been fully exploited with great differences across countries. A combination of country specific factors (proximity to EU markets, socialist heritage in automotive industry, skilled labour and privatization policies) coupled with strategies of automotive MNCs have generated different country patterns and very different effects on industry. Improvements in productivity and technology transfer in both embodied (equipment upgrading) and disembodied form (know how) are significant in countries with large FDI in automotive industry. Employment effects are positive in particular in automotive suppliers industry. Privatization policy in early 1990s which was followed by a policy of attracting Greenfield FDI on the eve of EU accession is crucial in explaining country differences in FDI presence. Automotive investors have foreseen EU accession and in that respect, EU membership will not bring to changes in trend but possibly a deepening of the automotive cluster in central Europe. However, whether this will happen or not will depend on the ability of CEECs to develop sector specific policies which would support the upgrading of local automotive suppliers. i

42 Summary The EU automotive industry has embraced Central and Eastern Europe (CEE) as its market and as production location, but the potentials have not been exhausted. In absolute terms, during the 1990s the production of cars stagnated at around 14 million units in Western Europe, while it increased from 1.8 to 2.3 million in the CEE countries. The shifting balance in production between Western and Eastern Europe is merely a reflection of the production shift towards emerging markets in the automotive industry. The restructuring of the CEE automotive industry has been entirely foreign-led. Foreign investment enterprises have substantially increased their share in employment in the car industry. In the Czech Republic, in Hungary and Poland foreign firms accounted for about 70% of overall employment by the end of the 1990s. Competition dynamics have led to growth of some producers and to a relative decline of others. Škoda and other VW subsidiaries are the most successful. Daewoo-FSO has faced problems since 1999 as a result of financial difficulties with its Korean parent company. Similarly, Fiat Poland encountered problems with its parent company in Italy and was forced to restructure and cut down its production. The latest investments have come from non-european companies Toyota and Hyundai who have associated with the two main French manufacturers, Renault and PSA Peugeot Citroën, as well as KIA, which builds greenfield plants in the Czech Republic and in Slovakia. The effects of FDI on growth, restructuring and employment in industry are positive. FDI has resulted in increased specialization in the automotive industry at the European level. The integration of Central and Eastern Europe into a network of major European automotive multinational corporations (MNCs) has made it possible to produce different models in different countries and to reorganize the value chain in a way that has created bigger value added for MNCs. Only a minority of activities have been relocated from Western Europe, the majority of internationalization took the form of expansions and extensions, which suggests that the EU enlargement has been a positive sum game in the automotive industry. Those CEE countries that have attracted FDI in this industry have benefited through preserved employment, increased productivity and export and through a great potential for developing a local supply base. The combination of country-specific factors (proximity, socialist heritage in the automotive industry, skilled labour, privatization) coupled with strategies of automotive MNCs have generated different country patterns and very different effects on industry. The privatization policy in the early 1990s, followed by the policy of attracting greenfield FDI on the eve of EU accession, has been crucial in explaining the country differences in FDI presence. ii

43 Development and integration has been most profound in the Czech Republic and Hungary. It is facing difficulties in Poland and has just started in Slovakia and Romania. Productivity in the automotive industry is well above the industry average and company evidence points to large productivity gains. In that respect, the arrival of large assemblers has produced quite substantial effects, which in a next stage need to deepen through further development of the local supply base. Regarding the prospects for further industry restructuring, they depend mainly on improvements among local suppliers. We expect a further arrival of first-tier suppliers and a deepening of the local supply base. However, whether or not this will happen will depend on the CEE countries ability to develop sectorspecific policies to support the upgrading of local automotive suppliers. A crucial policy issue is whether the current national and EU policies are addressing this next stage of industrial upgrading in the CEE automotive industry. Most of the CEE countries have been active through FDI policy to attract automotive MNCs. This policy focus has become far from sufficient for industrial upgrading, which requires integration between FDI and vocational training and innovation policies. In order to assist industrial upgrading, the CEE countries should take into account the network character of local and global companies. This has already been recognized (implicitly or explicitly) through the national subcontracting programmes (Czech Republic) and the Hungarian integrator programme, which aim to integrate domestic firms with foreign firms through supply linkages. Inter-firm linkages, which have emerged through automotive value chains, should be further deepened. Job and retraining grants as tools of FDI/subcontracting and innovation policy should be expanded throughout the region, possibly linked to Structural Funds programmes. This should be complemented with clustering policies and the promotion of learning networks which would closely connect suppliers and assemblers. iii

44 Slavo Radošević and Andrew Rozeik Foreign direct investment and restructuring in the automotive industry in Central and Eastern Europe 1 Introduction Since 1989 industrial restructuring of central and Eastern Europe (CEE) has been greatly dependent on foreign direct investment (FDI). These have brought capital, technology, know how and access to foreign markets. Equally, local markets, cost advantages and a skilled labour force have attracted foreign investors. Privatization policy and the EU accession have greatly facilitated this process. The automotive industry 1 is a leading sector in terms of its importance for industry restructuring in CEE. Moreover, it is one of three sectors, in addition to electronics and the clothing industry, in which CEE has become a global production location. Hence, this sector represents very interesting case for understanding how globalization of CEE as an automotive market and production location has affected these economies and industry in particular. Effects of globalization of CEE in the automotive industry, its determinants as well as its future prospects are still largely unexplored 2. This study is of the review type and aims to explore the major effects of FDI on industry restructuring of the CEE automotive industry. In particular, we are interested if automotive companies have exploited value creation potential of CEE? Which factors explain the scale and depth of automotive industry restructuring in the CEE? What are the economic effects of restructuring of automotive industry in terms of employment, trade and technology? What has been the role of national and EU policies in shaping FDI and restructuring in the CEE automotive industry? Our results suggest that the value creation potential of CEE as global automotive location has not yet been fully exploited. This process has most progressed in the Czech Republic and Hungary, is consolidated in Slovenia and has started in Slovakia and Romania. It is faced with difficulties in Poland while other CEE countries are largely bypassed by automotive networks. A combination of country specific factors (proximity to EU markets, socialist heritage in the automotive industry, skilled labour and privatization policies) 1 2 For the purpose of this paper, the automotive industry is defined using NACE classification DM (manufacture of transport equipment) code 34 (manufacture of motor vehicles, trailers and semi-trailers), and SITC classification 78 (road vehicles including air-cushion vehicles). The terms automobile, car, motor, and motor vehicle will be used here interchangeably to mean the same thing. There have been several papers written on the CEE automotive industry (see for example Havas 1997, Havas, 2000, Sadler et al. 1993, Sadler and Swain 1994, Swain 1998, van Tulder and Ruigrok 1998, Richet and Bourassa 2000, Pavlínek 2002, Pavlínek and Smith 1998). These are mostly written from an economic geography or industry perspective. To date there has been very little written on the globalization of CEE automotive industry from industry restructuring perspective. 1

45 coupled with strategies of automotive MNCs have generated different country patterns and very different effects on the industry. Improvements in productivity and technology transfer in both embodied (equipment upgrading) and disembodied form (know how) are significant in countries with large FDI in the automotive industry. Employment effects are positive in particular in the automotive suppliers industry. Privatization policy in the early 1990s, which was followed by policy of attracting Greenfield FDI on the eve of EU accession, was crucial in explaining country differences in FDI presence. Automotive investors have foreseen EU accession and in that respect, EU membership will not bring to changes in trends but possibly deepening of automotive clusters in central Europe. However, whether this will happen or not will depend on the ability of CEECs to develop sector specific policies which would support upgrading of local automotive suppliers. The restructuring of the automotive industry evolves around large investors and is based on a good understanding of competitive dynamics and micro-strategies. In terms of an analytical perspective, this requires the blending of economic and international business perspectives which has been followed in this paper. This paper is organized as follows. Part 2 positions CEE within the globalization process of the automotive industry and try to estimate value creation potential of CEE as production location for MNC. Ultimately, it is this value creation potential that will determine the scope and pace of restructuring in the region. Part 3 analyses major aspects of restructuring the automotive industry employment, productivity, trade and FDI effects. Section 4 analyses the industry from a micro perspective including how MNCs have addressed the issue of local supply base. Part 5 analyses key policies and how they have shaped globalization and restructuring of the CEE automotive industry. Conclusions summarize the main arguments and derive a few policy ideas. 2 Globalization of the automotive industry and the emergence of Central and Eastern Europe as a global production location The opening and integration of CEE automotive markets and industry coincided with the emergence of new model of competition in world automotive industry. The Sloan Foundation study on automotive industry (Sturgeon and Florida, 1999) concludes that the automotive industry has transited from an older "domestic" model of competition to a new global model. The key features of the old and new models are summarized in Table 1. The key drivers of transition from the domestic to global model of competition are modularization and supplier outsourcing. Increased complexity and capital costs of assembly have forced assemblers to deverticalize and shift part of design development to first tier suppliers. Once vertically integrated assemblers have become the node of networks to which they outsource a wide range of tasks. 2

46 Table 1 Old and new models of competition in automotive industry Old model of competition Domestic model (competition based on exporting from home country supply-base) Emerging markets as dumping grounds for old models and production equipment Export-led industry (firms from different countries compete mainly through markets) New model of competition "Global" model (day-to- day production functions are organized on a regional and global basis) Emerging markets as locations for building leading-edge productive capacity Network-led industry (each major firm is producing within each major market) Source: Based on Sturgeon and Florida, 1999, p Costs of design and produce are shared across common modules which can be built in a variety of different models. For example, VW has shared costs of its platform for Škoda with Audi and VW models. Sturgeon and Florida (1999) point out that the introduction of modular assembly may reduce minimum scale economies and greater specialization within an MNC network. Therefore, the issue of which parts are to be globalized and which localized become one of the central concerns of automakers. A crucial part of this issue is the sharing of responsibility between assembler and part suppliers. Sturgeon and Florida (1999) point to the rise of the global supplier. This leads to deverticalization (by automakers) and vertical integration (among first tier suppliers) that in combination with globalization is helping to create a new global supply-base capable of supporting the activities of final assemblers on a worldwide basis (ibid, p. 115). The EU automotive industry is in the process of embracing CEE as its market and as its production location. The issue is whether EU OEM producers will succeed in making use of this opportunity. As Table 1 suggests, the key challenge is that it is not any more excellence in production that matters but excellence in network management, i.e. how to govern spatially dispersed networks of plants, affiliates, and suppliers (ibid, p. 113). Prospects of EU accession have induced EU OEMs to rapidly enter CEE markets (VW), or deepen their presence (Fiat, Renault). Also, Asian OEMs (Suzuki, Daewoo) have been attracted by prospects of accessing EU markets via CEE. The entry of EU OEMs coincided with their expansion to other emerging markets like VW s expansion to Latin America or Renault s expansion to Turkey. Additionally, EU accession has facilitated OEMs to build regional integration strategies i.e. to gradually integrate CEE, in particular central Europe, in their production network. For OEMs this was strategy which they developed by integrating Spain in their production network since the 1980s. Initially, CEE played role mainly as market but increasingly OEMs followed "build-where-yousell" strategy (Sturgeon and Florida, 1999). However, small size of the CEE market led 3

47 them to gradually expand production and use CEE as production location to serve all the EU market. 2.1 CEE in the world automotive industry After a drop from 1990 to 1991, CEE automotive production has been continuously rising from 1.67mn units to 2.3m (2000). The share of CEE in the world production of cars has increased from 4.7% (1991) to 5.6% (2002). In between 2000 and 2002 its share has actually declined by 0.4% points. In comparative terms, this rise has not been exceptional as production in other emerging markets, in particular Asia, has increased much more, from 6.23% (1991) to 13.7% (2000). The share of Western Europe in world production of cars has decreased from 39% (1991) to 36.0% (2002) (Table 3). In absolute terms, during the 1990s production of cars in Western Europe has been stagnating around 14m units (Table 3). The shifting balance in production between West and East Europe is merely a reflection of production shift towards emerging markets in automotive industry. Table 2 Share of production of cars in West and East Europe, West 39.2% 39.2% 34.3% 37.7% 37.2% 37.7% 37.6% 36.2% 37.2% 36.0% East 3 5.0% 4.7% 5.9% 4.8% 5.3% 5.8% 5.6% 6.0% 5.7% 5.6% Source: Based on EIU, OICA. Table 3 Production of cars in global regions, (in thousand units) Western Europe NAFTA Japan Asia (excl Japan) Eastern Europe Other Markets Total Source: EIU, OICA. However, the shift towards emerging markets has its limitations in industry characteristics, in particular in scale economies (advantages of concentration); sensitivity to labour costs, and in the share of transportation costs. As a result, the trade-sales ratio in automotive industry was 42% in 2000 which is significantly behind consumer electronics (118%) and 3 Data does not include Russia and CIS 4

48 apparel (77%) but ahead of steel industry (33%)(McKinsey, 2003, ch3, p. 19). This is further compounded by legal and regulatory restrictions (trade barriers), and by organizational characteristics of the industry (firm strategies and union contracts). In addition, economics of the automotive industry suggest that different auto parts have very different relocation sensitivities 4. Nevertheless, physical proximity to Western Europe and liberalization of trade ahead of accession worked strongly in favour for CEE as a low cost source automotive production location. 2.2 Value creation potential of CEE for auto MNC The integration of CEE into global MNC networks may generate significant new value for EU automotive producers. McKinsey (2003, ch.3) points out that in the auto sector over USD 150 billion in cost savings and at least another USD 170 billion of revenue could result if the barriers to industry restructuring could be overcome. Together these two opportunities represent roughly 27% of the USD 1.2 trillion industry (p. 29). If we assume that the share of CEE industry in the world automotive industry is 6% this would generate a value creation potential for CEE of around USD 9 billion in cost savings and USD 10 billion in liberalization effects. An important policy and management issue is how that potential could be realized. Its realization is essentially a micro issue which requires restructuring of the value chains of leading automotive suppliers that are located in CEE, and integration and growth of local car parts suppliers. The McKinsey Global Institute has developed a taxonomy of types of global industrial restructuring which takes into account the new role of emerging markets in generating value creation in globalized industries (McKinsey, 2003). There are five types of global industrial restructuring: 1. Market entry: Companies enter new countries in order to expand consumer base using a very similar production model in the foreign country to the one they operate at home; This mode of restructuring has been present in CEE during the 1990s only through tariff jumping FDI like Opel s investment in Poland. However, liberalization of trade tariffs with 4 According to McKinsey (2003, ch. 3, p. 33) relocation sensitivity is the lowest for wiring harnesses, car radios, followed by radiators and major body stamping which have the highest relocation sensitivity. Criteria which are used in assessing relocation sensitivity are: bulk/value; ease of meeting quality standards obsolescence time, damage sensitivity, demand volatility and sunk costs 5

49 the EU has removed a source of profits for assemblers which gradually abandoned this strategy. 2. Product specialization: the entire production process of a product (components to final assembly) is located in a single location or region, with different regions specialized in different products and trading finished goods; Audi s investment in engine production in Győr (Hungary) from which it supplies the whole EU market or Fiat s production of cars in Poland for the entire European market are examples of this type of restructuring. 3. Value chain disaggregation: different components of one product (e.g. car engine brakes) are manufactured in different locations/regions and are assembled into the final product; 4. Value chain reengineering: after moving value chain steps to a new location, processes can be redesigned to capture further efficiencies/cost savings (e.g. capital/labour trade offs); Value chain disaggregation and reengineering are currently the most common in the CEE automotive industry as we indicate below. 5. New market creation: by capturing the full value of global activities firms can offer new products at significantly lower price and penetrate new market segments/geographies. This type of restructuring has the greatest value creation potential and has been fully implemented in the case of Škoda VW. Also, a forthcoming Logan model of Renualt Dacia is designed explicitly to generate such potential. The existence of all types of global industry restructuring in the CEE automotive industry shows that the full value creation potential of CEE has not been yet realized. A current situation could be seen as the outcome of foresight and strategic commitment of automotive assemblers coupled with the quality of business environment in CEE, especially quality of labour and local car parts suppliers. 3 Restructuring of the automotive industry in CEE: key features Initially, CEE has been seen by foreign investors as dominantly a market location and accordingly FDI has been perceived as mainly market seeking. Indeed, judging based on the market shares of automotive investors in CEE this has been the dominant but declining feature of the automotive industry. The liberalization of CEE markets, EU accession and discovery of the cost and skill advantages of CEE have developed the region into both market and production locations. 6

50 3.1 CEE as an automotive market Optimism regarding CEE as markets for cars has been based on initially relatively small and outdated car park. Producers have regarded region as a market for relatively cheap cars and at the lower end in terms of model ranges and component supplies (see Sadler et al. 1993, Sadler and Swain 1994, Swain 1998, van Tulder and Ruigrok 1998). Also, population growth in CEE is stagnant which does not bode well for car sales projections. So, the key driver of growth is purchasing power. On positive side, CEE has higher share of middle class than other regions of similar level of development. Figure 1 relates the initial level of car per 1000 pop in 1989 (axis x) with relative increases in period between (axis y). It shows that there is not tendency of convergence in car density across countries of different income level based on the initial density levels. Countries that initially had higher levels of cars density have further increased their car density and thus widening intra-regional differences. Figure 1 Relationship between initial level of cars per 1000 pop in 1989 and index of increase 1998/ Hun Increase in number of cars per 100 pop, 1998/ Alban Rom Blg Yug Pol CzRSk Number of cars per 1000 population, 1989 Source: Based on Pemberton and Puckering Data in absolute figures shows somewhat different picture. Figure 2 shows trends in registered cars in CEECs during the 1990s which suggest that only Poland recorded growth rates which could be considered as dynamic. There has been very little movement in some markets such as Hungary, Slovakia and Slovenia due to the relatively high number of cars per capita. 7

51 Figure 2 Registered cars in Central and Eastern Europe, ,000,000 Units 10,000,000 8,000,000 6,000,000 4,000,000 2,000, Source: National statistical yearbooks. Figure Car sales in Bulgaria Bosnia and Herzegovina Croatia Czech Republic Estonia Hungary Latvia Lithuania Macedonia Poland Romania Slovakia Slovenia Yugoslavia Ukraine 700, ,000 Number of units 500, , , ,000 Bulgaria Czech Republic Hungary Poland Romania Slovenia Slovakia 100, Source: National automotive agencies 8

52 Figure 3 further confirms the crucial importance of Poland, both in terms of market size as well as in terms of growth of car sales. In between , the average rate of car sales in Poland was 17.8% annually. Other markets have recorded significantly lower one digit rates which reflected lower rates of their economic recovery during the 1990s when compared to Poland. However, the decline in rates of GDP in Poland in early 2000 have led to declines in car sales. Penetration of cars in CEECs broadly reflects their income levels (see Figure 4). This suggests that expansion of market seeking FDI has its limits in still low income levels of these economies 5. Figure 4 Relationship between GDP per capita and number of cars per 1000 population, LU 600 Number of cars per 1,000 population LT EE BGHR LV PL SK HU YU RO PT SI CZ EL ES IT DE FR AT SE BE UK FI NL DK IE ,000 20,000 30,000 40,000 50,000 60,000 GDP per capita ($ PPP at current prices) Source: UNECE, 2003, The Statistical Yearbook of the Economic Commission for Europe Lower income levels of CEE means that their markets are very price sensitive and with limited purchasing power. Expenditure per household in CEE is between 30-70% of the EU15 average. A low expenditure per household is a strong regional feature as variations in this respect are much smaller than variations within the EU15 (see Figure 5). However, prices of cars are only 7 percentage points lower than in the EU15. So, limited purchasing power is a constraint for market seeking FDI in CEE. This has forced all assemblers to 5 There are some interesting country differences (Slovenia vs Greece) which would require further research to be explained. 9

53 increasingly orientate themselves towards exports which raises the issue of competitiveness of CEE in the global automotive industry. Figure 5 Relationship between household consumption and price level of passenger cars, EU15 = 100, DK Price of passenger cars RO BG TK CZ HU LV SK EE PL LT PT ES SI EL NL UK IT FR AT BE DE LU IE SE FI Houshold consumption Source: Eurostat, Statistics in Focus, Relative prices for new passenger cars in EU, EFTA, Acceding and Candidate Countries for 2002, Theme 2, 65/ CEE as an automotive producer Socialist heritage in automotive industry Before 1989, the CEE automotive industry developed along three lines (Richet and Bourassa, 2000; Pavlínek, 2002): Manufacturing based on indigenous development, technology, and car making traditions of the pre-wwii period (Czechoslovakia, GDR) Production developed using Western licences (Poland, Romania, Yugoslavia) Components production without automobile assembly operations (Hungary, Bulgaria, Albania) The opening of CEE has not led to radically new trends in the automotive industry but it did reinforce market position of already developed capacities. FDI projects since 1989 have reinforced the three former production patterns that were observable under socialist times: the Czech Republic (and East Germany) are specializing in automotive production, with 10

54 Poland, Romania and Slovenia increasing their production capacities based in part on the former links and cooperation licenses with western manufacturers, while components manufacturing has remained the core activity in the Hungarian automotive industry. In the long-term perspective growth of car sales during the 1990s are comparable to rates of sales during the period for Poland and are above rates of this period for other countries, except Bulgaria (Figure 6). However, during the 1980s all CEECs, except Yugoslavia and Romania, had either negative (Bulgaria, GDR, Poland) or very low rates (Czechoslovakia). The deterioration in terms of car production during the 1980s led to large pent up demand in early 1990s, despite strong declines in household incomes. Recovery in demand has been accompanied by increasing variety of models and price levels. Figure 6 Socialist car production in CEECs, , , ,000 Bulgaria No. of units 250, , ,000 Czechoslovakia GDR Hungary Poland Romania 100,000 Yugoslavia 50, Source: National statistical yearbooks All of the major indigenous automotive manufactures were taken over by foreign manufacturers, through privatization, as the large state owned enterprises were broken up and sold to foreign investors (Table 4). However, creative destruction was not without victims as some companies were liquidated (such as Wartburg, Trabant, FSO) or are in process of prolonged degradation (Crvena Zastava). 11

55 Table 4 The fate of major indigenous socialist manufacturers after 1989 Company Location Fate VEB Automobilwerk Eisenach (Wartburg) VEB Sachsenring Automobilwerke Zwickau (Trabant) Autombilové Závody Národný Podnik (AZNP) Škoda Tatra, národní podnik Fabryka Samochodów Małolitrażowych (Small Car Factory) FSM Fabryka Samochodów Osobowych (Passenger Car Factory) FSO Uzina de Autoturisme Piteşti (Piteşti Car Factory) Dacia Oltcit Industrija Motornih Vozil (IMV) Zavodi Crvena Zastava (Red Star Factory) ZCZ East Germany (Eisenach) East Germany (Zwickau) Czechoslovakia (Mladá Boleslav) Czechoslovakia (Kopřivnice) Poland (Bielsko Biała) Poland (Warsaw) Romania (Piteşti) Romania (Craiova) Yugoslavia (Novo Mesto) Yugoslavia (Kragujevac) Joint venture with GM Opel Opel-AWE Personenwagen GmbH formed on 26 March Plant closed on 10 April Plant facilities reopened as Opel-Werk on 23 September 1992 Closed down on 30 April Plant facilities taken over by Volkswagen AG as subsidiary Volkswagen Sachsen GmbH Taken over by Volkswagen AG. Plant called Škoda Auto a.s. Factory at Pribor closed and car production moved to Kopřivnice in Car production stopped in same year Taken over by Fiat S.p.A. Plant called Fiat Auto Poland S.A. Company liquidated on 18 January Agreement signed between Daewoo and FSO in November Taken over by Daewoo. Plant called Daewoo-FSO Motor S.A. Taken over by Renault S.A. Plant called Automobile Dacia S.A. Citroën withdrew from Oltcit SA joint venture in Government decision No. 499/1991 to form S.C Automobile Craiova SA. Taken over by Daewoo. Plant called Daewoo Automobile România S.A. Subsidiary Revoz transformed into public company Revoz d.d. in June Taken over by Renault S.A. Taken over by Nucarco on 4 October Joint venture called Zastava Motor Works Source: Authors The role of the automotive industry in CEE manufacturing Judging based on shares in main economic indicators, the automotive industry plays a very important role in manufacturing industry in three central European economies (the Czech Republic, Slovakia and Hungary), a moderate role in Slovenia, Poland and Romania, and a marginal role in Baltic economies (see Figure 7). In the leading group, the automotive industry employs in between 3.7% to 6.4% of manufacturing labour force but generates from 8-10% of value added. In terms of turnover, the automotive industry has a very similar share in the leading group of 12-13%. By the end of 1990s, investment in this industry in Hungary and Slovakia were 23% and 15% respectively. 12

56 Figure 7 Share of automotive industry in manufacturing, 2001 (Czech Republic, 2000) 25% % 20% 15% 10% Turnover Value added at factor cost Personnel costs Gross fixed investment 5% Employees 0% CZ SK HU SI PL RO EE L V L T Source: Eurostat, Cronos A combination of market size factors and inherited competencies in automotive assembly coupled with strategies of foreign investors has led Poland and the Czech Republic to be the two leading production locations (Figure 8). Trends across countries can be clearly traced to investment decisions of individual investors. Figure 8 Car production in CEECs, , ,000 Number of units 500, , , ,000 Czech Republic Hungary Poland Romania Slovakia Slovenia Yugoslavia 100, Source: National automotive agencies and OICA 13

57 Fiat dominated the Polish market until Daewoo entered via the takeover of a domestic producer by the end of the 1990s. However, due to Daewoo s and Fiat s troubles Polish production numbers have been stagnant since the end of the 1990s. Although being four times smaller market in terms of population the Czech Republic has become the leading producer in the region. Recent investment in a new plant by Toyota and PSA will eventually make the Czech Republic one of the largest European automotive manufacturers. As a production and supplier location, the Czech Republic is geographically close to the major European OEMs operating in Germany and France, and has also been favoured by many producers and suppliers due to its facilities, manufacturing history, low costs and technical capabilities. Slovakia has emerged only recently as an important production location. VW investment and recent investments by PSA and Kia have expanded the production base greatly. Once these plants are up and running by 2006/7 Slovakia will have the largest per capita production globally (150 cars per 1,000 inhabitants in 2007, compared with some 90 cars in the current global leader, Belgium, in 2002) (EIU, 2004). Throughout the 1990s the Romanian automotive industry operated as non-restructured, stagnant and state owned. This coupled with a very limited local market due to low purchasing power led to declining volumes. However, in early 2000 Renault/Nissan started a turnaround of Dacia and have launched entirely new model of car for emerging markets with large share of local content. If this project succeeds it will have important spillover effects on industry restructuring in Romania and may launch Romania as a new global low cost production location. Also, Continental and Michelin have made large investment in Romania to produce tyres. With a history of car and component manufacture stretching back to the 1950 s, Slovenia continues to play a leading role in CEE production. Its only producer, Revoz, is part of the Renault group. Baltic countries play a very small role within CEEC car production, in part due to the small local market size as well as lack of endogenous producers. Interestingly, Baltic countries are not used as production bases for neighbouring Russia. However, component manufacturers are setting up bases within Estonia, Latvia ad Lithuania Following the break-up of Yugoslavia (Serbia and Montenegro) in 1991, its largest manufacturer, Zastava, was faced with large scale restructuring. The following war with Croatia saw production in Yugoslavia nearly wiped out. The Zastava plant was heavily bombed by NATO in 1999, destroying the production line and forcing it to completely close 14

58 its operations. Reopened a year later, production was resumed, but at a fraction of its potential. Croatia does not have any car producers, but concentrates on component manufacture. There is very little domestic production of automotive parts, and almost no domestic production of car accessories. Local companies such as Cimos, AD Plastik, Elcon and Prevent produce a small number of parts for the French manufacturers Citroen, Peugeot and Renault, which entitles these OEMs to tax benefits. VW has a share of 58% since 1998 in its Bosnian operations, following a long running assembly joint venture with a UNIS-TAS which dates back to Having suffered from the Bosnian war, it reopened again with full production in 2000, after the war damage had been repaired. Essentially a CKD plant, with a possible capacity of 5,000, this remains the only automotive and component company in the country. VW will remain minor operation in the foreseeable future Bulgaria, Macedonia, Albania are peripheral to the CEE automotive industry. In Bulgaria, there has been some history of assembly of Russian cars (VAZ s and Moskvitch s), but only at the 15,000-20,000 unit level. Today there is marginal production, although component manufacture has been growing steadily since In Albania, under socialism there was no car or component manufacture as car sales and production were banned, which is reflected in today s lack of any production. All cars are imported. Turkey is production source for both Middle East and Europe. For Renault, Turkey is important location as it continues to benefit from EU custom union for global sourcing of its models. Fiat too has major production facilities to produce its global 178 car, under the local brand Tofas. Development of Russian automotive industry will have important effects on prospects for growth and restructuring in CEE. For 10 years, there have been numerous attempts by foreign investors to set foot on the Russian market but still with meagre success.however, some joint ventures are gradually progressing driven greatly by protected local market and high growth of domestic demand. Russian domestic producers benefited from the 1998 financial crisis as imports have become more expensive which boosted their profitability though the technological lag remains considerable. Avtovaz and Gaz are two biggest domestic producers. GM, Fiat, Renault and several other MNCs have joint venture agreements with Russian assemblers which are subject to delays and doubts. Nevertheless, with the return of growth and increased purchasing power automotive companies will gradually increase their investments in Russia which may have profound effects on the CEE automotive industry in terms of competition as well as cooperation. 15

59 Ukraine, with a population of 50 million, has also large market potential. However, economic difficulties have led to new car sales of only 10% of neighbouring Poland.(PwC 2002). In recent years, production has risen significantly, aided by strength of the local market and based on the number of SKD assembly operations designed to avoid import duties. With economic recovery and institutional stabilization Ukraine may emerge as an important competitor and partner of the CEE automotive industry despite its poor records of FDI, so far. This brief overview of the CEE as markets and production location shows that production capacities in automotive industry are currently highly concentrated on central Europe (the Czech Republic, Poland, Slovakia and Slovenia). New FDI in the Czech Republic and Slovakia will further increase their role in regional production networks. Romania is emerging as possibly an important new location. Other CEE countries operate mainly as markets though some of them are gradually becoming involved in supplier industries. Russia and Ukraine are still outside CEE production networks and attract mainly market seeking FDI. However, the return of growth to these countries may change this picture and may bring new competitive dynamics in the region Employment: OEMs and suppliers The employment structure in the automotive industry is different from value added or turnover structure (Figure 5) which points to important differences in productivity and in restructuring among CEECs. This primarily relates to a very high share of employment in this sector in Romania of (Figure 9) which has its effects in terms of productivity differences (see Figure 8). This large employment when compared to low production shows that the sector has not yet been restructured. Poland, the Czech Republic and Romania which employ together people are major employers in this sector. The Hungarian automotive sector, when compared to the Czech Republic, a country of approximate size is much smaller. Trends in employment between assemblers and parts suppliers show divergent tendencies (see Table 5). By the end of the 1990s and early 2000 assemblers have been reducing employment while suppliers have been expanding. Reduction of employment by assemblers is particularly strong in Poland which reflects problems with restructuring of Daewoo investment. Increases in Latvia reflect very low base level. High reductions in employment in Bulgaria in both subsectors reflect decline of industry driven by the absence of FDI. Nevertheless, declines are much stronger in the assembly than in the supplier sector. Figure 9 16

60 Number of employees in the automotive industry in CEE, 2001 BG, 3444 EE, 1471 LT, 313 SI, 7060 SK, LV, 671 HU, CZ, RO, PL, Source: based on Eurostat, New Cronos Table 5 Changes in employment in the CEE automotive industry 6 Number of persons employed 7, 2001/1998 Assemblers Suppliers BG 32.5% 74.5% HU 90.9% 106.3% CZ 158.8% Number of employees 8, 2001/1999 HU 109.4% 112.6% LV 252.8% 90.6% PL 56.5% 100.4% SI 97.7% 110.7% SK 116.0% Source: For 1998, EC, Business in candidate countries. Facts and figures, Theme 4, Panorama of the EU, Eurostat Higher rate of employment in supplier s sector than in assembly suggest that spillovers effects which start from assemblers are starting to make effects. As assemblers try to source more local components this induces foreign and domestic suppliers to enter or expand which creates demand for labour. It is not a coincidence that rises of employment An increase of number of employees in assembly in Hungary and their decrease in number of persons employed reflects changing working practices rather than clear employment trend. Number of persons employed is the total number of persons who work in the observation unit as well as outside working persons who belong to the unit and are paid by it. It includes all persons who are on the payroll of the enterprise, whether they are temporarily absent, part time, seasonal or home workers, apprentices, etc. The number of persons employed excludes manpower supplied to the unit by other enterprises and persons carrying out repair and maintenance work in the enquiry unit on behalf of other enterprises. The number of employees is defined as those persons who work for an employer and who have a contract of employment and receive compensation in the form of wages, salaries, fees, etc. 17

61 in suppliers sector is the strongest in the Czech Republic and Slovakia, two countries with the biggest rise in FDI in the automotive assembly. FDI have substantially increased their share in employment in car industry (see Figure 10). In Czech R, Hungary and Poland foreign firms have controlled by the end of 1990s around 70% of overall employment. Figure 10 Share of foreign affiliates in employment in car industry % / / Czech R Hungary Poland Slovenia Estonia Legend: For Czceh R, Hungary, Poland: Motor vehicles and trailers, 1993 and 1999 For Slovenia, motor vehicles and trailer, 1995 and 2000 For Estonia, Motor vehicles and transport equipment, 2000 Source: based on Radosevic et al (2003) Figure Change in employment in car industry and Poland Czech R Hungary Slovenia Estonia Total Foreign affiliates Note: see figure Source: based on Radosevic et al (2003) Table 6 18

62 Škoda Superb suppliers, 2002 Company Borg Warner Turbo System/Garrett Valeo Denso Freudenberg Bayer Bosch Hella Beru ZF Cikautxo Bosch Koyo Leoni Draexlmaier Autoliv Bosch Vitro Faurecia Autoliv SLI miniature lighting Johnson Controls Grammer Edscha Bohemia Edscha Bohemia SLI miniature lighting Visteon Eberspächer Bosch FAG Bridgestone/Firestone Hutchinson Doga ZF Behr-Hella Thermocontrol Bosch Federal-Mogul ZF Dana Fritz Winter Eisengiesserei TI Electronics Bosch Hutchinson Muhr & Bender Part Turbocharger Motors for anti-lock brakes Air conditioning compressor Radial shaft seals: Gearbox and engine Front grille Alternator Headlamps Glow plugs (diesel only) Automatic transmission Cooling hoses Valve springs Hydraulic power steering Engine harness (tier 2 supply) Harness system Chest airbags Wipers Complete electrical system Instrument panels Front seatbelt Door illumination Seating system Headrest Door hinges Door latches Boot lighting Rearlight cluster Exhaust system ABS Rear Bearings Tyres Door seals Door, roof and module reinforcements Chassis components Climate control system Navigation system Pistons Clutchset Clutch pipe Brake discs Camshaft and crankshaft sensors Engine management Heater/radiator hose Stabiliser bars Source: Automotive News Europe From employment perspective it is important that FDI have either preserved (Poland and Estonia) or created (Czech R, Hungary, Slovenia) new employment (see Figure 11). In 19

63 Poland, FDI have increased share in employment by while the overall employment went down by However, in Czech R, Hungary and Slovenia employment have either been preserved or increased not only in FDI firm but also in domestically controlled firms. This suggest that new jobs have been created among suppliers as indicated in Table 5. However, data like these may be capturing only a part of the overall employment effects of automotive assemblers. As the automobile produced have become more complex, it has resulted in a greater share of supplies coming not just from the automotive component sector (NACE 3430), but also from other sectors such as the rubber and plastic, paint, glass, metal producing, textile, and electrical sectors. Indeed, the automotive component sector is too narrow in terms of coverage, which is recognized by the NACE 3430 classification as it excludes engine and tyre manufacture, most electrical and electronic components (classified as NACE 3161), as well as glass, plastic or certain castings and other metal parts (Havas A, 2000b). This would seem to confirm the knock on effect the industry has in terms of employment. Table 6 shows the range of suppliers of new Skoda Superb model. Although we do not know which of these suppliers have established local production and which are only importers we can indirectly conclude that the employment effects are far from confined on narrowly defined automotive sector Productivity the automotive industry is the driver of industry restructuring and productivity changes in this sector have cumulative effect across the whole manufacturing industry. Figure 12 shows that in countries where FDI has entered on large scale in the automotive industry (Slovakia, Hungary, the Czech Republic, Poland) productivity of this sector is significantly above the manufacturing average. In countries where FDI is not present in the automotive industry (Bulgaria, Latvia, Lithuania) productivity in this sector is below the manufacturing average. It is striking to see that labour productivity in Slovakia and Hungary in the automotive industry is 2 times above the manufacturing average. In the Czech Republic it is 1.6 times above average 9. The Slovenian automotive sector is also more productive when compared to the average of manufacturing though the overall level of productivity in Slovenia is the highest among the CEECs. Although the share of FDI in the Polish automotive industry is high its productivity level is still relatively lower than would be expected, especially when compared to Slovakia and Hungary. Again, this reflects problems in the restructuring of Daewoo investments in Poland. In Romania, productivity in the automotive sector is somewhat below the manufacturing industry which reflects a largely restructured sector. This has been changing with Renault s investment in Dacia. Differences between countries in terms of labour costs are much smaller than differences in productivity which has strong effects on competitiveness among countries as reflected in 9 Estonian automotive sector is very small and data may reflect few very productive foreign firms. 20

64 relative unit labour costs. For example, in Hungary and Slovakia where productivity is 2 times above the manufacturing level relative labour costs are only 1.5 above average. This convergence in labour costs has strong negative effects on competitiveness of the Romanian automotive sector. Figure 12 Productivity, unit personnel costs and unit labour costs in manufacturing of motor vehicles compared to manufacturing (= 100), % % of ma nu fac 200% tu re Apparent labour productivity* 150% 100% 50% SK HU EE CZ PL SI RO BG L V L T Unit personnel costs** Unit labour costs*** 0% *Value added at factor cost/number of employees **Personnel costs/number of employees *** Unit personnel costs/unit labour costs Source: based on Eurostat, New Cronos Productivity is significantly higher in the assembly sector when compared to parts suppliers. For example, in Hungary assembly is 3 times more productive, and in Poland 1.8 times (see Figure 13). Again, differences in wages are much less significant which points to big differences in competitiveness in favour of assemblers. This may reflect the capital intensive nature of assembly as compared to parts suppliers as well as still nonrestructured supply chains and absence of high quality domestic suppliers. In Slovenia, differences in productivity between assemblers and suppliers are much less pronounced which leads to similar competitiveness of both segments of the automotive industry. This further reinforces the point of higher competitiveness of Slovenian domestic automotive firms when compared to the rest of manufacturing. Figure 13 21

65 Productivity, unit personnel costs and unit labour costs of assemblers, car parts suppliers = 100, 2001 In % of car parts supp liers 400% 350% 300% 250% 200% 150% *Value added at factor cost/number of employees **Personnel costs/number of employees 100% 50% 0% Hungary Poland Slovenia Unit labour costs*** *Value added at factor cost/number of employees **Personnel costs/number of employees ***Unit personnel costs/labour productivity Source: based on Eurostat, New Cronos Although lower productivity of car parts suppliers could be attributed to higher capital intensity of this sector, case study evidence suggest that it may be also partly due to weak competencies of local suppliers. For example, Sperling (2004) cites the opinions of automotive executives who hold that the CEE supplier industry is not yet able to deliver to world standards. We also explore this issue through the local content data (see below). Data on productivity of individual manufacturers expressed as number of cars per employee are very imperfect measures as they contain differences in vertical integration and nature of operations. Nevertheless, if available over time they are useful proxies for understanding trends in productivity. A compilation of data for several OEM producers in CEE shows (a) high and rising productivity for fully fledged producers (Fiat, Revoz, Suzuki) and (b) low and stagnant productivity for screwdriver type assembly operations (Opel, Audi) (Figure 14). A part of the differences in productivity is due to differences in capacity utilization. Aggregate data suggest that CEE like other region is experiencing excess capacity. A shift to global strategies which have now embraced all major emerging markets has led to overextension i.e. to similar investments at the same time. Although in CEE the overcapacity is localized on few producers it still is overexpansion. Overcapacity has been exacerbated by 22

66 slowdown in growth in Poland, by far the largest Central European economy. Worldwide industry utilization is estimated to be 70-75% (McKinsey, 2003, ch2). Our data for sample of CEE producers shows the simple average capacity utilization of 62% and weighted capacity utilization of 77.5% in 2001 which is well within the world average (Table 7). Figure 14 Manufacturer productivity, Cars per employee Source: compiled from company annual reports Table Estimated production plant capacity in CEECs, 2000 Manufacturer Plant Country Production Estimated capacity Škoda Auto A.S. Audi Hungária Motor Kft. Magyar Suzuki Rt. Daewoo-FSO Motor S.A. Fiat Auto Poland S.A. Opel Polska S.p. z.o.o Volkswagen Poznań Sp. z.o.o. Automobile Dacia S.A. Daewoo Automobile România S.A. Volkswagen Slovakia A.S. Revoz d.d. Zastava Automobila A.D. Daewoo Lublin/Nysa/Zeran Poland 85, , Rodae Romania 14, , Fiat Bielsko-Biala Poland 56, , Tychy Poland 236, , Ford Plonsk Poland 4,525 20, GM/Opel Gliwice Poland 97, , Warsaw Poland 1,800 7, Renault Novo Mesto Slovenia 122, , Pitesti Romania 55, , Suzuki Esztergom Hungary 77,250 70, Volkswagen Bratislava Slovakia 153, , Gyor Hungary 56,776 55, Kvasiny Czech Republic 22,705 40, Mlada Boleslav Czech Republic 320, , Poznan Poland 47,582 80, Vrchlabi Czech Republic 63,591 80, Source: Automotive World, World Automotive Manufacturing, November 2001 Capacity Utilization (%) 23

67 3.2.5 Trade Trade in motor vehicles and parts of CEE with the EU15 makes an important share of the overall trade of the CEE. This trade has risen from EUR6.8 billion (1993) to EUR24 billion (1999) (see Table 8). However, this 3.5 times increase is highly skewed towards Germany whose trade with the CEE has increased 4.2 times. This had led to an increase in the German share of EU15-CEE automotive trade from 43% to 52%. Shares of Italy and France have declined though in absolute terms they have increased by 2.6 and 2.3 times respectively. Table 8 Trade in road vehicles and parts (imports + exports) EU-15 candidate countries, EUR million 1993 Shares 1999 Shares Index EU % % 3.52 Germany % % 4.23 Italy % % 2.60 France % % 2.31 Spain % % 6.90 UK % % 3.17 Austria % % 4.81 Belgium % % 2.73 Netherlands % % 3.21 Sweden % % 3.09 Finland % % 3.04 Denmark % % 2.79 Greece % % 2.23 Portugal 9 0.1% % 8.44 Ireland 2 0.0% % Source: based on Eurostat, the automotive industry and candidate countries, Statistics in Focus, Theme 6-1/2001. Automotive trade is good example of very developed intra-industry trade. Germany dominates in both export and import (see Figure 12). However, there are noticeable differences among CEECs in terms of orientation on three major EU15 destinations (Germany, France and Italy) (see Figure 15). These differences reflect origins of large MNC investors. Exports and imports from the Czech Republic, and Hungary are very much oriented towards Germany due to the VW Group (Table 9). The Polish automotive sector, which has Fiat as one of the top investors, is oriented in export towards Germany and Italy but less so in import from Italy. This suggests that Fiat is purchasing car parts from other EU countries and install them into final products in Poland. Slovakia has very strong orientation 24

68 towards Germany (92%) in import but much less so in export where it is oriented also towards Italy. Slovenia and Turkey have balanced orientation towards three major automotive producers which reflects limited presence of assemblers from Germany and strong orientation of these countries as car parts suppliers towards all three major destinations. Figure 15 EU-15 exports and imports to/from CEE and Turkey, 1999, in EUR million Italy France Germany CZ-X CZ-M HU-X HU-M PL-X PL-M SK-X SK-M SI-X SI-M TK-X TK-M Source: Based on Eurostat, the automotive industry and candidate countries, Statistics in Focus, Theme 6-1/2001 Table 9 Shares in automotive exports and imports to/from CEE and Turkey of three major partner countries (= 100%), 1999 Export Import CZ HU PL SK SI TK CZ HU PL SK SI TK Germany 79% 89% 53% 76% 47% 36% 77% 88% 57% 92% 36% 55% France 13% 1% 5% 0% 28% 41% 14% 6% 19% 6% 46% 30% Italy 8% 10% 42% 23% 24% 23% 9% 6% 24% 3% 19% 15% Source: ibid In summary, trade data strongly reflects FDI and strategies of respective MNCs. They also show that automotive producers have departed from exclusively market entry restructuring and moved towards product specialization and value chain disaggregation. 25

69 3.2.6 Foreign Direct Investments (FDI) Trade and industry data show that the CEE automotive industry is driven entirely by FDI. It is one of the most FDI intensive sectors in CEE. In central Europe, share of automotive industry FDI stock in the overall manufacturing FDI in 2000 was between 10-15% (see Table 10) 10. However, FDI is very much concentrated on a few production locations. Their strong concentration in central Europe strongly contrast with their absence in Bulgaria, south east Europe and the Baltics. The heavy concentration of FDI on central Europe (Poland, Hungary, the Czech Republic, Slovakia and Slovenia) and its absence in other CEE countries has reached such proportions that it will remain a strong sectoral feature for the foreseeable future. This already has important effects in terms generating spillover effects on industries which are linked to automotive industry and on manufacturing in general. Table 10 FDI stock and share of automotive industry, 2000 FDI Stocks, 2000 in USD Share of motor vehicles Country Total Manufacturing Motor vehicles in total in manufacturing Poland % 15.7% Czech R % 16.6% Hungary* % 9.7% Slovenia % 11.5% Lithuania** % 6.8% Bulgaria % 0.4% Latvia* % 0.1% * includes other transport equipment **2001 Source: UNCTAD, World Investment Directory, accessed on April 15, Top automotive investors are most often the major investors in terms of sales (Table 11). Assemblers are much bigger investors when compared to car parts suppliers which reflect industry characteristics rather than strategies of MNCs. Five major MNCs dominate the CEE automotive industry. These are VW, Renault, GM, Fiat and Daewoo. They are all concentrated in Central Europe and mainly reflect past inherited capacities from the socialist period. It is important to recognize the strong presence of MNCs in Russia but which has not yet resulted in any significant investments. However, Russia is one of the most important emerging markets in the automotive industry and we may expect major FDI in future in this sector. This may have significant effects on the situation and pattern of networking in the automotive industry in Central Europe. 10 Table 9 does not fully capture presence of FDI in Hungary which is mainly in car parts industry. 26

70 Table 11 Top automotive MNCs in CEE, Ranking Country ranking Company Type Host economy Home economy Sales ($mn) 1 1 Skoda Automobilova Assembly Czech R Germany Audi Hungaria Motors Engines Hungary Germany Fiat Auto Poland Assembly Poland Italy Centrum Daewoo Assembly Poland Korea Revoz Assembly Slovenia France Volskwagen Poznan Assembly Poland Germany Opel Magyarorszag Jamugyarto Assembly Hungary United States Volkswagen Slovakia Assembly Slovakia Germany Suzuki Assembly Hungary Japan General Motors Poland Assembly Poland United States Automobile Dacia Assembly Romania France Daewoo Automobile Romania Assembly Romania Korea Renault Polska Assembly Poland France Opel South East Europe Assembly Hungary United States Skoda Auto Slovensko Assembly Slovakia Germany Opel Polska Assembly Poland United States NABI Car parts Hungary United States Ford Poland Assembly Poland United States Skoda Auto Poland Assembly Poland Germany Siemens Automobilovi Technika Car parts Czech R Germany Visteon Hungary Car parts Hungary United States Autopal Car parts Czech R United States Daewoo Motor Polska Assembly Poland Korea Lear Automotive Car parts Hungary Austria Porsche Slovenia Car parts Slovenia Austria Robert Bosch Car parts Czech R Germany Volvo Poland Assembly Poland Sweden Daewoo Avia Car parts Czech R Korea Porsche Slovakia Car parts Slovakia Austria Debica Car parts Poland United States BP Cycle Car parts Serbia/Monten Germany Daewoo Motor Slovakia Assembly Slovakia Korea Norma Car parts Estonia Sweden/US Sachs Slovakia Car parts Slovakia United Kingdom Scania Eesti Car parts Estonia Sweden Ilta Kiev?? Ukraine Switzerland Cimos Buzet Car Parts Croatia Slovenia 36 Showa Aluminium Car parts Czech R Japan --- AvtoVAZ Assembly Russia United States Avtoframos Assembly Russia France BMW Avtotor Kaliningrad Assembly Russia Germany Daewoo Tagenrog Assembly Russia Korea Ford Vsevolozhsk Assembly Russia United States Nizeghorod Motors Assembly Russia Italy ZIL AMO Assembly Russia Sweden AvtoZAZ-Daewooo Assembly Ukraine Korea/Russia MAZ - MAN? Belarussia Germany/Russia Source: UNCTAD, World Investment Directory, accessed on April 15,

71 Most of the big FDI Investors in the industry are from Germany whose share in FDI is similar to the share in EU-CEE trade in the automotive industry (see Table 12). This suggests that trade in the CEE automotive industry is FDI led rather than being driver of FDI. Opel, GM and Ford have integrated CEE plants into their European but not global operations. Table 12 Sales of top MNCs in the CEE automotive industry by home country, 1999/2000 (USD million) Germany United States Italy France Korea Japan Austria Sweden/US Switzerland 4.8 Source: Based on UNCTAD, World Investment Directory, accessed on April 15, What are the direct effects of FDI on the CEE automotive industry? FDI is the key driver of growth and restructuring in this sector. In that respect, FDI has led to extensive investments through upgrading equipment and reorganization of the production process. This has increased the capital intensity of assembly, improved management practices and has initiated the process of building a local supply base (see below). However, the most important mechanism for ensuring positive spillovers in local economies is investments in human capital, i.e investments in vocational training. By building local know how and by ensuring mobility of the labour force some of these investments may be reemployed in other firms or activities. Although we do not have systematic evidence on this aspect of FDI case study evidence points that these are substantial investments. A good example in this respect is Renault investments in training of labour in its Dacia Pitesti plant (Romania) ahead of production of the new car for emerging markets the Logan. A total of over a million hours of training have been given in manufacturing and support functions in production, management, IT, and so on. More than 450 Dacia employees received training abroad at courses lasting several weeks 11. Our estimates of Renault figures suggest that the overall training is equivalent of 40% of total number of workers of plant receiving full time training for half year. This figure roughly correspond to a survey of foreign managers carried out for the EBRD which estimate that foreign investment enterprises 11 Dacia group Renault. Logan: conquering new world markets and boosting the Renault group s profitable growth, Press Release, June 2,

72 would need around 6 months of training to achieve a level of productivity comparable to Western European workers 12. OEM producers have invested in vocational training as a part of the overall investment package. However, it is not certain whether they will want to sustain costs in training of local suppliers which is the key bottleneck for generating skills related spillovers. In that respect, there is room for coordinated action by local industry associations, government, training providers and investors. 4 Micro view: OEMs 13 in CEE key drivers of restructuring? The dynamics of industry restructuring in a highly oligopolistic sector such as the automotive industry can be better understood by analysing micro dynamics of competition. As argued by Porter at al (WEF, 2004) and McKinsey (2003) sources of productivity are at micro level. In this section we: (a) analyse market and production trends in the automotive industry at individual company level, and (b) analyse what has been the extent of local sourcing and what is the scope for building a local supply base in CEE. 4.1 Market and production dynamics at firm level Figure 16a shows that European automotive companies basically dominate the CEE market. Within the Big Four (VW, Renault, PSA, Fiat) VW is by far the largest player. This may be partly attributable to the higher import tariffs that have to be paid for third countries imports. As one would expect, there is the greatest share of sales by companies that are producing in those respective countries (e.g. VW group has the largest share in the Czech Republic and Slovakia, Suzuki in Hungary, Renault in Romania). In the countries with no production plants by the major manufacturers (e.g. Bulgaria, Croatia and Macedonia) there is a more even market distribution. Also, in Poland which is the biggest CEE market and with the biggest number of manufacturers, market is relatively evenly spread (Figure 16b). This is greatly due to weaknesses of Fiat and Daewoo. VW, which has entered later than the other manufacturers, is gaining market share helped in part by its strong presence in neighbouring countries. The monopoly which existed previously under socialism, restricted choice, and took some time to undo. Benefiting from unleashed competition consumers are now able to select from many different suppliers and models, including increased imports. The inefficient producers have been driven to the margins of the market. Individual national markets are significantly less dominated by national producers (e.g. Škoda in the Czech Republic and Slovakia, Daewoo-FSO in Poland, etc). This has led to increased market fragmentation. There is not market homogenization in the region in terms of vehicle type either. For See EBRD Transition Report 2000 Original equipment manufacturers 29

73 example, in Hungary, Poland, Slovenia and the Baltics, saloon cars are popular. However, in the Czech Republic, Slovakia and Poland, hatchbacks are more popular. Figure 16a Market shares of key players (market share >4%) in the CEE markets Slovenia Slovakia Romania Poland Macedonia Hungary Czech Republic Croatia Bulgaria 0 VW Renault PSA Fiat GM GM Daew oo Ford Suzuki Toyota Source: Based on Automotive World Automotive Quarterly Review Q Figure 16b Market shares of key players in the CEE markets 100% 90% Other 80% VW Toyota 70% Suzuki Renault Percentage 60% 50% 40% 30% 20% PSA Proton Porsche Nissan Mitsubishi MG Rover Mazda Hyundai Honda 10% GM Daewoo GM 0% Fuji HI Ford Bulgaria Croatia Czech Republic Hungary Macedonia Poland Romania Slovakia Slovenia Fiat DC BMW Source: Based on Automotive World Automotive Quarterly Review Q

74 In countries with a lower purchasing power, manufacturers that sell cheaper models seem to perform better, i.e. have higher shares. Simultaneously, upmarket companies such as BMW, Mercedes-Benz (part of DaimlerChrysler) and Porsche have also built a visible presence serving growing affluent groups. Differences in market presence highly correlate with production trends of individual OEM firms (Figure 17). The expanding presence of VW, and decline in production figures of Fiat and Daewoo FSO dominate production trends. So far, other players are significantly smaller though we may expect that expanding activities of Dacia Renault and new investments of Hundai, and PSA/Toyota may change this picture in the medium term. Figure 17 Production of CEE OEMs, Units 500, , , , , , , , ,000 50,000 Source: Company statistics Škoda Auto A.S. Audi Hungária Motor Kft. Magyar Suzuki Rt. Daewoo-FSO Motor S.A. Fiat Auto Poland S.A. Opel Polska S.p. z.o.o Volkswagen Poznań Sp. z.o.o. Automobile Dacia S.A. Daewoo Automobile România S.A. Volkswagen Slovakia A.S. Revoz d.d. Zastava Automobila A.D. It has been estimated that the minimum economies of scale in automotive assembly is 250,000 vehicles per year (McKinsey, 2003, ch2, p. 5). By using that criterion only 2 CEE producers are over this threshold Škoda and VW Slovakia. Fiat Poland was producing well over this threshold until problems in the company forced a cutback in production to under 250,000 after Even in terms of capacity, it is only these three plants that are capable of producing above this level. However, the development of common platforms and sourcing has ensured that a bigger number of plants are efficient than would be suggested based on volume criterion only. One of the important factors of Daewoo s difficulties is that it has been unable to share common platforms and sourcing as it had to cut ties with parent company. 31

75 Competitive dynamics have led to growth of some producers and relative decline of others. Škoda is the most successful of all the CEE OEMs. Daewoo-FSO has faced problems since 1999 as a result of financial problems with its Korean parent company. Similarly, Fiat faced problems with its parent company in Italy and was forced to restructure and cut down its production. A latest shift in competitive dynamics have come from the non-european companies - Toyota and Hyundai who have associated with the two main French manufacturers, Renault and PSA Peugeot Citroën, and will build greenfield plants in the Czech Republic and Slovakia. 4.2 OEMs and building of a local supply base OEMs and first tier suppliers are key agents of building a local supply base in CEE through their linkages with local suppliers. Hence, from an industrial restructuring perspective it is important to understand what are drivers and obstacles in spreading the local supply base. Systematic evidence on local suppliers issues is not widely available. In continuation, we will try to develop a picture of these issues based on evidence at firm level. Practically all major component firms have established their subsidiaries in CEE by taking over local companies or more commonly through greenfield investments. This is usually because they are requested to follow their clients in their strategic moves. Equally, assemblers are dependent on quality suppliers as costs of components and modules make great share in total costs. First tier suppliers are being given greater global roles and have been transforming their relationship with second tier suppliers. Individual evidence in CEE suggests that there are problems with local suppliers in terms of quality and ability to develop more complex sub-assemblies. High tech and high valueadded components originate from western firms. Indigenous component suppliers are seen of lower technological ability and quality (Pavlínek, 2002). However, there are also successful cases of increased local content and supplier development. A good example in this respect is Magyar Suzuki who due to small scale follows a single source strategy. Thus it has strong incentives to nurture a local supply base. As Havas (2000) reports Suzuki together with its Japanese suppliers conducts a thorough technological and financial audit, covering literally every single aspect of doing business from purchasing inputs through production methods and machinery, to accounting, sales and management, broadly defined. Then joint efforts are made to improve the selected supplier s technical level and economic performance, when needed. Table 13 from Havas (1997) shows the rise of local content to the level which was needed for meeting EU rules of origin criteria (60%). This protective measure was introduced to circumvent use of the CEEC as a backdoor to access the western market with cheaper products. 32

76 Table 13 Distribution of value-added at Magyar Suzuki (%) October 1992 December 1993 December 1994 December 1995 December Magyar Suzuki Hungarian Suppliers Local content (1+2) EU Suppliers* EU Content (3+4) Japanese Suppliers Total (5+6) * including associated member countries Source: Havas (1997). Local content is highly dependent on the nature of assembly operations. Opel operations in Hungary in early 1990s were of tariff were SKD operations with very low though rising local content. However, the most relevant examples in this respect are two major investors in CEE VW and Fiat (Tables 14 and 15). Table 14 Škoda Auto suppliers and supply volume, Number of suppliers Supply Volume (CZK billion) Czech Other Total Czech Other Total 1991* 8** (74.8) 2.7 (25.2) 10.7 (100) ** (82.2) 2.9 (17.8) 16.3 (100) (61.8) 130 (38.2) 340 (100) 14 (66.7) 7 (33.3) 21 (100) (62.4) 105 (37.6) 279 (100) 14.0 (71.4) 5.6 (28.6) 19.6 (100) (49.3) 211 (50.7) 416 (100) 16.1 (69.7) 7 (30.3) 23.1 (100) (31.5) 595 (68.5) 869 (100) 45.5 (66.4) 23 (33.6) 68.5 (100) (30.8) 627 (69.2) 906 (100) 42.7 (64.7) 23.3 (35.3) 66 (100) (25.5) 884 (74.5) 1187 (100) 60.6 (66.5) 30.5 (33.5) 91.1 (100) (23.5) 943 (76.5) 1233 (100) 65.4 (66.6) 32.8 (33.4) 98.2 (100) * since 16/04/1991. ** Czech Republic and Slovakia. Source: Škoda auto annual reports. VW Škoda has increased the number of local and foreign suppliers. The number of foreign suppliers had outweighed the number of local suppliers giving impression of a relative decline of local content. However, in terms of supply volume, local supply volume was still 33

77 around two-thirds of the total, i.e relatively unchanged. A similar trend of a declining share of local suppliers and increasing share in the number of foreign suppliers can be found in the case of Fiat as well (Table 14). Also, though the number of suppliers was drastically reduced by 2000 to one-third of its level in 1992, its level of local content has only reduced slightly, and is still significant at being over 50 percent (see Table 15). Table 15 Number of Fiat Auto suppliers, Polish Foreign Total Index of total suppliers (1992=100) Dec ' (65.3) 215 (34.7) 620 (100) 100 Sept ' (62.9) 208 (37.1) 560 (100) Dec ' (60.6) 193 (39.4) 490 (100) Dec ' (56.9) 188 (43.1) 436 (100) Dec ' (57.0) 177 (43.0) 412 (100) Dec ' (49.9) 196 (50.1) 391 (100) Dec ' (53.4) 170 (46.6) 365 (100) Dec ' (54.2) 163 (45.8) 356 (100) Dec ' (52.1) 123 (47.9) 257 (100) Source: Enrietti (2004). It is difficult to argue that a small and declining number of local suppliers are a regional feature. Also, we do not have evidence which would suggest what the regional average is as cases differ considerably. For example, the Suzuki case could be contrasted to VW Slovakia which had 1200 suppliers in 2000 of which only 2 were Slovak and fed directly to the plant (Weidokal and Stagles, 2002). However, a weak local supply base may be not entirely an endogenous problem of CEE. An increased reliance on modular systems and use of shared platforms enables MNCs to source from a much larger number of countries than before. For example, VWs move towards sharing common platforms across a wide and high model range led to it adopting a global sourcing strategy through which all of the automotive parts were unified into all models of the same size among its four brands (VW, Audi, Seat and Škoda), which allowed the company to eliminate a number of suppliers. Of course, if a supplier could meet the requirements of price and quality, it might be selected to supply all VW vehicle manufacturers. It seems that the CEE suppliers have been unable to meet these standards. As a result, the level of local content went down from 70% for the Škoda Felicia to 31% for the latest medium sized Octavia model (Pavlínek and Smith, 1998, p. 627). In conclusion, it seems that modularization and use of shared common platform have reduced incentives for assemblers to source locally and thus already reduced the low levels of local sourcing. This suggests the limits of industrial upgrading in the CEE 34

78 automotive industry and a clear area for policy action. High productivity improvements within assemblers have not been yet accompanied by the spread of a local supply base. The arrival of first tier suppliers is supposed to deepen automotive clusters and ensure local content with important effects on technology transfer and employment. A growth in employment in the supplier sector suggests that this process is underway. The issue is if there is would be the right role for policy to facilitate this process. 5 Policies influencing automotive industry restructuring in CEE Policy measures in different forms have strongly influenced restructuring and the role of FDI in the CEE automotive industry. We analyse this aspect by looking specifically at four policy instruments: tariffs, privatization and FDI policy, and clusters policy. 5.1 Tariffs Initially, tariffs were important motivational devices for attracting FDI in CEE countries. These were used by foreign investors as a way to ease their market position and thus compensate for investments that they had to incur. Naturally, some investors have used this opportunity and established tariff jumping FDI. Sometimes, these operations were established independently or as part of larger package like GM s investment in engine plant in Hungary (see Moran, 1998). Usually, these were shallow assembly operations of limited scale, with negligible employment and technology transfer effects and often with generous subsidies per job (ibid). Table 16 Import duties on EU imports, Poland Czech Republic Hungary Slovakia % Quotas New Used up to 1600 cc cc Over 2000 cc Source: Motor Business International 3 rd quarter 1996; 4 th quarter 1996; 4 th quarter 1997; CEEBIC. In all CEE countries, import duties in trade with the EU were phased out by 2001 which allowed unrestricted imports and had a knock on effect with increased sales (Table 16). Today, all new vehicles imported with a EUR1 certificate are exempt from duties. However, 35

79 this has put third countries at a competitive disadvantage in exporting to CEE. Import restrictions and tariffs remain in place for second hand cars which are in high demand due to low purchasing power in CEE. 5.2 FDI and privatization policies With the exception of Romanian Dacia and Serbian Zastava all other CEE automotive assemblers were privatized by 1994 and subsequently fully taken over by foreign owners (see Table 17). Table 17 Ownership changes in CEE manufacturers Company Main buyer Date Ownership level (%) Škoda Auto A.S. Volkswagen A.G. 16 April December December May Audi Hungária Motor Kft. Audi A.G. 18 February Magyar Suzuki Rt. Suzuki Motor Corp December May Daewoo-FSO Daewoo March Fiat Auto Poland S.A. Fiat S.p.A. 17 September June Opel Polska S.p. z.o.o GM Europe (Opel) October Volkswagen Poznań Sp. Z.o.o. Volkswagen A.G. 31 December December December Automobile Dacia S.A. Renault S.A. 29 September June March Daewoo Automobile România S.A. Daewoo 19 October Volkswagen Slovakia A.S. Volkswagen A.G. 12 March Revoz d.d. Renault S.A October December Source: Company press releases. For the national governments most of these investments were considered as strategic. Governments aimed to secure employment while at the same time offering direct (grants) 36

80 or indirect incentives like tax holidays. In most of the cases, bargaining between investors and local governments had positive sum effects through preserved employment and improved capacities. Only the Polish government was unlucky as its two crucial investors (Fiat and Daewoo) entered into trouble for different reasons 14. However, privatization was just the beginning of policy for attracting investors in the automotive industry. The quality and generosity of government investment packages have shown to be a very important determinant of greenfield FDI in the automotive industry. It may be not surprise that both Czech and Slovak governments have managed to attract new large scale investments primarily through quality of their investment support. For example, in Slovakia the program for development of the automobile industry was the result of enforcement of the Government Resolution from July Within the government strategy framework the institute of the SR Government Plenipotentiary for automobile production development was established, assigned with coordination of the whole implementation of the Program and reporting back to Government (Sario 2002). Other incentives such as tax holidays, tax relief, investment incentives, and incentives for the establishment of industrial parks have been also introduced. In Hungary there exist a number of key investment incentives for automotive producers and suppliers. These include tax benefits for development, tax-free investment reserves, reduction of the costs of wages, subsidies to establish company premises, direct infrastructural subsidy, subsidies to create jobs, training subsidies, subsidies for intellectual investments, the construction of ring roads around university towns, local benefits, and the implementation of a one-stop-shop system (ITD Hungary 2003). Polices as these led to an intra-regional investment race as increasingly countries have been drawn into fierce competition with one. Recently in 2003 and 2004 Slovakia won out to Poland to host PSA Peugeot s new facilities at Trnava and Kia s facilities at Zilina due to poorer infrastructure, weaker investment incentives, complicated and unpredictable tax system and heavier administrative burdens in Poland. This rush of investments have been generated by the fact that the CEECs will no longer be able to offer investors such generous tax breaks once they join the EU. The question is how the national automotive industries will develop without giving incentive to manufacturers, and with only EU regional aid being available. Even after EU accession CEECs will continue to be heavily dependent on FDI for industry upgrading. EU accession will not eliminate the need for competition in attracting FDI. In some ways, this competition may even increase further. Instead of trying to limit competition for FDI between the EU regions the EU should exploit contests for FDI between regions as a mechanism to improve the business environment in the weakest 14 Also, the Polish government seemed to be behind in competitive bidding race for Greenfield investments. 37

81 regions. As Kuznetsov (2002) points out the existing competition has important limiting features in terms of its protagonists and instruments as it is limited to large MNCs and national governments. Instead of implicitly accepting this, the EC should make these contests public as an incentive device for private and public actors to come together to develop innovative solutions to improve the investment climate on a sub-national level (ibid) with matching grants from the EU level. In this way contests could serve as an incentive device for local government and domestic firms to engage in meaningful joint actions and reforms; as a coordination device to coordinate activities at national and EU level under the umbrella of private-public competitiveness projects; and as a mechanism to share policy knowledge Clusters policy Most OEMs request their suppliers to follow them into any new markets they enter. Once one company sets up operations in a country, a follow-the-leader strategy is usually adopted, resulting in a number of producers and suppliers being networked and integrated into the local economy. Supplier parks have been set up around most of the CEE OEMs. Indeed, it has been estimated that around 75% of all producers and suppliers are to be found in a 200km radius around Poland, the Czech Republic, Slovakia and Hungary (McKinsey, 2002). Cluster policies are now to be found in CEECs. An example of this is the PANAC (Pannon Automotive Cluster) cluster in Hungary. Formed in 2000, this was the common initiative of the major car and component producers and suppliers, which created a country-wide industrial network for the automotive industry. Its goals are: To increase the ability of the supplier companies to join the global supplier networks of the automotive industry To help its partners to increase complexity of their products, thus being able to improve their position in the supply-chain To foster the development of a nation-wide automotive strategy To be a well accepted reference-provider on the partner companies (both to the government and domestic or international buyers) Internationalization of the network This is all achieved through providing access to industry news and information; providing trainings, workshops, and specialists events; performing technology transfer services; performing individual company assessment; leading industry-wide benchmarking activity and facilitating inter-company learning (Pannon Automotive Cluster, 2004). 15 Based on Kuznetsov (2002) 38

82 The Automotive Cluster of Slovenia (ACS) also provides links between members, supporting synergy with suppliers of machines, tools, manufacturing, design, logistics and other services. It also promotes joint members' activities to improve products and operations in R&D, production, quality assurance and to achieve and maintain business excellence. However, these examples seem still to be exceptions rather than rule and we may expect that CEE governments will increasingly prioritize policies for embedding FDI into their local suppliers networks. The Hungarian program Integrator is a good example of this type of policy support. 6 Conclusions Through its numerous production and knowledge linkages to other sectors and its substantial role as generator of exports, value added and employment restructuring of the CEE automotive industry will have important effects on the overall industry restructuring in the region. Restructuring of the CEE automotive industry has been entirely foreign led which has determined the focus of our study which has reviewed the major effects of FDI on industry restructuring. The overall effects of FDI on growth, restructuring and employment in industry are positive. FDI has led to increased specialization in the automotive industry at European level. Integration of CEE into a network of major European automotive MNCs enabled different models to be produced in different countries and the reorganization of the value chain in a way that has created bigger value added for MNCs. Only a minority of activities have been relocated i.e shifted abroad and local facilities closed. The majority of internationalization took the form of expansions and extensions which suggests that the EU enlargement has been a positive sum game in the automotive industry. Those CEE countries that have attracted FDI in this industry have benefited through preserved employment, increased productivity, export and through great potential for development of a local supply base. The biggest beneficiaries of this change were consumers who benefited through increased variety of supply and reduced prices. EU accession has further ensured that these benefits are generated through increased competition and the abolition of tariffs and non-tariff barriers. However, the value creation potential of CEE as global automotive location has not yet been fully exploited. This process has been the most developed in the Czech Republic and Hungary. It is faced with difficulties in Poland and has started in Slovakia and Romania. 39

83 While great changes have been made in assembly sub-sector we may expect further changes in car parts supplier sector through the further arrival of first tier suppliers and the deepening of the local supply base. A combination of country specific factors (proximity, socialist heritage in automotive industry, skilled labour and privatization) coupled with strategies of automotive MNCs have generated different country patterns and very different effects on industry. The CEE automotive industry is highly concentrated on central Europe (the Czech Republic, Slovakia, Hungary, Slovenia, Poland) with great potential benefits in terms of clustering of supplier networks. Improvements in productivity and technology transfer in both embodied (equipment upgrading) and disembodied form (know how, vocational training) are significant in countries with large FDI in the automotive industry. Productivity in industry is well above the industry average and company evidence points to large productivity gains. In that respect, the arrival of large assemblers has produced quite substantial effects which in the next stage needs to deepen through further development of the local supply base. Privatization policy in early 1990s which was followed by policy of attracting Greenfield FDI on the eve of EU accession was crucial in explaining country differences in FDI presence. Automotive investors have foreseen EU accession and in that respect, EU membership will not bring to changes in trend but possibly deepening of automotive cluster in central Europe. However, whether this will happen or not will depend on the ability of CEECs to develop sector specific policies which would support the upgrading of local automotive suppliers. Regarding prospects for further industry restructuring, they depend mainly on improvements among local suppliers. From a policy perspective it is crucial to ensure that local content is growing. A crucial policy issue is whether current national and EU policies are addressing this next stage of industrial upgrading in CEE automotive industry. In relation to the automotive industry, most of the CEECs have been active through FDI policy. This policy focus has become far from sufficient for industrial upgrading which requires integration between FDI and vocational training and innovation policies. In order to assist industrial upgrading CEE should take into account the network character of local and global companies. This has been already been recognized (implicitly or explicitly) through the National Subcontracting programmes (Czech Republic) and the Hungarian Integrator programme, which aim to integrate domestic firms with foreign firms through supply linkages. At EU level, we have proposed the somewhat heretical idea of EU wide FDI contests (see section 5.2.) whose underlying rationale is to integrate FDI and innovation 40

84 policies. In the automotive sector, whose restructuring in CEE is dominantly FDI led, this may be a way forward. Inter-firm linkages which have emerged through automotive value chains should be further deepened. Job and retraining grants as tools of FDI/subcontracting and innovation policy should be expanded throughout the region, possibly linked to Structural Funds programs. This should be complemented with clustering policies and promotion of learning networks which would closely connect suppliers and assemblers. We believe that the best employment policies are those that support upgrading of local suppliers. So far, growth and restructuring in the automotive sector was driven by actions and strategies of MNCs in interaction with government FDI policies which have shaped the overall pattern of restructuring. However, the next stage of industrial upgrading in the automotive industry will depend to a much greater extent on the quality of resources and ability of governments to coordinate policies at different levels. The shift in policy focus towards value chains complicates policy-making, as value chains cannot be fully supported only from the national level. They require multi level governance support at national, regional and EU levels. In addition, the focus on supply chains is not sufficient as upgrading based on value chains may be related to technology, skills and national innovation systems rather than direct production chains. The main challenge at national level is to coordinate diverse policies with very different life cycles covering very different constituencies. Government capability to integrate policy objectives and actions from the different tiers of government (EU, national, regional) will be essential for promoting industrial upgrading through industrial networks I automotive industry. This administrative capability will distinguish between losers and winners in the next stage of industry restructuring. 41

85 Bibliography Allen T, 2001, The Automotive Industry and Candidate Countries, Eurostat: Statistics in Focus, External Trade, Theme 6, 1/2001 Anuarul Statistic al Republicii Socialiste România (Statistical Yearbook of the Socialist Republic of Romania), various years, Bucureşti: Direçia Centrală de Statistică Anuarul Statistic al României (Romanian Statistical Yearbook), various years, Bucureşti Bucharest: Comisia Nationalâ pentru Statisticâ Asociatia Producatorilor si Importatorilor de Automobile APIA (Association of Automotive Manufacturers and Importers) ( Automotive World, 2001, World Automotive Manufacturing, November Central & Eastern Europe Business Information Center CEEBIC ( EBRD, 2000, Transition Report, London: European Bank for Reconstruction and Development EIU, 2004, Executive Briefing: Slovakia, Economist Intelligence Unit EIU data accessed through ITDH, 2003, Automotive Industry, Budapest: The Hungarian Investment and Trade Development Agency (ITDH) Enrietti A, 2004, Fiat Auto Poland and its Suppliers, In Faust, M, U. Voskamp, V. Wittke (eds.) European Industrial Restructuring in a Global Economy: Fragmentation and Relocation of Value Chains, SOFI, Gottingen University, Gottingen, Havas A, 1997, Foreign Direct Investment and Intra-industry Trade: The Case of the Automotive Industry in Central Europe, in Dyker, D (ed.) The Technology of Transition, Budapest: Central European University Press, pp Havas A, 2000, Changing Patterns of Inter- and Intra-Regional Division of Labour: Central Europe s Long and Winding Road, in Humphrey J, Lecler Y, and Salerno M S (eds.), Global Strategies and Local Realities, The Auto Industry in Emerging Markets, London: Macmillan Press Ltd. See also: ( Havas A, 2000b, Local Regional and Global Production Networks: Reintegration of the Hungarian Automotive Industry, in Hirschhausen and Bitzer (eds.), The Globalisation of Industry and Innovation in Eastern Europe: From Post-Socialist Restructuring to International Competitiveness, Cheltenham: Edward Elgar, pp ) Kuznetsov, Y. (2002): Waking up, Catching up and Forging Ahead, Institutional Agenda for Knowledge-Based Growth in Russia, World Bank, mimeo Latvijas statistikas gadagrāmata (Statistical Yearbook of Latvia), various years, Riga: Latvijas Republikas Valsts Statistikas komiteja Lithuanian Statistical Yearbook, various years Leetmaa P, 2003, Relative Prices for New Passenger Cars in EU, EFTA, Acceding and Candidate Countries for 2002, Eurostat: Statistics in Focus, Economy and Finance, Theme 2, McKinsey & Company, 2003, New Horizons: Multinational Company Investment in Developing Countries, Chapter 3, Auto Sector Case Study, McKinsey Global Institute McKinsey & Company, 2002, Business Consequences of EU enlargement: Major Change or Non-Event?, McKinsey Contribution to the European Business Summit 42

86 Magyar Gépjárműipari Szövetség (Association of the Hungarian Automotive Industry) ( Magyar statistikai évkönyv (Statistical Yearbook of Hungary), various years, Budapest: Központi Statistszkai Hivatal Moran, H. Theodore (1998), Foreign Direct Investment and Development. The new policy agenda for Developing countries and economies in transition, Institute for International Economics, Washington. Motor Business International, 1996, A Review of the Motor Industry in Poland: EU Export Base of the Future?, The Economist Intelligence Unit, 3 rd quarter, Chapter 4, pp Motor Business International, 1996, The Czech Republic: Shaping Up for an International Role, The Economist Intelligence Unit, 4 th quarter, Chapter 6, pp Motor Business International, 1997, The Vehicle Industry and Market in Hungary: Prospects for Growth, The Economist Intelligence Unit, 4 th quarter, Chapter 3, pp Organisation Internationale des Constructeurs d'automobiles (OICA) ( Pannon Automotive Cluster, General Introduction, Spring 2004, ( Pavlínek P, 2002, Restructuring the Central and Eastern European Automobile Industry: Legacies, Trends, and Effects of Foreign Direct Investment, in Post Soviet Geography and Economics, Vol. 43, No. 1, pp Pavlínek P and Smith A, 1998, Internationalization and Embeddedness in the East-Central European Transition: The Contrasting Geographies of Inward Investment in the Czech and Slovak Republics, Regional Studies, Vol. 32, No. 7, pp Pemberton M and Puckering D, 1999, World Auto Atlas & Directory, Financial Times Management Report, London: Financial Times Automotive Porter, M., Schwab, K., Sala-i-Martin, X., Lopez Claros, A. (2004) The Global Competitiveness Report , World Economic Forum, Oxford University Press PwC, 2002, East European Automotive Outlook 2002 Q1 Quarterly Executive Briefing, PricewaterhouseCoopers Radosevic, Slavo (2003) Foreign direct investment and its effects on employment in Central Europe, Transnational Corporations, Vol. 12, No. 1(April 2003), pp, Renault, 2004, Logan: conquering new world markets and boosting the Renault group s profitable growth, Press Release, June 2, 2004 Revoz, various years, Letno Poročilo Richet X and Bourassa F, 2000, The Reemergence of the Automotive Industry in Eastern Europe, in Hirschhausen and Bitzer (eds.), The Globalisation of Industry and Innovation in Eastern Europe: From Post- Socialist Restructuring to International Competitiveness, Cheltenham: Edward Elgar, pp Rocznik Statystyczny (Statistical Yearbook of Poland), various years, Warsaw: Glówny urzad statystyczny Sadler D, Swain, A and Hudson R, 1993, The Automobile Industry and Eastern Europe: New Production Strategies or Old Solutions?, Area, Vol. 25, No. 4, pp Sadler D and Swain A, 1994, State and Market in Eastern Europe: Regional Development and Workplace Implications of Direct Foreign Investment in the Automobile Industry in Hungary, Transactions of the Institute of British Geographers, New Series, Vol. 19, No. 4, pp Sdružení Automobilového Prŭmyslu (SAP) (Automotive Industry Association) ( 43

87 Sperling, Hans Joachim (2004) Going East a Volkswagen version of Globalisation, In Faust, M, U. Voskamp, V. Wittke (eds.) European Industrial Restructuring in a Global Economy: Fragmentation and Relocation of Value Chains, SOFI, Gottingen University, Gottingen, Sturgeon T and Florida R, 1999, The World that Changed the Machine: Globalisation and Jobs in the Automotive Industry, A Study by Carnegie Mellon University and the Massachusetts Institute of Technology, Final Report to the Alfred P. Sloan Foundation Statisticheski godishnik na Narodna Republika Blgariia (Statistical Yearbook of the Republic of Bulgaria), various years, Sofiia: T S Centralno Statistichesko Upravlenie Statisticheski godishnik Republika Blgariia (Statistical Yearbook of the Republic of Bulgaria), various years, Sofia Statistická ročenka Československé socialistické republiky (Statistical Yearbook of the Czechoslovak Socialist Republic), various years, Praha: SNTL - Nakladatelství techniké literatury Statistická ročenka Česke republiky (Statistical Yearbook of the Czech Republic), various years, Praha: Český spisovatel Štatistická ročenka Slovenskej republiky (Statistical Yearbook of the Slovak Republic), various years, Bratislava: Štatistický úrad Slovenskej republiky, Veda, vydavatelstvo Slovenskej Akadémie vied Statistički godišnjak Jugoslavije (Statistical Yearbook of Yugoslavia), various years, Beograd: Savezni zavod za statistiku Statistički godišnjak/ljetopis Federacije Bosne I Hercegovine (Statistical Yearbook of the Federation of Bosnia and Hercegovina), various years, Sarajevo: Federalni zavod za statistiku Statistički godišnjak na Republika Makedonija (Statistical Yearbook of the Republic of Macedonia), various years, Skopje Statistički ljetopis (Statistical Yearbook of the Republic of Croatia), various years, Zagreb: Državni zavod za statistiku Republike Hravtske Statistični letopis Republike Slovenije (Statistical Yearbook of the Republic of Slovenia), various years, Ljubljana: Statistični urad Republike Slovenije Statistika aastaraamat (Statistical Yearbook of the Republic of Eastonia), various years, Tallinn: ESA Statistisches Jahrbuch der Deutschen Demokratischen Republik (Statistical Yearbook of the German Democratic Republic), various years, Berlin: VEB Deutscher Zentralverlag für Staatliche Zentralverwaltung für Statistik Statystychnyi shchorichnyk Ukrainy (Statistical Yearbook of Ukraine), various years, Kyiv: Tekhnika Съюзъ нa Вносителите нa Автомобили в Бългаия SVAB (Union of Automobile Importers in Bulgaria) ( SARIO, 2002, Sector Analysis: Automobiles, Trailers and Semitrailers Production, Bratislava: Slovak Investment and Trade Development Agency (SARIO) Škoda, various years, Škoda Annual Report Sljivic N and Wilczynska A, 1997, The Polish Automotive Industry, Financial Times Management Report, London: Financial Times Automotive Swain A, 1998, Governing the Workplace: The Workplace and Regional Development Implications of Automotive Foreign Direct Investment in Hungary, Regional Studies, Vol. 32, No. 7, pp UNCTAD, 2004, World Investment Directory, Geneva UNECE, 2003, The Statistical Yearbook of the Economic Commission for Europe 2003, UN: Vienna 44

88 van Tulder, R and Ruigrok W, May 1998, European Cross-National Production Networks in the Auto Industry: Eastern Europe as the Low End of European Car Complex, Working Paper 121, BRIE ( Weidokal M and Stagles J, 2002, Central and East European Automotive Overview, Price Waterhouse Coopers, Autofacts Združenie automobilového priemyslu SR (ZAP SR) (Automotive Industry Association) ( 45

89 Economic Restructuring and Labour Markets in the Accession Countries Research Project commissioned by EU DG Employment, Social Affairs and Equal Opportunities, Contract No. VC/2003/0367 Coordinated by in cooperation with Alphametrics and The Vienna Institute for International Economic Studies Foreign direct investment in the new Central and Eastern European member countries by Ingo Geishecker* September 2004 *) German Institute for Economic Research (DIW), Berlin

90 Foreign Direct Investment in the new Central and Eastern European Member Countries Ingo Geishecker (DIW Berlin) September 2004 Executive Summary Since the fall of communist regimes across Central and Eastern Europe and the transition towards a market economy manufacturing industries in these countries have undergone an impressive reorientation towards western markets. Since the early 1990 s countries of the European Union are the most important trading partners and also largest investors in Central and Eastern Europe. This development has been significantly fostered by the signing of the Europe Agreements between the European Union and most of the Central and Eastern European countries in the mid nineties with the subsequent abolishment of most trade barriers and the prospect of joining the European Union. A particularly important role for economic restructuring and integration is played by multinational firms and their investment and production activities in Central and Eastern Europe. This paper aims to access the state of economic integration between the major Central and Eastern European Accession Countries and the current EU member states. We will analyse the activity of multinational firms in the Central and Eastern European Accession countries in terms of foreign direct investment and production sharing and derive conclusions about the potential for further economic integration and restructuring. Foreign direct investment essentially represents activities of multinational firms. Up until recently the theoretical literature on multinational firms was divided into models concerned with horizontal and respectively vertical multinationals. Most recent contri- 1

91 butions to the theoretical literature integrate both strains of the theoretical literature. In the so coined knowledge-capital-model three types of firms (national firms, horizontal multinationals and vertical multinationals) can arise endogenously depending on the characteristics of the home and foreign country. Horizontal multinational firms will be dominant if countries are similar in size and relative endowments and trade costs are sufficiently high. On the other hand, if countries are different in relative endowments but similar in size then vertical multinationals will be the dominant firm type in the model, since factor price differences motivate geographical fragmentation of production. Notably trade costs have a detrimentally different impact on horizontal and vertical multinationals, with rising trade costs fostering horizontal multinationals activities but impeding activities of vertical multinationals. This point is crucial when considering the impact of trade liberalisation and therefore of utmost interest when assessing the future development of foreign direct investment towards Central and Eastern European countries after joining the European Union. FDI in CEE showed significant growth through the 1990 s indicating a growing engagement of multinational firms in these countries. Furthermore in all observed Central and Eastern European countries foreign direct investment towards the end of the 1990 s was mainly geared towards the non-manufacturing sector. At the same time figures on exported intermediate goods suggest that manufacturing industries in the Central and Eastern European countries were able to establish vertical links with manufacturing industries in the European Union right from the beginning of the 1990 s. There is however no indication that the subsequent inflow of foreign capital has significantly fostered vertical integration. The evidence so far therefore indicates that foreign direct investment and production sharing in Central and Eastern Europe are only loosely related. Foreign direct investment appears to be primarily undertaken by horizontal multinational firms. In order to examine this hypothesis more thoroughly we assess the determinants for foreign direct investment in Europe in an multivariate gravity model. In accordance with the knowledge-capital-model we evaluate the role of overall market size, market size differences, endowment differences, trade costs and unobserved characteristics as deter- 2

92 minants for bilateral FDI. This model also forms the basis to simulate future developments of FDI in CEE. In the existing empirical literature usually aggregated FDI stocks are analysed. Arguably, the determinants of FDI should, however, differ between different economic sectors. To test this hypothesis we decompose FDI into manufacturing and non-manufacturing FDI and asses the determinants separately. To summarise the empirical findings, the determinants of FDI in non-manufacturing as opposed to manufacturing industries significantly differ from each other. The parameter estimates for FDI in non-manufacturing are similar to the ones from the aggregated model, and indicate a dominating role of horizontal multinational firms in non-manufacturing industries. In manufacturing industries, however, support for the horizontal model of FDI is much weaker. First, the coefficients of market size and market size differences are significantly smaller (in absolute terms) than in non-manufacturing. Second, the insignificance of the coefficient on the endowment difference measure indicates that low-wage seeking investment may indeed play a larger role in manufacturing than in non-manufacturing. Third, trade and FDI in manufacturing are not found to be substitutes. To clarify, we do not find strong support for the vertical model of FDI, rather FDI in manufacturing is somewhat more ambiguous in its nature than FDI in non-manufacturing, encompassing horizontal and vertical elements simultaneously. On the basis of the estimated gravity model for manufacturing and non-manufacturing FDI we can simulate the future development of FDI under different convergence scenarios which also later allows us to simulate future effects of FDI. Our first scenario is what can be described as modest convergence. We assume that endowment differences 1 converge with an annual rate of 3% (beta convergence). Accordingly, in this scenario it would take about 25 years to halve an initial endowment difference. Figure 1 depicts the simulation results for this modest convergence scenario. While in most countries predicted FDI in non-manufacturing is fairly flat and predicted FDI in manufacturing significantly grows there are some interesting country specific patterns. In the Czech Republic our results for the modest convergence scenario indicate that 1 Country GDP per head compared to Non-CEC average of GDP per head. 3

93 both non-manufacturing and manufacturing FDI will grow impressively over the next 20 years. For Bulgaria, however, one has to be less optimistic. Our simulations indicate that manufacturing as well as non-manufacturing FDI will decline over the next two decades. For Romania the picture is mixed. While in our modest convergence scenario FDI in manufacturing significantly grows, FDI in non-manufacturing industries is expected to decrease at least for the next 11 years. Generally speaking, our simulations indicate that with the exception of the Czech Republic the GDP per head growth rates implied by the modest convergence scenario are not sufficient to generate enough non-manufacturing FDI to overcompensate the decline in non-manufacturing FDI due to falling trade costs. For Bulgaria, the only country with falling manufacturing FDI, the decline in trade costs is not sufficient to compensate the decline in manufacturing FDI due to decreasing endowment differences. Our second scenario can be best described as optimistic. We now assume a convergence rate of 4% (beta convergence). That implies that an initial endowment difference is halved after approximately 16 years. Again, we assume that total GDP grows at the same rate as GDP per head and follow the same trade cost trends as in the modest scenario. The results of this simulation are depicted in Figure 2. In this optimistic scenario the development of FDI looks somewhat more promising. For half of the analysed countries we expect significant increases in non-manufacturing FDI while for the other half nonmanufacturing FDI is at least not decreasing and remains fairly flat. With the exception of, again, Bulgaria manufacturing FDI significantly grows in all CEC s. To summarize and conclude, our empirical results suggests that the bulk of multinationals activities worldwide is of the horizontal type, hence mainly is aimed at penetrating foreign markets rather than economising on factor price differences across Europe. However, when we distinguish between FDI in manufacturing and non-manufacturing industries we identify important differences in the determining factors. While it seems save to conclude that FDI in non-manufacturing is of the horizontal type, support for the horizontal model for FDI in manufacturing is much weaker. This has important implications for FDI towards CEE and leads us to expect a different growth path for 4

94 manufacturing and non-manufacturing FDI. We simulate FDI in manufacturing and nonmanufacturing industries until 2020 under a modest and a optimistic convergence scenario. With the exception of Bulgaria we predict strong increases in manufacturing FDI for both, the modest and the optimistic, convergence scenario. This indicates an increasing presence of vertical multinational enterprises in CEC s in the future. With regard to non-manufacturing FDI the picture is somewhat mixed. With modest convergence we predict that non-manufacturing FDI remains flat or even decreases. An exception in this regard is the Czech Republic, where non-manufacturing FDI grows at comparable rates as manufacturing FDI. Under somewhat more optimistic convergence assumptions FDI in non-manufacturing industries is expected to show more robust growth. 5

95 Figure 1: FDI simulation, modest convergence Predicted FDI inward stock BULGARIA CZECH REP. ESTONIA LATVIA LITHUANIA POLAND ROMANIA SLOVAK REP Year FDI Manufacturing FDI NonManufacturing, 6

96 Figure 2: FDI simulation, optimistic convergence Predicted FDI inward stock BULGARIA CZECH REP. ESTONIA LATVIA LITHUANIA POLAND ROMANIA SLOVAK REP Year FDI Manufacturing FDI NonManufacturing, 7

97 Foreign Direct Investment in the new Central and Eastern European Member Countries Ingo Geishecker (DIW Berlin) September 2004 I Introduction Since the fall of communist regimes across Central and Eastern Europe and the transition towards a market economy manufacturing industries in these countries have undergone an impressive reorientation towards western markets. Since the early 1990 s countries of the European Union are the most important trading partners and also largest investors in Central and Eastern Europe. This development has been significantly fostered by the signing of the Europe Agreements between the European Union and most of the Central and Eastern European countries in the mid nineties with the subsequent abolishment of most trade barriers and the prospect of joining the European Union. A particularly important role for economic restructuring and integration is played by multinational firms and their investment and production activities in Central and Eastern Europe. This paper aims to access the state of economic integration between the major Central and Eastern European Accession Countries and the current EU member states. We will analyse the activity of multinational firms in the Central and Eastern European Accession countries in terms of foreign direct investment and production sharing and derive conclusions about the potential for further economic integration and restructuring. Section II discusses the determinants of multinational activity and provides the 1

98 theoretical background. Section III gives an account of recent developments of foreign direct investment and production sharing. Section IV introduces the empirical model and presents the estimation results. In section V we apply the model to estimate the potential for foreign direct investment in Central and Eastern Europe and discuss some related literature. Section VI concludes. II Theoretical Background Foreign direct investment essentially represents activities of multinational firms. Up until recently the theoretical literature on multinational firms was divided into models concerned with horizontal and respectively vertical multinationals. Horizontal multinational firms can be characterised as multi-plant firms that produce similar goods and services in different countries. Models on horizontal multinational firms include seminal contributions by Markusen (1984), Horstmann and Markusen (1987) and Markusen and Venables (1998). The establishment of horizontal multinationals is essentially determined by a trade-off between economies of scale and trade costs. If trade costs were non-existing there would only be single-plant firms serving all markets from one location fully utilising economies of scale. In this type of models trade and multinational activities are substitutes. Vertical multinationals on the other hand are multi-plant firms that geographically fragment production into stages with different factor intensities economising on factor price differences between countries. Models on vertical multinationals include Helpman (1984) and Helpman and Krugman (1985). Vertical multinationals in this class of models arise due to differences in relative factor endowments between countries. Trade costs are impediments to the establishment of multinational firms. As opposed to the implications of horizontal models, trade and vertical multinationals activities are complements. Most recent contributions to the theoretical literature by Markusen (2002) and Markusen (1997) integrate both strains of the theoretical literature. In the so coined knowledgecapital-model three types of firms (national firms, horizontal multinationals and vertical 2

99 multinationals) can arise endogenously depending on the characteristics of the home and foreign country. Horizontal multinational firms will be dominant if countries are similar in size and relative endowments and trade costs are sufficiently high. The intuition behind is that if countries are very dissimilar in size but similar in endowments the larger country will be preferred as location and costly capacities in the smaller country will be avoided, hence there is no scope for horizontal multinationals. On the other hand, if countries are different in relative endowments but similar in size then vertical multinationals will be the dominant firm type in the model, since factor price differences motivate geographical fragmentation of production. If for instance the home country is skilled-labour abundant relative to the foreign country it is profitable to establish plants in the foreign country that specialise in low-skill intensive production stages. That holds as long as trade costs from the host country to the home country are sufficiently low. Notably trade costs have a detrimentally different impact on horizontal and vertical multinationals, with rising trade costs fostering horizontal multinationals activities but impeding activities of vertical multinationals. This point is crucial when considering the impact of trade liberalisation and therefore of utmost interest when assessing the future development of foreign direct investment towards Central and Eastern European countries after joining the European Union. According to the knowledge-capital-model the reduction of trade barriers between the European Union and Central and Eastern European accession countries since the mid 1990 s ceteris paribus should have resulted in a significant lowering of horizontal multinationals activities in Central and Eastern European countries. On the other hand trade liberalisation through the Europe Agreements ceteris paribus should have fostered vertical multinationals activities in Central and Eastern European countries. However one has to be cautious when interpreting the models implications since it disregards institutional factors and investment risk that are also likely to determine multinational activities in Central and Eastern Europe. Finally, according to Neary (2002) one should expect rising foreign direct investment from outside the trading bloc of the European Union as the common market becomes 3

100 larger and more homogenous. III Trends in foreign direct investment and trade Foreign direct investment in the Central and Eastern European countries has grown significantly since the early 1990 s. For the analysis of economic integration and restructuring one should however focus not on the flows of foreign direct investment but on the stock of foreign owned capital as it is the stock that is employed in production and therefore is a better representation of multinationals activities. 1 As becomes clear from Table 1 the European Union (EU15) has become the largest investor in Central and Eastern Europe. For most countries by 1997 more than 50 % of the inward stock of foreign direct investment came from the European Union. This share has grown fast for all Central and Eastern European countries with foreign direct investment from the European Union accounting for up to 95 % of the total inward stock of foreign direct investment in the case of the Czech Republic. Against the backdrop of significantly lowered trade barriers through the Europe Agreements since the mid 1990 s these relative increases in the foreign direct investment stock from the European Union do not correspond well with the implications of the knowledge-capital-model for horizontal multinationals as discussed in the previous section because trade barriers and horizontal multinationals activities do not appear to be complements in praxis. Improvements of the institutional framework and lowering of investment risk, that are not captured by the model, may indeed play an important role in explaining the rise in foreign direct investments. When implementing the empirical application of the knowledge-capital-model one has therefore to think of ways to control for these additional determinants of multinational activities. Furthermore it is important to identify in which economic sector foreign direct investment is channelled. Clearly, the stages of the privatisation process in Central and 1 Furthermore, while there exist no equilibrium concept for FDI flows it can be argued that the stock of FDI can reach an equilibrium value. 4

101 Table 1: Share of foreign direct investment stock from European Union (EU15) in % Year Bulgaria* Czech Republic Estonia Hungary* Lithuania Latvia Poland Slovak Republic Slovenia Source: Eurostat New Cronos for all years for which data is available, *data from national sources, own calculations Eastern Europe have had a substantial impact on the sectoral allocation of foreign direct investment. The important point to keep in mind for our analysis is that multinational enterprises that are actually active in the manufacturing sector potentially play a more important role for the restructuring of manufacturing as opposed to multinational enterprises in the tertiary or primary sector. We employ two statistical sources to analyse the sectoral composition of inward FDI in CEE. When focusing only on the three major accession countries, the Czech Republic, Hungary and Poland, a comprehensive database that contains the necessary information is available from the OECD International Direct Investment Statistics Yearbook. While the starting years vary for the different countries the most recent information on the stock of foreign direct investment is available for 2000 for the Czech Republic and Poland and for 1998 for Hungary. A wider range of countries is covered by Eurostats New Cronos database. The most recent observations in this database also date to back to 2001 while the starting year is However the direct investment stock in Hungary is not available from this source and one can only distinguish manufacturing and total investment. 2 Figure 1 depicts the development of total FDI and FDI in the manufacturing industries for 6 CEE countries utilising the New Cronos data. For the Czech Republic, the real total stock of foreign direct investment more than tripled between 1997 and For FDI in manufacturing growth was, however, somewhat slower so that by 2001 more than 60 % of the inward stock of FDI are in non-manufacturing 2 The data base also contains FDI in the Service industry but data particularly for CEE is incomplete. 5

102 industries. Similarly, in Poland and the Slovak Republic FDI in non-manufacturing accounted for about 65 % of total FDI while in the Baltic states Estonia, Latvia and Lithuania even more than 75 % of total FDI can be attributed to non-manufacturing investments. FDI in these countries clearly follows a pattern of strong increases in total FDI but relatively low and declining shares of manufacturing FDI. Based on OECD data a very similar picture can be drawn. Figures 2-4 show the shares of FDI in services and manufacturing for the Czech Republic, Poland and Hungary. Again, one can observe a strong decline in the relative importance of manufacturing FDI. While this is not to downplay the role of multinational firms in the manufacturing industry one has to keep in mind that the most dynamic sector for multinational firms activities is non-manufacturing. Parallel to this surge in the foreign capital stock, have Central and Eastern European Countries been increasingly integrated with the Common Market? Although we have no direct measure of the extent of integration. We can derive some conclusions from the analysis of trade data. Looking at the overall development of Exports to countries of the European Union and the development of Exports decomposed into consumption goods, capital goods and intermediate goods we can establish to what extent Central and Eastern Europe were able to penetrate the western market and whether they were able to establish links with western manufacturing industries in vertically disintegrated production chains. Figures 5 to 7 show the development of total merchandise exports in real terms for the three major Central and Eastern European Countries. With the notable exception of the Czech Republic total Exports grew significantly. For Hungary there has also been an significant increase in the share of exports directed towards the European Union. Exports towards the European Union accounted for 55% in 1993 and 60% in 2001 of all exports. For the Czech Republic and Poland the share of Exports towards the European Union actually fell of remained virtually unchanged over the same period. Against the background of the substantial reduction in trade barriers between Central and Eastern European countries and the European Union one would have expected stronger and more uniform increases of exports towards the European Union. 6

103 Nevertheless, as can be seen in Figures 8 to 10 the largest part of exports consisted of intermediate goods, suggesting that manufacturing industries in Central and Eastern European countries indeed were well integrated in vertical production chains. However the extent of integration into these productions chains appears to have been strong right from the beginning of the 1990 s and has at best increased only slightly since then. Summarising the analysis so far, the real inward stock of FDI in CEE showed significant growth through the 1990 s indicating a growing engagement of multinational firms in these countries. Furthermore in all observed Central and Eastern European countries foreign direct investment towards the end of the 1990 s was mainly geared towards the non-manufacturing sector. With regard to economic integration through trade, figures on exported intermediate goods suggest that manufacturing industries in the Central and Eastern European countries were able to establish vertical links with manufacturing industries in the European Union right from the beginning of the 1990 s. There is however no indication that the subsequent inflow of foreign capital has significantly fostered vertical integration. The evidence so far therefore indicates that foreign direct investment and production sharing in Central and Eastern Europe are only loosely related. Foreign direct investment appears to be primarily undertaken by horizontal multinational firms. This hypothesis can be examined more thoroughly when assessing the determinants for foreign direct investment in Europe in the empirical model in the next section. This model also forms the basis to model future developments of FDI in CEE. IV Econometric Model and Results Since the mid 1990 s a growing body of empirical literature assessing the determinants of multinationals activities has been emerging. The models used, mostly follow the design of gravity equations initially geared towards the analysis of trade flows. Seminal contributions to the formulation of gravity trade models were made by authors such as Linnemann (1966), Bergstrand (1985), Bergstrand (1989), Bergstrand (1990) and Hummels and Levinsohn (1995). 7

104 Recent applications of the gravity model for the analysis of the determinants of multinationals activities include Brainard (1997), Eaton and Tamura (1996), Brenton, DiMauro and Lücke (1999) and Egger and Pfaffermayer (2004). While the above studies made valuable contributions to our understanding of the determinants of multinational firms activities they lack the complexity to accurately model horizontal and vertical multinationals activities at the same time. It is however important to simultaneously assess determinants of horizontal and vertical activities as the aggregated country level data that is available does not allow to disentangle these activities a priori. In their seminal paper Carr, Markusen and Maskus (2001) formulate a model specification that is in line with the theoretical implications of the knowledge-capitalmodel discussed in section II. The empirical findings of Carr et al. (2001) indeed support the knowledge-capital-model: Similarity in size indeed results in higher multinationals activities, which indicates the importance of horizontal multinational firms. At the same time, however, dissimilarities in endowments increase multinationals activities, which underlines the role of vertical multinational firms. Most recently this specification was criticised by Bloningen, Davies and Head (2002). They show that Carr et al. (2001) miss-specify the terms that measure differences in skilled labour endowments which results in biased coefficients. After correcting for this specification error Bloningen et al. (2002) find strong support for the horizontal model of multinationals activities and reject the knowledge-capital-model. We expand on Carr et al. (2001) and Bloningen et al. (2002) and estimate following basic two-way panel model where endowment differences are correctly specified: F DI ijt = α + β GS GDP sum ijt + β GD GDP diff ijt + β ED ENDOW diff ijt (1) + β GED (GDP diff ENDOW diff) ijt + β T CI IMP T jit + β T CE EXP T ijt + β T ED (IMP T jit ENDOW diff) + δ t + ϑ ij + ɛ ijt where F DI ijt denotes the total inward stock of foreign direct investment from country 8

105 j in country i at time t in real terms. GDP sum denotes overall bilateral market size 3 and GDP diff the difference in market size. 4. According to the knowledge-capital-model discussed in section II we expect total market size to have a positive impact on foreign direct investment. Dissimilarities are however expected to lower foreign direct investment according to the model. ENDOW diff denotes endowment differences, which we approximate by absolute GDP per capita differences 5 following Baltagi, Egger and Pfaffermayr (2003). The term IMP T denotes trade costs for imports from country j. 6 Since in the model horizontal foreign direct investment and trade are substitutes we expect a positive impact of trade cost on foreign direct investment. The term EXP T 7 on the other hand denotes export costs. If the coefficient is significant at all, we expect a negative sign as vertical foreign direct investment should be negatively affected by export costs. The model also includes an interaction term of size and endowment differences (GDP diff ENDOW diff) to take account of model nonlinearities. If countries are different in size and endowments simultaneously Carr et al. (2001) argue that foreign direct investment is expected to be lower. In addition we interact import costs and endowment differences. Following Carr et al. (2001) we expect a negative coefficient as FDI that is driven by endowment differences (low-wage seeking) is hampered by trade costs. Finally we decompose the error term into time specific components δ t and following Hummels and Levinsohn (1995) bilateral fixed components ϑ ij. The remaining error term ɛ ijt is assumed to be idiosyncratic. Allowing for bilateral fixed effects as well as common time effects allows us to avoid omitted variable bias by comprehensively controlling for macro economic influences and institutional and cultural factors such as investment regulations and language that may determine foreign direct investment. Table 2 column 1 presents the estimated coefficients for the basic model. Market size 3 GDP sum ijt = ln(gdp it + GDP jt ) 4 GDP diff ijt = GDP it GDP jt ). 5 ENDOW diff ijt = GDP capita GDP it capita. jt 6 Trade costs are calculated as the ratio of imports at values including cost, insurance, freight (cif) and exports at values free on board (fob). IMP T ijt = IMP cif 7 EXP T ijt = IMP cif fob jit /EXPijt. ijt /EXP fob jit. 9

106 indeed has a positive impact on foreign direct investment. At the same time dissimilarity in size significantly hampers foreign direct investment as one would expect from the theoretical model. Endowment differences are found to have a significant negative impact on foreign direct investment within the sample while trade costs have a positive impact on the inward position of foreign direct. The interaction term of endowment and size differences is found to have a negative sign, indicating that if countries are dissimilar in size and endowment simultaneously foreign direct investment is discouraged ceteris paribus. The coefficient on the interaction term of export costs and endowment differences is found to be insignificant. The estimated parameters correspond well with foreign direct investment that is dominated by horizontal multinationals. Particularly, endowment differences are found to be detrimental to foreign direct investment within our sample, indicating that FDI indeed appears not to be lowwage seeking. So far we have looked at aggregate inward stocks of FDI only, which is, as previously discussed, dominated by FDI in non-manufacturing industries. We re-estimate Equation 1 for manufacturing and non-manufacturing FDI separately to assess whether there exist important differences in the determinants of FDI. Table 2 columns 2 and 3 present the estimated coefficients for non-manufacturing and manufacturing FDI respectively. Again market size is statistically significant and has the expected positive sign in both specifications. However, in manufacturing bilateral market size is much less important in determining FDI (compare the coefficient of 0.01 in Table 2 column 3 to 0.03 in column 2). Similarly, differences in market size significantly lower FDI in manufacturing and non-manufacturing industries. The impact is, however, much weaker in manufacturing than in non-manufacturing (compare the coefficient of in Table 2 column 3 to in column 2). With regard to endowment differences, FDI in manufacturing and non-manufacturing react differently. In non-manufacturing industries endowment differences are found to be detrimental to FDI: An increase of the difference in GDP per head by on Euro lowers the total inward stock of FDI by Euro. For manufacturing industries, however, the coefficient is not significant, indicating 10

107 that endowment differences do not uniformly influence FDI in a negative way. With regard to the impact of trade costs the picture is also ambiguous. While for FDI in non-manufacturing we clearly identify a positive impact of trade costs that indicates a substitutive relationship between FDI and trade as predicted by the horizontal model, trade costs in manufacturing do not significantly impact on FDI, at least not in a uniform positive or negative way. With regard to the coefficients of the interaction terms results again differ somewhat between manufacturing and non-manufacturing. FDI in general is lower between countries that simultaneously differ in their endowments and market size. The effect is, however, much more pronounced for FDI in non-manufacturing. Interaction of trade costs and endowments differences on the other hand does not yield significant results. Summarising, the determinants of FDI in non-manufacturing as opposed to manufacturing industries significantly differ from each other. The parameter estimates for FDI in non-manufacturing are similar to the ones from the aggregated model, and indicate a dominating role of horizontal multinational firms in non-manufacturing industries. In manufacturing industries, however, support for the horizontal model of FDI is much weaker. First, the coefficients of market size and market size differences are significantly smaller (in absolute terms) than in non-manufacturing. Second, the insignificance of the coefficient on the endowment difference measure indicates that low-wage seeking investment may indeed play a larger role in manufacturing than in non-manufacturing. Third, trade and FDI in manufacturing are not found to be substitutes. To clarify, we do not find strong support for the vertical model of FDI either, rather FDI in manufacturing is somewhat more ambiguous in its nature than FDI in non-manufacturing, encompassing horizontal and vertical elements simultaneously. V Potential for foreign direct investment The gravity model estimated in the previous section forms the basis to assess the potential for further foreign direct investment in the Central and Eastern European new member 11

108 Table 2: Regression of knowledge-capital-model Total Non-Manufacturing Manufacturing GDPsum [12.92]*** [11.52]*** [11.32]*** GDPdiff [9.90]*** [8.79]*** [8.78]*** ENDOWdiff [2.81]*** [3.32]*** [0.47] ENDOW diff GDP diff [6.43]*** [6.15]*** [4.13]*** IMPT [2.01]** [1.94]* [1.21] EXPT [1.64] [1.53] [1.20] ENDOW diff2 EXP T [0.54] [0.38] [0.82] Constant [4.56]*** [3.90]*** [4.56]*** Observations R Notes: t-statistics in parentheses; regressions include full set of time, pair fixed effects. Coefficient on interaction term ENDOW diff GDP diff *

109 states. There exist a number of studies that apply gravity model analysis to estimate the potential for foreign direct investment or trade flows towards Central and Eastern Europe (e.g. Buch, Kokta and Piazolo (2003), Schumacher (1997)). Most of these studies rely one cross sectional data and compare the within or out of sample predictions from a gravity model that represent the normal patterns of foreign direct investment or trade with the actual investment or trade flows. However, Breuss and Egger (1999) dismiss this strain of the literature on the grounds that the prediction performance of cross sectional gravity models in general is extraordinary weak, preventing any meaningful conclusion on the extent or existence of trade and investment potentials. Furthermore, Egger (2002) shows that even with panel data potential-estimates based on the comparison of withinsample predictions of trade (or FDI) and actual trade (or investment) flows simply reflect miss-specifications of the model. If the model is well specified, then the residuals should be white noise. However, if one can indeed identify a systematic pattern in the residuals which is frequently coined unused potential then the model assumptions are violated. Brenton et al. (1999) follow a different approach utilising cross sectional and time series data on outward direct investment stocks of 11 major investing countries. Their gravity model contains a measure of destination country income and size, bilateral distance and a indicator variable for economic freedom and is extended by dummies for two sets of CEE countries (first and second round candidates for European Union membership). The main finding is, that for the majority of investor countries the estimated coefficients on the Central and Eastern Europe dummy variables are non-negative. Brenton et al. (1999) therefore conclude that foreign direct investment towards CEE has already reached or even exceeded levels that one would expect among market economies. However the approach of Brenton et al. (1999) can be criticised since the empirical specification of the used gravity model for foreign direct investment is not well based on the theory of multinational firms and the empirical model fails to control for unobserved bilateral effects which is likely to lead to biased coefficients. Görg and Greenaway (2003) estimate the potential of FDI towards CEE with an econometric specification that is somewhat better suited to incorporate vertical and horizontal FDI according to the knowledge-capital- 13

110 model. Utilising panel data for outward and inward FDI for the United Kingdom (UK) over the period the authors estimate the parameters of the gravity equation for a sample of EU countries with a random effects model and subsequently use them for an out of sample prediction of FDI from the UK to CEE. Based on this exercise Görg and Greenaway (2003) conclude that while there is still some potential for increased FDI in the Service industry, FDI in manufacturing has already reached normal levels. The problem with this empirical approach is that it relies on the assumption that bilateral unobserved effects are indeed random and not correlated with the explanatory variables. This assumption may be to restrictive and needs to be tested. If indeed unobserved bilateral effects are correlated with some of the explanatory variables random effects estimation yields inconsistent estimates. Choosing the wrong econometric specification can therefore result in fundamentally biased potential-estimates as is shown in Egger (2002) for the estimation of trade potentials. As previously discussed in Section refsec:model1 we reject the random effects model and estimate fixed bilateral effects specifications. Accordingly, estimating FDI potentials along the lines of Görg and Greenaway (2003) is not appropriate 8. Using the gravity model for identifying unused potentials for FDI therefore is simply not possible. However, the gravity model can nevertheless be useful to simulate future FDI under different scenarios for the development of relative GDP or transport costs. In what follows we do exactly that. Utilising the estimated gravity model parameters we predict foreign direct investment into CEC s for different convergence scenarios. Estimating the model separately for manufacturing and non-manufacturing FDI is crucial for such an analysis as the impact of changes in endowment differences and trade costs on FDI is essentially detrimental as can be seen in Table 2. Our first scenario is what can be described as modest convergence. We assume that endowment differences 9 converge with an annual rate of 3% (beta convergence). Ac- 8 Neither can we follow Brenton et al. (1999) as country group dummies cannot be estimated in addition to bilateral fixed effects. 9 Country GDP per head compared to Non-CEC average of GDP per head. 14

111 cordingly, in this scenario it would take about 25 years to halve an initial endowment difference. To keep things simple we also assume that total GDP grows at the same rate as GDP per head, which amounts to holding population figures constant. From the model we would expect that declining endowment differences due to faster growth in CEC s lead to increasing FDI in non-manufacturing while for FDI in manufacturing the effect is negligible. Increased overall market size and decreased dissimilarity in market size has however an unambiguous positive effect on FDI in non-manufacturing and manufacturing. With regard to trade costs we extrapolate the development of trade costs over the last years in CEC s and Europe. We assume that CEC s costs for imports from the EU15 and the new member states decline with an annual rate of 1% while import costs from other countries remain unchanged. With regard to exports from CEC s into the EU15 or other new member states we expect that costs decline with an annual rate of 0.6%. From the model we know that declining trade costs have a detrimental effect for manufacturing and non-manufacturing FDI. Decreasing trade costs should lower non-manufacturing FDI as the incentives for horizontal multinationals are lowered. Manufacturing FDI that corresponds to vertical multinational enterprises is, however, fostered. Figure 11 depicts the simulation results for this modest convergence scenario. While in most countries predicted FDI in non-manufacturing is fairly flat and predicted FDI in manufacturing significantly grows there are some interesting country specific patterns. In the Czech Republic our results for the modest convergence scenario indicate that both non-manufacturing and manufacturing FDI will grow impressively over the next 20 years. For Bulgaria, however, one has to be less optimistic. Our simulations indicate that manufacturing as well as non-manufacturing FDI will decline over the next two decades. For Romania the picture is mixed. While in our modest convergence scenario FDI in manufacturing significantly grows, FDI in non-manufacturing industries is expected to decrease at least for the next 11 years. Generally speaking, our simulations indicate that with the exception of the Czech Republic the GDP per head growth rates implied by the modest convergence scenario are not sufficient to generate enough non-manufacturing FDI to overcompensate the decline in non-manufacturing FDI due to falling trade costs. For Bulgaria, the only country with 15

112 falling manufacturing FDI, the decline in trade costs is not sufficient to compensate the decline in manufacturing FDI due to decreasing endowment differences. Our second scenario can be best described as optimistic. We now assume a convergence rate of 4% (beta convergence). That implies that an initial endowment difference is halved after approximately 16 years. Again, we assume that total GDP grows at the same rate as GDP per head and follow the same trade cost trends as in the modest scenario. The results of this simulation are depicted in Figure 12. In this optimistic scenario the development of FDI looks somewhat more promising. For half of the analysed countries we expect significant increases in non-manufacturing FDI while for the other half nonmanufacturing FDI is at least not decreasing and remains fairly flat. With the exception of, again, Bulgaria manufacturing FDI significantly grows in all CEC s. VI Conclusion Since the early 1990 s Central and Eastern European countries have seen a significant increase of the inward stock of FDI. Overall FDI sofar is dominated by investment in non-manufacturing industries. In order to predict future FDI in CEE we assess the determinants of FDI on the basis of a gravity model. Our results suggests that the bulk of multinationals activities worldwide is of the horizontal type, hence mainly is aimed at penetrating foreign markets rather than economising on factor price differences across Europe. However, when we distinguish between FDI in manufacturing and non-manufacturing industries we identify important differences in the determining factors. While it seems save to conclude that FDI in non-manufacturing is of the horizontal type, support for the horizontal model for FDI in manufacturing is much weaker. This has important implications for FDI towards CEE and leads us to expect a different growth path for manufacturing and non-manufacturing FDI. We simulate FDI in manufacturing and non-manufacturing industries until 2020 under a modest and a optimistic convergence scenario. With the exception of Bulgaria we predict strong increases 16

113 in manufacturing FDI for both, the modest and the optimistic, convergence scenario. This indicates an increasing presence of vertical multinational enterprises in CEC s in the future. With regard to non-manufacturing FDI the picture is somewhat mixed. With modest convergence we predict that non-manufacturing FDI remains flat or even decreases. An exception in this regard is the Czech Republic, where non-manufacturing FDI grows at comparable rates as manufacturing FDI. Under somewhat more optimistic convergence assumptions FDI in non-manufacturing industries is expected to show more robust growth. 17

114 References Baltagi, Badi H., Peter Egger, and Michael Pfaffermayr, A generalized design for bilaterl trade flow models, Economics Letters, 2003, 80, Bergstrand, Jeff H., The Gravity Equation in International Trade: Some Microeconomic Foundations and Empirical Evidence, Review of Economics and Statistics, 1985, 67 (3), , The Generalized Gravity Equation, Monopolistic Competition, and Factor Proportions Theory in International Trade, Review of Economics and Statistics, 1989, 71 (1), , The Heckscher-Ohlin-Samuelson Model, the Linder Hypothesis and the Determinants of Bilateral Intra-Industry Trade, Economic Journal, 1990, 100 (403), Bloningen, Bruce A., Ronald B. Davies, and Keith Head, Estimating the Knowledge-Capital Model of the Multinational Enterprise: Comment, Working Paper 8929, National Bureau of Economic Research (NBER) Brainard, Lael S., An Empirical Assessment of the Proximity-Concentration Trade-off Between Multinational Sales and Trade, American Economic Review, 1997, 87 (4), Brenton, Paul, Francesca DiMauro, and Matthias Lücke, Economic Integration and FDI: An Empirical Analysis of Foreign Investment in the EU and in Central and Eastern Europe, Empirica, 1999, 26, Breuss, Fitz and Peter Egger, How Reliable Are Estimations of East-West Trade Potentials Based on Cross-Section Gravity Analysis?, Empirica, 1999, 26, Buch, Claudia M., Robert M. Kokta, and Daniel Piazolo, Foreign direct investment in Europe: Is there redirection from the South to the East, Journal of Comparative Economics, 2003, 31, Carr, David L., James R. Markusen, and Keith E. Maskus, Estimating the knowledge-capital model of teh multinational enterprise, American Economic 18

115 Review, 2001, 91 (3), Eaton, J. and A. Tamura, Japanese and US Exports and Investment as Conduits of Growth, Working Paper 5457, National Bureau of Economic Research Egger, Peter, An Econometric View on the Estimation of Gravity Models and the Calculation of Trade Potentials, World Economy, 2002, 25, and Michael Pfaffermayer, Foreign Direct Investment and European Integration in the 1990s, The World Economy, 2004, 27 (1), Görg, Holger and David Greenaway, Is there a Potential for Increases in FDI for Central and Eastern European Countries Following EU Accession?, in Heinz Herrmann and Robert Lipsey, eds., Foreign Direct Investment in the Real and Financial Sector of Industrial Countries, Heidelberg: Springer, 2003, pp Helpman, Elhanan, A Simple Theory of Trade with Multinational Corporations, Journal of Political Economy, 1984, 92 (3), and Paul R. Krugman, Market Structure and Foreign Trade, Cambridge: MIT Press, Horstmann, Ignatius J. and James R. Markusen, Strategic Investments and the Development of Multinationals, International Economic Review, 1987, 28, Hummels, D. and J. Levinsohn, Monopolistic Competition and international trade: Reconsidering the evidence, Quarterly Journal of Economics, 1995, 110 (3), Linnemann, H., An Econometric Study of International Trade Flows, Amsterdam: North-Holland, Markusen, James R., Multinationals, Multi-Plant Economies, and the Gains from Trade, Journal of International Economics, 1984, 16, , Trade versus Investment Liberalization, Working Paper 6231, National Bureau for Economic Research 1997., Multinational Firms and the Theory of International Trade, Cambridge: MIT Press, and Anthony J. Venables, Multinational Firms and the New Trade Theory, Journal of International Economics, 1998, 46 (2),

116 Neary, Peter J., Foreign Direct Investment and the Single Market, The Manchester School, 2002, 70 (3), Schumacher, Dieter, Perspektiven des Außenhandels zwischen West- und Osteuropa: Ein dissagregierter Gravitationsansatz, Report, DIW Berlin

117 VII Figures 21

118 Figure 1: FDI inward position by country CZECH REP. ESTONIA LATVIA LITHUANIA POLAND SLOVAK REP YEAR real FDI in manufacturing real FDI Total Graphs by country Million Euro 22

119 Figure 2: Czech Republic FDI inward position, sector shares Shares in % YEAR share of manufacturing share of services Figure 3: Poland FDI inward position, sector shares Shares in % YEAR share of manufacturing share of services 23

120 Figure 4: Hungary FDI inward position, sector shares Shares in % YEAR share of manufacturing share of services 24

121 Figure 5: Czech Republic Exports in real terms Mill. US$ YEAR total to the European Union Figure 6: Hungary Exports in real terms Mill. US$ YEAR total to the European Union 25

122 Figure 7: Poland Exports in real terms Mill. US$ YEAR total to the European Union Figure 8: Czech Republic Exports by Type in Shares 60 Share in % YEAR consumption goods intermediate goods capital goods not classified 26

123 60 Figure 9: Hungary Export by Type in Shares Share in % YEAR consumption goods intermediate goods capital goods not classified Figure 10: Poland Export by Type in Shares 60 Share in % YEAR consumption goods intermediate goods capital goods not classified 27

124 Figure 11: FDI simulation, modest convergence Predicted FDI inward stock BULGARIA CZECH REP. ESTONIA LATVIA LITHUANIA POLAND ROMANIA SLOVAK REP Year FDI Manufacturing FDI NonManufacturing, A Data Data on the inward stock of foreign direct investment was collected from the New Cronos data base of Eurostat and the OECD International Direct Investment Statistics Yearbook. For the gravity model we utilise New Cronos data only, since this database provides FDI data simultaneously differentiated by industry and source country and just has been updated. Where information on the inward stock on foreign direct investment was missing we imputed values by the respective outward stock of foreign direct investment from the partner country. The time period covered spans from 1997 until While coverage 28

125 Figure 12: FDI simulation, optimistic convergence Predicted FDI inward stock BULGARIA CZECH REP. ESTONIA LATVIA LITHUANIA POLAND ROMANIA SLOVAK REP Year FDI Manufacturing FDI NonManufacturing, 29

126 Table 3: Bilateral Sample Reporting country Number of partners for which FDI data is available Austria 12 Belgium-Luxembourg 8 Bulgaria 2 Switzerland 5 Czech Republic 17 Germany 18 Denmark 14 Spain 11 Estonia 11 Finland 22 France 17 Greece 12 Ireland 9 Italy 16 Japan 8 Lithuania 2 Latvia 16 Netherlands 16 Norway 3 Poland 17 Portugal 15 Romania 2 Slovak Republic 17 Sweden 10 Turkey 3 United Kingdom 20 USA 17 of total FDI is good in the New Cronos data base, coverage of FDI in manufacturing industries is less satisfactory. However, in order to make our results comparable across manufacturing and non-manufacturing we only use observations for which both, total FDI as well as FDI in manufacturing are known. This yields a total of 1561 bilateral observations. Table 3 gives an overview over the country/partner sample. Information on GDP and population was obtained from the World Banks World development indicators data base. All variables are expressed in real values, the respective deflators (capital stock deflator and GDP deflator) were also obtained from the World Bank. Trade data was obtained from the OECD Commodity Trade database. In order to decompose trade into trade with intermediate, capital and consumption goods we applied a concordance between the standard trade classification (SITC) and the United Nations broad economic categories classification supplied by Eurostat. Trade costs were calculated as a moving average of the ratio between cif (cost insurance freight) imports and fob (free on board) exports utilising information from the respective past three years. 30

127 Economic Restructuring and Labour Markets in the Accession Countries Research Project commissioned by EU DG Employment, Social Affairs and Equal Opportunities, Contract No. VC/2003/0367 Coordinated by in cooperation with Alphametrics and The Vienna Institute for International Economic Studies The skill-bias of foreign direct investment in Central and Eastern Europe by Ingo Geishecker* September 2004 *) German Institute for Economic Research (DIW), Berlin

128 The Skill-Bias of Foreign Direct Investment in Central and Eastern Europe Ingo Geishecker (DIW Berlin) September 2004 Executive Summary Foreign Direct investment(fdi) towards Central and Eastern European countries (CEC s) has grown substantially over the last decade and is in general seen as a key factor for transition countries eventual success in the process of catching up with the West. FDI have spurred policymakers and economists interest alike, since it is not only believed to raise tax revenues and employment but also has the potential for technology transfer towards CEC s. However, despite the positive perception of FDI, relatively little research has been actually conducted on the labour market effects of FDI. Particularly, little is known about the potential skill bias of FDI in CEC s. This work aims to fill this gap in the literature by assessing the role of FDI for skill upgrading utilising industry level panel data for CEC s. By assessing the role of FDI for the skill composition of manufacturing industries and utilising predictions for future FDI the ultimate goal of the paper is to provide benchmark simulations for the changing demand for workers s skills under different convergence scenarios in CEC s. A small but growing body of literature looks at the impact of foreign direct investment on receiving countries labour demand for low and high skilled labour. One can identify essentially two different strains of literature, one relating foreign direct investment to 1

129 international outsourcing, and the other focusing on ownership effects and technology transfers. Interestingly, the empirical methodology applied in existing studies is very similar no matter whether the authors are interested in the effects of international outsourcing or technology transfer through FDI. As a result causality between international outsourcing, FDI and skill biased labour demand is hard to establish on the basis of these studies. Although FDI and international outsourcing are related, the two only partly overlap. Even if one shares the view that FDI form the basis for international outsourcing, in practice outsourcing is also possible purely on the basis of subcontracting without substantial FDI. At the same time, empirical studies analysing the determinants of FDI suggest that a large share of FDI is actually of the horizontal type and aimed at serving host country markets instead of establishing vertical linkages. To clarify what labour demand effect is attributable to international outsourcing and knowledge spill-overs through FDI we expand on the existing literature and control for both determinants simultaneously. We estimate industry level employment share equations for low-, medium and high-skilled workers separately utilising a large three-way country/industry panel for CEC s. Our results suggest a significant impact of inward FDI on the relative demand for medium-skilled workers. Low- and high-skilled workers appear to be positively affected by inward FDI, the effect is however not statistically significant. International outsourcing, proxied by imported intermediate inputs, also exerts a negative impact on the relative demand for medium skilled workers and increases the employment share of high-skilled workers. Accordingly we find evidence for an skill-upgrading effect of international outsourcing. A one percent increase in imported intermediate inputs results in a 4 percentage point increase in the share of high skilled workers and a 6 percentage point decline in the share of medium skilled workers. The effects of FDI are somewhat more ambiguous. FDI lowers the employment share for medium-skilled workers. We can however not identify a significant positive effect of FDI on the employment share of high-skilled workers, indicating that to some extent FDI also increases employment of low-skilled workers. This result is not consistent with an 2

130 unequivocal technology spill-over effect that is biased towards high-skilled workers but rather suggestive of an increased disparity of required skills at the workplace. The reason for this development might lie in a mismatch of occupational skills of workers, acquired in the socialist system, and new skill requirements of multinational enterprises and their production network in CEE. Under changing production technology, former medium-skilled workers face a depreciation of their skills being downgraded to elementary occupations while at the same time demand for higher ranking occupations such as technicians and professionals is rising. Based on the parameters estimates from the previous regressions we assess the magnitude of the effect of FDI on the skill composition of the labour force that is to be expected in the future. In order to do so, we combine predicted inward stocks of FDI in manufacturing from the gravity model estimated previously with the point estimates from the regression. Since our model controls for a wide range of observable industry specific characteristics as well as country and industry unobserved characteristics we can interpret the coefficient of FDI as being representative for the average effect of FDI on the skill composition even for countries that initially were not included in our sample due to data constraints but for which we have estimates of future manufacturing FDI inward stocks (Bulgaria, Czech Republic, Estonia, Latvia, Lithuania, Poland, Romania and Slovakia). Figure 1 depicts the cumulated marginal effect of FDI on the demand for mediumskilled workers under two different convergence scenarios holding all other determinants constant. From our estimations follows that a one percent increase in the inward stock of manufacturing FDI (log percentage) ceteris paribus lowers the demand for medium skilled workers by 2 percentage points. With the exception of Bulgaria, manufacturing FDI is expected to rise significantly in all CEC s. Accordingly our simulations indicate an expected increase in the relative demand for medium-skilled workers in Bulgaria and expected decreases for the Czech Republic, Estonia, Latvia, Lithuania, Poland, Romania and Slovakia. In general the effects of FDI on the demand for medium-skilled workers are moderate cumulatively altering the share of medium-skilled workers by less than 3 percentage points over 16 years. Somewhat more pronounced effects are to be expected 3

131 for Latvia and Romania, where under optimistic convergence assumptions the expected increase in manufacturing FDI until 2020 is particularly high so that the share of mediumskilled workers falls by 6 and 4 percentage points respectively. In a further simulation we take changes in other important determinants of the skill composition into account. For the analysis we assume growth rates for manufacturing output that are consistent with modest and optimistic convergence respectively. Imported intermediate inputs extrapolated using country level average growth rates between 1998 and Since we now also take changes of other variables the FDI into account our sample is constrained to the countries for which we have predicted manufacturing FDI stocks and at the same time data on manufacturing output (from the New Cronos enterprise data). Figures 2 and 3 show the respective simulation results for the Czech Republic, Estonia, Latvia, Lithuania and the Slovak Republic. As becomes apparent FDI and outsourcing work at the same direction, both significantly lowering the demand for medium-skilled workers. Changes in output, however, counteract these effects significantly increasing the employment share of medium-skilled workers. As a result the overall changes in medium-skilled workers employment shares are fairly modest below 3 percentage points. An exception is Latvia where significant predicted increases of outsourcing activity and FDI lower the employment share by up to 8 percentage points. To conclude, FDI is an significant determinant of the skill-composition in CEC s and is biased against medium-skilled workers. The magnitude of this effect is, however, modest and partly offset by other factors. All in all we expect the employment share of mediumskilled workers to decrease by less than 3 percentage points between 2005 and Somewhat exceptional is however, the development in Latvia. Here strong increases in FDI and outsourcing activity could lead to a decline in the medium-skilled employment share of up to 8 percentage points. 4

132 Figure 1: Cumulated effect of FDI on demand for medium skilled workers, ceteris paribus Change in percentage points BULGARIA CZECH REP. ESTONIA LATVIA LITHUANIA POLAND ROMANIA SLOVAK REP. Modest convergence scenario Optimistic convergence scenario 5

133 Figure 2: Cumulated marginal effects on demand for medium skilled workers, modest convergence Change in percentage points Czech Rep. Estonia Latvia Lithuania Slovak Rep. Through changes in output Through changes in imported intermediate goods Through changes in FDI Total effect 6

134 Figure 3: Cumulated marginal effects on demand for medium skilled workers, optimistic convergence Change in percentage points Czech Rep. Estonia Latvia Lithuania Slovak Rep. Through changes in output Through changes in imported intermediate goods Through changes in FDI Total effect 7

135 The Skill-Bias of Foreign Direct Investment in Central and Eastern Europe Ingo Geishecker (DIW Berlin) September 2004 I. Introduction Foreign Direct investment(fdi) towards Central and Eastern European countries (CEC s) has grown substantially over the last decade and is in general seen as a key factor for transition countries eventual success in the process of catching up with the West. FDI have spurred policymakers and economists interest alike, since it is not only believed to raise tax revenues and employment but also has the potential for technology transfer towards CEC s. However, despite the positive perception of FDI, relatively little research has been actually conducted on the labour market effects of FDI. Particularly, little is known about the potential skill bias of FDI in CEC s. This work aims to fill this gap in the literature by assessing the role of FDI for skill upgrading utilising industry level panel data for CEC s. By assessing the role of FDI for the skill composition of manufacturing 1

136 industries and utilising predictions for future FDI the ultimate aim of the paper is to provide benchmark simulations for the changing demand for workers s skills under different convergence scenarios in CEC s. Section II. provides an overview over the relevant theoretical and empirical literature. While section III. discusses the data section IV. introduces the empirical modelling framework and presents the econometric findings. In section V. we simulate the future skill composition under different convergence scenarios. Finally section VI. summarises and draws some conclusions. II. Literature A small but growing body of literature looks at the impact of foreign direct investment on receiving countries labour demand for low and high skilled labour. One can identify essentially two different strains of literature, one relating foreign direct investment to international outsourcing, and the other focusing on ownership effects and technology transfers. A seminal contribution to the outsourcing literature is Feenstra and Hanson (1997) who analyse the impact of foreign direct investment on Mexican wage inequality. In the theoretical model underlying the empirical analysis which was initially developed in Feenstra and Hanson (1996), the output of each industry is produced combining many inputs that differ in their skill requirements. The model further assumes that there are two regions, North and South, that differ in their endowments with skilled and unskilled 2

137 labour and factor prizes are not equalised. In equilibrium, the North, being relatively better endowed with skilled labour, specialises in the production of skill-intensive inputs while the South has an advantage in the production of less skill intensive inputs. Foreign direct investment from the North to the South, or any increase in the Southern capital stock relative to the North, lowers relative production costs in the South and shifts part of input production with lower skill intensity from the North to the South. Due to FDI the South gains a cost advantage in production stages with higher skill intensity that were previously carried out in the North. In that sense, FDI can be seen as a prerequisite for international outsourcing. In the empirical application Feenstra and Hanson (1997) estimate wage bill share equations for non-production workers utilising regional and industry level data. Due to data availability, FDI is captured by the ratio of the number of maquiladoras to the total number of manufacturing establishments in a region. Summarising the key findings, FDI (international outsourcing) can explain more than 50% of the increase in the regions non-production wage bill share. The second, larger strain of the literature argues that the impact of FDI is not only confined to outsourcing and focuses on ownership advantage and technology transfers of multinational enterprises. Multinational enterprises, as a prerequisite for their success abroad, possess more advanced production technology with higher skill requirements that reflect their firm-specific assets and ownership advantages (see Caves (1996), Dunning (1988)). The presence of multinationals therefore has a direct effect on the demand for high skilled labour in the host country. In addition demand for high skilled labour 3

138 may be affected more indirectly by technology spill-overs through FDI. The theoretical literature puts a strong emphasis on the importance of multinational firms activities or FDI for introducing and diffusing new technologies in host countries (e.g. Findlay (1978), Das (1987), Wang and Blomström (1992)), however the empirical evidence for this effect is mixed. Figini and Görg (1999) assess the impact of multinational firms activities on wages for high and low skill workers empirically for the case of Ireland. Their results from a random effects regression for a panel of manufacturing industries over the period suggest a significant inverted U-shape relationship between multinational firms employment share and the relative wage for high skilled workers. 1 Bloningen and Slaughter (2001) analyse the impact of FDI on relative demand for high skilled labour for the case of the U.S. Contrasting the results of Figini and Görg (1999) for Ireland the authors find zero or even negative correlation between multinational activity and skill upgrading in U.S. manufacturing industries. Interestingly, the empirical methodology applied in the above studies is very similar no matter whether the authors are interested in the effects of international outsourcing or technology transfer through FDI. Authors such as Feenstra and Hanson (1997) interpret the coefficient on their FDI measure as evidence for the skill bias of international 1 Reinterpreting an endogenous growth model of Aghion and Howitt (1998) Chapter 8, Figini and Görg (1999) expect an inverted U-shape relationship. The argument is that in the early stage of multinational firms presence wage inequality rises as domestic firms learn by imitating multinationals and increase their demand for high skilled labour. Later, however, all firms have adapted the new technology, demand for initially low skilled labour is zero while so defined low-skilled workers have become more skilled through learning-by-doing. 4

139 outsourcing while authors such as Figini and Görg (1999) and Bloningen and Slaughter (2001) interpret the coefficient of their FDI measure as evidence for the impact of technology transfer. Clearly, on this basis causality between international outsourcing, FDI and skill biased labour demand is hard to establish. Although FDI and international outsourcing are related, the two only partly overlap. If one shares the view of Feenstra and Hanson (1997), FDI form the basis for international outsourcing. However, international outsourcing is also possible purely on the basis of subcontracting without substantial FDI. At the same time, empirical studies analysing the determinants of FDI suggest that a large share of FDI is actually of the horizontal type and aimed at serving host country markets instead of establishing vertical linkages (see Brainard (1997), Ekholm (1997), Geishecker (2004)). We want to expand on the existing literature by simultaneously assessing the impact of technology transfer through FDI and international outsourcing on the skill composition of manufacturing employment utilising a large industry panel data base for seven Central and Eastern European countries that recently have joined the European Union. The empirical literature on the effects of multinational firms activities, FDI and outsourcing in Central and Eastern European host countries at best can be described as sparse. One of the few studies is Egger and Stehrer (2003). With a panel of fourteen manufacturing industries over the period for the Czech Republic, Hungary and Poland, the authors regress the wage bill share for non-manual workers on measures for international outsourcing and FDI, with FDI being however not at the center of the analysis. Outsourcing is captured by imports and exports of intermediate goods. Multinational 5

140 activity is proxied by the share of multinational firms in the total number of firms. The results from a dynamic panel regression indicate that international outsourcing in fact significantly lowers the skill specific wage gap in Central and Eastern European manufacturing, contradicting predictions of the model by Feenstra and Hanson (1997). With regard to the effect FDI on the relative demand for low skilled workers results are ambiguous, in general suggesting a positive impact of FDI on the demand for low skilled workers which is however not always significant depending on the estimated specification. Our empirical approach will depart from the Egger/Stehrer study in two important ways: First, instead of using some proxy for multinationals activities, we will asses the role of FDI more carefully by including a direct measure of the stock of foreign capital in our regressions. Furthermore we will also control for the effect of domestic capital which Egger and Stehrer (2003) have excluded from their analysis. Second, instead of only distinguishing between manual and non-manual workers, we apply a more disaggregated skill definition, differentiating between high, medium, and low skilled workers based on industry level information on educational attainment and occupational placement of the workforce. III. Data In order to differentiate between low, medium and high skilled workers we utilise a special distribution of disaggregated tables from Eurostats European Labour Force Survey. This data contains information on employment by disaggregated occupational categories (ISCO 6

141 Table 1: Skill Definitions Skill group Occupation ISCO-88 Code High Skilled Legislators, Senior Officials, Managers 1 Professionals 2 Technicians and Associate Professionals 3 Medium Skilled Clerks 4 Craft and Related Trade Workers 7 Plant and Machine Operators and Assemblers 8 Low Skilled Elementary Occupations 9 88, one digit) and industry (NACE Rev.1, two digit). On this basis we define workers as high, medium and low skilled applying the skill definition as described in Table 1 and calculate the respective shares in total employment for each industry. Data on the inward position of FDI are derived from national statistics and were kindly provided by the Vienna Institute for International Economic Studies (WIIW). Data on industry output and gross investment in tangible goods was derived from Eurostats New Cronos Database. Capital stocks are constructed by the perpetual method using investment flow data. 2 Import and export data were obtained from the United Nations (UN) Comtrade database. Following Fontagné, Freudenberg and Ünal-Kesenci (1997), we decompose trade into intermediate and final goods trade by matching the United Nations classification of Broad Economic Categories (BEC) with the Standard International Trade Classification (SITC Rev. 3) utilising a concordance provided by Eurostat. Subsequently final and 2 For the construction of the initial capital stock we assume growth rates of investment of 0.1% and a depreciation rate of capital of 0.2%. 7

142 intermediate goods trade is allocated across NACE Rev. 1 industries using standard concordances. IV. The econometric model The aim of the analysis is to asses the role of FDI and international outsourcing for skill upgrading in manufacturing industries in CEE. Most empirical applications concerned with skill upgrading express relative demand for employment of different skill groups as cost share equations derived from a translog cost function (e.g. Berman, Bound and Machin (1998), Berman, Bound and Griliches (1994), Feenstra and Hanson (1996), Morrison- Paul and Siegel (2001)). However as reliable data on the wage bill at the industry level is not available for our sample of CEC s, we instead follow Machin and Reenen (1998) and Anderton and Brenton (1999) and estimate employment share equations of the form: S s ijt = α 0 + Σ s β s ln w s ijt + β Y ln Y ijt + β C ln C ijt + β F DI ln F DI ijt + β IMP F ln IMP f ijt + β IMP M ln IMP m ijt + β EXP F ln EXP f ijt + β EXP M ln EXP m ijt + δ t + λ j + µ i + ɛ ijt (1) where s indexes she skill level, i industry, j country and t time. w s denotes wages for high, medium and low skilled workers respectively, Y and C denote output and capital and F DI represents the stock of foreign direct investment. Imports (Exports) of final and intermediate goods are denoted by IMP f (EXP f ) and IMP m (EXP m ). The remaining error term is decomposed into a time specific component δ t, a country specific component 8

143 λ j, an industry specific component µ i and an idiosyncratic error term ɛ ijt. Due to the limited data availability for CEC s we cannot observe wages differentiated by skill and industry. Furthermore it can be argued that wages are in fact not exogenous in such an equation since relative demand and wages are in general determined simultaneously. We therefore reformulate the model, excluding the wage terms and capturing changes in relative wages by allowing for country specific general time effects υ jt. 3 S s ijt = α 0 + β Y ln Y ijt + β C ln C ijt + β F DI ln F DI ijt + β IMP F ln IMP f ijt + β IMP M ln IMP m ijt + β EXP F ln EXP f ijt + β EXP M ln EXP m ijt + υ jt + δ t + λ j + µ i + ɛ ijt (2) We estimate the model with OLS and TOBIT as the employment shares are constrained to lie between zero and one. The respective results for low-, medium-, and highskilled workers are reported in Table tab:ols and 3. As previously discussed, by including the stock of inward FDI and intermediate trade imports and exports in our model we can simultaneously assess and discriminate between the FDI and international outsourcing as determining factors for relative labour demand. Our results suggest a significant impact of inward FDI on the relative demand for medium-skilled workers. Low- and high-skilled workers appear to be positively affected by inward FDI, the effect is however not statistically significant. International outsourcing, proxied by imported intermediate inputs, 3 This is in line with a long run equilibrium view of the economy, where industry wage differentials do not persist. 9

144 also exerts a negative impact on the relative demand for medium skilled workers and increases the employment share of high-skilled workers. Accordingly we find evidence for an skill-upgrading effect of international outsourcing. A one percent increase in imported intermediate inputs results in a 4 percentage point increase in the share of high skilled workers and a 6 percentage point decline in the share of medium skilled workers. The effects of FDI are somewhat more ambiguous. FDI lowers the employment share for medium-skilled workers. We can however not identify a significant positive effect of FDI on the employment share of high-skilled workers, indicating that to some extent FDI also increases employment of low-skilled workers. This result is not consistent with an unequivocal technology spill-over effect that is biased towards high-skilled workers but rather suggestive of an increased disparity of required skills at the workplace. The reason for this development might lie in a mismatch of occupational skills of workers, acquired in the socialist system, and new skill requirements of multinational enterprises and their production network in CEE. Under changing production technology, former medium-skilled workers face a depreciation of their skills being downgraded to elementary occupations while at the same time demand for higher ranking occupations such as technicians and professionals is rising. V. Simulation Based on the parameters estimates from the previous regressions one can assess the magnitude of the effect of FDI on the skill composition of the labour force that is to be 10

145 expected in the future. In order to do so, we combine the predicted inward stock of FDI in manufacturing from the gravity model estimated in Geishecker (2004) with the point estimates from equation 2. Since our model controls for a wide range of observable industry specific characteristics as well as country and industry unobserved characteristics we can interpret the coefficient of FDI as being representative for the average effect of FDI on the skill composition even for countries that initially were not included in our sample due to data constraints but for which we have estimates of future manufacturing FDI inward stocks (Bulgaria, Czech Republic, Estonia, Latvia, Lithuania, Poland, Romania and Slovakia). Figure 1 depicts the cumulated marginal effect of FDI on the demand for mediumskilled workers holding all other variables in equation 2 constant. From our estimations follows that a one percent increase in the inward stock of manufacturing FDI (log percentage) ceteris paribus lowers the demand for medium skilled workers by 2 percentage points. With the exception of Bulgaria, manufacturing FDI is expected to rise significantly in all CEC s (see Geishecker (2004)). Accordingly our simulations indicate an expected increase in the relative demand for medium-skilled workers in Bulgaria and expected decreases for the Czech Republic, Estonia, Latvia, Lithuania, Poland, Romania and Slovakia. In general the effects of FDI on the demand for medium-skilled workers are moderate cumulatively altering the share of medium-skilled workers by less than 3 percentage points over 16 years. Somewhat more pronounced effects are to be expected for Latvia and Romania, where under optimistic convergence assumptions the expected increase in manufacturing FDI until 2020 is particularly high so that the share of medium-skilled workers falls by 6 11

146 and 4 percentage points respectively. So far we have simulated country wide effects of rising FDI holding all other factors that shape the demand for medium-skilled workers constant. However, other determining factors also change over time and partly have significant detrimental effects on the skill composition. From our regression analysis we know that these factors include output and outsourcing intensity. The question is now how the demand for medium-skilled workers changes when we take changes in these factors into account. For the analysis we assume growth rates for manufacturing output that are consistent with modest and optimistic convergence respectively. Imported intermediate inputs extrapolated using country level average growth rates between 1998 and Since we now also take changes of other variables the FDI into account our sample is constrained to the countries for which we have predicted manufacturing FDI stocks (see Geishecker (2004)) and at the same time data on manufacturing output (from the New Cronos enterprise data). Figures 2 and 3 show the respective simulation results for the Czech Republic, Estonia, Latvia, Lithuania and the Slovak Republic. As becomes apparent FDI and outsourcing work at the same direction, both significantly lowering the demand for medium-skilled workers. Changes in output, however, counteract these effects significantly increasing the employment share of medium-skilled workers. As a result the overall changes in medium-skilled workers employment shares are fairly modest below 3 percentage points. An exception is Latvia where significant predicted increases of outsourcing activity and FDI lower the employment share by up to 8 percentage points. We now proceed to simulate the effects of FDI on the skill composition at the industry 12

147 level for each country. For this exercise we allocate aggregated country level FDI predictions on the basis of the lagged average contribution of the particular sector to aggregated FDI growth, thus extrapolating past sectoral FDI growth trends adjusted for overall FDI growth. Industry output was extrapolated in a similar way, following past sectoral growth trends adjusted for overall GDP growth according to different convergence scenarios (See Geishecker (2004).) Sectoral capital stocks and trade in final and intermediate goods were extrapolated using past average growth rates. What becomes clear from the simulations depicted in Figures 4 to 13 is that in all countries most industries show a falling share of medium-skilled workers. However, this is not true for each and every industry. In the Czech Republic for instance, positive demand effects through the growth of output are strong enough to counteract negative demand effects through FDI and outsourcing activity in 5 out of 13 industries. This highlights that, while in general one has to expect a shift away from medium-skilled workers, certain industries might experience continuing growth of the medium-skilled employment share depending on the sectoral development of output, FDI and outsourcing. VI. Conclusion In this paper we evaluate the role of FDI and international outsourcing for shaping the skill-composition of labour demand in CEC s. Instead of distinguishing skills on the basis of manual and non-manual work, we expand on the existing literature by applying a more differentiated skill grouping utilising occupational data from the European Labour Force 13

148 Survey. Our results suggest a significant impact of inward FDI on the relative demand for medium-skilled workers net of international outsourcing. Low- and high-skilled workers appear to be positively affected by inward FDI, the effect is however not statistically significant. This result is thus not consistent with an unequivocal technology spill-over effect that is biased towards high-skilled workers. Furthermore, we find evidence for an skill-upgrading effect of international outsourcing. A one percent increase in imported intermediate inputs results in a 4 percentage point increase in the share of high skilled workers and a 6 percentage point decline in the share of medium skilled workers. Based on the parameter estimates we assess the magnitude of the effect of FDI on the skill composition of the labour force that is to be expected in the future. In order to do so, we combine predicted inward stocks of FDI in manufacturing from the gravity model estimated previously with the point estimates from the regression. Our simulations indicate an expected increase in the relative demand for medium-skilled workers in Bulgaria and expected decreases for the Czech Republic, Estonia, Latvia, Lithuania, Poland, Romania and Slovakia. However, with the exception of Latvia the effects of FDI on the demand for medium-skilled workers are only moderate. In a further simulation we take changes in other important determinants of the skill composition into account. FDI and outsourcing work at the same direction, both significantly lowering the demand for medium-skilled workers. Changes in output, however, counteract these effects significantly increasing the employment share of medium-skilled workers. As a result the overall changes in medium-skilled workers employment shares are fairly modest below 3 percentage points. An exception is Latvia where significant predicted increases of out- 14

149 sourcing activity and FDI lower the employment share by up to 8 percentage points. In a further exercise we simulate skill composition changes at the industry level. While it true that in general relative demand for medium-skilled workers is declining some industries might nevertheless significantly expand their relative demand for medium-skilled workers depending on the sectoral development of FDI, outsourcing and production. 15

150 References Aghion, P. and P. Howitt, Endogenous Growth Theory, Cambridge: MIT Press, Anderton, Bob and Paul Brenton, Outsourcing and Low-Skilled Workers in the UK, Bulletin of Economic Research, 1999, 51 (4), Berman, Eli, John Bound, and Stephen Machin, Implications of Skill-Biased Technological Change: International Evidence, Quarterly Journal of Economics, 1998, 113 (4), ,, and Zvi Griliches, Changes in the demand for skilled labor within U.S. manufacturing: evidence from the annual survey of manufacturing, Quarterly Journal of Economics, 1994, 109 (2), Bloningen, Bruce A. and Matthew J. Slaughter, Foreign-Affiliate Activity and U.S. Skill Upgrading, Review of Economics and Statistics, 2001, 83 (2), Brainard, Lael S., An Empirical Assessment of the Proximity-Concentration Trade-off Between Multinational Sales and Trade, American Economic Review, 1997, 87 (4), Caves, R. E., Multinational Enterprise and Economic Analysis, Cambridge: Cambridge University Press, Das, S., Externalities, and Technology Transfer through Multinational Corporations: A Theoretical Analysis, Journal of International Economics, 1987, 22 (1/2), Dunning, J. H., Explaining International production, London: Unwin Hyman, Egger, Peter and Robert Stehrer, International Outsourcing and the Skill-specific 16

151 Wage Bill in Eastern Europe, The World Economy, 2003, 26 (1), Ekholm, Karolina, Factor Endowments and the Pattern of Affiliate Production by Multinational Enterprises, CREDIT Working Paper 97/19, University of Nottingham Feenstra, Robert C. and Gordon H. Hanson, Foreign Direct Investment, Outsourcing and Relative Wages, in Robert C. Feenstra, Gene M. Grossman, and D. A. Irwin, eds., The Political Economy of Trade Policy: Papers in Honor of Jagdish Bhagwati, Cambridge, Massechusetts: MIT Press, 1996, pp and, Foreign Direct Investment and Relative Wages: Evidence from Mexico s Maquiladoras, Journal of International Economics, 1997, 42 (3-4), Figini, Paolo and Holger Görg, Multinational Companies and Wage Inequality in the Host Country: The Case of Ireland, Weltwirtschaftliches Archiv, 1999, 135 (4), Findlay, R., Relative Backwardness, Direct Foreign Investment and the Transfer of Technology, Quarterly Journal of Economics, 1978, 92 (1), Fontagné, Lionel, Michael Freudenberg, and Deniz Ünal-Kesenci, Statistical Analysis of EC Trade in Intermediate Products, Luxembourg: European Commission, Geishecker, Ingo, Foreign Direct Investment in the new Central and Eastern European Member Countries, Background Paper, Report for the European Commission on Structural Change in Eastern Europe Machin, Stephen and John Van Reenen, Technology and changes in skill 17

152 structure: Evidence from seven OECD countries, Quarterly Journal of Economics, 1998, 113 (4), Morrison-Paul, Catherine J. and Donald S. Siegel, The Impacts of Technology, Trade and Outsourcing on Employment and Labor Composition, Scandinavian Journal of Economics, 2001, 103 (2), Wang, J. and M. Blomström, Foreign Investment and Technology Transfer: A Simple Model, European Economic Review, 1992, 36 (1),

153 A Tables Table 2: OLS regression of labour demand equations Low-skilled Medium-skilled High-skilled ln Y [1.41] [2.69]*** [1.74]* ln C [0.52] [0.05] [0.52] ln F DI [1.34] [2.05]** [1.08] ln IMP m [1.64] [3.11]*** [1.94]* ln IMP f [0.60] [1.36] [0.94] ln EXP m [0.05] [0.02] [0.04] ln EXP f [0.30] [0.10] [0.15] Year [1.50] [1.97]* [0.82] Year [0.32] [0.04] [0.21] Year [0.17] [0.60] [0.74] Constant [1.08] [6.92]*** [0.30] Observations R Notes: t-statistics in parentheses. Regressions include full set of country dummies, interactions of country and time. B Figures 19

154 Table 3: Tobit regression of labour demand equations Low-skilled Medium-skilled High-skilled ln Y [1.79]* [2.96]*** [1.85]* ln C [0.24] [0.04] [0.58] ln F DI [1.55] [2.27]** [1.12] ln IMP m [1.60] [3.42]*** [2.04]** ln IMP f [0.85] [1.48] [1.00] ln EXP m [0.02] [0.02] [0.01] ln EXP f [0.36] [0.09] [0.20] Year [1.36] [2.16]** [0.90] Year [1.60] [3.55]*** [2.08]** Year [0.25] [0.65] [0.86] Constant [1.38] [7.58]*** [0.32] Observations Log likelihood Notes: t-statistics in parentheses. Regressions include full set of country dummies, interactions of country and time. 20

155 Figure 1: Cumulated effect of FDI on demand for medium skilled workers, ceteris paribus Change in percentage points BULGARIA CZECH REP. ESTONIA LATVIA LITHUANIA POLAND ROMANIA SLOVAK REP. Modest convergence scenario Optimistic convergence scenario 21

156 Figure 2: Cumulated marginal effects on demand for medium skilled workers, modest convergence Change in percentage points Czech Rep. Estonia Latvia Lithuania Slovak Rep. Through changes in output Through changes in imported intermediate goods Through changes in FDI Total effect 22

157 Figure 3: Cumulated marginal effects on demand for medium skilled workers, optimistic convergence Change in percentage points Czech Rep. Estonia Latvia Lithuania Slovak Rep. Through changes in output Through changes in imported intermediate goods Through changes in FDI Total effect 23

158 Figure 4: Skill simulation Czech Republic, modest convergence DA: Food products, beverages, tobacco DB: Textiles, textile products DC: Leather, leather products DE: Pulp, paper; publishing, printing DF: Coke, refined petroleum, nuclear fuel DG: Chemicals, man made fibres DH: Rubber, plastic products DI: Non metallic mineral products DJ: Basic metals, fabricated metal products DK: Machinery, equipment n.e.c DL: Electrical, optical equipment DM: Transport equipment DN: Manufacturing n.e.c low skilled medium skilled high skilled 24

159 Figure 5: Skill simulation Czech Republic, optimistic convergence DA: Food products, beverages, tobacco DB: Textiles, textile products DC: Leather, leather products DE: Pulp, paper; publishing, printing DF: Coke, refined petroleum, nuclear fuel DG: Chemicals, man made fibres DH: Rubber, plastic products DI: Non metallic mineral products DJ: Basic metals, fabricated metal products DK: Machinery, equipment n.e.c DL: Electrical, optical equipment DM: Transport equipment DN: Manufacturing n.e.c low skilled medium skilled high skilled 25

160 Figure 6: Skill simulation Estonia, modest convergence DA: Food products, beverages, tobacco DB: Textiles, textile products DC: Leather, leather products DE: Pulp, paper; publishing, printing DF: Coke, refined petroleum, nuclear fuel DG: Chemicals, man made fibres DH: Rubber, plastic products DI: Non metallic mineral products DJ: Basic metals, fabricated metal products DK: Machinery, equipment n.e.c DL: Electrical, optical equipment DM: Transport equipment DN: Manufacturing n.e.c low skilled medium skilled high skilled 26

161 Figure 7: Skill simulation Estonia, optimistic convergence DA: Food products, beverages, tobacco DB: Textiles, textile products DC: Leather, leather products DE: Pulp, paper; publishing, printing DF: Coke, refined petroleum, nuclear fuel DG: Chemicals, man made fibres DH: Rubber, plastic products DI: Non metallic mineral products DJ: Basic metals, fabricated metal products DK: Machinery, equipment n.e.c DL: Electrical, optical equipment DM: Transport equipment DN: Manufacturing n.e.c low skilled medium skilled high skilled 27

162 Figure 8: Skill simulation Latvia, modest convergence DA: Food products, beverages, tobacco DB: Textiles, textile products DE: Pulp, paper; publishing, printing DG: Chemicals, man made fibres DH: Rubber, plastic products DI: Non metallic mineral products DJ: Basic metals, fabricated metal products DK: Machinery, equipment n.e.c DL: Electrical, optical equipment DM: Transport equipment DN: Manufacturing n.e.c low skilled medium skilled high skilled 28

163 Figure 9: Skill simulation Latvia, optimistic convergence DA: Food products, beverages, tobacco DB: Textiles, textile products DE: Pulp, paper; publishing, printing DG: Chemicals, man made fibres DH: Rubber, plastic products DI: Non metallic mineral products DJ: Basic metals, fabricated metal products DK: Machinery, equipment n.e.c DL: Electrical, optical equipment DM: Transport equipment DN: Manufacturing n.e.c low skilled medium skilled high skilled 29

164 Figure 10: Skill simulation Lithuania, modest convergence DB: Textiles, textile products DC: Leather, leather products DE: Pulp, paper; publishing, printing DG: Chemicals, man made fibres DH: Rubber, plastic products DI: Non metallic mineral products DJ: Basic metals, fabricated metal products DK: Machinery, equipment n.e.c DL: Electrical, optical equipment DM: Transport equipment DN: Manufacturing n.e.c low skilled medium skilled 30

165 Figure 11: Skill simulation Lithuania, optimistic convergence DB: Textiles, textile products DC: Leather, leather products DE: Pulp, paper; publishing, printing DG: Chemicals, man made fibres DH: Rubber, plastic products DI: Non metallic mineral products DJ: Basic metals, fabricated metal products DK: Machinery, equipment n.e.c DL: Electrical, optical equipment DM: Transport equipment DN: Manufacturing n.e.c low skilled medium skilled high skilled 31

166 Figure 12: Skill simulation Slovak Republic, modest convergence DE: Pulp, paper; publishing, printing DG: Chemicals, man made fibres DH: Rubber, plastic products DI: Non metallic mineral products DJ: Basic metals, fabricated metal products DK: Machinery, equipment n.e.c DL: Electrical, optical equipment DM: Transport equipment DN: Manufacturing n.e.c low skilled medium skilled high skilled 32

167 Figure 13: Skill simulation Slovak Republic, optimistic convergence DE: Pulp, paper; publishing, printing DG: Chemicals, man made fibres DH: Rubber, plastic products DI: Non metallic mineral products DJ: Basic metals, fabricated metal products DK: Machinery, equipment n.e.c DL: Electrical, optical equipment DM: Transport equipment DN: Manufacturing n.e.c low skilled medium skilled high skilled 33

168 Economic Restructuring and Labour Markets in the Accession Countries Research Project commissioned by EU DG Employment, Social Affairs and Equal Opportunities, Contract No. VC/2003/0367 Coordinated by in cooperation with Alphametrics and The Vienna Institute for International Economic Studies Employment, education and occupation structures: a framework for forecasting by Robert Stehrer* Last update 10 September 2004 *) The Vienna Institute for International Economic Studies (wiiw)

169 Executive summary This paper introduces a model for forecasting changes in employment levels and structures (by sectors, occupational categories and educational attainments). The model is based on the following ideas: As the NMS have lower productivity levels as compared to the EU-15, the scope for technical change and catching-up is quite large. Thus, if these countries converge to the EU-15 productivity levels at given trajectories, real income levels are also changing, which implies changes in demand and thus output structures by Engel curve effects. The latter are modelled as convergence to the EU-15 output structures. These factors, i.e. changes in productivity levels and output shares, imply changes in the level and structure of employment. For making the forecasts we estimated the speed of convergence in productivity levels by sectors and the sectoral output shares econometrically from a larger country sample. From these estimates and the initial levels, forecasts of future developments for productivity levels and output shares can be calculated, which was done for the period A further decomposition with respect to occupational categories and educational attainment levels allows to forecast labour demand with respect to these groups. The most important results may be summarized as follows: (1) In terms of aggregate employment levels, the more advanced NMS (Slovenia, Czech Republic, Slovak Republic and Hungary) are already now, or will be in the next few years, in a phase of rising employment levels although not at very high rates (i.e. unemployment rates or inactivity rates remain almost stable). Another group of countries, Poland and the Baltic states, show modest decreases in employment levels with structural adjustments taking place mainly due to the high initial share of employment in the agricultural sector. Finally, Bulgaria and Romania face severe declines of employment levels in the next decade (about 15%) mainly caused by high productivity increases (due to the high initial gap) and the high initial share of output and employment in the agricultural sector. (2) The structural shifts of employment patterns across sectors are very similar across countries: Employment shares in Agriculture and Manufacturing are decreasing and the employment shares in the service sectors are increasing. Although the dynamic patterns are quite similar across countries, there are large differences in the magnitudes of these changes. (3) With respect to occupational categories, the group suffering most from the ongoing restructuring are the Blue-Collar High-Skilled and Blue-Collar Low-Skilled workers. For the first group an absolute decrease of demand is predicted for all countries; for the Blue-Collar Low-Skilled workers a decrease in demand is predicted for all countries except the more successful NMS (Czech Republic, Hungary and Slovak Republic). i

170 Demand for the other groups (White-Collar workers) are in most cases absolutely rising or at least stable. (4) With respect to educational attainment groups also a clear picture emerges: The group suffering most in relative terms are the low educated employees. Demand for this group is slightly decreasing in the successful NMS Czech Republic, Hungary and Slovak Republic and modestly decreasing in Slovenia and Estonia. A strong decrease in demand for the low educated employees is predicted for the other countries whereas demand for the other groups (medium educated and high educated) remains more or less stable or is even absolutely increasing. ii

171 Robert Stehrer Employment, education and occupation structures: a framework for forecasting 1 Introduction The economies in the new member states (NMS) the Czech Republic, Hungary, Slovak Republic, Slovenia, Poland, and the Baltics Estonia, Lithuania and Latvia and of candidate countries CC-2 Bulgaria and Romania have undergone rapid structural changes with respect to changes in the sectoral structure (unbalanced growth) and uneven productivity dynamics. This has also implied changes in the structure of employment and of labour demand with respect to sectoral employment shares, changing demand for occupational groups and skills (educational levels). Further, in the period of these changes demand for labour dropped in most countries implying either high and persistent unemployment rates or rising inactivity rates. This decrease in demand for labour was mainly caused by rapid technological catching-up processes to the EU-15 levels as well as changes in the sectoral structure of the economies. Although most of the economies performed relatively better with respect to total GDP growth (as e.g. the EU-15 countries) in this period the GDP growth was not enough to compensate the decrease in employment due to technical and structural change. Although the main emphasis of the paper is in forecasting future developments we also discuss some of the historical trends with respect to productivity and structural developments in this paper. These forces (total GDP growth, productivity catching-up and structural change) are also the main components of the model introduced in this paper which is used as a forecasting framework for future developments. The basic idea is as follows: As all these countries have lower productivity levels as compared to the EU-15 the scope for technical change and catching-up is quite large (for an early reference see e.g. Gerschenkron, 1952; the same idea can also be found in the recent convergence literature; see e.g. Barro and Sala-i-Martin, 1995). Thus, if these countries converge to the EU-15 productivity levels at a given trajectory, real income levels are also changing which implies changes in demand and thus output structures by Engel curve effects (i.e. in the case of non-homothetic preferences). Although we do not model this link (between real income levels and expenditure patterns) explicitly in this paper we allow for adjustment processes of sectoral output shares to the EU-15 structures. Here also some potential country-specific deviations from a common pattern may be considered (e.g. due to existing comparative advantages, welfare state policies, etc.) which are discussed at the end of the paper. Both these changes in productivity levels and output shares imply changes in the level and structure of employment. For making the forecasts we estimate the speed of convergence in productivity levels by sectors and the sectoral output shares econometrically from a larger 1

172 country sample. Knowing the speed of convergence and the initial levels then allows to forecast developments in productivity levels and output shares which in a next step then allows to calculate the forecasts for labour demand by sectors and at the aggregate level. The paper goes as follows. In section 2 we introduce the methodology for the scenario analysis, i.e. the model and the econometric analysis used in the study. Further the data sources and classifications are sketched. In section 3 the framework is applied at the aggregate level and forecasts (also including sensitivity analysis) for total labour demand are provided. In section 4 this framework is generalized to allow an application for a multisectoral economy which also allows to analyse sectoral developments. The scenarios not only provide information for development of total labour demand but also labour demand by sectors. Further a decomposition analysis is presented which allows to analyse the most important factors in the restructuring process (productivity developments, structural change, etc.). Finally, in section 5 a further breakdown of the data with respect to occupations and educational levels is analysed which allows forecasts of labour demand for occupational groups and by educational attainment. 2 Methodology and data 2.1 Modelling and estimation of convergence At various stages in this part of the project we estimate a catching-up model with respect to productivity levels and output shares. For this we shall use a framework suggested by Verspagen (1991) which is introduced here in terms of productivity catching-up. This approach will also be used when modelling the dynamics of employment over the next decade. Productivity in this study refers to labour productivity, i.e. output at constant prices divided by labour input. For data reasons we cannot include other factors of production (e.g. capital) into the analysis. The (labour) productivity gap of a country c with respect to a reference country L is expressed in logarithmic terms as c L c G = ln( y / y ) (1) c where G denotes the productivity gap, y c is productivity in country c, and y L is productivity in the country L to which we refer as the productivity leader. The growth rate L L L L of productivity in the leader country is exogenously given with yˆ = y& / y = γ. The rate of productivity growth in the follower country has an exogenous component and a catchingup term and is specified as yˆ c c c c c = γ + βg exp( G / δ ) (2) 2

173 c c where β denotes the catching-up parameter and exp( G / δ ) is a learning capability c (depending on the gap and a learning parameter δ ). Differentiating the technology gap with respect to time and substituting into equation (2) yields the dynamics of the gap L L c c d y y y L c c c c G& & & = ln = = ( γ γ ) βg exp( G / δ ). (3) c L c dt y y y Depending on the initial level of the technology gap and the learning parameter the country c either catches up or falls behind. In the case of δ convergence of the countries to the productivity levels of the leader country is assured. In this paper we only discuss this case. In this case equation (3) simplifies to & c c c (4) G = α βg c with α L c = γ γ. Under the assumptions above convergence is assured; however if the L c exogenous rates of productivity growth differ (where we assume that γ γ ) the follower country would stay behind the productivity level in the leading country at a constant rate. c Setting G & ~ c c = 0 this constant gap is given by G = α / β. In this case the growth rate of productivity in the follower country becomes ~ = γ + βg = γ + βα / β = γ c c c c c L ˆ (5) y i.e. productivity in the follower country grows at the rate of productivity in the leader country. Given the values for the differential of exogenous growth rates α c and the convergence parameter β the initial gap equation (4) determines the time trajectory of the convergence path. Solving the differential equation the productivity level in the follower country at time t is given by c L c ( G ( t) ) = y exp( G ( )) c L y ( t) = y ( t) exp 0 γ t (6) where G c (t) is determined by equation (4). In the case of a constant gap the productivity c L c c level is then given by y = y exp( α / β ). One can easily see that for equal exogenous productivity growth rates the follower country converges to the productivity level of the c L leader country, i.e. y = y. In econometric terms this means that one has to estimate equation (4) to get estimates for the parameters α and β. As we estimate equation (4) across countries we assume that these parameters are the same for all catching-up countries. The time trajectories of productivity convergence across countries then only differ as the initial productivity gaps 3

174 are different. For the estimation of equation (4) one has to estimate the long run motion of c the gap G. For each country in the growth rate of the gap is estimated as G + c c c = φ t c (7) c c c where φ is taken as a measure of the growth of the gap, as G& = φ. This procedure has the advantage that it uses all the data available and not only the first and last observation. Inserting into equation (4) above the convergence parameter is estimated by regressing the growth rate on the initial value of the gap c φ c c = α. (8) βg 0 This framework was introduced for convergence in productivity levels. A similar approach is used when studying convergence of sectoral value added shares. In this case instead of the productivity level of the leader we use the arithmetic mean of value added shares of the EU-15 as the target level for estimation of the speed of convergence in shares Dynamics of labour demand The framework sketched above yields estimates of the speed of convergence in productivity levels and value added shares. Given the initial levels of productivity and value added shares in the follower countries the time trajectories for these variables are determined. Let us first discuss labour demand at the aggregate level; the application for a multisectoral framework is discussed in section 4.3. c c Labour demand is determined by the labour input per unit of output l = 1/ y (i.e. the c inverse of labour productivity) times the volume of output at constant prices Y, i.e. c c c L = l Y (9) Taking derivates with respect to time the growth rate of labour demand can be expressed c c c as Lˆ = lˆ + Yˆ and is thus determined by productivity growth and output growth. Productivity growth is already determined from the analysis above; the second component is total GDP growth. Under the assumptions of full employment and a constant work force (i.e. constant participation rate and constant population) the growth rate of the economy c c c would be determined by the growth rate of the labour input coefficient as 0 = Lˆ = lˆ + Yˆ 1 There is a wide literature on the econometric applications when studying convergence processes and the above framework was heavily discussed for econometric reasons and a number of alternative estimation methods (e.g. time series models, dynamic panel estimations, etc.) were provided. For this study we have also used a dynamic panel framework but the results with respect to the speed of convergence (i.e. the implicit half-time) does not differ very much. So we have decided to stick to the simple framework introduced above as this is in line with the model outlined above. 4

175 or Yˆ lˆ c c =. (In this sense the model would by a standard neoclassical growth model introduced by Solow (1956); for an overview over the economic growth literature see Barro and Sala-i-Martin, 1995). However, this simulation strategy seems not to be appropriate for the economic dynamics ongoing in the NMS as, first, unemployment rates are partly quite high, second, part of the population not yet in the work force could start working if labour demand is rising (i.e. participation rates are not constant) and thus the supply of labour is elastic, and third, hidden unemployment in parts of the economies (e.g. in agriculture) means again that labour supply is not a constraint to economic growth. Thus a more appropriate modelling strategy is to assume that the total GDP growth rate is exogenously determined (which could be influenced by fiscal and monetary policies in the countries, growth rates of trading partners, etc.). Throughout the study we shall apply this assumption for total GDP growth and provide sensitivity analysis of the scenarios with respect to different GDP growth rates. 2.3 Data The data we use for this study are the new version of the OECD STAN database and the LFS database for employment data for NMS and CC-2. The OECD STAN database provides data for value added at constant prices and employment for a larger sample of countries and over a longer time period where in general we use data from 1975 onwards. In this study we only include the old member states (EU-15) i.e. the EU-15 countries for determination of the catching-up parameters. Equivalent data for the NMS and the CC-2 is provided by the National Accounts (taken from the wiiw database). These countries are included from 1995 onwards. Detailed data on employment are taken from the LFS which are available from 1998 onwards. The sectoral breakdown used in this study is presented in appendix Table A-1. As one can see the economy is divided into seven sectors ranging from Agriculture to Public Services. This breakdown was mainly determined by statistical and data availability reasons. Table A-2 provides information on the sample of countries. From the old EU member states Ireland is missing for data reasons and partly we included also Norway in the sample. From the new member states (NMS) we have not included Malta and Cyprus. In a later stage of this study we also use a breakdown of employment by occupational categories (ISCO-88) and educational levels (ISCED) which are given in Table A-3 and A-4. 5

176 3 Aggregate productivity convergence and labour demand As mentioned above we start with the simple labour demand equation L = ly where L denotes labour demand, l is labour input per unit of output (value added) i.e. the inverse of labour productivity - and Y denotes total output (in our case value added). In terms of (constant) growth rates this can be written as Lˆ = lˆ + Yˆ. Under the assumption that productivity converges as modelled above one can calculate first the critical value of output growth to keep labour demand constant, or use these estimates to produce forecasts for labour demand given the growth rate of output. For the latter case we shall present the scenarios and some sensitivity analyses below. Let us however, first, compare labour demand, productivity and value added growth in the old EU member states and the new member states (including Bulgaria and Romania) and, second, estimate the speed of catching-up at the aggregate level. 3.1 Aggregate productivity convergence Table 3.1 reports the average growth rates of value added, value added productivity (value added per employed person), employment and output for the EU-15 and for each of the NMS and CC-2. For the EU-15 the growth rates are calculated for the time period from 1975 (or later depending on data availability) to 2002 whereas for the NMS and CC-2 we the period 1995 to 2002 was considered. The second part of the table reports the growth rates of these variables for all countries in the period 1997 to < Table 3.1 Growth rates of output, employment and productivity > One can see that for the EU-15 output growth exceeded - in most cases - productivity growth to a small extent; the difference can be seen in the employment growth rates. On average output was growing over the total period at a rate of 2.5 per cent, productivity at a rate or 2 per cent which results in employment growth of 0.5 per cent. 2 For the NMS productivity growth was higher than output growth on average which leading to lower employment. The exceptions are Hungary and Slovenia only. One can also see that productivity growth in the NMS and CC-2 is higher than in the EU-15 which implies that catching-up in productivity levels is taking place. In the period the productivity growth rate of the NMS was almost five per cent and even six percent for the CC-2; the total GDP growth rate was lower at about four per cent per year. Although the latter growth rate exceeds the growth rate of the EU-15 it does not suffice to compensate for falling labour demand due to productivity catching-up. 2 One has to note that we do not distinguish between full and part-time employment and in this dataset we use only the number of employees. 6

177 For calculation of the scenarios we have to estimate the coefficient of productivity convergence β. For this we calculate the gap (value added per employed person) as defined in equation (1) and regress this measure on a linear time trend (see equation (7)) which is done for each country separately. This yields an estimate of the motion of the gap c φ which is used as the dependent variable in equation (8). For the initial level of the gap c G 0 we used the first year available for each country of the EU-15 and 1995 (or later if not available) for the NMS and CC-2. We dropped Luxembourg from the sample as this country has very high productivity levels but data are not available for the whole period which causes econometric problems. From the remaining countries the productivity leaders turned out to be France and Belgium over the period considered. Table 3.2 presents the results from the cross country regressions. In column (1) we used the whole sample whereas in column (2) we dropped some countries which performed badly in terms of convergence over the period considered. These countries are Germany (after reunification), Portugal, Latvia, and the CC-2 (Bulgaria and Romania). < Table 3.2 Estimation of convergence > The regressions show a R 2 of 0.65 for the first and 0.87 for the second estimation. The estimated coefficients for convergence are for the total sample and for the reduced sample and are highly significant in both cases. The half time of convergence (i.e. the time period used to close the gap to half of the initial gap) can be calculated as ln 0.5 / β. Inserting the point estimates above, the implicit half time is 23 and 16 years respectively. These estimates suggest faster convergence than e.g. the study by Barro and Sala-i-Martin (1999) which suggest a half-time for conditional convergence of about 27 to 30 years for a much wider sample of countries. As the countries included in our sample are quite homogenous and as the endowments of the NMS and CC-2 with physical and human capital is at sufficient level thus that technology transfer can take place easily our estimates seem to be reasonable for the productivity catching-up process of these countries. Table 3.3 presents an overview of population size and value added data for the NMS, the CC-2 and the EU-15. This table also provides information on the value added per capita and per employed person, respectively, as well as the gap to the weighted average of the EU-15. In terms of value added per employed person the countries farthest behind are Bulgaria and Romania reaching about a quarter of the EU-15 level only; the countries closest to the EU-15 are the Czech Republic, Hungary and Slovakia reaching about fiftyfive per cent and Slovenia with about 65 per cent of the EU-15 average. (In this way these countries outperform Greece and Portugal in the ranking). < Table 3.3 Initial gaps > 7

178 3.2 Implications for labour demand and employment The implications for labour demand under the assumption that the countries follow the specific path of productivity convergence specified above are sketched in a schematic way in Figure 3.1. Productivity growth rate in NMS Output growth rate in NMS Output growth rate of EU-15 Productivity growth rate in EU-15 t Ti Figure 3.1 Productivity convergence and labour demand For the EU-15 we assumed a constant long-run growth rate of productivity and output depicted on the vertical axes. As was shown above (see Table 3.1) output growth was higher than productivity growth leading to generation of employment. Under the assumptions of the model for the follower countries (i.e. convergence of productivity to the EU-level and equal exogenous productivity growth rates) the productivity growth rate of the NMS and CC-2 is relatively high at the beginning of the catching-up period (and highest for countries showing the largest gap) but as the gap is closed over time decreases over time. The effect on labour demand and employment then depends on the growth rate of output (GDP). If productivity is growing faster than GDP, demand for labour will decrease. Even if the GDP growth rates are higher in the NMS and CC-2 (as indicated in the figure) it is thus likely that productivity growth exceeds GDP growth at the beginning, leading to negative employment effects. As already discussed above this was the case for most of the NMS and CC-2 countries over the last decade. However, at a point in time countries may enter the phase where employment is created; in the figure this is indicated with t *. Figure 3.2 shows the productivity growth rates using the initial gap of the year 2002 and a convergence parameter of β = Further we assumed a long-term productivity growth rate of the EU-15 (to which the growth rate of the follower countries converge) of L γ =

179 BG CZ EE HU LT LV PL RO SI SK Figure 3.2 Implied productivity growth rates One can see that the projected productivity growth rates range from (Bulgaria) to (Slovenia) depending on the initial gaps of productivity (see Table 3.3 above). Under the assumption of a GDP growth rate of four per cent as a benchmark scenario for medium-term performance one can see that only some countries (especially Slovenia, Czech Republic, Hungary and Slovakia) moves to the stage of employment growth in the period to In all other countries the productivity growth rate remains above this benchmark of 4 per cent over the whole simulation period (until 2012) and thus one has to expect a quite long period of jobless growth or even job destruction. Two countries (Estonia and Poland) are expected to reach this benchmark at the end of the simulation period. (Another way of interpretation of this figure is the following: The numbers show the GDP growth rates that would have to be reached in order to keep employment at constant levels. The pressure for high GDP growth rates diminish over time when the gap to the EU- 15 is closed and thus the potential for productivity catching-up becomes smaller.) Let us mention potential caveats of this analysis. The most important caveat is that it assumes that output growth and productivity growth are independent from each other. The relationship between these two variables can go in either direction with a positive mutual influence: higher output growth may imply a higher productivity growth (i.e. the Kaldor- Verdoorn effect) and a higher productivity growth rate may imply a higher output growth rate (e.g. via export multipliers, etc.). Further there is also a relationship between the level of employment and effective home demand. Creation of jobs may thus lead to higher overall GDP growth even at lower productivity growth which would bring a Keynesian aspect into this analysis which is not captured in the simple framework above. To take account of these relationships in a detailed manner is, however, beyond the scope of this study and would involve a more explicit dynamic model and structural relationships (for a dynamic model which is similar to the framework introduced here see Landesmann and Stehrer (2002) for a closed economy and Landesmann and Stehrer (2003) for an open 9

180 economy framework). However, the essential point of the problem of jobless growth caused by productivity convergence is clear from the simple framework introduced above. Concluding, some of the countries considered are on the verge of creating employment and for some of them one has to expect further losses in employment over a longer period. In the next section we shall provide some figures of future employment levels based on the calculations above. 3.3 Levels of employment and labour demand Aggregate scenarios Using the framework introduced above we now present four scenarios for the dynamics of the aggregate employment levels for each country. As argued above there are two crucial parameters in this framework: the growth rate of GDP and the growth rate of labour productivity depending on the exogenous growth rate of productivity in the EU-15 (as the productivity leader) and the convergence parameter β and the levels of the productivity gaps of the NMS. For the first variable we show scenarios with 4 per cent and 5 per cent growth rate of GDP (which is in line with the past growth experience of these countries as shown in Table 3.1 above); for the second variable, we assume convergence parameters of β = and β = , respectively, which are in line with the econometric estimates reported above. Table 3.4 presents the forecasted employment levels for the four scenarios which are calculated under the assumptions given above. < Table 3.4 Scenarios for aggregate employment levels > In the first scenario (modest GDP growth and modest speed of convergence) only the Czech Republic, Hungary, Slovenia and Slovakia succeed in creating employment but only at very low rates. The most successful country is Slovenia where employment rises by about 5 per cent in the period 2002 to All other countries experience according to this scenario further losses in employment. These losses amount to more than 10 per cent of the employment level in 2002 for Latvia, Romania and Bulgaria (with more than 15 per cent). In the second scenario the GDP growth rate is assumed to be at 5 per cent per year; one can see that this increase of the GDP growth rate of one percentage point has a quite strong effect on labour demand and most countries show higher employment levels at the end of the simulation period than in The only exceptions are Latvia, Bulgaria and Romania, but losses of employment are in this case less than ten per cent. 10

181 In the third scenario (modest GDP growth and higher speed of productivity convergence) we assumed a convergence parameter of β = (which is similar to the estimated coefficient for the reduced sample). In this case all countries with the exception of Slovenia will experience losses in employment until Finally under the assumption of a higher GDP growth (five per cent per year) and the higher speed of convergence a number of countries shall again experience positive employment effects over the longer run and for most countries even higher employment levels at the end of the simulation period as compared to the first scenario are observed. 3.4 Policy issues From these considerations one policy conclusion may be to aim at a higher GDP growth rate which shall reduce the problem of jobless growth over a long time period. However, the experience so far has been that these countries do not succeed in maintaining sustainable and high growth rates. Further most of the countries have to account for constraints in their macroeconomic policies as most of them have high budget deficits on the fiscal side or want to qualify for accession into the euro area (which imposes constraints on the conduct of monetary and fiscal policies). Finally, GDP growth in these countries is also constrained by the slow dynamics of the European economy which could be a source of growth via exports. A second option would be to lower the productivity growth rate which would lead to less negative employment effects (given the growth rate of output). However, first of all this might not be a variable that is easily influenced by economic policy and, furthermore, this option may not be recommendable as, first, the future effects of high productivity growth may outweigh the short term losses and, second, the international performance of these countries would suffer as the productivity gaps would remain high for a long period of time. One could further think of increasing the productivity growth rate which in the first periods would imply even higher negative employment effects, however the countries may enter the region of employment creation earlier. But as the third scenario shows this speeding up of productivity convergence must go in hand with a higher overall GDP growth such that these economies can create jobs in the medium run. 4 The sectoral dimension The aggregate analysis above may however hide important issues with regard to the sectoral structure of the economy and changes of the structure of output. In this section we thus reformulate the framework above such that the sectoral dynamics in the economies play a role (for a more elaborate theoretical model see Stehrer, 2002a, for a closed economy and Stehrer, 2002b, for internationally integrated economies; this framework was 11

182 extended by Landesmann and Stehrer, 2004, to allow for non-homothetic preferences). The sectoral dimension is important as a destruction of jobs in a particular sector due to productivity growth and changes in demand may imply high adjustment costs to workers (e.g. geographical mobility, requirement of new skills, etc.). Furthermore as skill intensities and occupational structures differ across sectors this dimension become an important issue when studying the development of labour demand with respect to educational levels and occupational categories. As already mentioned above we distinguish seven sectors (see Table A-1). From the methodological point of view the additional variable to be considered is the share c of the particular sectors in the economy which denoted by α with α = and its c i i i 1 dynamics over time. Further, labour productivity changes at different rates in the particular sectors because of a different exogenous (i.e. the sectoral labour productivity growth rates of the leaders), sector specific convergence parameters and sector specific gaps in productivity levels. Let us discuss this issue first. 4.1 Productivity convergence at the sectoral level Dynamics of productivity, output and employment For the sectoral convergence patterns of labour productivity we can use the same framework as introduced in section 2 above; the only difference is that we have to index the equations (1) (8) with an index i for the different sectors. To give a first overview we present data on productivity growth, output growth and employment growth for two groups of the old EU member states (where we separated the cohesion countries Spain, Greece and Portugal), the NMS and CC-2 in Table 4.1. For the EU-12 (EU-15 without cohesion countries) and EU-3 (cohesion countries) the whole period whereas for the NMS and CC-2 only the period is considered. < Table 4.1 Sectoral productivity, output and employment growth rates > Let us refer mainly to the growth rates in employment, which results from the difference between the growth rate of output and productivity. Of course, the overall impact of a sector on total employment also depends on the relative size of this sector in the economy which shall be discussed below in more detail. In terms of growth rates the most important changes occurred in the agricultural sector (AB). This is the case for the EU-12 and EU-3 countries as well as most of the NMS, less so, however, for the CC-2. The average growth rate is about 0.03 for the EU countries (over the long period), but partly much higher (in absolute terms) for the NMS with growth 12

183 rates ranging from 0.04 (Hungary) to 0.09 (Slovakia). Exceptions to this are Poland and Romania with growth rates of 0.30 and Bulgaria with only This shedding of labour out of agriculture (AB) was mainly caused by rather high productivity growth rates in this sector lying above the country average of labour productivity growth (exceptions to this are Estonia and Romania) but also by slow output dynamics which was below the growth rate of the total GDP. Estonia, Latvia, and Slovenia even experienced negative output growth rates. Only the Czech Republic and Bulgaria have a higher output growth rate in the agricultural sector than in the total economy. Negative growth rates in output can be observed for Estonia, Latvia, Slovenia, and Romania. For the industrial sector (CDE) consisting of Mining and Quarrying (C), Total Manufacturing (D) and Electricity, Gas and Water (E) the losses in employment are less dramatic in terms of growth rates. The growth rates of employment in this sector are negative for almost all countries (the exception is Hungary) and quite high (in absolute terms) for Lithuania, Poland and the CC-2 Bulgaria and Romania. For this sector the growth rates of productivity are also rather high and sometimes higher than for agriculture, which was partly compensated by higher output growth rates as well. An exception is the Czech and Slovak Republic where productivity as well as output growth was low. For construction (F) the evidence is rather mixed: Czech Republic, Estonia, Lithuania, Poland, Slovak Republic and Bulgaria and Romania experienced negative whereas the remaining countries (Hungary, Latvia and Slovenia) experienced positive employment trends. For the first group this was partly caused by negative output growth. In the fourth sector Trade, Repairs and Hotels (GH) only Czech Republic, Lithuania, Poland and Romania show a negative trend mainly caused by high productivity growth rates, as output growth in this sector is positive for all countries. A similar picture can be seen for sector Transport (I) where output growth is positive and relatively high. Losses in employment which occurred in Czech Republic, Estonia, Latvia, Poland, Slovakia and Bulgaria thus are mainly caused by high productivity growth. Relatively low output growth but even lower productivity growth can be observed for Hungary, Latvia and Slovenia. In Romania output growth is almost zero so that even relatively low productivity growth results in losses in employment. The next sector Business Services (JK) consisting of Financial Intermediation (J) and Real Estate, Renting and Business Activities (K) is often regarded as a sector for potential job creation. This was especially the case for Hungary, Latvia, Poland and Slovakia so far with employment growth rates of about five to six per cent. The other countries have positive growth rates as well (the exceptions are Latvia with a slightly negative growth rate and Romania with 0.04) and given the negative employment growth rates of the total economy Finance and Business Services can be seen as one of the job creating sectors also in the NMS. 13

184 Finally, Public Services (LQ) was growing in employment terms in the Czech Republic, Latvia, Slovenia and Romania but at very modest rates. In most cases thus productivity growth outweighed the modest output growth in this sector. The only exception is the Czech Republic, where productivity was declining faster than output. Compared to the long run estimates of growth rates for the EU-12 and EU-3 one can see that the NMS and CC-2 have on average higher productivity growth rates in all sectors with the exception of Construction (F) for the NMS and Transport (I). Output growth rate is notably higher in Industry (CDE), Trade, Repair and Hotels (GH) and Business Services (JK) in the NMS and in most sectors for the CC-2. However, as productivity growth is greater than output growth in most sectors this results in negative employment effects. Thus a similar framework as used for the total economy (and graphically summarized in Figure #3.1) may be applied for each of the sectors Sectoral productivity levels and potential for catching-up The next step is to estimate the speed of convergence of labour productivity for each sector (if convergence occurs at all) and to forecast the productivity growth rates according to equations (4) and (2) above. Before going over to the estimation of the convergence parameters we look at the productivity gaps for each sector and country which in combination with the sector specific convergence parameters β. determine the growth rates. Table 4.2 shows the level of value added per employed person and in percentage of the EU15. i < Table 4.2 Value added productivity > For a better overview the productivity levels in percentage of EU-15 are plotted in Figure 4.1. Here we have also ranked the countries with respect to the gap of the total economy (the economy closest to the EU-15, Slovenia, is ranked first). The ranking of the countries is Slovenia, Czech Republic, Hungary, Slovakia and Poland which are above the mean, whereas Estonia, Lithuania, Latvia, Romania and Bulgaria are below the mean. The productivity levels as percentage of EU-15 ranges from 67 per cent (Slovenia) to 26 per cent (Bulgaria). This ranking, however, only partly shows up in the productivity levels for the particular sectors. 14

185 SI CZ HU SK PL EE LT LV RO BG AB CDE F GH I JK LQ Total Figure 4.1 Value added productivity in per cent of EU-15 With respect to the particular sectors we first compare the average levels of productivity across sectors. The best performing sector on average is Trade, Repairs and Hotels (GH) which reaches a level of more than 65 per cent of the EU15 level; this sector is followed by Business services (about 60 per cent). The next two sectors are Agriculture (AB) and Construction (between 52 and 55 per cent, respectively), and finally Industry (CDE), Transport (I) and Public Services (LQ) with levels ranging between 43 and 45 per cent. However, these means hide large country differences to which we turn next. For Agriculture (AB) the Czech Republic, Hungary and Slovakia reach more than 80 per cent of the EU level. In between are Estonia with 58 per cent and Slovenia with about 39 per cent. The other countries show levels of about 30 per cent. The productivity levels for the industry sectors (CDE) are much closer together in a range between 55 per cent for Slovenia and 24 per cent for Bulgaria. Here also the ranking as given for the total economy applies more or less, the only notable exception is Poland. In the construction sector (F) one can see no clear picture with regard to the overall ranking of countries and range is from 86 per cent (Slovenia) to about 30 per cent for the Czech Republic. With regard to Trade, Repairs and Hotels (GH) which shows the highest productivity level relative to the EU15 on average (almost 70 per cent) the ranking applies in the sense that countries with higher overall productivity levels have also a higher productivity level in this sector. The main exception to this is Hungary with a level of less then 60 per cent (compared to 80 to 100 per cent for the other leading countries). The country with the lowest productivity level is again Bulgaria reaching less than 30 per cent. In the Transport 15

186 sector (I) the levels are again closer together and the overall ranking applies more or less (only Hungary performs relatively better). In Business Services (JK) there is a group of five countries (Slovenia, Czech Republic, Hungary, Slovak Republic and Romania) with a level of more than 75 per cent relative to the EU15. Here also Estonia performs reasonably well with about 60 per cent. Finally, in Public Services (LQ) the overall ranking also applies but with the exception of the Czech Republic. The leading country is Slovenia with almost 90 per cent and the least ranked country again Bulgaria with more than 20 per cent Convergence at the sectoral level These results mean that there is potential size for productivity catching-up especially in Industry (CDE), Public Services (LQ) and Transport (I) for almost all countries whereas for the other sectors the scope for catching-up varies widely across countries. These different structures of the economies with respect to sectoral productivity gaps are accounted for in our framework as a larger gap implies higher productivity growth in the convergence equations. We next turn to the estimation of the speeds of convergence in the particular sectors similar to the aggregate case discussed in section 3. Let us discuss the results for each of the sectors. 3 Table 4.3 reports the estimates of the speed of convergence for the particular sectors where we present the results for various samples (having partly dropped countries which performed particularly bad or well over the period). < Table 4.3 Estimates of sectoral convergence > For agriculture (AB) we found no significant convergence for the whole sample. However, Romania, Slovenia and Greece were falling behind during the period observed. From these countries Romania and Slovenia showed a volatile performance in terms of catchingup over the period and thus the estimate of the growth rate of the gap may not be reliable. Greece on the other hand seems to be at a low productivity convergence path. Dropping these countries from the sample gives the estimate reported in Table 4.3, column (1). Still the estimate is not significant. The reason for this is that another group of countries (Austria, Portugal, Poland, Latvia and Lithuania) exhibit very slow growth rates although having a very high gap. For Austria data are not reliable (as already mentioned above). Latvia and Lithuania again have a very volatile dynamics of the gap, whereas Poland and Portugal seem to be on a lower productivity convergence path. One reason for this (as well as for Greece) may be that these sectors still play a role as large sectors in which workers 3 For other studies considering convergence at the sectoral level see Bernard and Jones (1996) and the discussion on this in S rensen (2001) and the reply by Bernard and Jones (2001). A further study on productivity convergence in service sectors is Gouyette and Perelman (1997). 16

187 not elsewhere employed find jobs. Dropping these five countries from the sample then yields a significant convergence parameter of For the industry sector (CDE) the convergence parameter of is significant for the whole sample. For Construction (F) the gap does not play a role at all for productivity growth. Even when dropping the outliers Latvia and Slovakia which were falling back quite fast the convergence parameter is insignificant. We have nonetheless reported the regression for this sector in Table 4.3. For Trade, Repair and Hotels (GH) there is a group of countries which are falling back quite rapidly. These countries are Bulgaria, Romania, Hungary and Germany (including the eastern part). Dropping these countries from the sample the coefficient of convergence becomes and significant at the 10 per cent level. The detailed results are reported in Table 4.3 in column (1). Further there Lithuania and Sweden show a quite rapid productivity development which is much faster than the average over the countries. Finally, Norway turns out to be the leader over the whole period (and thus no convergence can be measured). Dropping these three countries from the sample as well significant convergence at a rate of is found (for details see Table 4.3, column 2). In the Transport sector (I) there is significant convergence for the whole sample although Romania and Slovenia are falling back over the period considered. The convergence parameter for the whole sample is and is highly significant. Dropping the outliers Romania and Slovenia the coefficient of convergence becomes (again a highly significant). The detailed results for this regression are reported in Table 4.3, column (1). Again there is a group of countries which performed much better than the average, namely Greece, Estonia and Germany. Dropping these countries from the sample as well yields a highly significant coefficient of In this case the regression also has a quite high fit (the R 2 becomes 0.79); these results are reported in column (2). In Business Services (JK) the whole sample shows convergence at a rate of but only significant at the 10 per cent level. In this sector Poland was falling behind and for Bulgaria almost no convergence at all takes place although the initial gaps for these countries are quite high. Dropping these two countries from the sample the coefficient of convergence becomes and highly significant. Here one has however to mention that overall productivity growth in this sector is very low (and even negative for the EU15 average as can be seen in Table 4.1). Finally, in Public Services sector convergence is also significant for the whole sample. The Czech Republic may be seen as an outlier as this country falls back over the period. Dropping it from the sample yields a little higher coefficient of convergence and a higher 17

188 overall fit as can be seen in Table 4.3. The coefficient of convergence is also high at a level of Summarizing, we have found significant convergence in productivity levels for all sectors with the exception of Construction (F), at least when removing some countries from the sample. The performance of these countries may be seen as caused by country-specific characteristics or particular developments in the period observed which is particularly the case for some of the NMS and CC-2 countries. Not taking into account of the outliers, the coefficients of convergence are particularly high for Business Services (LQ) and Transport (I) (n the first estimation reported in column (1)) at a level of almost 0.04 (implying a halftime of about 17 years), in a medium range for Industry (CDE), and Public Services (LQ) with a parameter of about 0.30 (implying a half-time of about 23 years), and at a lower level for Agriculture (AB) and Trade, Repair and Restaurants (GH) with a parameter of about 0.02 (implying a half-time of about 35 years). 4.2 Convergence of GDP shares The structure of output The sectoral level of employment not only depends on productivity and its movement over time but also on the share of output of the particular sector in the economy. Let us first look at the sectoral structure of the NMS and CC-2 compared to the EU-15 average. Table 4.4 presents figures with regard to the sectoral structure in terms of value added and employment shares for the EU-15, the NMS and CC-2 for the year Figure 4.2 presents the same information graphically. Given the productivity levels in 2002 and the output shares the employment levels and shares are determined. For completeness these employment shares are also reported Table 4.4 and are shown graphically in Figure 4.3. In these tables and figures the countries are again ranked according to their aggregate productivity level compared to the EU-15 as already introduced above. < Table 4.4 Sectoral output and employment shares > 18

189 SI CZ HU SK PL EE LT LV RO BG EU AB CDE F GH I JK LQ Figure 4.2 Sectoral output shares SI CZ HU SK PL EE LT LV RO BG EU AB CDE F GH I JK LQ Figure 4.3 Sectoral employment shares Let us start with the output shares. In Agriculture (AB) all countries have higher output shares as compared to the EU-15. Closest to the EU-15 is Slovenia with only about three percent, followed by a group consisting of the Czech Republic, Hungary, Slovakia, Poland and Estonia with shares of slightly about five per cent. Lithuania and Latvia have a share of about nine and eight per cent respectively. The countries with the highest shares are the CC-2 (Bulgaria and Romania) with a share between fifteen and almost twenty per cent. 19

190 With respect to the output share in Industry (CDE) again all countries have shares above the EU-15 average. Romania holds the highest share (more than thirty-five per cent), followed by Slovenia and the Czech Republic with about 33 per cent. Hungary and Poland also still have shares above thirty per cent (as compared to 23 per cent of the EU-15) whereas the remaining countries (Slovakia and the three Baltics Estonia, Lithuania and Latvia). Estonia with less than 25 per cent has the lowest shares of all NMS and the CC-2. In Construction (F) no clear pattern emerges: The EU-15 average is at about six per cent; below this are the Czech Republic, Hungary, the Slovak Republic, and Bulgaria); slightly above the EU-15 average is Slovenia and little higher are Poland, Estonia and the three Baltic countries. For the service industries again no clear pattern can be found for Trade, Repair and Hotels (GH) and Transport (I). In the sector Trade, Repair and Hotels (GH) the Czech Republic and Latvia are about three percentage points above the EU15 average of fifteen per cent; Poland and Lithuania have the highest output share with more than twenty per cent. Clear below the EU15 average are Hungary and the CC-2 Bulgaria and Romania. Similarly, no clear-cut pattern can be seen in Transport (I) where the EU-15 average is of about nine per cent. Whereas Slovenia, Poland, and Romania are below this average (with about seven per cent) the other countries have higher shares. The highest share in this sector has Latvia with more than 15 per cent. In the remaining two sectors all countries have lower output shares as compared to the EU15. In Business Services (JK) the output share of the EU-15 is 25 per cent and thus highest compared to all other sectors. Closest to this is Estonia with about 20 per cent; the Czech Republic, Hungary and the Slovak Republic have about 18 per cent followed by Slovenia with 15 per cent. In Poland, Latvia, Lithuania, and the CC-2 the shares are only about ten per cent, i.e. 15 percentage points below that of the EU-15. Finally, in Public Services (LQ) only Slovenia comes close the EU-15 share of 21 per cent. The output shares of the Czech Republic, Romania and Bulgaria are at only slightly more than ten per cent, that of the other countries ranging in between. As already mentioned above the productivity level and the sectoral output shares determine the sectoral employment shares (which also are reported in Table 4.4 above) and graphically shown in Figure 4.3. Whereas in terms of output structure the ranking with respect to the size of the sector in the economy for the EU-15 is Business Services (JK), Industry (CDE), Public Services (LQ), Trade, Repair and Hotels (GH), Transport (I), Construction (F) and finally Agriculture (AB) the most important sector in terms of employment shares is Public Services (LQ) with an employment share of about 30 per cent, followed by Trade, Repair and Hotels (GH) with almost 20 per cent and Business Services (JK) with 15 per cent. The remaining sectors are of similar size with five to seven per cent. Given the productivity performance the NMS and CC-2 have as well higher 20

191 employment shares in Agriculture (AB) and Industry (CDE), show a mixed pattern for Construction (F) and Transport (I). The employment shares are lower in the services sectors: but, whereas in Trade, Repair and Hotels (GH) the difference is rather small (with the exception of Romania), the difference is quite large in Business Services (JK) and partly also in Public Services (LQ). In the latter sector it is particularly high for Romania and Bulgaria. Under the assumptions of productivity convergence and convergence in output shares employment shares must convergence as well; one thus can expect a major shift out of Industry (CDE) and for some countries Agriculture (AB) towards the Business and Public Services (JK and LQ). Given our framework the speed of this employment restructuring depends on the speeds of productivity convergence in the particular sectors (determined by the coefficients of convergence and the initial levels) and the convergence behaviour of output shares to which we turn next. Further, the restructuring in terms of employment shares of an economy may also occur at different paths: In the extreme cases, a reshifting of employment can take place by job destruction in the sectors having higher than average shares (i.e. Agriculture and Industry) without creation of new jobs in the other sectors (particularly in services). This kind of restructuring would imply social costs in terms of high unemployment rates or an increase in the inactivity rates. The second extreme would be job creation in the services sectors and may imply even rising activity rates for the economy as a whole. The framework we use in this paper will us tell also something on job creation/destruction in the sectors. Before, however, we have to estimate the speed of convergence of output shares and if this takes place at all Convergence in output shares One has first to notice that the data for the EU-15 countries show less convergence in shares than in productivity levels. Although there are some common trends, the dynamics of the shares exhibit hysteresis effects, i.e. shares converge if at all at very low rates. But there are also common trends (e.g. a decline in the share of agricultural output). This means that we find convergence using the concept of β -convergence for some sectors but do not find σ -convergence (which is not reported here). 4 There are some reasons for this: first, countries may have different structural patterns as the endowments with natural resources (including tourism) differ across countries, the building of sectoral clusters (e.g. finance activities, industrial zones,...) implies different specialization patterns of countries, etc. Secondly, non-linearities in the dynamics of the shares could imply that common trends are observed across countries, but the variance of the shares of particular industries across countries first rises and later falls (e.g. when this follows a S-shaped pattern over time). Thus, also we use the concept of β -convergence for the dynamics in the shares as 4 Note that β -convergence does not imply σ -convergence, but σ -convergence would imply β -convergence. 21

192 well, we have to be cautious when using the estimates in the scenario analysis later, especially we will have to take into account some specificities of the NMS and CC-2 in their adjustment paths in doing sensitivity analysis. For the dynamics in shares we estimate a similar equation as for productivity convergence. The only difference is that instead of the leader country we take the arithmetic mean of shares of the EU-15 as the benchmark. The equation implies that countries/sectors with above average shares are expected to have a decline in this share and countries/sectors with below average shares are expected to have rising shares. In Table 4.5 we report the results of the regression for each sector. Again we have dropped some outliers from each of the regressions. Of course, if these outliers are NMS or CC-2 we shall take care of this in the scenarios reported later. < Table 4.5 Estimation results for convergence of shares > In Agriculture (AB) the coefficient is significant and implies a half-time of 25 years. The equilibrium share is 2.14, although one has to notice that the constant is not statistically different from zero. In this regression we have dropped Bulgaria which shows a rather high growth rate of output. In Industry (CDE) we dropped Hungary, Romania and Slovenia. Hungary has an exceptionally high growth rate of output whereas Romania and Slovenia are characterized by high initial shares but low albeit positive growth rates. On the other hand, Greece which was also dropped has a very low initial share and a negative growth rate. The regression then gave a highly significant coefficient of (implying a halftime of 14 years) and an equilibrium share of about 25 per cent. For construction (F) we find no significant coefficients at all even after dropping the outliers Bulgaria and Czech Republic. In the next sector, Trade, Repair and Hotels (GH), we find a significant coefficient at the ten per cent level implying a half-time of 35 years. For this we had to drop a number of NMS showing quite high (positive) growth rates of output shares in this sector. For Transport (I) we find again a highly significant coefficient of convergence. Here the outliers are Bulgaria, the Czech Republic, Germany, and Estonia with quite high growth rates. The implied half-time is 27 years and the equilibrium share is about ten per cent. In Business Services (JK) no convergence takes place at all and output shares vary widely also across EU15 countries. Finally, in Public Services (LQ) significant convergence at a half-time of about 24 years is found and the equilibrium share is at about 20 per cent. Here again we dropped a number of NMS (the Czech Republic, Estonia, Hungary, Latvia, and Poland) experiencing negative growth rates and starting from low initial shares as compared to the EU15 average as discussed above. Summarizing, there seems to be a tendency of convergence in shares with the exception of two sectors (Construction (F) and Business Services (JK)). Convergence takes place at rates implying half-times of about 25 years; an exception to this is Industry (CDE) where 22

193 the implied half-time is only 14 years. For the scenario analysis and a forecasting period of ten years this means that if a country has a higher output share in a particular sector of 10 percentage points it would have a higher output share of about 7.5 percentage points after the ten years. Further in the scenarios we also take into account the specificities of the countries which show up in the data so far. 4.3 Scenarios for sectoral labour demand and implications for aggregate employment levels Similar to the aggregate framework we start with the a simple labour demand equation Li = liαiy where L i denotes labour demand in sector i, l i is labour input per unit of output (value added) i.e. the inverse of labour productivity - and Y denotes total output (in our case value added). Additionally we have to take into account the sectoral structure of the economy; this is done by the share of a particular sector i in total GDP. In terms of growth rates this equation can be written as Lˆ = lˆ + ˆ α + Yˆ. The growth rates of the input coefficient and value added shares are determined by the convergence dynamics; for total GDP growth we use as a base scenario a constant growth rate of four per cent per year. Table 4.6 summarizes the values for exogenous growth rates and convergence parameters used in the simulations. The initial gaps have been discussed above (see Tables 4.2 and 4.4). The level of employment is then given by L ( t) = L (0) exp( Lˆ dt). The aggregate employment level is then given by i i i L ( t) = L ( ). i i t i i i < Table 4.6 Parameter values used in scenarios > For the interpretation of the results we shall also refer to a decomposition of the changes according to a shift-share analysis. As L = l α Y (i.e. employment equals labour input per i i unit of output times the output share times total GDP) changes in employment can be decomposed in the following manner: i L i = l α i + ( αi,2002 Y + αiy αi Y ) + li ( αi,2002 Y + αiy αi i, 2002Y2002 li,2002 Y ) with liαi, Y... productivity effect l... structural effect i, 2002 αiy2002 li, 2002α i,2002 Y.. output effect li, 2002 αi Y... structural x output effect li αiy productivity x structural effect l α Y... productivity x output effect i i,

194 The change in the employment level of sector i can be decomposed into changes due to labour productivity changes (or change in the input of labour per unit of output). The change in output can itself be decomposed into a change in total GDP at constant shares and changes of the sectoral share of output. These two effects are referred to as the structural effect and the output effect. Further there are some mixed effects: the structuraloutput effect, the productivity-structure effect and the productivity-output effect. The seventh term is a mixed term which is l α Y. Summarizing across sectors gives the aggregate effect of the terms. i i, 2002 < Table 4.7 Decomposition analysis ( ) > Table 4.7a and 4.7b presents the decomposition analysis for the EU-15 countries, the NMS and CC-2. In Table 4.7b the changes of the various components relative to the aggregate employment level in 1997 have been calculated. Further these relative changes has been divided by the number of years. The scenarios are calculated from 2002 to We first present the implications for total labour demand and then discuss the implications for the sectoral employment dynamics Aggregate employment patterns Let us start to discuss the aggregate employment patterns which emerge from the scenarios. Table 4.8 presents the results for the NMS and CC-2 for the years 2002, 2007 and < Table 4.8 Aggregate labour demand dynamics > In most countries total labour demand is decreasing over this period. The only exceptions are the most advanced NMS Czech Republic, Hungary, Slovenia and the Slovak Republic which show an increase of labour demand between four to six per cent (2012 compared to 2002). For Estonia and Poland we find only small losses in labour demand which drops only by about two per cent relative to the year More severe losses can be expected in Latvia and Lithuania where labour demand drops by about ten per cent. Finally, Bulgaria and Romania experience severe losses of employment of more than 15 per cent as compared to the year In absolute levels this means that Romania looses about one million jobs from and half a million jobs from ; the figures for Bulgaria are and For the four most advanced countries (Czech Republic, Hungary, Slovenia and the Slovak Republic) increases in employment levels can be expected even in the shorter run (i.e. in the period ) according to the simulations. Estonia and Poland show losses in employment in the first period but rising levels in the second half of the simulation period. 24

195 Figure 4.4 presents the historical as well as the projected time series for the ten countries; with the level of employment in 2002 set equal to 1. 5 One can see that the more successful NMS show a U-shaped pattern whereas the other countries are on a downward trend with respect to employment levels. These aggregate figures hide the structural adjustment processes which are underlying the net gains and losses in jobs. Thus we turn next to the structural patterns of the employment dynamics Sectoral patterns of employment dynamics Figure 4.5 shows the evolution of the employment shares for the ten countries where also the historical data from 1997 to 2002 are included. Underlying these graphical representations Table 4.9 presents the absolute number of the persons employed, the changes in absolute terms between 2002, 2007 and 2012 and the employment relative to the year 2002 for all countries and sectors. < Table 4.9 Sectoral changes (Levels) > BG CZ EE HU LT LV PL RO SI SK Figure 4.4 Trends in employment levels < Figure 4.5 Trends in employment shares > 5 Data for Lithuania and Romania were adjusted in levels due to breaks in the time series. 25

196 Let us discuss each country in turn. Let us start with the four most successful and advanced NMS (Czech Republic, Hungary, Slovenia and Slovak Republic). The decomposition analysis for the scenarios is presented in Table 4.10 for the aggregate effects. Appendix Table A.5 presents further data on the projected value added shares, productivity levels and the decomposition analysis by sectors. < Table 4.10 Decomposition analysis for scenarios > Czech Republic The model predicts that the largest shake out of labour in absolute terms occurs in the manufacturing sectors where in the first period more than jobs and in the second period more than jobs will be lost. This amounts to loss of about 20 per cent of jobs in the manufacturing sector (relative to 2002). A similar loss in relative terms is in the transport sector (I), although the absolute numbers are smaller due to the smaller number of employed persons in this sector. Over the whole period about jobs will get lost in this sector. Additionally, about employees are dismissed in agriculture which thus looses about 15 per cent of the employed persons in All other sectors are creating employment: in absolute terms the far largest job creator is the public services sector (LQ) in which employment increases of about in each of the two subperiods are expected; this is followed by the Business Services sector (JK) which creates jobs in the first and more than jobs in the second subperiod. This sector is closely followed by the Trade, Restaurants and Repair with slightly less high absolute numbers of job creation. Finally, in the Construction sector (F) to jobs shall be created in the two subperiods. In relative terms the Business services sector (JK) is the most important with an increase in jobs of about 50 per cent over the whole period. This is followed by construction (F) with more than 30 per cent, Trade, Repair and Restaurants (GH) with more than 20 and finally Public Services (LQ) with slightly less than 20 per cent. Figure 4.5 shows the dynamics of the employment patterns also compared with the EU15 (the shaded boxes). One can see that the Czech Republic will have a higher employment share in the manufacturing sector (CDE) also in the medium run (about 24 per cent as compared to 17 per cent in EU-15) though the share is dramatically falling from more than 30 per cent. The share is also higher in Construction (F) of about 5 percentage points and even rising. A much lower share can be observed in Business Services (JK) and Public Services (LQ) where in 2012 the shares are about five percentage points below the EU average despite the increases in employment in these sectors. Convergence to the EU shares can be observed in Trade, Repair and Restaurants (GH) and Transport (I), whereas in Agriculture (AB) the share tends to a lower level than the EU average. What are the main driving forces behind these shifts? Table 4.10 reports the shift-share analysis for the particular sectors and the total economy. At the economy level about 26

197 jobs will be created over the period This net increase results from a loss of employment due to productivity increases of more than 1.5 million persons, GDP growth accounts for the creation of more than 2.3 million jobs, whereas about jobs are created by shifts in the structure of the economy (towards more labour intensive sectors). Also the mixed terms, especially the term which accounts for changes in productivity and change in output are of considerable size. The lower part of the table shows the changes relative to total employment demand in In all sectors the most important items behind the changes are employment losses due to productivity growth and employment gains due to total GDP growth. Changes in the sectoral composition of the economy play a minor role in most cases with the exception of Public Services (LQ). Hungary Hungary shows a similar overall dynamic pattern as the Czech Republic, however starts from different levels. The main differences are that the manufacturing sector (CDE) starts with lower shares and also has a lower employment share at the end of the period. Construction (F) and Trade, Repair and Restaurants (GH) have slightly higher shares. Finally, Public Services (LQ) show a little higher share than in the Czech Republic, but still below the EU15 level; similarly, the share in Business Services (JK) is about five percentage point below the EU15 in In absolute terms the largest shake-out of labour occurs in Manufacturing (CDE) where more than jobs will be lost over the period Employment losses are also expected in the Transport (I) with a loss of about jobs and agriculture with a loss of about jobs. All other sectors are creating jobs where the most important in absolute terms are Trade, Hotels and Repair (GH) and Business Services (JK) with about jobs. A slightly lower magnitude is expected for the Public Services sector (LQ) with creation of about jobs and less important in absolute terms is Construction (F) with a rise in labour demand of about jobs. In relative terms the most important employment creating sector is Business Services (JK), followed by Construction (F) and Hotels (GH). The least important employment creating sectors is Public Services (LQ). For the employment shedding sectors it turns out that all of them loose about 20 per cent of employment as compared to The most important factor for lower employment demand is again productivity growth which accounts for more than one million losses in employment over the whole period. This is more than compensated by employment growth due to GDP growth which creates about two million jobs. Structural shifts count only for an increase of about jobs in the economy. These shifts account for employment growth even less than in the Czech Republic which partly explains the lower job creation in Hungary. 27

198 Slovak Republic The Slovak Republic shows again a similar pattern as the two countries already considered. In Manufacturing (CDE) about jobs will be lost over the period and little bit less than jobs are lost in Agriculture (AB) and Transport (I), respectively. The other sectors are creating jobs where the most important sector in absolute terms is Trade, Hotels and Restaurants (GH) and Business Services (JK) where in each of these labour demand increases of about jobs. The remaining two sectors are also important with a higher labour demand of about jobs in Public Services (LQ) and about jobs in Construction (F). In relative terms, Business Services (JK) is the most important job-creating sector where labour demand increases of 50 per cent, followed by Construction (F) with 30 and Trade, Restaurants and Hotels (GH) with 20 per cent increase from The labour shedding sectors are very similar in relative terms as each of them looses about 20 per cent of jobs over the simulation period. More than 600,000 jobs are lost due to productivity growth but more than one million jobs are created by GDP growth (which assumed to be four per cent p.a.). The change in the structure of the economy has a slightly negative effect on total employment changes. Slovenia Of these four most successful countries Slovenia starts with a relatively high share of employment in agriculture (about ten per cent in 2002) which decreases to the EU15 level over the period. Also the employment share of almost 35 per cent in manufacturing is relatively high compared to the other countries already discussed. Given these facts the shares of employment are lower in Construction (F) and mainly in Public Services (GH) when compared to the other countries. In relative terms the employment losses in agriculture are higher than in the other countries (about 30 per cent of the level in 2002), whereas losses in Manufacturing (CDE) with about 20 per cent are at a comparable level and employment destruction in Transport (I) is even lower with about 15 per cent. On the other hand, job creation in Public Services (LQ) with 25 per cent is more than double compared to the other countries and at the same level for Business Services (JK) with about 50 per cent. Employment creation in relative terms is also somewhat higher in Construction (F) and Hotels and Restaurants (GH). In absolute terms the most important labour shedding sector are Manufacturing with jobs and Agriculture (AB) with jobs. The most important employment creating sector are Public Services (LQ) with more than jobs, Hotels and Restaurants (GH) with jobs and Business Services (JK) with more than jobs. The less important sector is Construction (F) with less than jobs. Productivity growth is again the most important reason for job losses, although slightly less important than in the other countries which reflects the fact that Slovenia is already closer 28

199 to the EU15 productivity levels on average. Structural change has a slightly higher effect on aggregate employment levels as compared to the other countries which mainly reflects the employment losses in Agriculture (AB) for which employment shares drop from ten to five per cent. But still the effect of structural change is negligible compared to the other components of the shift-share analysis. Let us next come to the group of countries which should experience a decline in employment levels according to the simulations. These are Poland and the three Baltic countries Estonia, Latvia and Lithuania. Estonia Let us start with Estonia as the structure of this country in 2002 is similar to the ones discussed before. At the end of the simulation period the share in Manufacturing (CDE) is relatively low with 20 per cent; the share in Public Services (LQ) and Business Services (JK) becomes high compared to the other countries. Labour shedding is highest in Manufacturing (CDE) with a loss of jobs, followed by Transport (I) in which employment is reduced by and Agriculture (AB) with a reduction of about jobs. The most important job creating sectors are Business Services (JK) and Hotels and Restaurants (GH) where labour demand is rising by more than jobs. Labour demand in Construction (F) will increase by about jobs and the least important sector in absolute terms is Public Services (LQ) where an increase of about is expected. In relative terms the fall in employment in the labour shedding sectors is higher than in the countries discussed before and is less than between 25 and 30 per cent of the level in On the other hand, job creation is lower than in the other countries and is 30 per cent above the level in 2002 in Business Services (JK), 27 per cent in Construction (F), 13 per cent in Trade, Hotels and Restaurants (GH) and only five per cent in Public Services (LQ). This reflects the fact that productivity catching-up is even more important as the initial productivity gap is higher on average. The decomposition analysis shows that the productivity effect is larger than in the other countries discussed so far. The effects of changes in shares are rather small but positive. However, the total effect on employment demand remains negative. The following countries are different as they start with a relatively high share of employment in agriculture and thus follow a somewhat different pattern than the countries before. 29

200 Poland Poland starts off with a share of employment in Agriculture (AB) of slightly less than 20 per cent, a share of 25 per cent in Manufacturing (CDE) and relatively lower employment shares in Public Services (LQ). The scenarios show that in 2012 the share in Agriculture (AB) is still at a level of about 13 per cent, has declined in Manufacturing to 19 per cent and mainly risen in Trade, Hotels and Restaurants (GH), Business Services (JK) and Public Services (LQ). In relative terms job destruction is highest in Agriculture (AB) where labour demand is more than 35 per cent below the level of In the other two sectors with job destruction the relative decreases are similar to the other countries with 17 per cent in Manufacturing (AB) and 20 per cent in Transport (I). In absolute terms this implies that one million jobs is lost in Agriculture (AB), more than half a million in Manufacturing (CDE) and additionally jobs are lost in Transport (I). This decline in labour demand is not compensated by increases in other sectors in the first period; the net effect on employment is however positive in the second period. Here the most important sectors in absolute terms are Trade, Hotels and Restaurants (GH) and Public Services (LQ) with an increase of more than jobs each. About jobs shall be created in Business Services (JK) and Construction (F). The far most important source of job losses are again increases in productivity levels which accounts for a loss in demand of almost five million jobs. Total GDP growth contributes however more than 6.7 million jobs, whereas the total effect of a change in the structure accounts for a loss of about jobs. However, the interaction term is rather low (-2.1 million) such that the net effect becomes negative. Latvia and Lithuania These two countries start off from rather similar positions. The main difference is that Lithuania has a higher share of employment in Agriculture (AB) 17 per cent in Latvia and 21 per cent in Lithuania but lower shares mainly in Transport (I) and Business Services (JK). The decrease in labour demand is highest in Agriculture (AB) in relative terms (about 35 per cent will get lost) and also in absolute terms ( in Lithuania and in Latvia). The second most important sector in absolute terms is Manufacturing with in Lithuania and in Latvia. More important in relative terms but less so in absolute terms is Transport with a decrease of about jobs in Lithuania and about in Latvia. In Public Services (LQ) Latvia will additionally loose jobs, whereas employment in this sector in Lithuania remains more or less stable. For the other sectors there are only slight differences in relative terms. 30

201 With respect to the main causes of the employment decline the same as for Poland can be said and shall not be repeated here. Let us next turn to the countries which according to the scenarios shall experience the most severe losses in employment, namely Bulgaria and Romania. Bulgaria Bulgaria starts off with almost equal shares of employment in Agriculture (AB) and Manufacturing (CDE) but relatively low shares in Public Services (LQ). Relative to the other countries also the share in Trade, Hotels and Restaurants (GH) are low with 13 per cent. In absolute terms the by far most important sector with regard to losses in employment is Agriculture (AB) where more than jobs will get lost over the period. This is followed by Manufacturing (CDE) where a loss of about jobs is expected and Transport (I) with In absolute terms the creation of jobs is rather small: the most important are Construction (F) and Trade, Hotels and Restaurants (GH) with less than jobs created, Business Services (JK) with about and finally Public Services (LQ) with only In relative terms this means that employment drops by more than 40 per cent in Agriculture (AB), about 30 per cent in Transport (I) and 30 per cent in Manufacturing (CDE). The most important gains in jobs in relative terms are in Construction (F) with about 30 per cent and less than 20 per cent in Business Services (JK). In Trade, Hotels and Restaurants (GH) the number of employees will rise at eight per cent and it will remain almost stable in Public Services (LQ). The effect of increases in productivity on employment is much stronger than in the other countries and in absolute terms is only slightly lower than the positive effect of total GDP growth. Again, the total effect of structural change is rather small. Romania Finally, the country with the highest share of employment in Agriculture (AB) is Romania with about 40 per cent. Correspondingly, the shares in all other sectors are much smaller. The dynamic pattern exhibits dramatic changes in employment structures: The share of employment in Agriculture (AB) drops from 40 to 25 per cent and rises from 10 to 15 per cent in Trade, Hotels and Restaurants (GH) and from less than 15 to more than 20 per cent in Public Services (GH). In the other sectors changes are less dramatic. The share of employment in Business Services (JK) is very small with about 3 per cent which remains below five per cent at the end of the simulation period. 31

202 According to this scenario employment decreases in Agriculture (AB) to about 57 per cent of the initial level, in Manufacturing (CDE) to about 73 per cent and in Transport to 75 per cent. In absolute terms this means that almost 1.5 million employees will loose their jobs in Agriculture (AB), almost in Manufacturing (CDE) and in Transport (I). Although relative increases in employment are sometimes high (e.g. 65 per cent in Business Services (JK) and 30 per cent in Construction (F)) in absolute terms the increases are rather small as the employment shares are rather small in these sectors. The most important job creating sector are Trade, Hotels and Restaurants (GH) with an increase of about jobs and Public Services (LQ) with about jobs. Further employment demand will rise in Construction (F) and Business Services (JK) by about jobs in each sector. Again the effect of productivity increases has the largest negative impact on employment demand. In this case similarly to Bulgaria the effect of total GDP growth is only slightly higher in absolute terms. Additionally, the effect of structural change is negative and stronger than in the other countries, although it is small relative to the other terms. 5 Changes in demand for occupations and educations 5.1 Labour demand by occupations In a next step we use the breakdown of the LFS data by occupations (BCLS; BCHS, WCLS,WCMS, WCHS) and analyse the dynamics of employment patterns by these occupational categories. At the aggregate level changes in the occupational structure can result from (i) changes in the sectoral structure of the economy (at constant occupational shares within sectors) and (ii) from changes of the occupational structures within sectors. For the latter we assume that the convergence parameter to the EU-15 occupational structures in 2002 are the same as for labour productivity convergence in the respective sectors. Before presenting the scenario let us shortly discuss the structure of occupations. Table 5.1 reports the occupational structures of the NMS and CC-2 and the mean of the EU-15 whereas Figure 5.1 shows the deviations from the EU15 mean. < Table 5.1 Occupational structures > < Figure 5.1 Occupational structures (deviations from EU15 mean) > In most countries the structure of occupations for each sector deviate less then ten percentage points from the EU-15 means. The only exception is Agriculture (AB) where 32

203 especially the Czech Republic, Hungary and the Slovak Republic have a much lower ratio of BCLS jobs. 5.2 Labour demand by educational levels Similarly to the occupational structures we report the employment shares of educational groups by sectors in Table 5.2 and the deviations from EU-15 means in Figure 5.2. < Table 5.2 Employment demand by education and sectors > < Figure 5.2 Educational structures (deviations from EU15 mean) > One can see that the medium educated persons (ME) are overrepresented relative to the EU-15 in all sectors which reflects the supply side differences between the NMS, CC-2 and the EU-15 countries. Finally, Table 5.3 presents educational shares within the five occupational groups and Figure 5.3 the deviations from the EU-15 mean. < Table 5.3 Educational shares by occupations > < Figure 5.3 Educational shares by occupations (deviations from EU15 mean) > One again finds that the low educated are underrepresented relative to the EU-15 mean whereas the medium educated employees are overrepresented again reflecting the supply side of the economies. For most countries also the high educated people are underrepresented in all occupational categories. These results mean that the supply side of the economy play a role in the educational structures by sectors and occupations. 5.3 Scenarios Let us now turn to the scenario results for occupational categories and educational groups. Here we report only the aggregate level and do not go into details with respect to the sectoral structures Demand for occupations Let us now come to the scenarios with respect to education levels and occupations. For occupations we assumed convergence to the EU15 mean using the same convergence parameters as for productivity. Figure 5.4 shows the results of the changes in employment 33

204 demand by occupations for the ten countries. The absolute numbers are presented in Table A.6 in the appendix. < Figure 5.4 Employment demand scenarios by occupations > The group which suffers most from the ongoing changes are the Blue-collar-high-skilled workers (BCHS). In the more advanced NMS, the Czech Republic, Hungary, Slovak Republic and Slovenia demand for this group decreases by about 10 percent, whereas demand for the other groups is increasing. The only exception here is Slovenia where demand for the Blue-Collar-Low-Skilled (BCLS) is also decreasing to about 95 per cent. The increase in demand is highest for the White-Collar-Low-Skilled (WCLS) and White- Collar-Medium-Skilled (WCMS) group with an increase of 20 per cent (little bit less for Hungary). A similar pattern is found for Estonia. For the other countries of the less advanced NMS (Poland, Latvia and Lithuania) demand for the Blue-Collar-Low-Skilled (BCLS) is falling dramatically to a level of 75 per cent relative to the year 2002 whereas demand for the Blue-Collar-High-Skilled workers is shrinking by about 10 per cent. Demand for the other groups is increasing (an exception is Latvia where demand for the White-Collar-High- Skilled (WCHS) is decreasing as well) at partly high rates. A similar pattern can also be seen in Bulgaria and Romania where demand for BCLS and BCHS workers decreases between 25 and 35 per cent. In Bulgaria employment for the other groups remain more or less constant whereas in Romania employment levels of the other groups are even rising. What causes these shifts in demand by occupations. Similar to above one can decompose the total change in the following way: L io = Lγ γ i io + L γ γ i io + Lγ γ i io + L γ γ i io + Lγ γ i io + L γ γ i io + L γ γ i io where L io denotes employment in sector i and occupation o, L is total employment, and γ i and γ io denote the share of employment in sector i and the share of occupation o in employment of sector i. Table 5.4 presents the results of this decomposition analysis (the number are again expressed relative to total employment in 2002; in this table the figures are not annualized). < Table 5.4 Decomposition of changes in demand for occupations > First, the change in demand for a particular occupational group due to overall changes in employment is proportional to its share in total employment. The change in demand due to 34

205 changes in the sectoral structure is negative for the groups BCLS and BCHS in all countries. For the countries with relatively lower share of employment in agriculture (the Czech Republic, Hungary, Slovak Republic and Slovenia) the effect of structural change is particularly strong for the BCHS workers, whereas in countries with a higher initial employment share in agriculture the effect of structural change is particularly strong for the group BCLS (especially so for Bulgaria, Romania, Lithuania and Poland). In all countries there is a positive effect of structural change on demand for WCLS and WCHS workers (where in most countries demand is increasing for this groups). Finally, changes in demand due to changes of the occupational structures are particularly strong and negative for the BCHS workers. Thus for this group demand is relatively decreasing because of changes in the sectoral structure of the economy and changes in the occupational structures within sectors Changes in demand for educational levels For reasons already discussed above we assume that the educational structure within sectors and occupations remains constant over time. This implies that changes in demand for educational groups results only from the dynamics of total employment, of structural changes and of changes in the occupational structures. Figure 5.5 presents the results for the educational groups. The levels and changes in absolute figures are presented in Table A.7 in the appendix. < Figure 5.5 Scenario: Employment demand by educational groups > The relative pattern of the dynamics of demand for educations is similar in all countries. The group which is suffering most from the ongoing changes are the low-educated persons; demand is relatively rising for high-educated persons in all countries. However there are differences across countries with respect to overall demand for educations. Whereas in the Czech Republic, Hungary and the Slovak Republic demand for HE and ME is increasing of about 15 and 5 per cent, respectively, demand for the LE is constant decreasing only slightly. The situation is already different for Slovenia, where demand for the LE is falling to 90 per cent of the 2002 level whereas demand for ME and HE is rising even more than in the other three countries discussed before. A similar picture also arises for Poland where demand for LE is falling to about 80 per cent, demand for the ME remains constant and demand for HE is rising to a 10 per cent higher level than in Estonia show the least differentiated picture with respect to relative developments. There are only slight increases in demand for HE and only slight decreases in demand for ME and LE. In Latvia and Lithuania demand for all groups is decreasing (demand for HE remains more or less constant in Lithuania) and even more so for the LE group. 35

206 Finally, the situation for the low educated persons in Bulgaria and Romania are even worse. In Bulgaria 30 and in Romania about 35 per cent of the low educated employees will loose their jobs. The situation is less dramatic but still very severe for the medium educated group where employment is 15 per cent in Bulgaria and 10 per cent in Romania below the 2002 level. Demand for the high educated group in Bulgaria will fall only of about three percent and is even rising in Romania. The decomposition analysis is similar to the analysis for occupational groups above as the share of educational groups within sectors and occupations are assumed to be constant and can thus be written as L ioe = γ ioe ( Lγ γ + L γ γ + Lγ γ + L γ γ + Lγ γ + L γ γ + L γ γ io ) i io i io i io i io i io i io i Summing up over sectors and occupations shows the changes for educational groups. These results are presented in Table 5.4 (again we report the changes relative to total employment in 2002 and not annualized). < Table 5.4 Decomposition of changes in demand for education > Similar to above the effect changes in overall labour demand the demand by educational groups is proportional to the overall employment shares. Let us therefore mainly discuss the effects of structural changes and the effects of changes in the occupational structures within industries. First, the effect of structural shifts on demand for the low educated workers is negative in all countries and particularly high (in absolute terms) for Romania and Bulgaria which have a high initial share of employment in agriculture (AB). For the medium educated group the effect of structural change are on demand is mixed: It is negative for Czech Republic, Estonia, Hungary, Lithuania and Slovak Republic and positive for the other countries. For the group of high educated employees the effect of structural change on demand is positive. 6 Conclusions This paper introduced a framework for forecasting employment levels and structures by sectors, occupations and educational attainments. The framework is based on the idea that countries having lower productivity levels have a higher potential for productivity growth and thus converge to the levels of the technological leaders (which is similar to the convergence processes formalized in the growth literature and already argued by Alexander Gerschenkron s idea of the advantage of backwardness ). The same convergence process was assumed for sectoral shares in value added. As productivity levels converge the real income levels of the follower countries get closer to that of the leader countries which then leads to similar expenditure structures (i.e. the income effects 36

207 on demand structures are important). For application of this framework we estimated aggregate and sectoral convergence parameters for productivity levels and value added shares. Given the initial values for the NMS and CC-2 the variables can be forecasted. These variables together with an assumed total GDP growth then allows to forecast levels of employment. This framework was then also extended for an analysis of the structural developments with respect to occupational categories and demand for educational attainments. The main results according to the scenarios can be summarized as follows: (1) In terms of aggregate employment levels the more advanced NMS (Slovenia, Czech Republic, Slovak Republic, and Hungary) are already or will be in the next few years in a phase of rising employment levels although not at very high rates (i.e. unemployment rates or inactivity rates remain quite high). Another group of countries, i.e. Poland and the Baltic states, do not show rising employment but show a modest decrease in employment levels with structural adjustments taking place mainly due to the high share of employment in agriculture. Finally, Bulgaria and Romania face a severe decline of employment levels in the next decade (about 15 per cent of employment in 2002) mainly caused by high productivity increases (due to the large gap) and the high initial share of agriculture. (2) The structural shifts of employment patterns are very similar across countries with lower shares in Agriculture and Industries and higher shares in the service sectors. (3) With respect to occupational categories the group suffering most from the ongoing restructuring are the Blue Collar High Skilled and Blue Collar Low Skilled workers. For the first group a decline in demand is predicted in all countries; for the BCLS a decreasing demand is predicted in all countries except Czech Republic, Hungary and Slovak Republic. Demand for the other groups are in most cases rising or at least stable. However, demand is also declining for the Blue Collar Low Skilled workers in Slovenia, (4) With respect to educational attainment groups there is a clear picture: The group suffering most in relative terms are the low educated employees. Demand for this group is almost stable in the successful NMS Czech Republic, Hungary and Slovak Republic and modestly decreasing in Slovenia and Estonia. A strong decrease in demand for the low educated persons is predicted for the other countries whereas demand for the other groups (medium educated and high educated) remains more or less stable or is even increasing. Of course, the scenarios depend on the assumptions with respect to convergence processes, total GDP growth etc. So further research has to tackle potential caveats of this approach and to undertake a number of sensitivity analysis with respect to the various 37

208 assumptions. Let us mention some of the adjustments which seem reason able to include in further research: (1) The model sketched in section 2 allows for a much wider range of potential convergence trajectories (e.g. by introducing the learning parameter) and would even allow for falling behind if the initial gap is too large. This was not yet used in this paper. Similarly one could allow e.g. a S-shaped pattern of technological catching-up. So one potential extension is to allow for different convergence patterns of productivity. (2) In the framework used so far we have not allowed for comparative advantage structures which would imply different sectoral output shares in the longer run. Such comparative advantage structures could arise because of endowments with natural resources (e.g. tourism, etc.), human capital stock, path-dependent structures (i.e. agglomeration effects as the automobile cluster), etc. Such structural aspects can be taken into account in the model by assuming target levels for sectoral value added shares. (3) Similarly, we have assumed that the share of the public sector becomes equal across countries over time. This may not be the case as the NMS may not converge to the typical welfare state structure as most of the EU-15 countries has done. This can again be taken into account by assuming different target levels for this sector. (4) Finally, we assumed a constant and exogenous growth rate of GDP in the scenarios. Although the average growth rate over the next decade can be at this level the assumption that it is the same for all countries is not justified as e.g. the initial shares differ (i.e. some countries have a higher initial share of a sectors with above average growth rates) and thus the aggregate growth rate may differ, the exposure to external markets is different (as e.g. some of the countries are already EU-members, the CC-2 are not), the fiscal and monetary policies differ across countries, etc. Such considerations can be easily taken into account by recalculating the scenarios with different growth rates. Although these caveats should be taken into account in further research the model seems to be robust in a qualitative sense to these potential adjustments within the period for which the scenarios were calculated (10 years) as given the convergence parameters there is no full catching-up to the EU-15 levels in productivity or shares such that country specific differences still matter. 38

209 References: Barro, R.J. and X. Sala-i-Martin (1999): Economic Growth, MIT Press, Cambridge/M. Bernard, A.B. and C.I. Jones (1996): Comparing Apples to Oranges: Productivity Convergence and Measurement Across Industries and Countries, American Economic Review, 86(5), Bernard, A.B. and C.I. Jones (2001): Comparing Apples to Oranges: Productivity Convergence and Measurement Across Industries and Countries: Reply, American Economic Review, 91(4), Gerschenkron, A. (1952): Economic Backwardness in Historical Perspective, in: Hoselitz (ed.): The Progress of Underdeveloped Areas, Chicago. Gouyette, C. and S. Perelman (1997): Productivity Convergence in OECD Service Industries, Structural Change and Economic Dynamics, 8(3), Landesmann, M.A. and R. Stehrer (2004): Income Distribution, Technical Change and Dynamics of International Economic Integration, wiiw Working Paper, forthcoming. Landesmann, M.A. and R. Stehrer (2003): Technology Diffusion, International Competition and Effective Demand, Revue d Économie Industrielle, 105(1), Landesmann, M.A. and R. Stehrer (2002): Technical Change, Effective Demand and Economic Growth, in Salvadori, N. (eds.): Old and New Growth Theories: An Assessment, ch. 9, Edward Elgar, Cheltenham. Solow, R.M. (1956): A Contribution to the Theory of Economic Growth, Quarterly Journal of Economics, 70(1), S renson, A. (2001): Comparing Apples to Oranges: Productivity Convergence and Measurement Across Industries and Countries: Comment, American Economic Review, 91(4), Stehrer, R. (2002a): Technical Progress in a Dynamic Input-Output Model with Heterogenous Labour, in: Cowan, R. and N. Jonard (eds.): Heterogenous Agents, Interactions and Economic Performance, Lecture Notes in Economics and Mathematical Systems Series, Ch. 20, Springer-Verlag, Berlin. Stehrer, R. (2002b): Dynamics of Trade Integration and Technological Convergence, Economic Systems Research 14(3), Verspagen, B. (1991): A New Empirical Approach to Catching Up or Falling Behind, Structural Change and Economic Dynamics, 2(2),

210 Tables see File Employment education occupation Tables.xls 40

211 Economic Restructuring and Labour Markets in the Accession Countries Research Project commissioned by EU DG Employment, Social Affairs and Equal Opportunities, Contract No. VC/2003/0367 Coordinated by in cooperation with Alphametrics and The Vienna Institute for International Economic Studies FDI and the skill composition of the workforce: the case of the electronics industry in Hungary by Kushal Kataria* and Harald Trabold* September 2004 *) German Institute for Economic Research (DIW), Berlin

212 Executive summary 1 Introduction As most industries in the accession countries, the electronics industry in Hungary underwent fundamental changes in the early 1990s as a result of the transformation from central planning to market coordination. The old firms were either closed, restructured or sold to investors. As a result, the industry experienced a substantial decline in production and employment in the early years of transformation. By 1993 growth in the electronics industry started to pick up and by the end of the 1990s the electronics industry in Hungary had developed into the most competitive and biggest of all accession countries. Multinational enterprises (MNEs) contributed substantially to this success by establishing new production facilities and buying some of the old firms. MNE activities can have substantial impact on a host country s economy, e.g. on production, foreign trade, and the labour market. In particular, as they transfer new technologies and organizational knowledge, they may change the supply and demand for skills, not only within their own affiliate but also in the rest of the economy. The purpose of the case study is to analyse the restructuring of the electronics industry in Hungary and to assess the impact of FDI inflows into the Hungarian electronics industry on the skill composition of the workforce. The second section of the case study looks at how and why the FDI inflow to Hungarian electronics industry, which was at its peak during the late 1990s, suddenly fell by a huge margin at the turn of the century. The third section elaborates on the link between MNE activities in the host country and skill composition of the workforce. The fourth section looks at how MNE activities changed the demand for skills in Hungary and its electronics industry, the fifth section focuses on skill supplies. Section 6 concludes the case study. 2 Changes in FDI Inflows to Hungarian Electronics Industry Hungary has quite a tradition in producing electrical and electronic goods. During the 1980s, a small albeit competitive computer industry grew and not only exported computers to the COMECON but also to some Western European countries. From 1989 onwards, the Hungarian electronics industry came face to face with international competition. As a result, production declined substantially in the early nineties. However, earlier than all other transformations countries, Hungary (and its electronic industry) benefited from FDI inflows. Many multinational enterprises established their production and assembling plants in Hungary in the 1990s. In the year 2000, four out of the five largest foreign affiliates in Hungarian manufacturing were in the electronics industry. 1

213 Several factors contributed to this early and significant inflow of FDI into Hungary and its electronics industry. First, Hungary had opened up its economy as early as 1972 to foreign investors, allowing them a minority participation (less than 50%) in a joint venture. Second, since 1968 Hungarian firms were allowed to work as subcontractors for Western firms. These business connections allowed them to establish trust, prove their reliance and acquire knowledge on foreign markets and business practices which in turn helped them to attract FDI after Third, a combination of low wages and skilled labour, fiscal and regulatory incentives and its history as a producer of electrical and electronic goods made Hungary a preferred target among the accession countries for investment in the electronics industry. Fourth, Hungary s privatization program, aimed at selling strategic state-owned enterprises, began much earlier than in other Eastern European countries (mainly to foreign investors). However, with the turn of the century, Hungary s electronics industry saw a rather dramatic turnaround. There was a sudden fall in FDI inflows. Many multinational enterprises closed their operations in Hungary and moved to other locations in Asia and Europe. A number of developments took place, globally and locally, in the end of 1990s and early 2000s that contributed to the sudden fall in FDI inflows, hence prompting the restructuring of Hungarian electronics industry. On the international level these factors include the burst of the so-called IT-bubble, the stagnating West European economy, and the events of September 11, 2001 which were followed up by the run-up to a war in Iraq. On the local level, several factors - such as rising unit labour costs, changes in the tax system and a shortage of qualified labour made the Hungarian electronics industry less attractive for MNCs. It seems that the stream of FDI pouring into Hungary and the other accession countries has stabilized. Even after enlargement it is unlikely that FDI flows will rise substantially, as the domestic factors remain rather unchanged and as the benefits of a worldwide recovery with respect to FDI inflows will largely accrue to China, India and several other Asian countries. Econometric estimates using gravity models of FDI also suggest, that the potential for additional inflows to Eastern Europe is limited 3 On the link between MNE activity and the skill composition in the host country It is widely accepted that many MNE activities have an impact upon the level, growth, quality and wages of their own and the home country s labour force. However, there is neither a general rule nor an established theory which could predict the effect of MNE activities on the skill composition of the labour-force. A hypothesis entertained quite frequently in the literature is that MNEs from high income countries investing in middle income countries raise the demand and supply for skills. Various studies point towards 2

214 such a positive relationship between FDI and the skill composition of the workforce. Sections 4 and 5 look at the factors behind structural skill change. 4 MNE activities and the demand for skills Compared to local firms, MNEs possess several advantages, e.g. ownership and internalization advantages. Ownership advantage comes in the form of knowledge assets, i.e., patents, proprietary technology, processes, etc. By transferring some or all of these assets to their affiliates, MNEs raise the demand for skills, as most of these assets can only be employed productively with highly qualified personnel. The transfer of these assets to the host country takes place in several ways. First, FDI by MNEs usually implies to the transfer of competency and technology to affiliates under multinational firms ownership and control. This can take the form of new production or management techniques, whose implementation can lead to a boost in demand for skilled workers. Second, MNEs may chose to establish or expand their R&D activities in the host country, thereby raising the demand for skills (see Table 4 for R&D activities of MNEs in Hungary). Third, besides endowing their affiliates with new technology, MNEs can also transfer new technology to domestic firms in the host country. This form of transfer can be market mediated or via non-market channels. The market mediated transfer can take the form of patent licensing, minority joint-ventures, subcontracting, transfer of knowledge assets to suppliers, etc. Fourth, new technologies can also reach domestic firms through non-market channels in the form of spillovers. Fifth, subcontracting proves to be an effective corporate adjustment strategy for a host of domestic manufacturers at the time of regime change. 5 MNEs and the supply of skills MNE activities do not only increase the demand for skills in the host country, they also impact on the supply side. This might occur directly through training of staff or provision of scholarships or indirectly through inter-firm labour mobility or a raise in the skill premium which creates an incentive to acquire (additional) skills. MNEs directly affect the skill composition of workforce by training their employees. This training affects most levels of the company hierarchy and can also extend to the MNEs suppliers, subcontractors and customers. On average, MNEs provide more training than their local counterparts. TNCs also support skill development by providing scholarships to individuals and by supporting the development of local educational institutions. They can also provide assistance and advice through membership and advisory boards, curriculum review committees, councils and senates. Nokia and Tateyama are examples of multinationals 3

215 operating in Hungary that support students and educational institutions and hence contribute to the process of skill development in Hungary (see Table 5). Some of the skills developed through training and education spill over to domestically owned firms as a result of employee turnover. This knowledge may or may not be firm specific. The level of mobility of labour force determines whether these spillovers take place at all and whether they remain within the region or spread through the entire economy. As established earlier, MNE activities in Hungary positively affected the demand and supply and thus the wages for skilled workers. Together with skill biased technological progress and institutional changes during transition, MNE activities resulted in an increase of the wages of the high-skilled relative to the low skilled. This rise in the skill premium can be observed in most transition economies and the response to the rising skill premium is evident from the level of education of the school leavers in Hungary. Table 6 reveals that the number of students with college and university degrees has increased steadily since It almost doubled during the decade after the fall of the iron curtain while it grew only 5% in the previous decade. A similar trend can be observed for the secondary school leavers whose number increased by more than 35%. On the other hand, the number of school leavers from primary or vocational schools declined substantially during the same period. 6 Conclusion Our analysis of MNE activities on the skill composition of the workforce for the case of the electronics industry in Hungary clearly shows that the demand for and the supply of skills increased as a result of various effects. The transfer of modern production technology, organizational knowledge and the relocation of R&D activities by MNEs to foreign affiliates in Hungary substantially increased the demand for skills. The role of spillovers is difficult to ascertain. While dubbed important in the theoretical literature, they are generally hard to measure empirically. In the case of Hungary and its electronics industry they don t seem to have played a major role in the past. Subcontracting by MNEs contributed to an upgrade of the technological and organizational capabilities of local firms, thereby raising the demand for skills at least in the subcontractor s plants. MNE activities also increase the supply of skills. They train their own employees, both in the host and home country and provide scholarships and other support to individuals. In addition, some MNEs have established training centres in which they provide training for employees of other firms. As trained staff moves from one firm to another they create spillover effects and raise the supply of skills outside the firm originally providing the training. The rise in skill premium which is at least partially a consequence of MNE activities provides an incentive to individuals to invest in human capital formation. At least in the case of Hungary there is no doubt that MNEs substantially contribute to the upgrading of skills of the workforce. 4

216 Kushal Kataria and Harald Trabold FDI and the skill composition of the workforce: the case of the electronics industry in Hungary 1 Introduction As most industries in the accession countries the electronics industry in Hungary underwent fundamental changes in the early 1990s as a result of the transformation from central planning to market coordination. The old firms were either closed, restructured or sold to investors. As a result, the industry experienced a substantial decline in production and employment in the early years of transformation. By 1993 growth in the electronics industry started to pick up and by the end of the 1990s the electronics industry in Hungary had developed into the most competitive and biggest of all accession countries 1. Multinational enterprises (MNEs) contributed substantially to this success by establishing new production facilities and buying some of the old firms. MNE activities can have substantial impact on a host country s economy, e.g. on production, foreign trade, and the labour market. In particular, as they transfer new technologies and organisational knowledge, they may change the supply and demand for skills, not only within their own affiliate but also in the rest of the economy. The purpose of this study is to analyse the restructuring of the electronics industry in Hungary and to assess the impact of FDI inflows into the Hungarian electronics industry on the skill composition of the workforce. The second section looks at how and why the FDI inflow to Hungarian electronics industry, which was at its peek during the late 1990s, suddenly fell by a huge margin at the turn of the century. The third section elaborates on the link between MNE activities in the host country and skill composition of the workforce. The fourth section looks at how MNE activities changed the demand for skills in Hungary and its electronics industry, the fifth section focuses on skill supplies. Section 6 concludes the paper. A major problem for the empirical analysis is that in most national and international statistics there is no such a thing as the electronics industry. An effort to overcome such difficulties with data is The Yearbook of World Electronics Data compiled by Reed Electronics Research (Reed, 2003). They define the electronics industry as comprising 10 major product groups 2 and collect data on production and trade from national and international sources. We use this source quite frequently in the second section as it offers 1 2 Reed (2003, p. 77). See Reed (2003, pp ) for a complete list of products covered by their definition of the electronics industry. Software is not included this definition. 5

217 internationally comparable data on the electronics industry. However, we complement this information from other sources whenever appropriate. 2 Changes in FDI Inflows to Hungarian Electronics Industry Hungary has quite a tradition in producing electrical and electronic goods. During the 1980s, a small albeit competitive computer industry grew and not only exported computers to the COMECON but also to some Western European countries. In addition, before the fall of the iron curtain, radios, (colour) TVs and radios were produced in Hungary (Reed, 2003, pp.77-78). From 1989 onwards, the Hungarian electronics industry came face to face with international competition. Outdated production technologies (by international standards), low productivity, lack of finance and limited bargaining power severely impeded the operation of the local manufacturers. As a result, production declined substantially in the early nineties. As in other accession countries, the old firms were either driven out of the market or underwent substantial restructuring (Radosevic, 2002, pp.2-4). However, earlier than all other transformations countries, Hungary (and its electronic industry) benefited from FDI inflows (Kaminski/Riboud, 2000, p. 5). Many multinational enterprises established their production and assembling plants in Hungary in the 1990s. 3 In the year 2000, four out of the five largest foreign affiliates in Hungarian manufacturing were in the electronics industry (UNCTAD, 2003a, table 88). This inflow of FDI helped Hungary s electronics industry to become internationally competitive and achieve record level growth rates. With an average annual growth rate of 70% between 1996 and 2000 it outperformed all other countries of the world by a wide margin (see table 1) and made it the largest producer of electronics in Central and East Europe. In the year 2001, 39% of CEEC s production and 50% of CEEC s exports of electronics originated in Hungary (Reed, 2003, pp ). Hence, electronics became one of the largest manufacturing industries in Hungary. Several factors contributed to this early and significant inflow of FDI into Hungary and its electronics industry (Kaminski/Riboud, 2000, pp. 4-5). First, Hungary had opened up its economy as early as 1972 to foreign investors, allowing them a minority participation (less than 50%) in a joint venture. Although these provisions had almost no impact on foreign investment and technology transfer until 1989, as multinationals mainly established their presence in Hungary. Their presence, however, became very useful when the iron curtain fell and these multinationals could start to invest and expand their operations in Hungary 3 In 2001, foreign owned enterprises in Hungaria s manufacturing sector had the highest share in employment and sales of the major accession countries (Hunya, 2004, pp.3-5). 6

218 immediately. Second, since 1968 Hungarian firms were allowed to work as subcontractors for Western firms. These business connections allowed them to establish trust, prove their reliance and acquire knowledge on foreign markets and business practices which in turn helped them to attract FDI after Third, a combination of low wages and skilled labour, fiscal and regulatory incentives 4 and its history as a producer of electrical and electronic goods 5 made Hungary a preferred target among the accession countries for investment in the electronics industry. Fourth, Hungary s privatisation program, aimed at selling strategic state-owned enterprises, began much earlier than in other Eastern European countries (mainly to foreign investors). This provided an excellent opportunity for MNEs to enter the electronics industry in Hungary. However, with the turn of the century, Hungary s electronics industry saw a rather dramatic turnaround. There was a sudden fall in FDI inflows. Many multinational enterprises closed their operations in Hungary and moved to other locations in Asia and Europe. According to UNCTAD (2003, p. 61), the electronics industry of CEE, both local and foreign, faces global overcapacity, sluggish demand and competition from East Asia, especially China. As a result electronics firms in Hungary started to restructure their operations. This resulted in closures of some firms and capacity expansion in others. Among the companies that have moved their operations out of Hungary, the most eminent one is IBM, which relocated to China, resulting in a loss of 3,700 jobs in Hungary. Other companies that have moved operations, partially or fully out of Hungary are Flextronics (to China), Philips (to China), Kenwood (to China) and TDK (to Ukraine). 6 On the other hand, some electronics companies like Samsung, Bosch and Alcoa invested in Hungary during the same time. 7 A number of developments took place, globally and locally, in the end of 1990s and early 2000s that contributed to the sudden fall in FDI inflows, hence prompting the restructuring of Hungarian electronics industry. Looking at the global level changes, the beginning of the 21 st century was characterised by the burst of the so-called IT-bubble which was the outstanding feature of the contraction of international trade (and also investment) in This led to a weaker demand not only for computer hardware and telecommunication products but also for other types of electronic products. The burst of the IT-bubble had dramatic repercussions for countries like Hungary, which have built up substantial capacities in the computer section of the Apart from the general incentives like low corporate tax rate, tax holidays, grants, subsidies and customs regime, Hungary offered a special subsidy for the development of regional electronic markets. (Radosevic, 2002, p. 50) This history in manufacturing was important in attracting FDI as it provided MNEs with the opportunity to find people with basic knowledge and skills required for production activities. (also see Ministry of Economic Affairs and ITDH) See table 2 for a list of companies which have reduced or terminated their operations in Hungary. See table 3 for a list of companies expanding or establishing operations in Hungary. 7

219 electronics industry 8 and which have specialized in export of electronic products. On a regional level, the stagnating West European economy 9 and specially the poor economic growth in Germany 10 have also been largely responsible for the falling FDI inflows to Hungary. The events of September 11, 2001 which were followed up by the run-up to a war in Iraq further raised uncertainties and added to the dismay of the global business community by delaying hopes of recovery. The year 2001 proved to be disappointing for the electronics industry world-over. As a result of a slowing world economy, falling component prices and very high manufacturers inventories the world electronics industry experienced a real market 11 fall of 17% in 2001 which was quite a change compared to the 14% growth in 2000 (Reed, 2003, p. 8). On the local level, several factors faded away which had helped the Hungarian electronics industry to gain an advantageous position in attracting FDI during the 1990s. First, a steady increase in unit labour costs is seen as one of the important reasons for MNCs moving their operations out of Hungary (Szanyi, 2003, p. 14). Unit labour costs in Hungary rose by 50% (in Euro terms) between 1999 and 2003 (DIW, 2004, p. 217). This increase compares unfavourably with other rival transition economies which during the same time showed either a decline (Latvia and Lithuania) or lower increase in unit labour costs (Estonia, Poland, Slovakia, Slovenia). In addition, China emerged as a major low-cost investment location for electronics firms. Second, in an effort to harmonise its legislation to EU norms as part of the accession process, Hungary made several structural and macroeconomic changes in the recent years. In May 2001, Hungary abandoned its crawling peg exchange rate regime, which allowed the forint to appreciate significantly by about 18% in less than a year. (Reed, 2003, p. 75). Third, during the same period, Hungary withdrew many of the fiscal and regulatory incentives given to MNEs, as these clashed with EU competition rules. Some of the popular incentives withdrawn include tax holidays for corporate income tax, and customs and tax exemptions to companies operating in the free-trade zones (Szanyi, 2003, p. 11). Fourth, the Hungarian electronics industry is largely export oriented. It exported nearly 67% of its production in 1999 (Radosevic, 2002, p. 14). A combination of increases in wages (and unit labour costs in Euro terms), strengthening of the Forint and an end to tax incentives eroded the cost competitiveness of export oriented MNEs operating in Hungary In 2000, computer hardware, including parts, accounted for 48% of the total output of Hungarian electronics industry (Reed, 2003, p. 77) 75% of Hungary s electronics exports go to the European Union. (Reed, 2003, p. 75) Since 1989, the largest number of foreign investors to settle in Hungary has come from Germany (Wood, 2002) The real market is largely an estimate of the apparent consumption of a country (production plus imports minus exports). It also excludes in-house production (Reed, 2003, p. 150). 8

220 As a result, many of them decided to move their operations to Asia and other East European Countries (Krudy, 2003). Fifth, the end of Hungary s privatisation program, which was one of the main driving forces behind FDI inflows, and the slow pace of infrastructure development are also contributing to stagnating or declining FDI inflows. Sixth, a major problem seems to be the diminishing supply of appropriately qualified labour in the country s main FDI locations (Szanyi, 2003, p. 14). The lack of qualified labour has been caused by the uneven regional spread of investment and low regional labour mobility, which resulted in excess demand for some skilled labour as a production input in a few regions while leaving similar inputs unexploited in other parts of the country (Hungarian Investment and Trade Development Agency, 2002a, p. 7). To sum up, it seems that the stream of FDI pouring into Hungary and the other accession countries has stabilised. Even after enlargement it is unlikely that FDI flows will rise substantially, as the domestic factors remain rather unchanged and as the benefits of a worldwide recovery with respect to FDI inflows will largely accrue to China, India and several other Asian countries. Econometric estimates using gravity models of FDI also suggest, that the potential for additional inflows to Eastern Europe is limited (Brenton et al., 1999, Geishecker, 2004). 3 On the link between MNE activity and the skill composition in the host country It is widely accepted that many MNE activities have an impact upon the level, growth, quality and wages of their own and the home country s labour force (Dunning, 1992, p. 349). However, there is neither a general rule nor an established theory which could predict the effect of MNE activities on the skill composition of the labour-force. Traditional trade theory, for instance, implies that FDI in developing countries would be located in the sectors which utilize the relative abundant factor intensively. This should increase the demand for low-skilled labour in the host countries of FDI, as this is their abundant factor. By the same token, demand for highly skilled labour falls and as a result the skill composition of the work force changes in favour of low-skilled labour. On the other hand, new trade theory implies that MNEs may transfer activities abroad, which are less skillintensive when compared to the home country average, but more skill-intensive when compared to the host country average. This leads to an increase in the skill-level of workers through the use of new technology, training and several other ways. Slaughter (2002), for instance, looks at how FDI affects the demand and supply of skills in an economy. He suggests that FDI has a positive effect on the demand and supply of skilled workers in host countries. 9

221 While scholars from the field of international business (IB) concede that the effects of FDI depend to some degree on factor endowments and the level of development, they also stress a number of additional factors which are responsible for changing the demand and supply for skills 12. UNCTAD (1999), for instance, presents the relationship between inward FDI and skill upgrading on a case by case basis taking into consideration various factors like the amount and type of FDI that the country receives, the strategies of the TNCs involved and home country features. This also implies that the industry to which the MNE belongs might play a crucial role. The effects of FDI in the food or retail sector on the skills demanded by a MNE should be quite different from an investment in the software, electronics of pharmaceutical industry of the home country. A hypothesis entertained quite frequently in the literature is that MNEs from high income countries investing in middle income countries raise the demand and supply for skills (see e.g. Kézdi, 2002). Various studies point towards such a positive relationship between FDI and the skill composition of the workforce. For example, Figini and Görg (1999) found that inward FDI was associated with increased demand for skilled workers in Irish manufacturing over Similarly, the study of Feenstra and Hanson (1995) shows that over the period from 1975 to 1988, FDI was positively correlated with the relative demand for skilled labour in Mexico. Such an upgrade of skills should especially occur, if MNEs belong to high-tech industries in which R&D, innovation and technology development are the main sources of a firm s competitive advantage. As Blomström and Kokko (2002) suggest such FDI inflows create a special potential for knowledge spillovers, leading to skill-upgrades in the economy. To sum up, theoretical considerations and empirical evidence available suggest that FDI in the Hungarian electronics industry should lead to a skill upgrade of the workforce. The next two sections look at the ways in which FDI and other MNE activities affect the demand and supply of skills, considering both the direct and indirect effects. We also provide empirical evidence whenever available. 4 MNE activities and the demand for skills Compared to local firms, MNEs possess several advantages, e.g. ownership and internalisation advantages. 13 Ownership advantage comes in the form of knowledge assets, i.e., patents, proprietary technology, processes, etc. By transferring some or all of these assets to their affiliates, MNEs raise the demand for skills, as most of these assets can only be employed productively with highly qualified personnel. The transfer of these assets to the host country takes place in several ways See e.g. Blomström and Kokko (2002) for a detailed discussion. See Dunning (1992, pp ) for a detailed exposition of these advantages. 10

222 4.1 Technology transfer to affiliates FDI by MNEs usually implies to the transfer of competency and technology to affiliates under multinational firms ownership and control (De Sousa and Richet, 2000). This can take the form of new production or management techniques, whose implementation can lead to a boost in demand for skilled workers. The fact that foreign affiliates purchased about 80% of all the machinery and equipment imported into Hungary indicates that they were the main force behind the spread of modern technology (Szanyi, 2002a, p. 13 / Eltetö, 2001, p. 16). Even the critics (Csath, 1996, p 262, quoted after Farkas, 2000), who claim that MNEs are bringing in technologies that are outdated in developed countries, agree that this technology is advanced when compared to the existing technologies in Hungary, thus raising the demand for high skilled labour. A similar view is expressed by Te Velde and Morrissey (2002) who argue that MNE production is more skill intensive than local firms 14. Therefore when they transfer their knowledge assets to affiliates, they demand higher skill levels from their employees than what the host-country firms demand. Apart from bringing new and more value added processes to Hungary, several MNCs are also upgrading their existing production. Samsung, for example, came to Hungary with the production of combined TV sets, but is now undertaking investments to produce new generation flat screen and plasma TV sets (Hungarian Investment and Trade Development Agency, 2003, pp 5). Other companies following similar pattern are Flextronics, EDS and Nokia (see table 4) 4.2 Research and development MNEs may chose to establish or expand their R&D activities in the host country, thereby raising the demand for skills. The recent developments in Hungary clearly show the contribution of MNCs in this respect. As table 4 indicates, a number of multinationals have started or expanded R&D activities in Hungary. Even those companies that originally came in as low cost manufacturers or assemblers are now undertaking R&D operations in Hungary. These include companies like Compaq, Hewlett Packard, Nokia, IBM and Flextronics (Hungarian Investment and Trade Development Agency, 2002, pp 11). According to Szanyi (2002a), MNCs accounted for 45% of the total industrial R&D in Hungary in In addition he claims that their share was increasing in the following years. Several other MNCs like General Electronics (GE), Flextronics, Bosch and EDS have also relocated their regional service centres to Hungary in the recent past (Hungarian Investment and Trade Development Agency, 2003a, p. 4). This trend is not limited to the electronics industry but extents to others as well (Hungarian Investment and Trade Development Agency, 2003a, p. 4) These upgrades in products, processes and 14 Foreign firms may have different skill intensity than domestic firms, pushing up the average skill intensity. This can be derived from the fact that in comparison to local enterprises MNEs pay higher wages to their employees in return for higher skills demanded and spend more on R&D (Slaughter, 2002, pp ). MNEs accounted for 45% of the total industrial R&D in Hungary (Szanyi (2002a)) 11

223 technologies by MNCs show that transfer of technology to Hungary has not been just a one time activity, rather it has been an ongoing process. It also means that apart from the demand for skilled workers, the level of skills demanded by these MNCs has also increased. 4.3 Technology transfer to domestic firms Third, besides endowing their affiliates with new technology, MNCs can also transfer new technology to domestic firms in the host country. This form of transfer can be market mediated or via non-market channels. The market mediated transfer can take the form of patent licensing, minority joint-ventures, subcontracting, transfer of knowledge assets to suppliers, etc. 4.4 Spillovers New technologies can also reach domestic firms through non-market channels in the form of spillovers 15. There are several channels through which spillovers take place. It can be via a range of informal contacts, exposure to affiliate products, reverse engineering, supplier-distributor networks or buyer-seller links. While these forms of spillovers mainly work along industry lines there are other channels through which technology spillovers can also take place. MNEs may increase efficiency among other firms by stimulating competition. They may impose minimum quality standards or strict delivery dates on their suppliers and subcontractors which forces them to implement better technology and increase managerial efforts. Thus the increased technology intensity of domestic firms, raises the demand for skilled workforce in the host economy. Studies for Hungary do not provide a clear picture with respect to the importance of spillovers. Günther (2002) finds almost no spillovers via buyer-seller links between affiliates of MNEs and domestic firms. In 2000, almost 90% of the raw materials and intermediate products bought by wholly foreign owned affiliates operating in Hungary came either from outside the country or from other foreign affiliates in Hungary. The main reason for this is the technological backwardness of local suppliers which results in problems with product quality and delivery on time (Günther, 2002, pp ). Similarily, Farkas (2000) argues that the new technology brought along by MNEs operate more or less isolated within the MNEs and does not spillover to other firms. Nonetheless, there have been positive spillovers of FDI in Hungary. Domestic firms entering the global production networks as suppliers experienced technological upgrades, since a technological match between MNC and its local suppliers is a prerequisite for cooperation. The role of local suppliers is evident 15 Technology or R&D spillovers are most often strictly defined as externalities, although more recent definitions also comprise of voluntary exchange of useful knowledge. (Dumont and Tsakanikas (2001)) 12

224 from the case studies (Réthi, 2001, GM, 2001, quoted by Szanyi (2002a)) indicating that Hungarian suppliers play an appreciable, if fairly small role in the foreign affiliates activities. Even though the contribution of local suppliers is small when compared to developed countries it is markedly higher than the developing countries. There is also anecdotal evidences from Suzuki and Rába (Szanyi, 2002a, pp 19) showing that multinational firms provide support and guidance to their local suppliers which among other things includes help with developing or adapting technology. Overall, it seems that spillovers through buyer-seller links do not play a major role in spreading technology. Hence, we may suspect that the impact on the demand for skills was comparatively low in the past in Hungary. 4.5 Subcontracting Subcontracting proved to be an effective corporate adjustment strategy for a host of domestic manufacturers at the time of regime change. Within the early years of transition, the share of subcontracting firms in exports was 30% and at that time it was the most common form of cooperation between Hungarian and foreign firms (Szanyi, 2002a,p. 10). Critics of subcontracting (like Sereghyova and Vesely, 1998 quoted after Szanyi, 2002a), say that it leads to the degeneration of the technological potential of the firm. On the other hand, Szanyi (2002b) provides empirical evidence that companies took up subcontracting deliberately and tried to use it as a primary source for modernization. Their success in this effort can be notices from the forward linkage effect of foreign capital on productivity. According to De Souza and Richet (2000, pp 329), 50% of Hungarian subcontractors experienced an increase in their efficiency and a large majority admitted that their clients helped them with production and quality problems. Although it is difficult to assess the precise impact empirically, we may safely assume, that subcontracting contributes to the increase in the demand for skills. 5 MNEs and the supply of skills MNE activities do not only increase the demand for skills in the host country, they also impact on the supply side. This might occur directly through training of staff or provision of scholarships or indirectly through inter-firm labour mobility or a raise in the skill premium which creates an incentive to acquire (additional) skills. We shall consider these factors in turn. 5.1 Training MNEs directly affect the skill composition of workforce by training their employees. This training affects most levels of the company hierarchy (Blomström and Kokko, 2002) and can also extend to the MNEs suppliers, subcontractors and customers. At times MNEs in 13

225 the same industry collaborate in offering training courses to each other s employees. Compared to the local firms, MNEs have well-developed routines, systems, materials for training and can transfer their trainers and employees across countries (UNCTAD, 1999). As a result, MNEs provide more training than their local counterparts. 16 Evidence from Hungary clearly supports this argument. A number of MNCs undertake training programs for their employees in Hungary see table 5. IBM, Nokia (Hungarian Investment and Trade Development Agency, 2002, p. 9) and Flextronics (Hungarian Investment and Trade Development Agency, 2002a, p. 9) have already opened up their training centres in Hungary. Other companies, at least during the initial stages of their market entry to Hungary, send their host country staff back to the home country. Samsung which sent engineers back to South Korea for training is a case in point (Radosevic, 2002, p. 56). Indian s Tata Consultancy Services (TCS), a newcomer to the Hungarian market, plans to train its employees partly in India and partly in Hungary ( Tata s Consultancy Services Office in Hungary (2001), p. 1). In addition to training their own employees, multinationals like IBM offer training opportunities to other companies (domestic and multinationals) who have no specialized training programs of their own (Hungarian Investment and Trade Development Agency, 2002, p 9). 5.2 Support for education TNCs also support skill development by providing scholarships to individuals and by supporting the development of local educational institutions. They may attract or induce training institutions from their home countries to set up similar establishments in host countries. (UNCTAD, 1999, p. 274)) They can also provide assistance and advice through membership and advisory boards, curriculum review committees, councils and senates. Nokia and Tateyama are examples of multinationals operating in Hungary that support students and educational institutions and hence contribute to the process of skill development in Hungary (see table 5). Microsoft Hungary, on the other hand, is helping education development activities by coordinating the work of the Education Initiative of American Chamber of Commerce s (AmCham) IT Committee aimed at providing IT companies and AmCham with a clearer picture of the IT needs and sponsorship possibilities in Hungarian higher education (Balázs, 2002). 5.3 Inter-firm labour mobility Some of the skills developed through training and education spill over to domestically owned firms as a result of employee turnover. This knowledge may or may not be firm specific. The level of mobility of labour force determines whether these spillovers take place at all and whether they remain within the region or spread through the entire 16 These training initiatives are at times highly firm specific and may not be of value to others. 14

226 economy. Knowledge spillovers can also take place as a result of MNCs influence on local competitors and unrelated firms that emulate their practices in employee training. The empirical relevance for this type of spill-over is inconclusive Skill premium As established earlier, MNE activities in Hungary positively affected the demand and supply and thus the wages for skilled workers. Together with skill biased technological progress and institutional changes during transition, MNE activities resulted in an increase of the wages of the high-skilled relative to the low skilled. This rise in the skill premium can be observed in most transition economies. It provides an incentive to individuals to acquire higher skills through education and/or training (Slaughter (2002)). It may also encourage governments to invest in higher education either alone or in conjunction with business associations (UNCTAD, 1999), thus resulting in increased supply of skills in the medium to long term. The response to the rising skill premium is evident from the level of education of the school leavers in Hungary. Table 6 reveals that the number of students with college and university degrees has increased steadily since It almost doubled during the decade after the fall of the iron curtain while it grew only 5% in the previous decade. A similar trend can be observed for the secondary school leavers whose number increased by more than 35%. On the other hand, the number of school leavers from primary or vocational schools declined substantially during the same period. 6 Conclusion Our analysis of MNE activities on the skill composition of the workforce for the case of the electronics industry in Hungary clearly shows that the demand for and the supply of skills increased as a result of various effects. The transfer of modern production technology, organizational knowledge and the relocation of R&D activities by MNEs to foreign affiliates in Hungary substantially increased the demand for skills. The role of spillovers is difficult to ascertain. While dubbed important in the theoretical literature, they are generally hard to measure empirically. In the case of Hungary and its electronics industry they don t seem to have played a major role in the past. Subcontracting by MNEs contributed to an upgrade of the technological and organizational capabilities of local firms, thereby raising the demand for skills at least in the subcontractor s plants. MNE activities also increase the supply of skills. They train their own employees, both in the host and home country and provide scholarships and other support to individuals. In addition, some MNEs have established training centres in which they provide training for employees of other firms. As trained staff 17 See Slaughter (2002, pp ) for details. 15

227 moves from one firm to another they create spillover effects and raise the supply of skills outside the firm originally providing the training. The rise in skill premium which is at least partially a consequence of MNE activities provides an incentive to individuals to invest in human capital formation. At least in the case of Hungary there is no doubt that MNEs substantially contribute to the upgrading of skills of the workforce. 16

228 References Balázs, E. (2002), IT and Telecom Companies Play Key Role in Supporting Higher Education, Business Hungary < Brenton, Paul, Francesca DiMauro, and Matthias Lücke (1999), Economic Integration and FDI: An Empirical Analysis of Foreign Investment in the EU and in Central and Eastern Europe," Empirica, 1999, 26, Blomström, M and A. Kokko (2002), FDI and Human Capital: A Research Agenda, Technical Paper, No.195, OECD Development Centre. Csath, M. (1996), Innovációs helyzetünk és EU~tagságunk: mit kellene tennünk? (Our Innovation Situation and EU Membership: What Should be Done?), Ipargazdasági Szemle (Industrial Economic Review), No. 1-3, pp De Souza, J. and X. Richet (2000), The Impact of Foreign Capital on Local Supply Companies: The Case of Hungary, Economic Systems, Vol. 24 (4), pp DIW (2004), EU-Osterweiterung: Klare Herausforderungen, unberechtigte Ängste, in: Wochenbericht des DIW, 17/2004, pp Dumont, M and A. Tsakanikas (2001), Knowledge Spillovers Through R&D Networking, Discussion Paper, No. 2001/02, University of Antwerp. Dunning, J. (1992): Multinational Enterprises and the Global Economy, Wokingham (UK). Eltétö, A. (2001), The Competitiveness of Hungarian Companies: A Comparison of Domestically Owned Firms and Foreign-owned Enterprises in Manufacturing, Working Paper, No.118, Institute of World Economics, Hungarian Academy of Sciences, Budapest. Farkas, P. (2000), The Effects of Foreign Direct Investment on R&D and Innovation in Hungary, Working Paper, No.108, Institute of World Economics, Hungarian Academy of Sciences, Budapest. Fazekas, K, et. al. (2003), Statistical Data, in: Fazekas, K. and Koltay, J., The Hungarian Labour Market: Review and Analysis, Institute of Economics, HAS, National Employment Foundation, pp , Budapest, Feenstra, Robert C. and Gordon H. Hanson (1995), Foreign Direct Investment and Relative Wages: Evidence from Mexico s Maquiladoras, Working Paper, No.5122, National Bureau of Economic Research, Cambridge. Figini, P. and H. Görg (1999), Multinational Companies and Wage Inequality in the Host Country: The Case of Ireland, Weltwirtschaftliches Archives, Vol. 135 (4), pp GE Homepage, < Geishecker, I. (2004), Foreign Direct Investment in the new Central and Eastern European Member Countries, mimeo, Berlin. GM (2001), Magyar beszállítók helyzete az autóipari és elektronikai multinacionális cégeknél (The Position of Hungarian Suppliers in Vehicle-Industry and Electronics TNCs). Budapest: Ministry of the Economy. Mimeo. Günther, J. (2002), Kaum Technologie-Spillovers durch Zuliefererkontakte ausländischer Tochtergesellschaften in Ungarn, in: Wirtschaft im Wandel, 13/2002, pp Hungarian Investment and Trade Development Agency (2001), The Hungarian Electronics Industry, Budapest. Hungarian Investment and Trade Development Agency (2002), The Hungarian Electronics Industry, Budapest. Hungarian Investment and Trade Development Agency (2002a), Competitiveness 2002: an International Comparison of the Competitive Advantages of Hungary, Budapest. Hungarian Investment and Trade Development Agency (2003), The Hungarian Electronics Industry, Budapest. 17

229 Hungarian Investment and Trade Development Agency (2003a), Regional Service Centres in Hungary, Budapest. Hunya, G. (2004), Impact of FDI on Employment in CEECs, Mimeo, Vienna. Kaminski, B and M. Riboud (2000), Foreign Investment and Restructuring: The Evidence from Hungary, Technical Paper, No.453, World Bank, Washington, D.C. Kézdi, G. (2002), Two Phases of Labour Market Transition in Hungary: Inter-Sectoral Reallocation and Skill- Biased Technological Change, Working Paper No. 2002/3, Hungarian Academy of Sciences, Budapest. Krudy, E. (2003), Jumping Ship: Are Multinationals Abandoning Hungary?, Business Hungary < Ministry of Economic Affairs (2000), Main Industrial Sectors, < Ministry of Economic Affairs and ITDH, Investment Records, < Radosevic, S. (2002), Electronic Industry in Central and Eastern Europe: The Emerging Production Location in the Alignment of Network Perspective, Working Paper, University College London. Reed (2003), Yearbook of World Electronics Data East Europe and World Summary, 7 th edition, Reed Electronics Research, Oxon (UK). Réthi, S. (2001), A háttéripar szerepe az átrendező dés folyamatában (The Role of Background Industry in the Restructuring Process). Budapest: Floreno Kft. Mimeo. Sereghyova, J. and L. Vesely (1998), Progress Linking the Enterprise Sphere of Central-European Countries in Transition into West-European Corporate Networks, Mimeo. Slaughter, Matthew J. (2002), Skill Upgrading in Developing Countries: Has Inward Foreign Direct Investment Played a Role?, Technical Paper, No.192, OECD Development Centre. Szanyi, M. (2002a), Spillover Effect and Business Linkages of Foreign-owned Firms in Hungary, Working Paper, No.126, Institute of World Economics, Hungarian Academy of Sciences, Budapest. Szanyi, M. (2002b), Subcontracting and Outward Processing Trade as a Form of Networking in Hungary, Working Paper, No.124, Institute of World Economics, Hungarian Academy of Sciences, Budapest. Szanyi, M. (2003), An FDI-Based Development Model for Hungary New Challenges?, Working Paper, No.141, Institute of World Economics, Hungarian Academy of Sciences, Budapest. Tata s Consultancy Services Office in Hungary (2001), Spotlight: News Magazine of the Hungarian Investment and Trade Development Agency, Budapest. Tateyama Homepage, < Te Velde, Dirk W. and O. Morrissey (2002), Foreign Direct Investment, Skills and Wage Inequality in East Asia, Overseas Development Institute, London. UNCTAD (1999), World Investment Report 1999: Foreign Direct Investment and the Challenges of Development, United Nations, New York and Geneva. UNCTAD (2003), World Investment Report 2003: FDI Policies of Development: National and International Perspective, United Nations, New York and Geneva. UNCTAD (2003a), World Investment Directory: Hungary FDI Profile, < > W.E.T. to Build Technical Development Centre (2001), Spotlight: News Magazine of the Hungarian Investment and Trade Development Agency, Budapest. 18

230 W.E.T. Automotive System (2002), Press Release No. 10/2002, < Wood, V. (2002), Outspoken Criticism: German Chamber of Commerce (DUIHK) Survey on Hungary, Business Hungary < 19

231 Tables Table 1 Production of Electronics in selected countries Average Annual Growth Rate Hungary % -2% Central Europe % 9% China % 17% Finland % -21% Ireland % -9% Source: Reed (2003); Radosevic (2002) Note: CEE includes Bulgaria, Croatia, Czech, Hungary, Poland, Romania, Slovakia & Slovenia. 20

232 Table 2 Companies Reducing Production or Relocating out of Hungary Date Name People Involved Activity Concerned Type of Action Other Location Involved 2000 Ericsson (Sweden) Mobile phones Relocation from Hungary China 2000 Mannesmann (Germany) Car audio plant Closure- Relocation from Hungary China 2001 Flextronics (Singapore) Subcontracting subassembly work Downsizing Ukraine May 2002 Flextronics International (Singapore) Kft X-box production Relocation from Hungary China October 2002 IBM Storage Products Kft. (USA) Hard Disk Drive Relocation from Hungary China October 2002 TDK Elektronika Kft. (Japan) Partial relocation from Hungary Ukraine December 2002 Kenwood Electronics Bretagne (Japan) S.A Car Radios Consolidation of Global Production Network (from 9 to 5). January 2003 Philips Magyarorszag (Netherlands) Kft -500 Cathode Ray Tube Monitors Relocation from Hungary China Source: Own compilation based on UNCTAD, 2003; Radosevic,

233 Table 3 Companies Expanding Production or Relocating to Hungary Date Name People Involved Activity Concerned Type of Action 2000 SMK (Japan) Mobile phone batteries Relocation to Hungary 2000 Sony (Japan) TV sets and monitors Relocation to Hungary 2000 Elcoteq (Finland) Acquired Nokia s monitor manufacturing plant in Pecs. February 2002 GE Hungary Rt. (USA) +500 (2004) GE Lighting s regional headquarters June 2002 Foxconn Hon Hai (Taiwan) July 2002 Sunarrow Hungary Kft. (Japan) September 2002 December 2002 December 2002 Electronic Data Syatems (USA) Elcoteq Magyarorszag Kft. (Finland) Electrolux Lehel Kft. (Sweden) 2002 Flextronics International Kft. (Singapore) Flextronics International Kft. (Singapore) Philips Magyarorszag Kft. (Netherlands) (by 2006) Spare parts for computers, mobile phones and other consumer electronics Expansion through acquisition New Capacity Other Location Involved UK UK (2004) Supplier to Nokia New Capacity +110 Regional Service Center New Capacity +250 Expansion of Capacity +400 Refrigerator Production Relocation Spain Telephone Modem Production In part, mobile phone production Relocation Hungary Expansion Capacity Expansion of Capacity 2003 GE Hungary Rt. (USA) +100 Lighting Bulb Production Expansion of Capacity January 2003 IBM Magyarorszagi Kft. (USA) to of Czech Republic +377 January 2003 Jabil Circuit Kft. (USA) +600 Relocation to Hungary January 2003 February 2003 February 2003 February 2003 Philips Magyarorszag Kft. (Netherlands) Samsung Magyar Elektromechanikai Rt. (South Korea) Robert Bosch Elektronika Kft. (Germany) Bosch Rexroth Kft. (Germany) April 2003 Samsung Magyar Elektromechanikai Rt. (South Korea) +330 Cathode Ray Tube Televisions Relocation Hungary Cathode ray tube production Relocation to Hungary to United Kingdom France Other European countries +250 Car Electronics Relocation Austria, Germany +400 Car Electronics New Capacity... Television production Expansion of Capacity Source: Own compilation based on UNCTAD, 2003; Radosevic, 2002; Hungarian Investment and Trade Development Agency (2002); Hungarian Investment and Trade Development Agency (2001) 22

234 Table 4 R&D activities and process upgrades by multinationals in Hungary Company R&D activities and process upgrades by MNEs Bosch (Germany) Compaq (USA) EDS (USA) Electrolux (Sweden) ELCOTEQ (Finland) R&D, supply and sales centre. Established R&D team in Hungary. IT and data processing services. Moved R&D to Hungary or started such activity in Hungary. Although Elcoteq s more established plants in Finland handle most challenging new products, Hungary has begun to take on some small industrial products fresh from the R&D lab. Ericsson (Sweden) Trafficlab (Traffic Analysis & Network Simulation Laboratory) started in Ericsson has similar research laboratories all over the world. Ericsson Hungary is the country s largest enterprise performing research in the field of software development, telecom and IT. Ericsson s research laboratory was established in close cooperation with Budapest Technical University. Ericsson Hungary s current R&D department was established on July 1, 1998 with the merger of the software development centre, which had been founded in 1992, and the research laboratory. Flextronics (Singapore) Fujitsu (Japan) General Electronics (USA) Hewlett Packard (USA) Horvath AG (Germany) IBM (USA) Knorr-Bremse (Germany) Motorola (USA) Nokia (Finland) Panasonic (Japan) Pan-Tel (Netherlands) Philips (Netherlands) Established R&D team in Hungary. Flextronics has Hungarian engineers developing and laying out circuits in support of its Austrian-based R&D centre. Set up an engineering development centre with 30 engineers. In the future, Flextronics Hungary would like to expand its suppliers activities so that it can offer its clients services with increasing complexity. European Administration and Accounting Centre. Undertaking research activities, in collaboration with universities, at Infopark. GE Lighting's Europe, Middle-East & Africa Headquarters are now located in Budapest, as well as its centre of excellence for Research & Development and the Machine Division. The company has recently developed a new European Customer Response Centre. Undertaking research activities, in collaboration with universities, at Infopark. Located its R&D division in Hungary. Undertaking research activities, in collaboration with universities, at Infopark. Research Centre for electronic brake systems. Moved R&D to Hungary or started such activity in Hungary. In 1998 the company established three R&D centres in Hungary. Two of the centres are concerned with IP networks and digital software solutions, while the 3 rd centre is Nokia s central R&D institution, the Nokia Research Centre. Undertaking research activities, in collaboration with universities, at Infopark. Undertaking research activities, in collaboration with universities, at Infopark. Moved R&D to Hungary or started such activity in Hungary. (Table 4 continued) 23

235 Table 4 (continued) Company Siemens Sony (Japan) Samsung (South Korea) TATA TCS (India) Tateyama Laboratory Hungary Ltd. (Japan) W.E.T. Automotive Systems Ltd. (Germany) Zenon (Greece) R&D activities and process upgrades by MNEs Moved R&D to Hungary or started such activity in Hungary. In 1998, additional major investments were implemented, among them a new IT system for the audio visual products division was installed. Samsung is planning to bring R&D to Hungary. Software Development and Distribution Centre. Operates only R&D centre in Hungary with no production facility. Established technology centre in Hungary in September Employs seventy engineers and technicians. Set up its R&D and regional service centre in Hungary. Sources: Own compilation based on GE Homepage; Hungarian Investment and Trade Development Agency (2003); Hungarian Investment and Trade Development Agency (2003a); Hungarian Investment and Trade Development Agency (2002); Hungarian Investment and Trade Development Agency (2002a); Hungarian Investment and Trade Development Agency (2001); Ministry of Economic Affairs (2000); Radošević (2002); Szanyi (2002a); Tata s Consultancy Services Office in Hungary (2001); W.E.T. to Build Technical Development Centre (2001) and W.E.T. Automotive System (2002). Table 5 Training and support for education IBM (USA) Established training centre in Infopark (Budapest), to help students perform their daily IT duties, and teach them how to use the various IT devices effectively in their jobs. An increasing number of companies who have no specialized training themselves rely fully on the training courses organized by the IBM Training Centre in providing in-service PC training. Nokia (Finland) Established Central European training centre in Sátoralijaújhely in The company maintains strong relations with higher education institutions working in Budapest and in rural areas. Nokia does not only sponsor students, but provides opportunities for students to prepare diploma works and also scholarships in the last year of their university education. Flextronics (Singapore) Operates a training and development centre in Zalaegerszeg. Runs 2-15 days training program for technicians and engineers. Attended by at least 40 people each month. Samsung (South Korea) TATA TCS (India) Tateyama Laboratory Hungary Ltd. (Japan) Cisco (USA) Hungarian engineers sent to South Korea to study production techniques. Hungarian employees receive training partly in India and partly in Hungary. Tateyama works in close cooperation with several Universities in Hungary, providing scholarships for both students and their supervisors. Among these institutions are the Budapest University of Technology and Economics as well as the University of Veszprém. Runs software academy in Hungary. Sources: Hungarian Investment and Trade Development Agency (2002); Hungarian Investment and Trade Development Agency (2002a); Radošević (2002); Tata s Consultancy Services Office in Hungary (2001); Tateyama Homepage. 24

236 Table 6 School leavers by level of education Year Primary school Specialized secondary school Vocational school Secondary school College and university ,809 2,646 46,586 43,167 14, ,891 3,241 50,483 52,573 15, ,614 3,375 51,558 53,039 15, ,907 3,890 55,412 54,248 16, ,287 3,810 62,451 59,646 16, ,200 6,302 60,040 68,607 16, ,857 7,285 55,617 68,604 18, ,333 6,991 50,066 70,265 20, ,529 6,414 47,795 73,413 22, ,708 4,895 41,973 75,564 24, ,651 3,995 38,871 77,660 25, ,302 2,460 36,362 73,965 27, ,500 a 72,200 b 28,300 b , ,500 a 70,441 29,746 a Specialized secondary schools included. Estimated data. b Estimated data Note: Primary school: completed the 8 th grade. Other levels: received certificate. Source: Fazekas (2003). 25

237 Economic Restructuring and Labour Markets in the Accession Countries Research Project commissioned by EU DG Employment, Social Affairs and Equal Opportunities, Contract No. VC/2003/0367 Coordinated by in cooperation with Alphametrics and The Vienna Institute for International Economic Studies Impact of FDI inflows on labour market differences in Hungary stylized facts and policy implications by Károly Fazekas* and Éva Ozsvald* September 2004 *) Institute of Economics, Hungarian Academy of Sciences, Budapest

238 Contents 1 Introduction 2 Impact of FDI on employment and wages 3 Impact of FDI on regional labour market differences 4 FDI policies References List of Tables and Figures Table 1 Owners equity of enterprises by industries, 2002 Table 2 Largest exporting foreign investment enterprises in Hungarian manufacturing, 2001 Table 3. Productivity of FIEs and DEs by industries, 2002 Table 4 Table 5 Number of employees in foreign investment and domestic enterprises in the corporate sector, 1993 to 2002 Employees of the corporate sector by the share of foreign ownership, per cent Table 6 Net job creation rates in the corporate sector, 1992 to 2002 Table 7 Number of employees and firms by industries, 2002 Table 8 Distribution of FIE and DE employees by classification of manufacturing industries based on the level of technology, 2002 Table 9 Changes of composition of FIE employment by industries, 2000 to 2002 Table 10 Employment change in different industries, 2000 to 2002 Table 11 Average gross earnings of employees by industries, 2002 Table 12 FDI stock by regions, 1995 to 2002 Table 13 Distribution of the number of FIEs across regions and counties, 1995 to 2002 Table 14 Table 15 Distribution of the number of FIE employees across regions and counties, 1995 to 2002 Changes of corporate employment in low- and high-employment micro regions between 1993 and 2002

239 Figure 1 Time path of FIEs employment in the corporate sector, 1993 to 2002 Figure 2 Net job creation rates in the corporate sector, 1992 to 2002 Figure 3 Figure 4 Figure 5 Spatial distribution of micro-regions in the four quartiles of employment rates, 1990 and 2001 Wage costs and productivity of firms settled in high-employment regions compared to firms settled in low-employment regions in manufacturing, 2002 Time path of the productivity gap between firms in manufacturing settled in low- and high-employment regions, 1993 to 2002

240 Summary Two thirds of FIE employees and less than 27% of DE employees worked in manufacturing in The share of employees working in high-technology manufacturing industries was two times higher in the group of FIEs (12.3%) than in the group of DEs. Only one third of FIE employees but more than a half of the DE employees in manufacturing worked in low-technology industries. During the 1990s FDI had a profoundly positive overall impact on the level of employment and wages and had a decisive role in labour market restructuring in Hungary. However, expansion of foreign firms substantially increased labour market differences in terms of employment and wages. More than 80% of net job creation of the corporate sector in Hungary between is attributed to the foreign enterprise sector. More recently, between the number of FIE employees decreased by 6% and the distribution of FIEs employees between economic sectors changed. The share of those employed in construction decreased while the share of those in manufacturing and trade increased. Within the manufacturing sector, the share of FIE employees working in the production of coke, refined petroleum and nuclear fuel, manufacturing of rubber and plastic products and of transport equipment increased while the share textile industry and of the electrical and optical equipment industry decreased. According to several experts the break in the decade long expansion of FDI expansion and the shift in the weight of industries mark a start of a new period of FDI activities in Hungary. Based on some evidence it is expected that instead of investments concentrating on massproduction in low-technology industries, the emphasis will shift to high value-added and advanced knowledge-based production using a skilled workforce and utilising Hungary s excellent geographical position. Employment statistics does not give much support to the above expectation. According to the latest data, the share of employment has not increased neither in high technology nor in medium high-technology industries. There is, however, a shift in FIE employment from low-technology to medium-low technology industries. We can see the highest net job destruction in the low-technology industries (-10.4%) while employment in medium- low technology industries increased by a remarkable 16% in two years. Ownership-specific differences in productivity are high and are clearly reflected in relative wages. Young-skilled workers employed by foreign enterprises have higher wages relative to their unskilled and skilled-old colleagues than have their counterparts in domestic firms. More efficient matching of new technologies and new skills in foreign than domestic enterprises was the driving force behind the appreciation of younger generations. The returns to skills in domestic firms started to follow the foreign-firm pattern lately. 1

241 Local labour markets are closed and fragmented in Hungary due to the relatively high cost of commuting and underdeveloped transport infrastructure. Grouping micro-regions into quartiles according to their employment rates gives a clear east-west, core-periphery division. The central agglomeration and regions along the main east-west transport routes in the direction of Graz and Vienna have the highest employment rates while most of the low employment regions are located at the periphery. Core-periphery division of microregions had become stronger over the 90s. While the intensity of job destruction features an equal regional distribution, the intensity of job creation concentrates on core areas. After 2000, net job destruction of FIEs was more concentrated in high employment regions increasing decreased regional employment differences. The effect was counterbalanced by the spatial distribution of job creation/job destruction of domestic firms. In sum, for the whole period between we can observe a net job destruction (-1% of the working age population) in the low employment regions and a net job creation (+11%) in the high employment regions. Due to the productivity advantage of high employment regions the same wage level means much large unit labour costs in the bad regions than in the good ones. Firms moving from the most developed regions to the less developed regions could save only 2-7% in wage cost. The regional gaps of productivity and unit labour cost have substantially increased over the last ten years. Increasing return to agglomerations constitutes an important part of the explanation. The higher the density of foreign firms in the high employment regions, the stronger the spillover effect towards domestic (and foreign) firms and, as a consequence, the higher the productivity advantages of these regions. The increasing density of FIEs has a significant positive effect also on the productivity of domestic firms. 2

242 Károly Fazekas and Éva Ozsvald Impact of FDI inflows on labour market differences in Hungary stylized facts and policy implications 1 Introduction Hungary s performance since the second half of the 1990 s was impressive both in terms of economic growth and the completion of the transition to a fully fledged market economy. The rise in industrial productivity and exports was, on average, faster between than in any European country, bar Ireland. Privatization as a main feature of system change is generally referred to as being the most successful among the transforming CEE countries. The considerable amount of foreign direct investment that the country received during the 1990 s played a major role both in the growth performance (via investments, restructuring and exports) and the efficient privatization programme of the country. In 2002 foreign owned businesses claimed 53.5% of owners equity of firms in the corporate sector. They produced half of the output with 25 per cent of the corporate sector employees. Their share of exports was in excess of 80 per cent and 14 big multinational companies produced almost 40 per cent of total Hungarian exports. FDI is highly concentrated in a few sectors of the economy. Foreign investment enterprises 1 possess the largest share in owners equity in manufacturing (77.6 %) and within manufacturing in the production of electrical and optical equipment (90.7 %) and transport equipment (97.0%) (Tables 1-2.) FDI had an outstanding role in the fast increase of productivity in Hungarian industry during the 1990 s. Table 3 shows productivity indicators of foreign investment enterprises (FIEs) and domestic enterprises (DEs) by industries in We can see a substantial productivity gap between foreign and domestic firms. The average gross sales per employee ratio was 2.6 times higher, the value-added per employee ratio was 1.6 times higher in foreign firms than in the domestic enterprises. In the case of manufacturing industries the FIEs/DEs productivity gaps were 3.3 and 1.5 sequentially. Studies on the impact of FDI on the Hungarian economy have extensively discussed the consequences of FDI inflows. Kalotay (2001, 2003) emphasises that massive inflow of FDI was an essential precondition for fast and successful privatization of formerly state owned enterprises. Szalavetz (1994, 1996), Csáki-Sass-Szalavetz (1996) and Major Vezzoni Szalavetz (1999) listed a number of positive multiplier effects that resulted from the dissemination of new organizational forms, more advanced management practices and 1 A foreign direct investment enterprise is an incorporated or non-incorporated enterprise in which an investor resident in another economy owns 10% or more of the ordinary shares or voting power for an incorporated enterprise (CSO 2004). 3

243 technological development. A number of empirical studies aimed at showing the different channels of technology transfer through FDI and its impact on the productivity growth of domestic firms in CEE countries. Papers discussing the direct and indirect spillover effects of FDI towards domestic enterprises in Hungary are not unanimous. Szanyi (2002) and Konings (2000) did not find convincing empirical evidences on positive spillover effects, while according to several other papers (Campos-Kinoshita 2000, Damijan-Knell-Majcen- Rojec 2003, Lücke-Szalavetz 1997, Schoors-Van der Tol 2001), Farkas (2002) these effects did occur in certain sectors and regions in Hungary. The purpose of this paper is to analyse the labour market effects of foreign direct investments that Hungary hosts. FIEs had an outstanding role in job creation within the corporate sector from the beginning of the transition. More than 80% of net job creation of the corporate sector between is attributed to the foreign enterprise sector. FIEs pay higher wages than domestic firms within the same industries. After controlling the composition and productivity effects these differences remain, although they become much smaller. FIEs had an important role in the re-valuation of the human capital. FIEs concentrated in the most developed urbanized regions close to the western portals. Location preferences of foreign firms contributed to the increase of regional differences of employment rates since most of the jobs created by FDI concentrated in high employment regions. At the level of local labour markets, spatial concentration of FIEs employment in the most developed regions has not decreased despite all the government s incentives. This paper is structured as follows: the introduction is followed by a chapter that concentrates on the employment and wage impacts of FDI. The third part deals with the role FDI plays in regional labour market differences. The fourth part gives an overview of the efforts that governments have taken so far in encouraging foreign capital towards the backward regions of Hungary. Data on the labour market impact of FDI are derived from two main sources: the CSO data base on foreign direct investment in Hungary and the FDI-Micro Regional data base of IE-HAS. 2 We also draw on the information collected from academic studies and experts materials listed in the References section of the paper. 2 Impact of FDI on employment and wages During the 1990 s FDI had a profoundly positive overall impact on the level of employment and wages and had a decisive role in labour market restructuring in Hungary. However, 2 The source of this database is the firm-level Balance-sheet Corporate Database of the HCSO. This covers all incorporated firms and practically all firms employing more than five persons. In the IE-FDI Micro regional Database a set of balance sheet data of all foreign and domestic enterprises was separately aggregated at NUTS-4 level of regions. Data covers all years between 1993 and

244 expansion of foreign firms substantially affected labour market differences in terms of employment and wages. Impact of FDI on the level of employment Table 4 and Figure 1 show that the number of FIEs and the employment they provided continuously increased over the 1990 s with the major part of the expansion occurring before After that, the employment in FIEs stagnated and then started to decline in The share of FIEs employment in the corporate sector increased from 16% to 27% between In 2002 (the last year in the CSO-Corporate Data Base) the share of FIEs employment was 25.4%. Data in the EO-Wage Survey show the same tendency (Table 5.) The share of FIEs employees in the sample was 30% in 2000 and decreased to 25% in Although it is impossible to make a clear distinction between jobs taken over by foreign firms via privatization and jobs created by business start-ups, most of the estimates show a positive net effect of job creation and job loss as a result of FDI inflows to Hungary. The impact of increasing foreign firm s sector on the aggregate level of employment can be measured calculating the net effects of job creation and job destruction carried out by new, terminated and existing foreign firms. Unfortunately, neither the CSO-Corporate Data Base, nor the EO-Wage Survey are suitable to identify the new and terminated firms in a given year. 3 As a second best solution, Kőrösi (1997, 2003) calculated the so called narrowly defined job creation and job destruction of FIEs and DEs on a sub-sample of the CSO-Corporate Data Base taking into account the data of only those companies that operated under the same registration number in both years. 4 Figure 2 and Table 6 show the net job creation rates of FIEs and DEs between On the basis of his calculations we can distinguish three periods. In the first three years of transition state owned and private domestic enterprises shed labour in huge quantities while FIEs produced relatively small job destruction rates. In the second period, between the job destruction continued albeit in the domestic sector at a slower rate while FIEs had a relatively large 4.1% per year job creation rate. After 2000, following the downswing of the business cycle job destruction speeded up again in the domestic sector while FIEs performed a net job destruction rate at an average of.5% in The registration number identifying the companies changed for a relatively large number of firms especially at the beginning of the period and this resulted in an upward bias in the indicators by increasing the number of firms that were apparently established or closed. The gross job creation rate is the total increase of the employment of all expanding companies divided by the total average employment of the industry. Similarly, gross job destruction is the total number of lay-offs divided by the total average employment of the firms. The difference between the two is the net job creation or destruction rate. 5

245 Impact of FDI on the industrial composition of employment Table 7 shows that there are substantial differences between the distribution of FIEs and DEs employment according to the industries in which they operate. In 2002, FIEs employees had a higher than average share in manufacturing (44 %) and in electricity, gas steam and water supply (30 %). Within manufacturing, FIEs had the highest share in the manufacturing of electrical and optical equipment (67 %), in the production of transport equipment (62 %) and in the manufacture of coke, refined petroleum and chemical products (64 %). According to conventional wisdom Hungary, having a relatively large pool of skilled labour was successful attracting FDI in high-technology industries. The distribution of employment by the technology level of industries 5 seems to justify this tendency. (Table 8) Two thirds of the FIEs employees and less than 27% of DEs employees worked in manufacturing in The share of those working in high-technology industries was two times higher in the group of FIEs than in the group of DEs (12.3 % and 6.2 % respectively). Only one third of FIEs employees and more than a half of the DEs employees in manufacturing worked in low-technology industries. The number of FIEs employees decreased by 6% between Behind the overall decrease we can observe shifts in the distribution of FIEs employees between the sectors of the economy.(table 9) The share of those employed in construction decreased while the share of those in manufacturing and trade increased. Within the manufacturing sector, the share of FIEs employees working in the production of coke, refined petroleum and nuclear fuel, manufacturing of rubber and plastic products and of transport equipment increased while the share in the textile industry decreased. According to several experts the break in the decade long growth of FDI expansion and the shift in the weight of industries mark a start of a new period of FDI activities in Hungary. Based on some evidence it is expected that instead of investments concentrating on massproduction in low-technology industries, the emphasis will shift to high value-added and advanced knowledge-based production using a skilled workforce and utilizing Hungary s excellent geographical position. The employment statistics, however, do not give much support to the above expectations. According to the latest data, the share of employment has not increased either in high technology or in medium high-technology industries. There is, however, a shift in FIEs employment from low-technology to medium-low technology industries. Table 10 shows the changing proportions of employment in the four manufacturing groups classified by the 5 Classification of industries follows the latest OECD classification of manufacturing industries based on technology. See OECD (2003). 6

246 level of technology. In high and medium high technology industries employment decreased by around the manufacturing average. We can see the highest net job destruction in the low-technology industries (-10.4 %) while employment in medium- low technology industries increased by a remarkable 16% in two years. Impact of FDI on wages Table 11. shows that FIEs pay higher wages than DEs across the industries. (See Fazekas Köllő, 1998, Csengődi Jungnickel Urban, 2003.) In 2002, average gross earnings of FIEs employees was more than one-and-a half times the average in the corporate sector. The wage gaps were 133% in manufacturing, more than 200% in construction and between 147% and 221% in different branches of the service sector. The picture gets more complex, however, if the differences in the average wages are examined in a desegregated way. Both structural differences (composition effects) and differences in capital intensity of typical domestic and foreign firms should be taken into account in this case. According to Köllő s (2002) calculation the most important factor causing earning differentials is the significantly higher productivity of the foreign enterprises which widens the foreign-domestic difference by nearly half of the total average wage difference. Köllő found that after clearing the effects of different composition and productivity of FIEs and DEs the pure foreign ownership effect decreased over the years of transition. He interpreted this tendency as follows: foreign businesses were willing to pay a risk surcharge when they first entered the unknown Hungarian labour market. This gave them an opportunity to pick and choose during the tricky time when they had to build up their staff from scratch. The disappearance of the surcharge if it was more than just an apparition brought about by changes in complex composition can suggest that this special situation has come to an end and that wages have dropped to market level. Impact of FDI on return to skill and experience in different stages of transition Several empirical studies (Kertesi and Köllő, 1999, 2002; Köllő, 2003; Kézdi, 2002, Tarjáni, 2004) investigated the impact of age (experience) and education on the wage differences in the corporate sector. They found a general rise in the returns to education between 1989 and 1992 when technological change was minimal and the forces of the market had just started to work. They explain this process as a mirror image of the collapse of demand for unskilled labour during transitional crises. Kertesi and Köllő (2001) estimated the productivity equation for domestic and foreign firms separately. The estimates cover the period during which the percentage of workers employed by foreign firms rose from 10 per cent to 40 per cent in the EO-Wage survey. After the transitional crises when market institutions were already at work and modern technologies started to flow in, the general appreciation of education stopped but 7

247 the returns to experience continued to decline. Technological renewal apparently contributed to the appreciation of young and educated labour in this period. Kertesi and Köllő found that these workers are paid higher wages and their skills are valued more in a modern environment. By contrast, neither productivity nor wages grew for the older cohorts of educated workers after Changes in the skill-related wage differentials at least partly reflect changes in relative productivity levels. The productivity yield that is attributed by the model to young-skilled labour input was growing rapidly in while the productivity of older skilled labour input was declining in to a point where after that it did not differ significantly from the productivity yield of unskilled labour. Kézdi (2002) found the same tendency of re-valuation of human capital during transition. Foreign direct investment and the more efficient matching of new technologies and new skills in foreign rather than domestic enterprises, was the driving force behind the appreciation of younger generations for half a decade. The returns to skills in domestic firms started to follow the foreign-firm pattern only at the end of the nineties. The ownership-specific differentials in productivity are clearly reflected in relative wages. Consistently with the predictions of the productivity model young-skilled workers employed by foreign enterprises have higher wages relative to their unskilled and skilled-old colleagues than have their counterparts in domestic firms. 3 Impact of FDI on regional labour market differences The Central Statistical Office provides macro-regional level 6 time series of the Labour Force Surveys and the national accounts. These data show that the decline in economic performance and employment during the transitional crisis was much more severe in disadvantaged rural regions of the East and Southwest than in the more urbanized Central and North-western territories. Regional employment and unemployment rate differences at the macro-region level, however, are not particularly large by international comparison and did not tend to increase during recent years. The problem is that in the case of Hungary macro- or meso-region level analyses of labour market indicators give a distorted picture. Due to the relatively high travel costs of commuting and the underdeveloped transport infrastructure local labour markets (LLMs) are closed and fragmented in Hungary. The size of LLMs fits more into the category of micro-regions. Grouping micro-regions into quartiles according to their employment rates gives a simple but clear picture of the winners and losers of transition at the level of LLMs. Maps 1-2 show the geographic situation of top and bottom quartiles (high and low employment 6 There are 7 statistical-planning regions (NUTS-2 units), 19 counties and the capital, Budapest (NUTS-3 level units), 150 statistical micro-regions (NUTS-4 level units) and 3,120 settlements (NUTS-5 level units) in Hungary. The average size of micro-regions is km 2, the average number of the local population is 77,279 and the average density of population is cap./km 2 On the NUTS classification see Eurostat (1995). 8

248 regions) in 1990 and in One can see a clear east-west, core-periphery division for both years. The central agglomeration and regions along the main east-west transport routes in the direction of Graz and Vienna have the highest employment rates while most of the low employment regions are located on the periphery, along the East-Slovakian, Ukrainian, Romanian and Croatian borders. Fazekas (2004) described how core-periphery division of micro-regions had become stronger over the 90s. Employment differences were mostly generated by demand side factors. After the collapse of the socialist economy the high intensity of job destruction was accompanied by dynamic job creation. (Kőrösi 2003) Research results show invariably that while the intensity of job destruction portrays an equal regional distribution, the intensity of job creation follows an uneven spatial pattern. Differences of spatial concentration of FIEs and DEs employment CSO FDI data base contains information on the distribution of FDI stock and of FIES employees at the level of macro regions and counties between (see Tables 12 to 14). Data show that both FDI stock and FIEs employees were highly concentrated in the most developed regions (Central Hungary, Central Transdanubia and Western Transdanubia). In 1995, 83% of the FDI stock and 72% of FIEs employees were located in these regions. On the level of macro regions the most important change that occurred between was that the share of the Central Region in FIEs employment decreased by 10 percentage points, while the share of Northern Hungary increased by the same degree. Unfortunately the CSO FDI data base is not suitable for measuring the impact of FDI on local labour markets for two reasons: (i) it contains information on the regional distribution of foreign and domestic firms only at the level of macro regions and counties; ii) in the CSO FDI database firms are classified into regions according to the settlement of the headquarters of the firms. This method, however, overestimates the spatial concentration of firms because premises located in different regions are taken into account as if they were located in the headquarters region (Hamar, 1999). We can investigate, however, the micro region level the distribution of foreign and domestic enterprises with the help of the IE-FDI micro-regional data base. It contains a set of balance sheet data of all foreign and domestic enterprises, separately aggregated at NUTS-4 level of regions. Since the balance sheets of the firms contain the settlement code and the number of employees of each establishment of enterprises, the bias found in the CSO FDI data base has been reduced by the redistribution of firms data between microregions in proportion to the branch s share in the total number of employees of the given firms. 9

249 Studies on spatial distribution of FDI (Hamar 1999, Fazekas 2001, Békés 2004) revealed that FDI inflows were highly concentrated in certain regions so it comes as no surprise that the concentration of FIEs jobs is higher than the concentration of the working age population and higher than the concentration of DEs employees. The difference between the concentration of FIEs and DEs jobs is, however, not particularly high. According to Fazekas (2004) calculations for the year 2002 the Gini coefficients of the working age population, DEs employees and FIEs employees were 0.50, 0.63 and 0.70 respectively. 17.1% of the working age population, 23.0% of the domestic firms employment and 23.5% of the foreign firms employment were concentrated in one region: in the capital of the country. The top quartile of the micro regions (37 regions) having the highest shares covered 61.1 per cent of the working age population, 73.3% of the DE s jobs and 78.3% of FIE s jobs in Using relative concentration indexes we could measure the difference between the spatial distribution of FIE s or DE s jobs and the distribution of a benchmark variable (the working age population in this case). Fazekas (2004) found that the relative concentration of FIE s jobs is the highest in most of the micro-regions along the Austrian border but that there are also several regions of the top quarter in the eastern part of the country as well. The relative concentration of DE s jobs does not show a similarily clear east-west division. We can have a more detailed picture of the determinants of spatial concentration of FIEs and DE s jobs by estimating the relative concentration of jobs by regressions using selected explanatory variables. Fazekas (2003, 2004) came to the conclusion that regional differences in unemployment rates of micro-regions have been determined by three main factors: the industrial past of the regions, the proximity to the western portals and the education level of the local labour force. Table 15 shows the size of employment changes in the corporate sector in two groups of micro regions. Table shows changes of FIEs and DEs employees as a percentage of the working age population. High employment regions and low employment regions refer top quarter and bottom quarter of micro regions ranked according to their employment rates. One can see that between job creation of FIEs in the two quartiles had a crucial impact on employment rates differences. In the high employment regions net job creation of FIEs was 8% of the working age population while it was only 1.8% in the low employment regions. After 2000 net job destruction of FIEs was more concentrated in high employment regions. It means that after 2000 spatial distribution of job destruction of FIEs decreased regional employment differences. This effect was counterbalanced by the spatial distribution of job creation/job destruction of domestic firms. In low employment regions DES jobs decreased as much as 4% of the working age population while DEs job creation in high employment regions was around 3% of the working age population. In sum, for the whole period between we can observe a net job destruction (- 1% 10

250 of the working age population) in the low employment regions and a net job creation (+11%) in the high employment regions. Impact of regional wage differences on spatial distribution of FIEs employment Elasticity of net earnings to local unemployment rates had reached 10% in the mid 1990 s. According to the literature it is a typical value in the developed market economies. (Köllő 2002, 2004) At that time the highest unemployment rate at micro region level was 8 times the lowest unemployment rate. Taking into account the effect of the 10% elasticity the estimated net earnings difference between the two regions was around 17%. The question arises: Why are the foreign and domestic investors alike so reluctant to relocate their activities towards the high unemployment/low employment regions? The regional differences of productivity and the unit labour costs of foreign and domestic firms explain a great deal of this reluctance. According to Köllő s (2003) calculations a considerable part of wage differences related to unemployment rates disappears if we take into account the large productivity gaps between the low and high unemployment regions. Because of the productivity gaps the same wage level means much larger unit labour costs in the bad regions than in the good ones. After clearing the effect of productivity gaps Köllő found that firms moving from the most developed regions to the less developed regions could save a meagre 2-7% in wage costs. Figure 4 shows Fazekas (2004) calculation on regional differences in wages, productivity and unit labour costs between firms in manufacturing operating in high and low employment regions. One can see that there are substantial regional differences in both FIEs and DEs groups. Wage costs are higher in high employment regions than in low employment regions but because of the high productivity the unit labour costs of firms operating in high employment regions is less than 80% of those settled in low employment regions. No doubt that besides region-specific factors (proximity, density of firms, externalities offered by urban agglomerations etc) the regional productivity gap has been influenced by a number of firm specific factors, such as sectoral composition, technologies and labour/capital ratio. Nevertheless, the time paths of regional gaps in the case of FIEs and DEs reveal a striking tendency. (Figure 5.) The regional gaps of productivity between firms settled in high and low employment regions have substantially increased in both groups over the last ten years. Factors behind the increasing wage and productivity gaps require a careful analysis which is beyond the scope of this paper. Nevertheless, we are convinced that increasing return to agglomerations constitutes an important part of the explanation. Regional spillover effects between firms could be an important element of agglomeration effects. The higher the 11

251 density of foreign firms in the high employment regions, the stronger the spillover effect towards domestic (and foreign) firms and, as a consequence, the higher the productivity advantages of these regions are. According to empirical evidence from Hungary the increasing density of FIEs has a significant positive effect on the productivity of domestic firms. 4 FDI policies FDI incentives in the golden age In the first phase of transition government policies related to FDI in a narrow sense significantly contributed to the attractiveness of Hungary as locational choice for the investments of TNCs, including a number of big-name blue-chip companies. For them generous incentives were offered through individual deals. As pointed out by Sass (2003b) in some cases even a monopoly position was assured for foreign investors. In sum, the Hungarian investment promotion system in its initial stage produced the desired goal but was lacking transparency, the normative and regional approaches. In the later part of the nineties this quantity-oriented, discretionary and privatization focused approach was giving way to more transparent and normative government policies with performance requirements. In theory, foreign owned companies settled in a less developed region could generate highly beneficial spillover effects, establish linkages with local companies and could play an outstanding role in reducing regional inequalities. Market forces, however, proved to be stronger than governments efforts to realize these gains from FDI. In the case of regional policy, agglomeration effects were much more influential in determining the location choice of foreign investors than government s fiscal (tax allowances with lower minimum investment requirements) and financial (grants and preferential credits) incentives to direct foreign investment into backward and high unemployment regions. This is not to say, however, that there were not some individual success stories of FDI attraction when some inventive proactive local government (Tatabánya, Rétság, Szekszárd) could overcome the general trend. Foreign firms due to the favourable tax conditions and predictable business environment also started to allocate some of their profits to Hungary. From 1997, the growing majority of foreign investment was not new equity investments but approximately 63% of them were reinvested earnings and to a less extent, intra-firm loans (Palánkai, 2003) The importance of this trend was recognized by investments promotional policies with a considerable time lag. The confusion was a result of some distorted analysis based on Hungarian balance of payments statistics which under the heading of FDI comprise equity investments and intercompany loans (from 1995) but leaves out reinvested earnings. 12

252 Antalóczy-Sass (2001, 2003), Palánkai (2003), OECD (2000; 2004); Ernst & Young (2003) and Sass (2003b) give comprehensive overviews on the most important elements of the FDI policies in the pre-accession period. The main incentive tools used by the government were fiscal subsidies, financial incentives, and the promotion of industrial free-trade zones and incentives and tax holidays offered by local governments. Industrial free-trade zones were a very effective FDI promotion scheme. Their unique features are described in detail in Antalóczy- Sass (2003). Exemptions from custom duties and value-added tax in these zones could reduce production costs substantially which explains the popularity of free-trade zones among export-oriented, assembly companies. Incentives and tax holidays offered by local governments played an important role attracting FDI in some urban regions (Székesfehérvár, Győr) These policies were mainly successful in the most developed regions so these measures increased rather than reduced the spatial concentration of FDI. Fiscal incentives: Hungary's corporate tax act provides for a ten-year, 100% corporate tax holiday for all large investments (more than HUF 10 billion or HUF 3 billion in designated underdeveloped areas). This allowance is currently valid until the 2011 tax year. This incentive exempts the company from Hungary's 18% corporate tax. The allowance was made available only in those years in which the revenues of the taxpayer increased by a certain amount. From the second year of putting the investment into operation, the allowance was only available if the taxpayer has hired at least 500 new employees since the year prior to the start of the investment (100 persons in underdeveloped areas). 5-year corporate tax relief for investors operating in industrial areas or preferential regions. 5-year 50% corporate tax relief for investors investing at least HUF 1 billion and for whose sales increased by 5% of the value of its investment within the same tax year. Some allowance provided for investors in the hotel business if their total sales increased at least by 25% or a minimum of HUF 600 millions. 1-year corporate tax relief for investors investing at least HUF 1 billion in any areas of Hungary. A fiscal incentive in the form of a deduction of taxable profits exists for employers who employ individuals who had previously been unemployed for at least six months Accelerated depreciation schedules were introduced in industrial zones. For investments in underdeveloped priority regions where unemployment exceeds 15% or in certain enterprise zones, the government offers regional tax allowances equal to the full corporate tax calculated on the basis of net revenues In enterprise zones, companies may receive tax relief for up to five tax years, as long as net annual sales increases exceed 1% per year. Up to 6% tax relief for machinery or infrastructure investment in preferential areas. Direct costs of R&D can be subtracted from the tax base.(palánkai, 2003); (OECD, 2004) 13

253 Financial incentives: One of the principal sources of direct financial assistance to businesses was the Targeted Allocation for the Development of the Economy. In 1996, the programme had HUF 15 billion at its disposal. Assistance might have involved repayable loans or non-repayable grants, which are awarded on the basis of a competitive tendering process. The specific forms of subsidy included interest-free loans, interest subsidies, interest-free payment in instalments for the customs duty payable on import leasing, bank loans refinanced by the National Bank of Hungary and equity participation by the State. The amount of the subsidy could not normally exceed one third of the combined costs of the investment in manufacturing facilities and infrastructure and could not be greater than HUF 200 million (with the exception of individual manufacturing investments exceeding HUF 1 billion). The National Technology Development Board provided high technology investment support of HUF 40 million for projects that could be realized within three years. Among the other main forms of assistance for investment was the Targeted Allocation for Regional Development. This fund provided assistance to companies that created new jobs and hired the unemployed in regions where unemployment exceeded 15%. All business organizations registered in Hungary, upon submission of economically viable proposals, were eligible for assistance amounting to HUF 700,000 per one new job created. One half of the funds was allocated in a decentralized manner, based on the decisions of the regional development councils of the individual counties concerned. (Source: Regional incentives in the form of provincial development funds are available in all of Hungary's 19 counties as long as the supported project creates at least 100 jobs. The incentive may take the form of a grant, loan, or interest support, and all allocations are used strictly for greenfield investment projects. Small and medium-sized enterprise (SME) incentives are available for foreign-owned firms that are registered in Hungary, in the form of interest support for capacity-increasing investments. The company must have fewer than 250 employees, HUF 4 billion in annual revenues and HUF 2.7 billion total assets. FDI promotion after accession to the EU Becoming a member of the EU has been the most important event that will determine the development opportunities of backward regions and will certainly have a strong influence on the investment decisions of foreign businesses. Hungary s accession to the EU will mean a more transparent, firmly rule-based and more predictable environment for FDI. Government s incentives related to FDI be it fiscal, financial or other unique features such as the industrial free trade zones had to be adjusted to be compatible with the EU requirements. Investment incentives are available to all enterprises registered in Hungary, regardless of the nationality of owners or location of incorporation. To comply with European Union rules, the government of Hungary no longer grants tax holidays based on investment volume. The EU will not deal with the tax allowances provided before 1 January 2003 but following that date the support intensity of the given investment had to be taken into account. 14

254 The level of support is expressed in terms of "investment intensity", an index expressing the ratio of tax allowances + grants + other financial incentives compared to the total cost of investment. The intensity ratio is set individually for each region and industrial sector. Thus, the intensity index is 35% for Budapest, 40% for Pest County, 45% for the Western part of Transdanubia and 50% for the other regions. For "sensitive" industries (e.g. automotive, mining, synthetic yarns) the intensity index is further reduced. Certain forms of state support, such as training and retraining grants, support for the building of infrastructure outside the factory premises, support for the establishment of industrial parks, the setting up of a reserve tax fund, costs of R&D, accelerated depreciation and certain deductions from the tax base would remain outside the cost elements added up when calculating the eventual investment intensity. For very large investments individual support packages can be granted by the Hungarian Government. (HTC 2004) There are four groups of companies regarding support intensity: Companies which invested before January 1st 2001 are eligible to receive investment allowances up to 75% of their total investments. Companies which invested between January 1st 2001 and December 31st 2002 are entitled to receive tax relief of up to 50% of their investments. 11 car manufacturers are placed in two further groups: Car manufacturers which invested before 2001 are entitled to receive tax relief of up to 30% of their investments. Car manufacturers which commenced their investment after 2001 have a tax relief of up to 20% of their investment. According to the agreements, companies can add investments completed before 31st December 2005 to the basis of the calculated tax allowances. (Palánkai, 2003) The fate of the most efficient Hungarian FDI scheme, the industrial free trade zones, however, is sealed since they are already stripped of their essence and those that do not disappear will function as general industrial parks. Smart Hungary Programme Since 1999 Hungary has been increasingly lagging behind in the regional competition for FDI. The reasons for this deterioration which include the macroeconomic mismanagement of the country and the regional competition in FDI incentives are multi-layered and are analysed extensively elsewhere. Here we touch upon the renewed efforts of the present government to create a new investment promotion strategy to revitalize FDI inflows. Smart Hungary, the Government's new investment incentive program which already incorporates the new EU-compliant tax rules, was launched in January Learning from the successful examples of investment promotion agencies in other parts of the world 15

255 the government wants the Hungarian Investment and Trade Development Agency (ITDH) to act as a "one-stop shopping" hub for foreign investors seeking government support subsidies and licenses. Corporate taxation remains the cornerstone of investment incentive strategy as shown by the following details: Corporate-tax decrease down to 16 % (from 18%), one of the lowest rates in continental Europe For major investors, 10-year-long development tax benefit up to 80% of the due corporate tax One of the more favourable VAT regulations in Europe Corporate tax benefit up to 25% of local tax Tax-free investment reserve Tax allowances for corporate R&D and innovation Tax-related incentive for adult training On the effectiveness of FDI incentives One stream of literature based on empirical research comes to the conclusion that agglomeration forces and proximity to export markets are the most important determinants of the site selection decision of investors. It follows from this that subsidy coming from FDI policies must be very large to overcome agglomeration and distance effects. In the case of Hungary we do not know of even approximate calculations which would relate the costs of incentives that aim at channelling FDI in the less developed regions to the social benefits of such diversions. We are inclined to believe that similarly to the experiences of other countries and regions, narrower FDI incentives play a lesser role in investment location decisions than do traditional determinants such as economic stability, human resources, infrastructure or distance. The EU fund for which Hungary will be eligible in the coming years will be earmarked for projects such as motorway construction, IT development etc. Once completed these projects could elevate the attractiveness of regions so far excluded from the interest of foreign investors. 16

256 References Antalóczy K. Sass M. (2000): Mûködõtõke-áramlások, befektetõi motivációk és befektetés-ösztönzés a világgazdaságban és Magyarországon. Közgazdasági Szemle, No. 5. pp Antalóczy, K. and Sass, M. (2003). Befektetésösztönzés és Magyarország csatlakozása az Európai Unióhoz. (FDI incentives and Hungary s EU-accession). Külgazdaság, (XLVII:4), pp Békés, G. (2004): Location of manufacturing FDI in Hungary: How important are business-to-business relationships? CEU, Budapest, mimeo. Campos, N. F. Kinoshita, Y. (2000). Foreign Direct Investment as Technology Transferred: Some Panel Evidence from the Transition Economies. William Davidson Institute (William Davidson Working Paper 438.), Csáki, Gy. Sas M. Szalavetz, A. (1996): A külföldi tőke modernizációs szerepe. Külpolitika, 2. Évf. 2. Sz. pp Csengődi, S. R. Jungnickel D.Urban (2003): Foreign Takeovers and Wages in Hungary. Paper prepared for the 4 th workshop of the fifth framework programme project FLOWENLA at CEPS, Brussels Damijan, J. P. Knell, M. Majcen, B. Rojec, M. (2003). Technology Transfer through FDI in Top-10 Transition Countries: How Important are Direct Effects, Horizontal and Vertical Spillovers? William Davidson Institute (William Davidson Working Paper 549.), Ernst & Young (2003): Hungary Challenges for Direct Taxation., EU+10 Tax Accession News. Ernst & Young monthly newsletter. Issue 7. Farkas, Péter The Effects of Foreign Direct Investment on R&D and Innovation in Hungary. Institute for World Economics, Hungarian Academy of Sciences Working Papers, No Fazekas K. (2000). The Impact of Foreign Direct Investment Inflows on Regional Labour Markets in Hungary. IWM Policy Project (SOCO Project Paper 77c) Fazekas, K. (2003): Effects of FDI inflows on regional labour market differences in Hungary, paper presented at the International Conference Reinventing Regions in the Global Economy organized by the Regiona1 Studies Association. 12 th -15 th April Pisa, Italy. Fazekas, K. (2004): Spatial concentration of domestic and foreign investment enterprises in Hungary, in Fazekas, K. and K. Koltay (eds) The Hungarian Labour market Review and Analysis, Institute of Economics, Budapest. Fazekas, K. Köllő, J. (1998): A külföldi érdekeltségű vállalatok munkaerő-keresletének jellemzői Magyarországon 1995-ben. [Characteristics of labour-demand of foreign interest companies in Hungary], in: Fazekas, K. (ed.): A munkaerőpiac és regionalitás az átmeneti időszakban. [Labour market and regionality in the period of transformation] MTA Közgazdaságtudományi Kutatóközpont, Budapest, pp Hamar J. (1999): A külföldi működőtőke beáramlás Magyarországon belüli területi jellemzői, (Regional characteristics of foreign direct investment in Hungary) Külgazdaság, Vol. XLIII. March. pp Hamar, J. (2001): A külföldi működőtőke és az egyensúlyi problémák. Inf-Társadalomtudomány. 46. Sz. pp Hamar J. (2001). A külföldi és hazai tőkével működő vállalatok szerepe a magyar iparban. Külgazdaság 65, HTC (2004): Laws and Regulations Related To Foreign Investments, Hungarian Trade Commission, London. 17

257 Kalotay, K. (2001): The Contribution of Foreign Direct Investment to Transition. Revisted. The Journal of World Investment., Geneva, Vol. 2. No. 2. June, pp Kalotay Kálmán (2003): Müködõtõke válságban? Közgazdasági Szemle, Vol 1. pp Kertesi G. & Köllő J. (1999): Economic Transformationand the Return to Human Capital. Budapest Working Papers on the Labour Market, 1999/6, MTA KTI-BKE, Budapest. Kertesi,G. & Köllő, J. (2001): Economic transformation and the revaluation of human capital - Hungary, , Labour Research Department, Institute of Economics, Hungarian Academy of Sciences, BWP. 2001/4, Budapest Kertesi, G. & Köllő, J. (2002) Labour demand with heterogeneous labour inputs after the transition in Hungary, and the potential consequences of the increase of minimum wage in 2001 and 2002, Institute of Economics, Hungarian Academy of Sciences, BWP. 2002/5., Budapest Kézdi, G. (2002): Two Phases of Labor Market Transition in Hungary: Inter-Sectoral Reallocation and Skill- Biased Technological Change Budapest Working Papers on the labour Market, BWP/2002/3. Labour Research Department, Institute of Economics, Hungarian Academy of Sciences Department of Human Resources, Budapest University of Economics and Public Administration. Könings, J, (2000): The Effect of Direct Foreign Investment on Domestic Firms: Evidence fromfirm Level Panel Data in Emerging Economies William Davidson Institute Working Paper 344 Kőrösi G. (1997). Labour Demand During Transition in Hungary. Ann Arbor : Michigan : William Davidson Institute. (William Davidson Institute, WDI Working Paper 116.) Kőrösi, G. (2003): Job Creation and Destruction in Hungary, CEU-Economics WP1/2003 Köllő, J. (2002): Wage evolution by economic sector. Effects of ownership, in: Fazekas, K. and K. Koltay (eds) The Hungarian Labour market Review and Analysis, Institute of Economics, Budapest. Köllő, J. (2004): Regional differences in Earnings and Wage Costs, in: Fazekas, K. and J. Koltay (Eds.) The Hungarian Labour Market Review and Analysis, IE-HAS Budapest Lücke, M. Szalavetz A. (1997). Export Reorientation and Transfer of Know-how and Technology the Case of Hungarian Manufactured Exports. Kieler Weltwirtscaftliches Institut. (Kieler Arbeitspapiere 801.) Major I. Vezzoni, C. Szalavetz A. (1999). Company Restructuring after Privatization in Hungary Between In: Privatization and Economic Performance in Central and Eastern Europe. / ed. J. Major Cheltenham: Edward Elgar. OECD (2000): OECD Reviews of Foreign Direct Investment. Hungary.OECD, Paris. OECD (2004): OECD Economic Surveys 2004 Hungary. OECD, Paris. Palánkai, T. (2003): Determining Factors of Environment of FDIs in CEE Associazione Universitaria di Studi Europei Pavia, Italy Sass, M. (2001): Foreign direct investment incentive policies in the Czech Republic Estonia, Hungary, Poland Slovakia and Slovenia. Paper prepared in the framework of the PHARE ACE project Government policies to enhance spillovers from FDI in CEECs. Sass, M. (2003a): Competitiveness and economic policies related to foreign direct investment. Ministry of Finance Working Paper No. 3. September. Sass, M. (2003b). The effectiveness of host country policy measures in attracting FDI: The case of Hungary. In: The Development Dimension of FDI: Policy and Rule-Making Perspectives. Proceedings of an Expert Meeting. UNCTAD, Geneva pp Sass, M. (2004): FDI in Hungary the first mover s advantage and disadvantage EIB Papers Vol. 9 No.2 18

258 Schoors, K. Van der Tol, B. (2001). The Productivity Effect of Foreign Ownership on domestic Firms in Hungary, SBS, Oxford University, Mimeo. Szalavetz (1996): A külföldi működőtőke befektetések multiplikátor hatása: mikrogazdasági példák makrogazdasági megközelítésben. Külgazdaság, 40. Évf. 2.sz. pp Szalavetz, A. (1994): Mit hoz mit visz a befektető? A külföldi tőke multiplikátor-hatása. Cégvezetés, 10. Sz. pp Szanyi, M. (2002): A külföldi tulajdonú cégek Magyarországon: Új fejlődési modell központi szereplői? Kihívások., 159.sz. október, MTA VKI, Budapest, p.11. Tarjáni, H. (2004): Estimating The Impact Of Sbtc On Input Demand Elasticities In Hungary, Mnb Working Paper 2004/3. Török, Á. (2004): Vállalati és állami szerep amagyar iparfejlődésben a kilencvenes évek után. Competitio, III. évf. 2. Sz. 19

259 Figure 1 Time path of FIEs employment in the corporate sector, 1993 to FIEs FIEs share (right scale) Note: Financial sector excluded. Source: CSO-FDI database. 20

260 Figure 2 Net job creation rates in the corporate sector, 1992 to % DEs FIEs Source: Calculation of Gábor Körösi. 21

261 Figure 3 Spatial distribution of micro-regions in the four quartiles of employment rates, 1990 and Note: White: top quartile, high-employment micro regions; Black: bottom quartile, low-employment micro regions. Source: HCSO Census 1990,

262 Figure 4 Wage costs and productivity of firms settled in high-employment regions compared to firms settled in low-employment regions in manufacturing, Foreign enterprises Domestic enterprises Wage costs Productivity Unit Labour Costs Note: Firms settled in low employment regions = 100%. Source: IE-FDI micro regional data base. 23

263 Figure 5 Time path of the productivity gap between firms in manufacturing settled in low- and high-employment regions, 1993 to y = 3,6252x + 142,43 R 2 = 0, % y = 3,6323x + 108,5 R 2 = 0, PG_FIEs PG_DEs Note: Productivity gap = (average productivity firms settled in high employment regions) / (average productivity of firms settled in low employment regions) *100 Productivity = net sales/employees Source: Fazekas (2004). 24

264 Table 1 Owners equity of enterprises by industries, 2002 Owners' equity of enterprises Code Industries, branches FIEs DEs Total FIEs share Billion HUF % A,B Agriculture, forestry and fishing C Mining and quarrying D Manufacturing DA Manufacture of food products, beverages DB Manufacture of textiles and textile products DC Manufacture of leather and leather products DD Manufacture of wood and wood products DE Manufacture of pulp, paper and paper products; publishing and printing DF-DG Manufacture of coke, refined petroleum products and nuclear fuel; of chemicals, chemical products, and manmade fibres DH Manufacture of rubber and plastic products DI Manufacture of other non-metallic mineral products DJ Manufacture of basic metals and fabricated metal products DK Manufacture of machinery and equipment n.e.c DL Manufacture of electrical and optical equipment DM Manufacture of transport equipment DN Manufacturing n.e.c E Electricity, gas, steam and water supply F Construction G Wholesale and retail trade; repair of motor-vehicles, and household goods H Hotels and restaurants I Transport, storage, post and telecommunications K Real estate, renting and business activities M Education N Health and social work O Other community, soc. and pers. services Activities TOTAL Source: CSO

265 Table 2 Largest exporting foreign investment enterprises in Hungarian manufacturing, 2001 Hungarian Affiliate TNC Source Country Industry Export Volume (million USD) Export Share (%) Audi Hungária Motor Kft Volkswagen Germany Motor vehicles IBM Storage Products Kft IBM USA Electronics Philips Magyarország Philips The Netherlands Electronics GE Hungary Rt. General Electric USA Electronics Opel Magyarország Járműgyártó General Motors USA Motor vehicles Kft. Flextronics International Kft Flextronics Singapore Electronics Alcoa Köfém Kft Alcoa USA Aluminium Products Suzuki Rt. Suzuki Motor Japan Motor vehicles NABI Rt. North American USA Motor vehicles Bus Samsung Electronics Magyar Rt. Samsung South Korea Electronics Electronics Electrolux Lehel Hűtőgépgyár Electrolux Sweden Household Kft equipment Visteon Hungary Kft Visteon USA Electronics Motor vehicles Delphi Packard Hungary Kft Delphi USA Motor vehicles Automotive Systems Egis Gyógyszergyár Rt. Servier France Pharmacy Total Source: Török (2004). 26

266 Table 3 Productivity of FIEs and DEs by industries, 2002 Gross Sales/Employees Gross value added/empl. Industries, branches FIEs DEs FIEs/DE s FIEs DEs FIEs/DE s Code Million HUF/ employee Million HUF/ employee A,B Agriculture, forestry and fishing C Mining and quarrying D Manufacturing DA Manufacture of food products, beverages and tobacco DB Manufacture of textiles and textile products DC Manufacture of leather and leather products DD Manufacture of wood and wood products DE Manufacture of pulp, paper and paper products; publishing and printing DF-DG Manufacture of coke, refined petroleum products and nuclear fuel; of chemicals, chemical products, and man-made fibres DH Manufacture of rubber and plastic products DI Manufacture of other non-metallic mineral products DJ Manufacture of basic metals and fabricated metal products DK Manufacture of machinery and equipment n.e.c DL Manufacture of electrical and optical equipment DM Manufacture of transport equipment DN Manufacturing n.e.c E Electricity, gas, steam and water supply F Construction Wholesale and retail trade; repair of G motor-vehicles, and household goods H Hotels and restaurants I Transport, storage, post and telecommunications K Real estate, renting and business activities M Education N Health and social work O Other community, social and personal services activities TOTAL Source: CSO

267 Table 4 Number of employees in foreign investment and domestic enterprises in the corporate sector*, 1993 to 2002 Number of employees* FIEs DEs Total FIEs share 1000 person % Note: FIEs = enterprises with higher than 10 % of foreign owned equity * Financial sector excluded Source: IE-FDI Micro regional data base Table 5 Employees of the corporate sector by the share of foreign ownership, per cent Foreign Ownership % Majority Minority % Note: : firms employing 10 or more workers; : firms employing 5 or more workers. Source: EO-Wage Survey. 28

268 Table 6 Net job creation rates in the corporate sector, 1992 to 2002 (%) DEs Sector Agriculture Mining and energy Manufacturing TCF Chemical Engineering Trade Construction Services Size: Large firms Medium sized Small firms DEs Total FIEs Sector Agriculture Mining and energy Manufacturing TCF Chemical Engineering Trade Construction Services Size: Large firms Medium sized Small firms FIEs Total Note: TCF stands for Textile, Clothing and Footwear. Source: Körösi (2004). 29

269 Table 7 Number of employees and firms by industries, 2002 Number of employees Number of enterprises Code Industries, branches FIEs DEs FIEs share % FIEs DEs FIEs share % A,B Agriculture, forestry and fishing , ,4 C Mining and quarrying , ,0 D Manufacturing , ,6 DA Manufacture of food products, beverages and , ,7 tobacco DB Manufacture of textiles and textile products , ,6 DC Manufacture of leather and leather products , ,1 DD Manufacture of wood and wood products , ,6 DE Manufacture of pulp, paper and paper products; , ,6 publishing and printing DF-DG Manufacture of coke, refined petroleum chemical , ,2 products, products and nuclear fuel; of chemicals and man-made fibres DH Manufacture of rubber and plastic products , ,6 DI Manufacture of other non-metallic mineral products , ,8 DJ Manufacture of basic metals and fabricated metal , ,8 products DK Manufacture of machinery and equipment n,e,c, , ,9 DL Manufacture of electrical and optical equipment , ,3 DM Manufacture of transport equipment , ,5 DN Manufacturing n,e,c, , ,3 E Electricity, gas, steam and water supply , ,2 F Construction , ,7 G Wholesale and retail trade; repair of motor-vehicles, , ,7 and household goods H Hotels and restaurants , ,2 I Transport, storage, post and telecommunications , ,4 K Real estate, renting and business activities , ,0 M Education , ,9 N Health and social work , ,0 O Other community, soc, and pers, Services activities , ,7 TOTAL , ,1 Source: CSO

270 Table 8 Distribution of FIE and DE employees by classification of manufacturing industries based on the level of technology, 2002 Manufacturing industries FIEs DEs FIEs DEs (ISIC Rev. 3) Person % High-technology industries Medium-high-technology industries Medium-low-technology industries Low-technology industries Total manufacturing Note: Classification of industries follows the latest OECD classification of manufacturing industries based on technology. OECD (2003). Source: CSO-FDI database. 31

271 Table 9 Changes of composition of FIE employment by industries, 2000 to 2002 Industry FIEs empl, 2000 FIEs empl, 2002 Change FIEs empl, 2000 FIEs empl, 2002 Change Person Person Person % % % Agriculture, forestry and fishing A Mining and quarrying CA-CB ,3 0,2-0,1 Manufacturing D ,6 59,7 1,1 + Electricity, gas, steam and water supply E ,2-0,8 Construction F ,1 1,5-1,6 - Wholesale and retail trade; repair of G ,1 15,9 1,8 + motor-vehicles, Hotels and restaurants H ,1 0 Transport, storage, post and I ,7-0,4 telecommunications Financial J ,5 4,3-0,1 Real estate, renting and business K ,3 6,1-0,1 activities Education M ,1 0,1 0 Health and social work N ,2 0,3 0,1 Other community, social and personal O ,7 0,9 0,1 services activities TOTAL Total Manufacturing food products, beverages and tobacco DA ,5-0,5 textiles and textile products DB ,8 8,9-1,8 - leather and leather products DC ,4 3,4 0 wood and wood products DD ,5 1,3-0,2 pulp, paper and paper products; publishing and printing DE ,1 3,1 0 coke, refined petroleum and nuclear fuel; DF-DG ,5 8,1 1,6 + of chemicals, rubber and plastic products DH ,5 5,7 1,2 + non-metallic mineral products DI ,1 3,3-0,9 basic metals and fabricated metal DJ ,4 6,8 0,4 products machinery and equipment n,e,c, DK ,9 electrical and optical equipment DL ,8 27,5-2,3 - transport equipment DM ,6 8,1 1,4 + n,e,c, DN ,2 2,4 0,2 Manufacturing total D Source: CSO

272 Table 10 Employment change in different industries, 2000 to 2002 Manufacturing industries FIEs Employees in 2000 Changes of Changes of FIEs employees FIEs employees between between FIEs Employees in 2000 FIEs Employees in 2002 (ISIC Rev. 3) Person Person % % % High-technology industries Medium-high-technology industries Medium-low-technology industries Low-technology industries Total manufacturing Non manufacturing Total corporate sector Note: Classification of industries follows the latest OECD classification of manufacturing industries based on technology. OECD (2003). Source: CSO-FDI database. 33

273 Table 11 Average gross earnings of employees by industries, 2002 Code Industries. branches FIEs TOTAL FIEs/ TOTAL HUF/year HUF/year) % A.B Agriculture. forestry and fishing C Mining and quarrying D Manufacturing DA Manufacture of food products. beverages and tobacco DB Manufacture of textiles and textile products DC Manufacture of leather and leather products DD Manufacture of wood and wood products DE Manufacture of pulp. paper and paper products; publishing and printing DF- Manufacture of coke. refined petroleum products and nuclear fuel; DG of chemicals. DH Manufacture of rubber and plastic products DI Manufacture of other non-metallic mineral products DJ Manufacture of basic metals and fabricated metal products DK Manufacture of machinery and equipment n.e.c DL Manufacture of electrical and optical equipment DM Manufacture of transport equipment DN Manufacturing n.e.c E Electricity. gas. steam and water supply F Construction G Wholesale and retail trade; repair of motor-vehicles. and household goods H Hotels and restaurants I Transport. storage. post and telecommunications K Real estate. renting and business activities M Education N Health and social work O Other community. social and personal services activities TOTAL Source: CSO

274 Table 12 FDI stock by regions, 1995 to 2002 % Regions Counties Budapest Pest Central Hungary Fejér Komárom-Esztergom Veszprém Central Transdanubia Győr-Moson-Sopron Vas Zala Western Transdanubia Baranya Somogy Tolna Southern Transdanubia Borsod-Abaúj-Zemplén Heves Nógrád Northern Hungary Hajdú-Bihar Jász-Nagykun-Szolnok Szabolcs-Szatmár-Bereg Northern Great Plain Bács-Kiskun Békés Csongrád Southern Great Plain Not allocated TOTAL Source: CSO (2004). 35

275 Table 13 Distribution of the number of FIEs across regions and counties, 1995 to 2002 % Regions Central Counties Budapest 48,6 49,5 50,5 51,6 52,8 53,8 53,9 52,7 Pest 3,7 3,8 3,5 3,5 3,4 2,9 2,7 2,5 52,2 53,3 54,0 55,1 56,2 56,7 56,5 55,2 Hungary Fejér 4,5 4,1 3,5 3,2 3,0 3,1 2,9 2,6 Komárom-Esztergom 1,1 1,0 1,0 0,9 0,9 0,9 0,9 0,9 Veszprém 1,4 1,3 1,4 1,5 1,4 1,4 1,3 1,4 Central 7,1 6,4 5,9 5,6 5,4 5,5 5,1 4,9 Transdanubia Western Gy.M.S. 6,8 6,6 5,4 4,6 4,1 3,6 3,2 2,6 Vas 2,0 2,0 1,9 1,8 1,7 1,6 1,6 1,6 Zala 5,0 4,9 4,8 4,8 4,8 4,6 4,3 4,3 13,9 13,5 12,1 11,2 10,6 9,8 9,0 8,5 Transdanubia Baranya 1,5 1,5 1,4 1,3 1,2 1,1 1,1 1,0 Somogy 1,0 1,0 1,1 1,2 1,1 1,1 1,0 1,0 Tolna 1,9 2,0 2,2 2,0 1,9 2,1 2,3 2,1 Southern Transdanubia 4,4 4,6 4,7 4,5 4,3 4,4 4,4 4,1 BAZ 0,6 0,6 0,6 0,6 0,6 0,6 0,6 0,6 Heves 6,0 6,3 6,5 6,7 7,1 7,5 7,5 8,0 Nógrád 2,5 2,6 2,6 2,3 2,3 2,0 2,0 1,8 Northern 9,1 9,5 9,6 9,6 10,0 10,1 10,1 10,4 Hungary Northern Great Plain Southern Hajdú-Bihar 1,2 1,3 2,2 2,8 2,3 2,6 4,5 6,7 JNSZ 1,1 1,0 1,1 1,1 1,0 0,9 0,9 0,9 SzSzB 1,1 1,0 1,0 1,0 1,0 1,0 0,9 0,9 3,3 3,3 4,3 4,9 4,3 4,5 6,3 8,6 Bács-Kiskun 2,6 2,6 2,7 2,7 2,7 2,7 2,6 2,6 Békés 3,0 2,9 3,0 3,0 3,1 3,0 2,9 2,8 Csongrád 4,3 3,8 3,6 3,5 3,4 3,2 3,0 2,9 9,9 9,4 9,4 9,2 9,2 9,0 8,6 8,3 Great Plain TOTAL 100,0 100,0 100,0 100,0 100,0 100,0 100,0 100,0 Source: CSI FDI database. 36

276 Table 14 Distribution of the number of FIE employees across regions and counties, 1995 to 2002 Regions Counties Budapest 47,0 46,6 44,0 42,4 40,9 40,1 39,9 38,8 Pest 3,4 3,2 2,8 2,8 2,8 2,5 2,4 2,3 Central Hungary 50,4 49,8 46,8 45,3 43,7 42,5 42,3 41,1 Fejér 2,7 2,6 2,5 2,6 2,5 2,7 2,6 2,6 Komárom-Esztergom 1,1 1,3 1,3 1,5 1,2 1,3 1,4 1,2 Veszprém 2,9 4,3 3,5 3,6 3,5 3,6 3,9 3,7 Central Transdanubia 6,8 8,1 7,3 7,7 7,2 7,6 7,9 7,4 Gy.M.S. 3,3 3,0 3,1 3,0 2,8 2,6 2,6 2,6 Vas 3,4 3,8 3,7 3,9 4,2 5,3 5,3 4,9 Zala 5,5 5,4 6,4 6,2 7,1 5,8 5,8 5,8 Western Transdanubia 12,2 12,2 13,2 13,0 14,1 13,8 13,7 13,3 Baranya 3,7 3,2 3,3 3,2 3,4 2,7 2,5 2,4 Somogy 2,6 2,5 2,3 2,4 2,4 2,3 2,4 2,4 Tolna 2,7 2,7 2,8 2,8 2,8 4,2 3,4 3,2 Southern Transdanubia 9,0 8,3 8,4 8,5 8,6 9,2 8,3 8,0 BAZ 0,9 0,9 0,8 0,9 0,8 0,8 0,8 1,0 Heves 5,1 6,2 7,6 8,3 8,7 9,9 10,9 13,3 Nógrád 1,8 1,7 1,7 1,8 1,9 2,7 2,7 2,7 Northern 7,7 8,8 10,2 11,0 11,4 13,4 14,4 17,0 Hungary Hajdú-Bihar 1,1 1,5 1,5 1,6 1,6 1,5 1,6 1,7 JNSZ 1,6 1,6 1,9 2,1 2,0 2,1 2,2 2,1 SzSzB 0,9 0,9 1,2 1,3 1,3 1,3 1,3 1,3 Northern 3,6 4,0 4,6 5,0 5,0 4,9 5,1 5,1 Great Plain Bács-Kiskun 4,1 4,1 4,5 4,7 5,0 3,9 3,9 4,1 Békés 3,2 2,5 2,7 3,0 3,0 2,8 2,5 2,3 Csongrád 2,9 2,3 2,3 2,1 2,0 1,9 1,9 1,7 Southern 10,2 8,9 9,5 9,9 10,0 8,6 8,4 8,1 Great Plain TOTAL 100,0 100,0 100,0 100,3 100,0 100,0 100,0 100,0 Source: CSI FDI database. 37

277 Table 15 Changes of corporate employment in low- and high-employment micro regions between 1993 and 2002* Low employment micro regions (Top quartile) High employment micro regions (Bottom quartile) Country Total DEs FIEs Total DEs FIEs Total DEs FIEs Total * As a percentage of the working age population of the regions. Note: Financial sector excluded. Source: IE-FDI Micro Regional Database. 38

278 Economic Restructuring and Labour Markets in the Accession Countries Research Project commissioned by EU DG Employment, Social Affairs and Equal Opportunities, Contract No. VC/2003/0367 Coordinated by in cooperation with Alphametrics and The Vienna Institute for International Economic Studies Agricultural regions and regional policy in Poland by Eugeniusz Kwiatkowski*, Paweł Gajewski* and Tomasz Tokarski** August 2004 *) Institute of Macroeconomics, University of Lodz **) Institute of Economics and Management, Jagiellonian University, Cracow

279 Eugeniusz Kwiatkowski, Paweł Gajewski and Tomasz Tokarski Agricultural regions and regional policy in Poland 1 Introduction Transition to a market economy in Poland caused various economic changes. In particular, in the labour market, the phenomenon of open unemployment appeared, replacing the phenomenon in centrally planned economies of over-employment and hidden unemployment. An important feature of Polish unemployment during the transition has been its strong regional variation. Differences in unemployment across regions is related, inter alia, to their economic structure, that is, to the importance of agriculture, industry and services in the regional economy. The goal of the paper is to present regional variations in unemployment in Poland, paying particular attention to unemployment in agricultural regions. The aim is not only to examine the extent of open unemployment in these regions, but to attempt to estimate the size of over-employment in agriculture. The potential consequences for unemployment of reducing over-employment in the different regions are also considered. In addition, state policy towards agricultural and rural areas is examined. The empirical analysis presented here is based on data for Polish voivodships. Data on employment and unemployment are available for the years , while data on the structure of employment and value added are available for the period The analysis breaks down regions into the 16 voivodships (NUTS 2 regions) which were defined for EU structural policy and other purposes in January, 1999 (see: map 1). Data for earlier years were aggregated into the same regional groupings. The paper begins by setting out regional variations in unemployment rates, the sectoral structure of employment and labour productivity. Secondly, it examines the extent of overemployment in agriculture in the different regions. Thirdly, it considers the potential impact on unemployment of reducing over-employment and estimates the investment required to create the jobs needed in other sectors to employ those at present working in agriculture for no other reason than they cannot find employment elsewhere. Fourthly it examines the challenge facing government policy to alleviate agricultural and regional problems. 2 Labour market developments in Polish regions The concern here is to analyse major developments in regional labour markets over recent years. Map 2 shows average unemployment rates in the Polish voivodships over the period As it can be seen, the highest rates occurred in the Warminsko- 1

280 Mazurskie voivodship (north-eastern Poland) and in the western voivodships of Zachodniopomorskie and Lubuskie. The lowest rates of unemployment, on the other hand, were recorded in the eastern part of the country, in Podlaskie, Lubelskie, Mazowieckie (where Warsaw is situated), Wielkopolskie and Malopolskie. In the remaining voivodships average unemployment was between 12.4% and 17.4%. High unemployment rates in Warminsko-Mazurskie, Zachodniopomorskie and Lubuskie (as well as Dolnoslaskie) can be explained by the liquidation of state-owned farms (socalled PGRs Panstwowe Gospodarstwa Rolne) at the beginning of the transition period, which caused very high structural unemployment. Because of the lack of jobs in other sectors and the low chances of finding a job, high long-term unemployment progressively emerged in these regions. Low open unemployment in Mazowieckie can be attributed to two main factors: first, the dynamic development of the service sector in Warsaw and surrounding areas; secondly, the high hidden unemployment in agricultural areas outside Warsaw. Mazowieckie, therefore, consists, apart from the capital city, chiefly of rural areas with private (mostly family-run) farms, in particular, in the former voivodships of Ciechanowskie, Ostroleckie and Siedleckie, in which, in 1996, the share of employment in agriculture exceeded 50% and which make up over half of the present land area of the new region. The low rates of unemployment observed in Podlaskie and Lubelskie also results to a large extent from over-employment or hidden unemployment in agriculture as those losing their jobs in industry or services and living in rural areas or have the possibility of moving to these areas take up work on the land in order to make a living. Low unemployment in Malopolskie, on the other hand, is in a large part a consequence of the delay in restructuring heavy industries. Table 1 gives a more detailed indication of unemployment rate developments in Polish regions. Over the period examined, there was an increase in unemployment in all 16 regions, though to very differing extents. Overall, unemployment rose by nearly 6 percentage points during these 7 years. The largest increases were in Lubuskie (11.3 percentage points) and Zachodniopomorskie (10.2 percentage points), while there were also above average rises in Warminsko-Mazurskie, Opolskie, Kujawsko-Pomorskie and Dolnoslaskie. The smallest increases in unemployment occurred in Mazowieckie and Wielkopolskie (2.9 and 3.9 percentage points, respectively), where the relatively dynamic metropolitan centres of Warsaw and Poznan are located. Map 3 shows the average sectoral division of employment by region over the years It indicates that the eastern regions of Lubelskie, Podlaskie, Swietokrzyskie and Podkarpackie are predominantly agriculture ones (the share in total employment 2

281 exceeding a third). The smallest proportion of employment in agriculture is in the western regions of Slaskie, Zachodniopomorskie, Lubuskie, Dolnoslaskie and Pomorskie. These differences in the share of employment in agriculture is related to the pattern of ownership before In northern and western parts of Poland, state-owned farms (PGRs) accounted for more than 40% of the farmland and their liquidation was accompanied by a big reduction in the number of people working in agriculture. On the other hand, in the eastern and central parts of the country in Lubelskie, Podlaskie, Podkarpackie, Swietokrzyskie, Lodzkie and Mazowieckie the share of land controlled by state-owned farms was 10% at most and meant that a large proportion of people worked in small, backward private farms. Lubelskie and Podlaskie, in particular, are both typical rural regions, distant from large metropolitan areas and with poor socio-economic infrastructure. Together they have the smallest shares of employment in industry in the country. By contrast, industry is a major source of employment in Slaskie, while in most other regions, services employ the majority of people. The highest shares of employment in services are in Zachodniopomorskie, Pomorskie Dolnoslaskie, Lubuskie, Warminsko-Mazurskie and Mazowieckie, in all of which over half of those in work were employed in the sector. Table 2 presents data on the division of employment between agriculture, industry and services in 1995 and in It indicates that shifts in employment between the three sectors were by no means similar across regions The largest changes occurred in Pomorskie, Slaskie, Swietokrzyskie, Warminsko-Mazurskie and Zachodniopomorskie, while the smallest ones were in Kujawsko-Pomorskie, Lubelskie, Mazowieckie and Wielkopolskie. Interestingly, two regions experienced shifts in the opposite direction to the general trend in Poland. In Opolskie and Wielkopolskie, the share of employment in agriculture did not decline and the share in services was lower in 2001 than in Table 2 enables typically agricultural regions to be identified. If an agricultural region is defined as one in which over 20% of employment is in agriculture (ie a share larger than the Polish average in 2001), the following agricultural regions can be distinguished: Lubelskie (38%), Podlaskie (37%), Podkarpackie (30%), Swietokrzyskie (30%), Malopolskie (25%), Mazowieckie (20%) and Opolskie (20%). All these regions except Opolskie are located in eastern or south-eastern Poland near the Ukrainian or Belarussian border. The following features are characteristic of rural areas and of the agricultural sector in Poland: a fragmented agrarian structure of farmland. In Poland, there are (according to data collected at the end of 1990s) over 2 million farms, but as many as 55% of these have a land area of less than 5 hectares and only 19% have more than 10 hectares. The 3

282 average size of farm at the end of the 1990s was only 7 hectares 1. Moreover, some 1.4 million of these farms (around 70%) were producing food solely for their own needs and only 600 thousand farms were selling produce into the market 2 ; labour market problems with large hidden unemployment and in regions, where PGRs were liquidated, high open unemployment rates; low household income levels of farmers and other people living in rural areas. At the end of 1990s, the incomes of rural households were on average 30% lower than those of urban households and all the indicators show significantly higher levels of poverty in rural areas than in towns or cities 3 ; a less-well educated population than in urban areas. In 1995, 55% of rural inhabitants had elementary or less than elementary education as against 31% in urban areas, while 2% had tertiary education as against 10% in urban areas. Table 3 shows differences in labour productivity (measured as value-added per person employed) across regions and changes in real terms between 1996 and 2001 (the period chosen is dictated by the availability of reliable data on regional value-added). It indicates that labour productivity increased in Poland as well as in each region over the period, with the largest rise in Mazowieckie, where the presence of a dynamic service sector was key. Big increases also occurred in Dolnoslaskie and Slaskie, where the largest decline in employment took place. Table 4 shows regional differences in labour productivity by sector over the same period. It indicates that labour productivity in agriculture was only around 15% of the level in services and only around 20% of that in industry and construction. Productivity in agriculture also varied markedly across regions. It was especially low in Malopolskie and Podkarpackie (EUR 890 per person employed in 2001) and only slightly higher in Swietokrzyskie (EUR 1,630), Lubelskie (EUR 1,400 and Podlaskie (EUR 1,690). All of these, as noted above, are in the eastern part of Poland with very high employment in agriculture. On the other hand, productivity was relatively high in Zachodniopomorskie (EUR 9,490 per person employed in 2001), Lubuskie (EUR 4,400), Warminsko-Mazurskie (EUR 4,370) and Wielkopolskie (EUR 4,420). The first three of these were dominated by PGRs before 1989 and have relatively low employment in agriculture, most of it in productive, large farms. Wielkopolskie, on the other hand, comprises traditionally well-developed and highly productive agricultural land National Human Development Report. Poland Rural Development, UNDP, Warsaw 2000, p. 12. Orlowski W., Przeciw Stereotypom, UKIE, Warszawa 2001, p. 74. Natioanl Human Development Report, 2000, pp

283 3 Over-employment in agriculture in the Polish Regions In order to estimate the scale of over-employment (or hidden unemployment) in agriculture across regions in Poland, the approach adopted is to relate labour productivity in these to that other economic sectors and to compare the resulting ratio with that in Spain and Italy. These countries were chosen for this purpose since they are relatively similar to Poland in terms of working population, have relatively high levels of structural unemployment and low mobility of labour and have levels of GDP per head below the EU-15 average. The estimation of hidden unemployment was based on a two-stage approach. The first stage is to estimate "potential" labour productivity in agriculture for each region from the following equation: ŷ i = y~ ŷ i (1) where: y~ i = ωy i (2) ŷ = ωy (3) and ω is relative labour productivity in agriculture of Spain or Italy in 1999 relative to labour productivity in other sectors ( ω equals for Spain and for Italy); y is labour productivity in the non-agricultural sector in Poland in the years ; y i is labour productivity in the non-agricultural sector in the i-th region (i=1,2,...,16) in the years Equation (3) defines the level of the "desirable" labour productivity in the i-th region in Poland on the assumption that the labour productivity in each region depends on the overall labour productivity in the non-agricultural sector and on the relationship of labour productivity in agriculture to labour productivity in the non-agricultural sector in Spain or Italy. This implicitly means that, on the basis of formula (3), the same "potential" labour productivity in agriculture is estimated for each Polish region. Such an approach has the advantage of not assuming that labour productivity in agriculture in the regions with high labour productivity in the non-agricultural sector (e.g. Mazowieckie, Slaskie or Pomorskie) is higher than in regions with low labour productivity in the non-agricultural sector (e.g. Lubelskie, Lodzkie or Warminsko-Mazurskie see Table 4). However, by the same token it means that because of this assumption, potential levels of labour productivity in agriculture are assumed not to be affected by regional differences in socio-economic development. Equation 2 can be introduced to allow "potential" labour productivity in agriculture to vary across regions. This postulates that the ratio of labour productivity in agriculture to that in 5

284 the non-agricultural sector in each region is equal to the ratio in Spain or Italy in Equation (1) calculates the geometric mean from "potential" labour productivity ω. This serves below to denote the "desirable" level of labour productivity in agriculture in each region. In the second stage, employment is calculated from the estimates of "potential" labour productivity given value-added in agriculture in the years Subtracting this hypothetical level of employment from the actual number of people working in the sector gives an estimate of over-employment in agriculture in each region. Estimates of this are shown in Maps 4 and 5. They indicate: the total level of over-employment in Polish agriculture in the years amounted to some million people, over 60% of those employed in the sector. the largest over-employment, measured by both Spanish or Italian variants, is in Mazowieckie ( thousand), Lubelskie ( ), Malopolskie ( ) and Podkarpackie ( ); in Zachodniopomorskie, where agricultural productivity is relatively high, on the other hand, there is estimated to have been underemployment rather than over-employment of around 3.8 thousand if measured by the Italian variant or over-employment of under 1 thousand on the Spanish variant; over-employment is estimated to be low in Lubuskie (4-6 thousand), Warminsko- Mazurskie (22-27 thousand) and Opolskie (28-32 thousand). 4 Macroeconomic effects of reducing over-employment Table 5 shows the effect of eliminating over-employment in agriculture on regional unemployment. It indicates that: average over-employment in Polish agriculture in the years was between 62% and 65% of employment in the sector, depending whether the Italian or Spanish variant is used; the largest overemployment in proportionate terms is in Malopolskie (82-83%), Podkarpackie (79-80%), Swietokrzyskie (75-77%), Lubelskie (73-75%), Mazowieckie (70-73%) and Podlaskie (68-71%). Except for Mazowieckie all the regions are located in eastern or south-eastern part of the country. the smallest extent of overemployment is in the post-pgr regions of Zachodniopomorskie (-7-1.5%), Lubuskie (11-18%) and Warminsko-Mazurskie ( %), all with high unemployment; adding hidden unemployment in agriculture to the existing levels of open unemployment increases the rate by around 11 percentage points. The largest 6

285 increases are in Lubelskie (by percentage points), Swietokrzyskie (22-23 percentage points), Podlaskie (22-23 points), Podkarpackie (21 points) and Malopolskie (18 points), all characterised by by small, pre-industrial farms. the smallest increases of unemployment are in Lubuskie (1-2 percentage points), Warminsko-Mazurskie and Dolnoslaskie (4 percentage points in each), while in Zachodniopomorskie, unemployment is changed by very little. Table 6 and Map 6 show the investment required to create the number of jobs needed outside agriculture to eliminate over-employment in the sector. This was computed on the assumption that to create one job, the same amount of capital is needed as the average capital available per person employed in the economy outside of agriculture in each region in The estimates indicate that: the total amount of additional fixed capital required to eliminate over-employment in agriculture is over EUR 55 billion; the greatest need for capital is in Mazowieckie, at over EUR 14 billion as a consequence of both large over-employment in agriculture and a high capital-labour ratio in non-agricultural sectors of the economy (especially in Warsaw); apart from Mazowieckie, Lubelskie and Podkarpackie and other regions located in eastern Poland have most need for capital because of the scale of over-employment in agriculture even though the capital-labour ratio is relatively low. 5 Sectoral and regional policy in rural areas in Poland Government policy over the period period towards agriculture and rural areas was aimed a 4 : Improving incomes of farmers accelerating changes in the structure of farms strengthening the development of rural areas, creating new jobs and improving education improving agricultural competitiveness. A number of institutions were charged with helping to implement this policy: the Agricultural Marketing Agency the State Treasury Agricultural Property Agency the Agricultural Restructuring and Modernisation Agency. 4 Information on implementing the agricultural policy of the Polish government, Ministry of Agriculture, Warszawa

286 The effectiveness of the policy is, however, very difficult to evaluate, though there are problems of identifying measures which have been unequivocally successful. The Agricultural Marketing Agency was established in 1990 in order to carry out government policy of intervening in agricultural markets to stabilise prices and maintain the income of agricultural producers. The agency buys and sells produce on the domestic and foreign markets and also manages stocks and provides credit guarantees. Expert opinion is that the intervention prices of grain were set at too high a level. In order to maintain farm prices, there are attempts to protect domestic farmers from imports. Such a policy does not encourage agricultural modernisation or changes in the structure of farms. It is worth noting in this regard, however, that the approved simplified system of direct payments from the EU to Polish farmers is also aimed at providing social aid rather than at promoting modernisation of agriculture. The State Treasury Agricultural Property Agency, established in 1992, is responsible for managing changes in ownership in agriculture. The agency s role is to take over land from liquidated state-owned farms (PGRs) and to undertake other tasks, such as: restructuring and privatising state treasury land used for agricultural purposes purchasing land from private owners, creating new jobs in connection with the restructuring of state agriculture 5. The main challenge for state policy for rural areas is to encourage the creation of new jobs in rural areas outside agriculture in order to provide productive employment for those working in small, unproductive farms. To achieve this, large capital outlays are needed as well as measures to improve the education and skill levels of the rural work force. So far, local initiatives aimed at attracting capital to rural areas do not seem to have been effective. A major problem is that young people from rural areas still make up a relatively small share of university students which limits the extent to which the quality of the rural work force is likely to improve. A new system of student grants is needed in order to increase the number of students who come from rural areas. The Agricultural Restructuring and Modernisation Agency (ARiMR) has the role of supporting measures aimed at accelerating change in agriculture and rural areas. It is intended to promote investment in agriculture and in rural services, to develop infrastructure and to encourage action to improve and extend education. The Agency manages funds from the state budget and external aid, which are used for: 5 See National Human Development Report. Poland 2000, p. 13 8

287 subsidising interest rates on loans subsidising infrastructure projects subsidising educational institutions providing credit guarantees 6. There is little evidence that the Agency s activities have been effective in achieving its objectives.. An important element of government rural and agricultural policy is the social insurance system for farmers. The system is administered by the Farmers Social Insurance Fund (KRUS) and is separate from the workers social insurance system. The KRUS is responsible for establishing rules governing the distribution and financing of retirement pensions and disability pensions, allowances and social benefits. The system is intended to provide social insurance for farmers and their families. Contributions are paid quarterly and are set at relatively low flat-rate level. Pension entitlements are also the same for everyone. Revenue from contributions, it should be emphasised cover only 5% of expenditure on pensions and the remaining amount is covered by state subsidies. As a result, the KRUS plays a social role and helps maintain the income of the rural population. Labour market policies both active and passive operate in agricultural regions, like other regions in Poland. It is worth noting, however, that passive policy is not effective in rural areas since farmers with areas of more than 2 hectares are not entitled to unemployment benefit. Although there is no restriction of a similar kind on active programmes, but these have a relatively limited effect because of their small scale (only 13%-15% of the unemployed participate in such programmes). Their development has been constrained and, in addition, only a small proportion of participants in programmes are people from rural areas. Institutions providing vocational training are scarce in such areas and subsidised jobs are limited by the lack of employers willing to participate in these schemes, while public works providing jobs for the unemployed are costly and do not necessarily lead to the latter finding long-term employment outside the public sector. 6 Conclusions and recommendations Agricultural regions face the most difficulty in overcoming structural problems. Low GDP per head, a poorly qualified work force, high open and/or hidden unemployment and low household income are all characteristic features of these regions. To transform the situation requires an extensive and coordinated policy as well as appropriate institutions and the involvement of local communities. The aim has to be to take measures to improve the quality of the work force in rural areas, in part by encouraging the teaching of relevant 6 Informacja o realizacji polityki rolnej rządu RP, Ministerstwo Rolnictwa I Rozwoju Wsi, Warsaw

288 skills, to modernise infrastructure, to develop continuing training to enable people to learn new skills and acquire new qualification to prepare them for jobs outside agriculture and to establish and improve vocational advisory services. The need is to develop rural areas in a multifunctional way and to establish the conditions for job creation in non-agricultural activities. The measures suggested include: improving technical infrastructure to encourage investment flows into rural areas; creating favorable conditions for attracting investment (including by training local authority staff and personnel in relevant institutions); increasing the financial support for investment credits provided by the AriMR; creating credit guarantee funds to assist investors in rural areas; promoting rural tourism. Active labour market policy needs to take account of the specific features of local areas, including their rural nature. Suggestions are to: develop training programmes for the unemployed and others to raise skill levels and increase employability; maintaining public work programmes to provide jobs for the unemployed despite their high costs given the difficulties of expanding subsidised jobs in rural areas; increasing loans for the unemployed who wish to start their own business and making loan conditions more favorable to those from rural areas. 10

289 Literature Blanchard, O., Commander, S., Coricelli, F. (1994): Unemployment and the labour market in Eastern Europe, OECD, Unemployment in Transition Countries: Transient or Persistent?, Paris. Boeri, T. (1995): Labour Market Flows and the Scope of Labour Market Policies in Central and Eastern Europe, Vienna, mimeo. Brdulak J. (2001) Problemy rozwoju regionalnego (Regional Development Issues), Materiały z VII Kongresu Ekonomistów Polskich-styczeń 2001 (Proceedings of the 7th Congress of Polish economists), Vol., Wydawnictwo PTE-Bellona, Warszawa. Fierla I. (2001) Narastanie przestrzennych dysproporcji rozwojowych w Polsce (Accumulation of Spatial Economic Disproportions in Poland) in: J. Brdulak (2001). Information on implementing state agricultural policy of the Polish government, Ministry of Agriculture, Warszawa Kabaj, M. (1995): Active Labour Market Policy and the Role of Employment Councils in Poland in Counteracting Unemployment, the 5th ILO Conference, Warsaw. Kaczorowski P., A. Rogut, T. Tokarski (2001) Sektorowe zmiany strukturalne w ujęciu regionalnym (Sectoral Changes by Regional Breakdown), Wiadomości Statystyczne nr 9/2001. Krajewski S., T. Tokarski (red.). (2002) Wzrost gospodarczy, restrukturyzacja i bezrobocie w Polsce. Ujęcie teoretyczne i empiryczne (Economic Growth, Restructuring and Unemployment. Theretical and Empirical Approaches), Katedra Ekonomii Uniwersytetu Łódzkiego, Monografie, Rozprawy, Raporty, Opracowania, Vol 6, Łódź. Kwiatkowska W. (2000) Zmiany struktury pracujących w Polsce w kontekście integracji z Unią Europejska (Chnages in the Employment Structure of Poland in the Context of her Intergration with the European Union) in Wzrost gospodarczy, restrukturyzacja i bezrobocie w Polsce. Ujęcie teoretyczne i empiryczne (Economic Growth, Restructuring and Unemployment. Theretical and Empirical Approaches), S.Krajewski, T.Tokarski, Kwiatkowski, E. (1993): Public Sector Adjustment Through Employment: Retrenchment Policies and Practicies in Poland, Occasional Paper, No. 20, ILO, Geneva. Kwiatkowski, E., Tokarski, T. (1995): Determinanty bezrobocia w Polsce w okresie transformacji (modele teoretyczne oraz próba ich weryfikacji), PAN, INE, zeszyt 11. Kwiatkowski E., L. Kucharski, T. Tokarski (2002) Elastyczność zatrudnienia w ujęciu sektorowym i regionalnym w Polsce (Elasticity of Employment in Poland by Sectoral and Regional Breakdown), Gospodarka Narodowa nr 4/2002. Kwiatkowski E., A. Rogut, T. Tokarski (2002) Prognoza popytu na pracę i stopy bezrobocia w Polsce oparta na analizie współczynników pracochłonności (Forecast of Demand for Labour Force and Rate of Unemployment in Poland based on Labour Productivity Ratios) in: S. Krajewski, T. Tokarski (2002). Kwiatkowski E., T. Tokarski (2000) Employment Structure and Employment Flexibility in Poland in Transition, International Review of Economics and Business, June Lehmann, H. (1995): Active Labour Market Policies in the OECD and in Selected Transition Economies, The World Bank, Policy Research Working Paper National Human Development Report. Poland Rural Development, UNDP, Warsaw OECD in Figures 2001, Statistics on the Member Countries, strona internetowa OECD, 11

290 Orłowski W. (2001), Przeciw stereotypom, UKIE, Warszawa Puhani, P.A., Steiner, V. (1996): Public Works for Poland? Active Labour Market Policies during Transition, ZEW, Discussion Paper No Rocznik Statystyczny (Statistical Yearbooks), GUS, Warszawa, various editions of the years Rogut A., T. Tokarski (2001) Regionalne zróżnicowanie płac w Polsce w latach dziewięćdziesiątych (Regional Diversification of Wages in Poland in the 90-s), in: J. Brdulak (2001). Rogut A., T. Tokarski (2002) Regional Diversity of Employment Structure and Outflows from Unemployment to Employment in Poland, International Journal of Manpower, No.?/2002. Rutkowski, M.(1990): Labour Hoarding and Future Unemployment in Eastern Europe: The Case of Polish Industry, Discussion Paper, LSE, No. 6. Tokarski T., A. Gabryjelska, P. Krajewski, M. Mackiewicz (1999) Determinanty regionalnego zróżnicowania PKB, zatrudnienia i płac (Determinants of Regional Diversification of GDP, Employment and Wages), Wiadomości Statystyczne nr 8/1999. Wzrost gospodarczy, restrukturyzacja i bezrobocie w Polsce. Ujęcie teoretyczne i praktyczne (Economic Growth, Restructuring and Unemployment in Poland) (2000), Materiały z Konferencji (conference proceedings), Katedra Ekonomii Uniwersytetu Łódzkiego, Łódź. Witkowski, J. (1994): Unemployment in Poland in the Period of Transition , CSO, Warsaw. 12

291 Maps and tables Map 1 Territorial division in Poland (NUTS 2 regions) POMORSKIE WARMINSKO-MAZURSKIE ZACHODNIOPOMORSKIE KUJAWSKO-POMORSKIE PODLASKIE MAZOWIECKIE LUBUSKIE WIELKOPOLSKIE LODZKIE LUBELSKIE DOLNOSLASKIE OPOLSKIE SWIETOKRZYSKIE SLASKIE PODKARPACKIE MALOPOLSKIE Map 2 Average unemployment rates, Unemployment rates do 21.3 (3) 14.9 do 17.4 (2) 13.9 do 14.9 (3) 12.4 do 13.9 (3) 10.3 do 12.4 (5) Source: own calculations on the basis of GUS data: 13

292 Table 1 Unemployment rates in voivodships, 1995, 2002 (in %) and changes (in percentage points) Voivodship Changes Dolnoslaskie Kujawsko-Pomorskie Lubelskie Lubuskie Lodzkie Malopolskie Mazowieckie Opolskie Podkarpackie Podlaskie Pomorskie Slaskie Swietokrzyskie Warminsko-Mazurskie Wielkopolskie Zachodniopomorskie POLSKA Map 3 Average sectoral structure of employment, ,5 0,1 A IC S Source: own calculations on the basis of GUS data: 14

293 Table 2 Sectoral composition of employment, 1995 and 2001 (in %) Voivodship Agriculture Industry and construction Dolnoslaskie Kujawsko-Pomorskie Lubelskie Lubuskie Lodzkie Malopolskie Mazowieckie Opolskie Podkarpackie Podlaskie Pomorskie Slaskie Swietokrzyskie Warminsko-Mazurskie Wielkopolskie Zachodniopomorskie POLAND Services Source: own calculations on the basis of GUS data: 15

294 Table 3 Value added per employee in EUR 000 at 2001 prices and exchange rates Voivodship Dolnoslaskie Kujawsko-Pomorskie Lubelskie Lubuskie Lodzkie Malopolskie Mazowieckie Opolskie Podkarpackie Podlaskie Pomorskie Slaskie Swietokrzyskie Warminsko-Mazurskie Wielkopolskie Zachodniopomorskie POLAND Source: own calculations on the basis of GUS data: 16

295 Table 4 Sectoral labour productivity, 1996 and i2001 (EUR 000, 2001 prices and exchange rates) Voivodship Agriculture Industry and construction Dolnoslaskie Kujawsko-Pomorskie Lubelskie Lubuskie Lodzkie Malopolskie Mazowieckie Opolskie Podkarpackie Podlaskie Pomorskie Slaskie Swietokrzyskie Warminsko-Mazurskie Wielkopolskie Zachodniopomorskie POLAND Services Source: own calculations on the basis of GUS data: 17

296 Map 4 Regional diversification of overemployment in agriculture, thousands of people (Italian variant) Ov eremployment in ' do 310 (3) 185 do 248 (1) 122 do 185 (2) 59 do 122 (3) -4 do 59 (7) Source: own calculations on the basis of GUS data: Map 5 Regional diversification of overemployment in agriculture, thousands of people (Spanish variant) Overemployment in ' do 320 (2) 145 do 268 (3) 103 do 145 (4) 36 do 103 (3) 0 do 36 (4) Source: own calculations on the basis of GUS data: 18

297 Table 5 Overemployment in agriculture and unemployment rates in Poland (average, ) Italian variant Spanish variant UR Overemployment UR with Overemployment UR with Voivodship In agriculture overemployment In agriculture overemployment % Thousands % of empl. in % Thousands % of empl. In % agr. agr. Dolnoslaskie Kujawsko-Pomorskie Lubelskie Lubuskie Lodzkie Malopolskie Mazowieckie Opolskie Podkarpackie Podlaskie Pomorskie Slaskie Swietokrzyskie Warminsko-Mazurskie Wielkopolskie Zachodniopomorskie POLAND UR-unemployment rate Source: own calculations on the basis of GUS data: 19

298 Table 6 Fixed capital required to eliminate overemploment in agriculture 7 Overemployment K/L in other sectors ('000 EURO) Capital needed (bln EURO) Dolnoslaskie Kujawsko-Pomorskie Lubelskie Lubuskie Lodzkie Malopolskie Mazowieckie Opolskie Podkarpackie Podlaskie Pomorskie Slaskie Swietokrzyskie Warminsko-Mazurskie Wielkopolskie Zachodniopomorskie POLAND K/L capital/labour ratio Source: own calculations on the basis of GUS data: 7 InTtable 6 and in Map 6 it is assumed that the level of overemployment in agriculture in Zachodniopomorskie voivodship is equal to zero. 20

299 Map 6 Need for fixed capital in voivodships (EUR billion) 11,1 do 14,6 (1) 7,4 do 11,1 (2) 3,7 do 7,4 (2) 0 do 3,7 (11) Source: own calculations on the basis of GUS data: 21

300 Economic Restructuring and Labour Markets in the Accession Countries Research Project commissioned by EU DG Employment, Social Affairs and Equal Opportunities, Contract No. VC/2003/0367 Coordinated by in cooperation with Alphametrics and The Vienna Institute for International Economic Studies Inter-industry labour mobility in Poland by Eugeniusz Kwiatkowski*, Paweł Kubiak* and Leszek Kucharski* December 2004 *) Institute of Macroeconomics, University of Lodz

301 Eugeniusz Kwiatkowski, Paweł Kubiak and Leszek Kucharski Inter-industry labour mobility in Poland 1 Introduction The transition to a market economy and integration with the European Union brought about various changes in the Polish labour market. Deep structural changes in output and employment are linked to adjustments to a changing pattern of demand for labour. The hidden unemployment and over-employment inherited from the centrally planned economy are in decline, as tight budget constraints become major principles on which the operation of companies is based. In addition, the different levels and growth rates of labour productivity and technical progress in different sectors combined with the changing patterns of output are giving rise to differing rates of growth of employment in different sectors. There is therefore an ongoing process of shifts in the sectoral division of labour. This study is aimed at providing an insight into inter-industry labour mobility in Poland over the transition period. In particular, the study is concerned to: describe trends in labour mobility between sectors and to identify the sectors where employment is declining and those where it is increasing; consider the socio-economic determinants of this mobility; examine government policies for promoting labour mobility. The statistical data on which the analysis is based come, on the one hand, from enterprise surveys conducted by the Central Statistical Office (GUS) and, on the other, from Polish Labour Force Surveys. Since there was a change in the classification of activities at the beginning of the transition period, the analysis is mostly confined to the period The paper, first, considers concepts and measures of inter-industry labour mobility; secondly, it analyses trends in labour mobility in the economy as whole and then in different sectors; thirdly, it examine the underlying socio-economic factor affecting the extent of labour movement; fourthly, it comments on government policies which influence mobility, both those which at present exist and those proposed and finally draws some conclusions from the analysis. 2 Concepts and measures of inter-industry labour mobility Labour mobility means the extent to which those in employment move from job to job in the labour market. Such movements encompass changes in places of work, occupations, places of residence and shifts between employment, unemployment and inactivity. Movements can be either voluntary, resulting from personal decisions to make a change 1

302 for career or other reasons, or involuntary, resulting perhaps from changing in economic circumstances. The concern here is with movements between different sectors or industries, which we term inter-industry labour mobility. In the analysis below, three rates are used to capture the extent of inter-industry mobility: the hiring rate, the separation rate and the turnover rate. These rates are estimated from aggregate data from the Central Statistical Office. The hiring rate (hr) is defined as: H h r= (1) L t 1 where H is the number of people taking up work in a given period less the number returning from child-care leave and Lt-1 is the number of people working at the end of the previous period. The separation rate (sr) is defined as: S sr= (2) L t 1 where S is the number of people leaving employment in a given period less the number granted child-care leave and unpaid vacations. The turnover rate (tr) is the sum of the hiring rate and the separation rate: t = h s (3) r r + r Labour turnover results in part from job turnover, which is the sum of rates of job creation and job loss. Although it is not possible to calculate job turnover rates for Poland from published statistical data, it is possible to calculate rates of employment change in particular sectors, inasmuch as the change in the stock of employment in a given period is the difference between jobs created and jobs eliminated. Inter-industry labour mobility can be measured by a number of indicators. Its relative extent over the economy as a whole can be measured by an index (wmg), calculated as: wmg = n i= 1 L t O j= 1,i j ij 100% (4) where O ij is outflows from employment in section i to employment in section j and L t is the number employed in the economy in period t. The index wmg indicates the proportion of the total employed who change sectors within a given period. 2

303 Mobility in a particular sector can be measured from indices of outflows from the sector and inflows into it. The index of outflows can be defined as follows: n O ij j= 1, i j mg ( sio) = 100% (5) L i where O ij is outflows from employment in sector i to employment in sector j within a year (the sum of flows occurring between consecutive years of the LFS) and L i is the average number employed in sector i in a particular year. 3 Trends in inter-industry labour mobility in the economy as a whole The concern here is to estimate the scale of labour mobility in the Polish economy. Figure 1 shows hiring, separation and labour turnover rates over the period It indicates, first, that there was an upward trend in mobility in the years (with the exception of 1997) and a significant decline in mobility in the years , when business conditions in the Polish economy worsened. Secondly, up until 1999, hiring rates were higher than separation rates while after 2000, the reverse was the case. Moreover, in the years , the hiring and separation rates both showed an upward trend, while n the years , they both declined. Figure 2 presents the trends in quarterly inter-industry mobility indices in Poland in the years , calculated on the basis of data on the economic activity of individuals from the LFS. 1 It shows a declining trend in mobility over the period analysed. 2 The mobility index remained much the same In the years , but then began to decline. This coincides with a slackening in the speed of restructuring in the Polish economy. Restructuring tends to mean that many people lose their job, but at the same time many find new employment, even though some of those losing their jobs become unemployed or economically inactive.. A reduction in the pace of restructuring tends to mean less rotation among the pool of workers. The decline in inter-industry mobility after 1998 was a result of a lower rate of economic growth and a significant deterioration in labour market conditions in Poland. The number of people voluntarily leaving their jobs fell because of problems finding new ones, while the increase in restructuring meant a rise in outflows from employment into unemployment. 1 Because of the break in the series in 1999, no data are included for this year. 2 The data from the LFS is likely to underestimate the scale of inter-industry mobility because those who go abroad and remain there and those living temporarily in hostels and similar accommodation are outside the scope of the survey. 3

304 4 Trends in inter-industry labour mobility by branch Over the period as a whole, separations exceeded people hired in mining and quarrying and agriculture (see Tables 1-2), sectors still characterised by over-employment and which, in the case of agriculture at least, has not undergone a process of restructuring. On the other hand, the number of people hired exceeded separations almost throughout the entire period in real estate, renting and business activities; research and development and public administration. Separation rates increased over the years in all the sectors, as a result of a deterioration in business conditions and increased restructuring. The highest rate of job turnover over the period occurred in financial intermediation, construction, mining and quarrying and other community, social, and personal services (Table 3). The changes in employment in mining took place at the same time as the Government secured the finance to fund severance pay for redundant miners. While the high turnover of jobs in construction is partly a result of the seasonal nature of work, it also a consequence of business conditions. Up until 1997, employment in the sector rose rapidly but as the business situation deteriorated after 1997, employment declined markedly. The lowest rate of job turnover occurred in education and agriculture, both sectors which have not been subject to restructuring. Table 4 presents the composite data on inter-industry mobility indices by sector over the years , calculated from the LFS.. The lowest rates of mobility over the period are found again in agriculture and education as well as in mining and quarrying; health care and social services, public administration and distribution. The smallness of outflows of people from agriculture to other sectors is partly a consequence of the low level of education of those concerned, while in the case of public administration, it partly reflects the relatively high average wages compared with those in other parts of the economy, In distribution, it may be a result of those employed in the sector not having the skills demanded by employers elsewhere in the economy. The sectors with the highest mobility indices were again financial intermediation, other community, social and personal services and construction, as well as real estate, renting and business activities and manufacturing. In the last, this is probably a result in part of increased restructuring. The changes in the economic system initiated in 1990 led to the near elimination of some manufacturing industries. Employment in some industries in 2000 was 60-80% lower than in 1989 (Table 5). The steepest decline in employment occurred in the mining of hard coal, iron and steel, electronic equipment and components, knitting products and various parts of the textile, clothing and footwear industry. The in mining and steel was a result of a reduction in domestic demand and losses of export market, while in the other industries, it was a consequence in part of an absence of consistent restructuring programmes. 4

305 5 Determinants of inter-industry labour mobility The determinants of inter-industry labour mobility can be divided into two groups. The first are linked to the demand for labour. The transformations taking place in Poland have led to changes in the structure of supply and demand in the market for goods and services and this has involved changes in the sectoral composition of demand for labour. The second are linked to labour supply. Different groups in the work force with different levels of education, ages and professions show different degrees of mobility. It is still the case that most people in Poland are still wedded to the notion of their working careers taking place in one place and line of work. 5.1 Labour demand effects Inter-industry mobility is linked to structural changes. These lead to a loss of jobs in some sectors and the creation of new ones in others. This process is particularly marked in Poland, since the structure of employment in the centrally-planned economy differed considerably from that in developed market economies, with a large share of employment in agriculture and, to a lesser extent, in industry and a small share in services 3. The transition has led to significant changes in the employment structure, with the proportion employed in services increasing and that in industry and agriculture declining, if only slightly in the latter case. These changes took place in the context of an overall reduction in employment in the Polish economy 4. Changes in the sectoral structure of labour demand also result from various other factors, such as: differing rates of technical progress in different sectors, different levels of competitiveness and differing responses by enterprises to actual and prospective changes in market demand. These differences are reflected in the pattern of value-added (table 6). In the first place, the share of agriculture in value-added declined throughout the period analysed 5. This decline, in combination with only a small fall in the share of employment in the sector is evidence of high hidden unemployment in rural areas and very low labour productivity. Secondly, the share of industry and construction in value-added also fell significantly, in a similar way to the decline in many developed market economies. Thirdly, the share of services in value added increased, though the gap in share between Poland and other EU countries is still wide. Fourthly, this comparison of sectoral shifts in the structure of value-added with those in the structure employment point to wide diversities in labour productivity (value-added per person employed) between sectors. 3 Kwiatkowska, Kwiatkowska, Kwiatkowska, 2000, Kwiatkowski, Kucharski, Tokarski,

306 Changes in the structure of labour demand within broad sectors of activity have also been pronounced. Within manufacturing,, the rates of change in value-added over the period were highly diverse (Table 7). Value-added declined in three sectors: leather (DC), textiles (DB) and basic metals (DJ). These three industries all showed a fall in employment, the fall being larger in the first two than in other parts of manufacturing. The largest increases in value-added were recorded in non-metallic mineral products (DI), paper and publishing (DE); furniture and manufacturing not elsewhere classified (DN) and office machinery and computers (DL). Growth in employment occurred only in four manufacturing industries:, two of them being paper and publishing (DE) and furniture and manufacturing not elsewhere classified (DN) the other two being rubber and plastics (DH) and wood and wood products (DD), This demonstrates the dependence of employment on output growth. 5.2 Labour supply effects The question: arises as to which characteristics of workers stimulate mobility and which hamper it. Movement from one sector of the economy to another often means a need to acquire new skills and professional qualifications as well as perhaps a different approach to work. In order to estimate the effect of such characteristics a multinomial logit model was used 6. This focused on the relationship between these different characteristics and the probability of a) moving between sectors,b) moving to another place of work in the same sector and c) moving from employment into unemployment. The characteristics examined were gender, age, education, marital status, the job performed, firm size, the size of the locality and the level of the socio-economic development of regions. The model was estimated on the basis of a sample of 192,229 observations from the LFS over the years The results suggest the following: men are more likely to move between sectors than women but also more likely to become unemployed. This may reflect the fact that women seem to be treated less well by employers and, therefore, have lower expectations about wages and working conditions; the likelihood of movement tends to decline with age, though there is no apparent relationship between age and inter-industry mobility for non-manual workers. The greater tendency for younger people to move may reflect their better education and their greater willingness to take up a challenge, as well as their higher chances of being dismissed; single people are more likely to move than married ones; 6 Kwiatkowski, Kubiak, Kucharski,

307 the level of education appears to have only a minor effect on inter-industry mobility. Women with secondary education are more likely to move than those with primary, vocational or tertiary education. In addition, the risk of unemployment for women without tertiary education does not depend significantly on whether they have secondary, vocational or primary education. Among men, those with secondary or tertiary education tend to have a higher inter-industry mobility, though this applies only to those in manual occupations and not to those in non-manual ones; labour mobility is greater in regions with a higher level of socio-economic development, with a more modern structure of the economy which facilitates movement. The risk of becoming unemployed is also lower; the size of the local area in which people live has little apparent effect on inter-industry mobility, though those living in small and medium-sized towns face the most risk of becoming unemployed; inter-industry labour mobility is related to the profession or occupation performed, though this relationship is stronger for men than women. Women in agriculture and in sales and service occupations have the lowest tendency to move, while the highest mobility among men is among skilled and less skilled manual workers; those employed in large firms (with over 100 employees) have lower mobility than those employed in small firms, which reflects their stronger market position and, therefore, their more stable jobs; there are no apparent differences in inter-industry mobility between those employed in the private sector and those employed in the public sector. 6 Policies for promoting labour mobility The above analysis demonstrates that inter-industry labour mobility is determined by a number of factors, including changes in the structure of the economy and the characteristics of different people in the labour market. The Government can, therefore, stimulate labour mobility by influencing the factors which affect it. Firstly, the Government can stimulate inter-industry labour mobility by implementing restructuring programmes in particular industries and sectors. The relatively weak budgetary constraints on state-owned enterprises meant that in the early years of transition employment in them was determined more by habits acquired during the period of centrally-planning than by the aim of increasing economic efficiency. This led to the maintenance of high hidden unemployment. The gradual tightening of budgetary constraints led to the collapse of many enterprises and even entire sectors, for example, light industry. For many sectors, however, restructuring programmes have been developed and implemented, including in hard coal mining, iron and steel, power engineering, the defence industry, gas, oil, shipbuilding and the food processing. These can be divided into 7

308 two groups. The first group consists of industries hard coal mining, iron and steel, the defence industry and shipbuilding in which production capacity built up during the period of central planning was excessive in relation to demand in the new market economy. The second group comprises industries power engineering, gas and oil in which forecasts of demand growth indicated a need for an expansion of capacity. The need to sell off auxiliary and non-productive assets and for a high level of investment to replace and modernise production equipment is common to both groups. The programmes of restructuring also encompass employment among other elements. The employees losing their jobs in restructuring enterprises receive benefits under social packages introduced by legislation. These are aimed at increasing their chances of finding a new job. A primary principle adopted in the recently implemented programme for the restructuring of the hard coal mining industry, one of the most difficult tasks in the Polish economy, was the avoidance of enforced collective redundancies. As part of the Miners Social Package, the intention is to restructure employment through both passive measures, in the form of the introduction of five-year leave for miners, and active measures, in the form of help with finding a new job and training. An employee who wants to change job will have the right to social benefits during the period of job search, access to a training programme, severance pay, and a loan to start their own business. It is expected that as a result around 118,000 people will leave the mining industry (Walewski, 2000). The evaluation of the sectoral restructuring programmes made by the Task Force for Structural Policy in Poland suggests the greatest successes were achieved in the power engineering industry and in food processing. The biggest failures seem to be shipbuilding and hard coal mining. Restructuring the latter is arguably almost exclusively a social rather than a technical or economic problem. The problem is of such a scale and importance that further restructuring will not be possible without restructuring the entire Slasko-Dabrowski region and the industry located there in order to make mining communities economically viable. The development of detailed and complex programmes of restructuring is urgently required, encompassing measures which not only reduce employment in mining but also contain pro-active measures aimed at re-employing those losing their jobs in the industry. These should have access to programmes which enable them to develop new skills and find other employment or to start up their own business. This includes vocational training and preferential loans. In the restructuring programmes carried out so far such measures were taken only on a small scale, with the focus on one-off cash payments to those made redundant and on encouraging some of the work force to become economically inactive. Secondly, the Government can also stimulate inter-industry labour mobility through the development of education and training. Empirical studies as well as human capital theory 8

309 demonstrate the importance of education as a determinant of adjustment in an unstable and changing labour market. People with better education tend to have a higher degree of mobility. The educational structure of the labour force in Poland is different from that in other OECD countries (see Table 8). Only 13% of people in Poland participated in continuing training in 1998 as opposed to 31% in OECD countries as a whole 7. The Labour Force Survey for 2003 shows that some 5% of people aged participated in some form of education or training in Poland in the four week preceding the survey as compared with almost 10% in the EU-15 and over 6% in the other new Member States. Moreover, there is marked regional diversity in the structure of education and training across the country, which is particularly evident in the disproportionate opportunities offered to those living in rural and urban areas. Raising the education level of the work force is a long-term process that calls for a number of changes in the education system as a whole. In terms of the urgent action which needs to be taken in this respect, initial and continuing vocational education and training are of prime importance. Measures to expand continuing training in Poland, and to encourage people to improve their skills and know-how, include an obligation on employers under the Labour Code to help workers increase their skills, tax relief on expenditure on training and the financing from the Labour Fund (funded by the State budget and employers contributions) of training for the unemployed, other job seekers and workers threatened by job loss. These various measures, however, are not necessarily consistent with each other and are not part of a coherent package. The lack of a good understanding of the demand for continuing training means that it is not possible to assess the extent to which existing measures of support are effective and to identify the barriers which might exist to the take-up of training initiatives and the changes which should be made. The supply-side of the market for training services is also poorly researched, which makes it difficult to control the quality of courses and to ensure that money spent on training is cost-effective. The statistics on education only cover education which takes place within the educational system, and complete statistics on education outside the system are still unavailable (Kotowska, 2003). Nevertheless the data which are available show that such activities are undertaken on a smaller scale in Poland than in other EU countries (see Table 9). The development of continuing training programmes and the widening and updating of the skills of the work force is, therefore, another area in which policy can increase labour mobility. Thirdly, the State can influence labour mobility through labour market policy. As noted above, training organised as part of active labour market programmes can serve to 7 In addition, according to the Continuing Vocational Training Survey conducted by Eurostat in 1999, 16% of employees working in enterprises in Poland received some training over the year as against an EU-25 average of 39%. 9

310 increase labour mobility. Other measures targeted at the unemployed, such as subsidised jobs or public works programmes, which can increase skills and provide temporary employment, or unemployment benefits, which can be used to fund business start-ups, are also important. All the measures carried out as part of labour market policy are financed from the Labour Fund, the size of which is small in relation to GDP (Table 10). The great majority of these funds go on statutory unemployment benefits (amounting 1.3% of GDP in 2001) and very little on active measures. In 2001, the latter amounted to under 0.1% of GDP as compared with expenditure on labour market training programmes of 0.3% of GDP in EU-15 countries (according to the Eurostat Labour Market Policy database), despite unemployment being only around half the level in Poland 8. This small level of funding means that there are relatively few participants in active programmes - the total representing under 8% of the unemployed in 2002 (see Table 11). The effectiveness of different types of measure, as indicated by the re-employment rate of participants, varies significantly. Moreover, the rate of re-employment only measures the gross effectiveness of programmes. In order to assess the effect of completing a programme on the ability to find a job, the labour market status of participants needs to be compared with that of non-participants. The difference between the indicates the net effect of a particular programme 9. Estimates of this net effect suggest that vocational training increases the chances of participants finding employment. However, this is not the case in respect of participants in training programmes organised by the public employment service, but only in respect of that funded by the people concerned themselves or by employers. The findings suggest that vocational training can help individuals increase their employability, but that it should be directed primarily at the short-term unemployed. Labour offices need to pay particular attention to the selection of people for training. The reasons for the lower effectiveness of training sessions organised by labour offices lie perhaps, on the one hand, in the organisation and quality of the programmes and, on the other, in the attitude of the participants. Estimates of the effectiveness of active programmes also suggest, however, that participants have a smaller risk of long-term unemployment than the unemployed who do not participate. This difference is most evident immediately following the completion of a programme and diminishes gradually over time. This analysis suggests that active measures to increase labour mobility should be developed further, especially in areas where restructuring is most pervasive. 8 See A. Malarska, 2000, p See E. Kwiatkowski,

311 7 Summary The above analysis lead to the following conclusions: in the period examined there was a change in the behaviour of rates of job turnover, as well as in hiring and separation rates. After 1998, all three rates showed a downward trend; inter-industry labour mobility in Poland was low throughout the period and between 1994 and 2002, there was a downward trend in mobility indices; the highest labour mobility over the period was in construction, distribution and hotels and restaurants, in part because of the seasonal nature of the work. The lowest labour mobility over the period was in education, mining and quarrying, electricity, gas and water, and health and social services; sectors with the highest inter-industry mobility over period were real estate, renting and business activities; financial intermediation; other community, social and personal services, construction, and manufacturing;. the lowest rates of inter-industry mobility were in agriculture, mining and quarrying; health and social services; education; public administration and distribution; inter-industry mobility could be increased through a coherent strategy involving complex restructuring programmes and the development of education and training and active labour market policy. Complex restructuring programmes mean the creation of conditions for the re-employment of those losing their jobs. They need, therefore, to involve measures such as training, loans for business start-ups or possibly support for migration to other regions. Such measures have been included in restructuring programmes in the past, but a lack of finance meant that the programmes were relatively small and most of this went on severance pay for those losing their jobs. The development of education and systems of continuing training is likely to lead to an increase in labour mobility. Although major changes have occurred in the system of education and higher education in recent years, financial obstacles to many young persons from poor families as well as those from rural areas and small towns continuing in education remain a problem. There is, therefore, a need to introduce grant schemes, educational loans and so on. The development of continuing vocational training does not necessarily go hand in hand with the development of the education system. Relatively little has been done in this area and there is a need for the development of training institutions. This calls for the creation of a monitoring system to identify labour market needs and systems for assessing the effectiveness of the measures introduced. 11

312 Active labour market policy, which can also increase labour mobility, is a separate problem. The limited finance which has been devoted to funding such a policy remains a basic constraint. There is, therefore, a need to increase expenditure on active measures and concentrate them in regions with difficult labour market conditions. As in the case of continuing training, there is also a need for establishing a system for monitoring the effects of the measures undertaken. 12

313 Bibliography KOTOWSKA I (2003), Kształcenie ustawiczne czynnikiem poprawy sytuacji na rynku pracy, Rynek Pracy, nr 12. KWIATKOWSKI E, KUBIAK P, KUCHARSKI L., Inter-industry and intra-industry mobility of labour in Poland in , [in:] Labour market flexibility in the wake of EU accession (2002), IPiSS, Warszawa KWIATKOWSKI E. (2002), Bezrobocie. Podstawy teoretyczne, Wydawnictwo Naukowe PWN, Warszawa KWIATKOWSKI E., KUCHARSKI L., TOKARSKI T. (2003), Elastyczność zatrudnienia w ujęciu sektorowym i regionalnym w Polsce [in:] Wzrost gospodarczy, restrukturyzacja i bezrobocie w Polsce. Ujęcie teoretyczne i empiryczne, edited by Kwiatkowski E., Tokarski T., Łódź KWIATKOWSKA W. (2000), Zmiany struktury pracujących w Polsce w kontekście integracji z Unią Europejską, [in:] Wzrost gospodarczy restrukturyzacja i bezrobocie w Polsce. Ujęcie teoretyczne i praktyczne, edited by Krajewski S., Tokarski T, Łódź LIPOWSKI A. (1997), Sektorowe programy restrukturyzacji w polskim przemyśle Próba oceny. [in:] Sektorowe programy restrukturyzacji i prywatyzacja majątku państwowego. Wybór ekspertyz, edited by Bochniarz H., Krajewski S., Zespół Zadaniowy ds. Polityki Strukturalnej w Polsce, Warszawa. LIPOWSKI A. (2000), Uwagi o restrukturyzacji polskiego przemysłu, [in:] Tworzenie zatrudnienia a restrukturyzacjia ekonomiczna, edited by S. Golinowska, M. Walewski, CASE, Warszawa MALARSKA A. (2000), Bezrobocie w Polsce w ujęciu regionalnym. Studium statystyczne, Wydawnictwo Uniwersytetu Łódzkiego, Łódź. NENEMAN J., SOWA M. (2002), Miękkie ograniczenie budżetowe w Polsce [in:] Wzrost gospodarczy restrukturyzacja i bezrobocie w Polsce. Ujęcie teoretyczne i empiryczne, edited by Krajewski S., Tokarski T., Łódź OLEJARZ T. (2000), Zmiany w przepisach o Funduszu Pracy, Służba Pracownicza nr 10/2000. Sektorowe programy restrukturyzacji i prywatyzacja majątku państwowego. Wybór ekspertyz (1997), edited by Bochniarz H., Krajewski S., Zespół Zadaniowy ds. Polityki Strukturalnej w Polsce, Warszawa. Statistical Yearbooks of the Central Statistical Office, different issues from the years Transformacja społeczno-gospodarcza w Polsce [Socio-economic transformations in Poland], Government Centre for Strategic Studies, Warsaw, July WALEWSKI M (2000), Przegląd rządowych programów restrukturyzacji górnictwa węgla kamiennego w latach , [in:] Tworzenie zatrudnienia a restrukturyzacja ekonomiczna, edited by S. Golinowska, M. Walewski, CASE, Warszawa 13

314 Figure 1 Hiring and separation rates and labour turnover indices in Poland, , in % The hire rate The separation rate The labour turnover rate Source: Statistical Yearbooks of the Central Statistical Office, different issues from the years , own calculations. Figure 2 Inter-industry mobility indices (wmg) in Poland, % 9% 8% 7% 6% 5% 4% 3% 2% 1% 0% Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q inter-industry flows Source: individual data from the Labour Force Survey, Central Statistical Office, own calculations. 14

315 Table 1 Hiring rates by sector (NACE Rev.1) in Poland, , in % Items Total Agriculture, hunting and forestry Fishing Industry in total - Mining and quarrying - Manufacturing activities - Electricity, gas and water production and supply Construction Trade and repair Hotels and restaurants Transport, storage and communication Financial intermediation Real estate, renting and business activities; research and development Public administration and defence; compulsory social security Education Health care and social protection Other service, community, social and personal activities Source: Statistical Yearbooks of the Central Statistical Office, different issues from the years

316 Table 2 Separation rates by sector (NACE Rev.1) in Poland, , in % Items Total Agriculture, hunting and forestry Fishing Industry in total - Mining and quarrying - Manufacturing activities - Electricity, gas and water production and supply Construction Trade and repair Hotels and restaurants Transport, storage and communication Financial intermediation Real estate, renting and business activities; research and development Public administration and defence; compulsory social security Education Health care and social protection Other service, community, social and personal activities Source: Statistical Yearbooks of the Central Statistical Office, different issues from the years

317 Table 3 Growth rates of employment by sector (NACE Rev 1) in Poland, , in % Items Total Agriculture, hunting and forestry and fishing Industry in total - Mining and quarrying - Manufacturing activities - Electricity, gas and water production and supply Construction Trade and repair Hotels and restaurants Transport, storage and communication Financial intermediation Real estate, renting and business activities; research and development Public administration and defence; compulsory social security Education Health care and social protection Other service, community, social and personal activities Source: Statistical Yearbooks of the Central Statistical Office, different issues from the years , own calculations. 17

318 Table 4 Quarterly average inter-industry mobility indices (mg(sio)) by sector (NACE Rev 1.) , in % Items Total Agriculture, hunting and forestry and fishing Industry in total - Mining and quarrying - Manufacturing activities - Electricity, gas and water production and supply Construction Trade and repair Hotels and restaurants Transport, storage and communication Financial intermediation Real estate, renting and business activities; research and development Public administration and defence; compulsory social security Education Health care and social protection Other service, community, social and personal activities Source: the outcome of the Labour Force Survey, own calculations. 18

319 Table 5 Manufacturing industries with biggest losses of employment, Symbol of the Industry Level of Loss of jobs Polish Classification of employment in 2000 Employment in 000 Decrease in 000 Activities (1989=100) Manufacture of flax products Manufacture of wool products Other mining and quarrying (quartzite, diatomite) 32.3 Consumer s electronics Electronic components Mining of chemical minerals (sulphur) Tanning of leather Manufacture of man-made fibres Manufacture of knitting products Manufacture of cotton products Manufacture of hosiery Manufacture of carpets and rugs Manufacture of silk products Manufacture of footwear Manufacture of tubes Manufacture of optical instruments and photographic equipment Mining of hard coal Manufacture of power machinery (turbines and boilers) Mining of metal ores Manufacture of engines and turbines Manufacture of medical instruments Manufacture of clocks and watches Manufacture of basic iron and steel and of ferro-alloys Manufacture of fertilisers Manufacture of railway locomotives and rolling stock Manufacture of cement Source: Transformacja społeczno-gospodarcza w Polsce [Socio-economic transformations in Poland], Government Centre for Strategic Studies, Warsaw, July 2002, p

320 Table 6 The structure of value-added by sector, , in % Sector: Agriculture 7,1 6,7 5,7 4,9 4,1 3,5 3,7 3,1 Industry 38,0 36,5 37,1 35,8 34,1 33,9 31,3 30,3 Services 54,9 56,8 57,2 59,3 61,8 62,6 65,0 66,6 Source: Kwiatkowski, Kucharski, Tokarski (2003) str. 234, statistical data of the Central Statistical Office, own calculations. 20

321 Table 7 Change in value-added at constant prices (VA) and employment (E) in manufacturing (1995=100) Subsection Manufacture of food products and VA 110,32 123,78 134,69 130,54 127,81 141,72 beverages (subsection DA) E 106,87 102,86 108,73 103,89 95,94 90,71 Manufacture of textiles and textile products VA 99,86 101,17 106,85 92,43 87,68 87,25 (subsection DB) E 95,39 88,49 84,71 74,04 66,54 58,78 Manufacture of leather and leather VA 115,45 113,76 96,09 82,69 82,12 78,83 products (subsection DC) E 98,69 89,55 76,43 62,52 55,55 48,76 Manufacture of wood and wood products VA 103,14 144,38 170,44 147,24 177,41 161,88 (subsection DD) E 106,48 107,18 117,09 113,76 118,81 106,31 Manufacture of pulp, paper and paper VA 114,29 137,67 156,78 185,59 183,41 183,02 products, publishing and printing (subsection DE) E 106,64 99,62 114,29 122,10 114,38 110,54 Manufacture of coke, refined petroleum VA 87,32 92,80 112,79 108,86 108,99 133,53 products and nuclear fuel (subsection DF) E 98,41 95,12 87,28 108,97 86,37 80,50 Manufacture of chemicals, chemical products and man-made fibres (subsection DG) VA E 97,60 99,15 100,68 91,03 99,01 88,72 99,15 84,03 100,44 75,72 106,50 71,31 Manufacture of rubber and plastic products VA 114,87 132,95 148,32 140,97 140,30 157,81 (subsection DH) E 110,71 107,72 129,77 138,92 148,95 133,74 Manufacture of other non-metallic mineral VA 151,42 173,69 192,47 201,35 223,56 221,64 products (subsection DI) E 102,35 97,31 100,83 95,30 93,70 83,69 Manufacture of basic metals and fabricated metal products (subsection DJ) VA 94,57 116,06 114,89 100,07 101,57 96,25 E 103,09 95,48 99,80 90,28 83,35 78,40 Manufacture of machinery and equipment not elsewhere classified (subsection DK) VA 105,69 109,22 108,48 96,11 101,07 101,30 E 96,19 88,67 85,36 77,23 69,80 62,87 Manufacture of office machinery and VA 121,04 140,37 157,73 156,71 161,79 171,26 computers (subsection DL) E 101,40 98,58 99,65 97,65 93,29 89,93 Manufacture of motor vehicles, trailers and semi-trailers (subsection DM) VA 109,25 121,98 152,63 136,41 128,49 121,36 E 97,76 92,18 90,22 90,04 80,70 73,06 Manufacture of furniture; manufacturing VA 112,05 140,30 165,58 168,18 170,43 174,37 not elsewhere classified (subsection DN) E 105,58 106,75 114,94 115,05 112,30 103,11 Source: statistical data of the Central Statistical Office, own calculations. 21

322 Table 8 Indicators of education and continuing training in Poland and OECD countries Indicators Poland 1998 OECD average 1999 Proportion of persons with at least secondary education level in 54% 62% population of persons aged Proportion of persons with tertiary education level in population of 11.3% 14% persons aged Proportion of participants in continuing training 13% 31% Number of points achieved in a test of reading with understanding 230* 269 according to OECD survey * Poland was ranked 17th of the 18 countries surveyed Source: Education at a glance, OECD indicators, 2001; Strategia rozwoju edukacji narodowej [The development strategy of national education], Warsaw Table 9 Training of employees Poland (1999) EU candidate countries (1999) EU countries (1993) Proportion of firms cofinancing training of employees 39% 11% 69% 18%-96% Proportion of employees who are offered training 33% 20% 53% 32%-63% Average duration of training per participant (in hours) Proportion of training costs in labour costs 0.8% 0.5% 1.9% 0.9%-3.6% Source: K.Nestler, E. Kaillis, First survey of continuing training in enterprises in candidate countries (CVTS2). Statistics in Focus, Eurostat, European Communities 2002; Continuing Training in Enterprises Fadsend Figures, A report of the results of the CVTS, European Commission, 1994 Table 10 Labour Fund expenditure in Poland, in Year Labour Fund expenditure Expenditures on active policy in million PLN % of GDP in million PLN % of GDP , , , , , , , , , , , , , , , lack of data Source: Bezrobocie w 2002 r. [Unemployment in 2002] (2003) p. 125, T. Olejarz (2000) p. 20, data for the year 2000 provided by the Central Statistical Office in Warsaw, own calculations. 22

323 Table 11 Effectiveness indicators of active measures undertaken by labour offices, Tasks No. of persons Re-employment No. of Re-employment rate (%) persons rate (%) Total (Share of ALMP participants in total 51.2 (23,6) (16,7) unemployment, %) 49.8 Training Subsidised employment Public works Reimbursement of contribution paid to the Social Insurance Institution Reimbursement for graduates Graduate traineeships Socially useful work Special programmes Loans in total for unemployed for employers Other tasks Tasks No. of persons Re-employment rate (%) No. of persons Re-employment rate (%) Total (Share of ALMP participants in total 48.6 (7,5) (7.8) unemployment, %) 48.5 Training Subsidised employment Public works Reimbursement of contribution paid to the Social Insurance Institution x X Reimbursement for graduates Graduate traineeships Socially useful work x X Special programmes Loans in total for unemployed for employers Other tasks Source: MGPiPS 23

324 Economic Restructuring and Labour Markets in the Accession Countries Research Project commissioned by EU DG Employment, Social Affairs and Equal Opportunities, Contract No. VC/2003/0367 Coordinated by in cooperation with Alphametrics and The Vienna Institute for International Economic Studies Active labour market policy in Bulgaria by Iskra Beleva* July 2004 *) Institute of Economics, Bulgarian Academy of Sciences, Sofia

325 Contents Introduction The importance of ALMP for reducing unemployment ALMP development Involvement of the unemployed in active measures Incentives and disincentives for unemployed to participate in active labour market programmes Labour office services The range of ALMP programmes Participants in training and re-training programmes Assessment of the efficiency of ALMP The monitoring and evaluation process Conclusions...19 Annex:...20 List of Tables and Figures Table 1 Bulgaria: expenditure on labour market policies from the Retraining and Unemployment Benefits Fund (% of total) Table 2 Broad types of ALMP as a share of total ALMP expenditure Table 3 Number of participants in active labour market programmes/measures by type of participation Table 4 Number of participants by type of ALMP programme Table 5 Division of unemployed between different types of training (% of total number of participants) Table 6 Number and % of people who entered employment as a % of total unemployment Table 7 Bulgaria: labour market flows,2003 Figure 1 Bulgaria: total, LTU and youth unemployment Figure 2 Number of unemployment and the number participating in active programmes Figure 3 Average unemployment benefit in relation to the average wage Figure 4 Bulgaria: unemployed eligible to unemployment benefits and total unemployment Figure 5 Number of unemployed who entered employment as a % of total unemployment

326 Iskra Beleva Active labour market policy in Bulgaria Introduction The priorities of labour market policy (LMP) during the years of transition changed with shifts in the process of economic development. The main changes in policy were directed at defining the criteria for unemployment compensation more precisely; reducing the incentive for people to claim benefit; and encouraging the unemployed to participate more actively in programmes for employment, training, retraining and self-employment. The Bulgarian Government Programme (2000) identified active labour market policy (ALMP) as one of its priorities, aimed at reducing unemployment and increasing social welfare. The New Social Policy Strategy (2003), 1 underlined the importance of ALMP for improving the functioning of the labour market. In accordance with the provisions of the Accession Partnership of 2002, the Bulgarian Government and the EC prepared a Joint Assessment Paper (JAP). The paper includes a detailed analysis of ALMP in force, identifies the advantages of the different programmes and makes recommendations for improvement. 2 In May 2004, the first annual Report on the progress in implementing the commitments made under JAP was prepared and discussed. In 2003, the Government adopted the Employment Strategy ( ). The strategy defines the path for achieving the goals of the European Employment Strategy in Bulgaria over the medium and long-term, namely: full employment, quality and productivity at work and social and regional cohesion. The Action Plan (2004) outlines in detail ALMP programmes and measures being undertaken in the short-term, as well as the direction for their further development. The Strategy and the Action Plan also include a package of new programmes and measures to be implemented for strengthening equal opportunities and incentives for life-long learning. The aim of this study is (i) to identify the importance of ALMP in Bulgaria for reducing unemployment; (ii) to outline the progress made in developing ALMP and (iii) to draw some conclusions on the future priorities of ALMP with respect to economic development and Bulgarian EU accession in A document of the Ministry of Labour and Social Policy and a component of the above-mentioned programme. Joint Assessment of Employment Priorities in Bulgaria, October 2002, EU and MLSP, Bulgaria. 1

327 1 The importance of ALMP for reducing unemployment The development of ALMP in Bulgaria during the transition has changed significantly in terms of priorities and overall design. The mechanisms and the instruments for its implementation have also been improved and continue to be so. Over the period , Bulgaria suffered high and persistent unemployment. In the early years of the transition ( ), as in many other transition countries, unemployment was a consequence of the radical economic reforms undertaken. In Bulgaria, the initial rise of unemployment was not reversed because of inconsistent economic reforms and the additional economic shocks (such as the Gulf war and the Yugoslavian embargo). As a result, unemployment remained high, characterized by high long-term unemployment and increasing youth unemployment (Figure 1). In 2003, employment was 33% below the pre-transition (1989) level, average annual unemployment for the period numbered 584 thousand people and the average annual unemployment rate was 16.6%. Figure 1 Bulgaria: total, long-term (LTU) and youth unemployment (YU) thousand LFS unemployment LTU YU \up to 24 years old The job creation potential of economic growth over the past few years (an increase of over 4% between 2000 and 2003) could not compensate for the significant job destruction over the preceding years. The number of jobs which were lost still exceeded those newly created. The growth of the private sector could not absorb the number of people looking for work. Thus, active labour market policy was, and still remains, an important factor for the re-integration of the unemployed, particularly of vulnerable groups among these, into work. In the first half of 2004, there has been a downward trend in unemployment, due to a large extent to active labour market programmes and subsidized employment, in particular. 2

328 2 ALMP development In Bulgaria, as in all other transition countries, labour market policy was introduced alongside economic and social reforms. At the beginning of the transition period, passive labour market policy prevailed. It was a time when unemployment emerged and exploded as a new social phenomenon. As a result, one of the main targets of labour market policy was to identify and register people who were unemployed. Since the legal framework was in a preparatory stage, policy decisions were made by way of Resolutions issued by the Council of Ministers, targeted at relieving social tension resulting from the high unemployment rate. 3 The institutions responsible for tackling unemployment were also in the process of being established and developed. The passive and reactive character of LMP prevailing at that time can be seen in the pattern of expenditure of the special fund set up, called the Retraining and Unemployment Benefits Fund (Table1). Table 1 Bulgaria: expenditure on labour market policies from the Retraining and Unemployment Benefits Fund (% of total) Total expenditure Active labour market policies Passive labour market policies Other expenditure Source: National Employment Service. The amounts collected from employers and employees contributions (initially at a ratio of 4:1) were used for funding labour market policy until In 1993, some 83% of total expenditure went on passive labour market policy. By 1998, this had declined to 57% but it then rose to 75 % in The importance of employment promotion as an element of labour market policy increased during the course of the reforms. At the end of 1997, the Council of Ministers approved the Unemployment Protection and Promotion of Employment Act (UPPEA). This specified the creation of new jobs and the increasing employability of job-seekers as priorities. But, high unemployment and the unemployment benefits to which it gave rise exhausted the limited financial resources, leading to further tightening of the conditions for eligibility to 3 CoM Resolution 57 on transfers and effective use of redundant labour force dates from 1989; CoM Resolution 102\1990 introduced price indication on benefits analogous to wage; CoM Resolution 110\1991 on solving urgent employment and unemployment issues; CoM Resolution 209\1992 on amending and supplementing the previous bylaws; CoM Resolution 135\1992 further lowered the level of benefits to 60% of the previous gross wage; CoM Resolution 659\1993 awarded successful completion of training with an additional lump sum equivalent to 15 % of unemployment benefits for the whole period of retraining. For more details see Beleva, Tzanov, Labour market flexibility and employment security, Employment paper 2001/30, ILO, pp

329 benefit so as to reduce the share of labour market expenditure going on passive measures and to expand the share going on active measures. After one year, however, it was evident that it was not possible to develop active labour market programme within the existing unemployment fund and in 2000, the government decided to finance a number of priority programmes from the state budget. Priorities continued to change and the above act was replaced by the Employment Promotion Act (EPA), which came into force in January This act clearly identified unemployment as a social risk and unemployment benefits as social insurance payments and the National Social Security Institute (NSSI) was placed in charge of paying benefits. The existing fund was renamed as the Unemployment Fund and is at present managed by the NSSI, while active labour market policy is governed by the Ministry of Labour and Social Policy together with the Employment Agency and is fully funded by the state budget and international organizations. Expenditure on ALMP rose from 0.07% of GDP in 1994 to 0.21% of GDP in 1999; 0. 28% in 2001 and 0.37% in In 2004, it doubled to 0.67%,an indicator of the priority given to ALMP in the past few years as a means of increasing the efficiency of labour and reintegrating the unemployed into work and the society. The improved targeting of ALMP along with larger funds resulted in a decline in unemployment among the priority target groups of long-term and youth unemployed (see Figure 1 above). 3 Involvement of the unemployed in active measures 4 Participation of the unemployed in ALMP depends to a large extent on the availability of finance. The above mentioned changes to the funding of ALMP were associated with an increase in the proportion of the unemployed participating in active programmes from 16.7% in 1995 to 33.6% in , a rise of 40% in terms of numbers (Figure 2). The inverse relationship between unemployment and the number participating in active labour market programmes is clearly evident, which suggests either that these programmes are effective in reducing unemployment or that the funding available for such programmes is itself inversely related to the number of unemployed who need income support. 4 5 The analysis is based on last three years data because as was explained in the text above there is a break in the time series due to the radical change in the legal base in the end of Unemployed included in ALMP to total registered unemployed people annual per year. 4

330 Figure 2 Number of unemployment and the number participating in active programmes Total unemp Unempl. included in ALMP Source: Agency for Employment, Information Bulletin for the respective years. 4 Incentives and disincentives for unemployed to participate in active labour market programmes There are two options for people losing their jobs to find work again. The first is to register as unemployed and to become eligible to participate in an active programme and obtain support in searching for a job The second is to enter shadow employment. There is also an option to move from unemployment to inactivity or to shadow employment. This is of significance in Bulgaria since many people who are unemployed and discouraged from actively looking for work move off the unemployment register into the informal or grey economy. Three factors motivate people to register themselves as unemployed: access to active labour market programmes, as noted above; entitlement to unemployment benefit and access to the social assistance system. Active programmes are of main importance for people who would like to get back into employment. The unemployment benefit system, and social assistance, is designed in some respects to provide an incentive for the unemployed people to participate in active programmes and actively to search for jobs. People who become unemployed as a result of mass lay-off associated with economic restructuring are eligible for a certain period of time (from 4 to 12 months) for unemployment benefits so long as they satisfy certain conditions. Eligibility to benefit and the amount receivable is related to the previous length of service and the length of payment of unemployment insurance; and the reason for unemployment whether because of lay-off, dismissal and so on. Unemployment benefit is set as a proportion of previous earnings in work (70 or 80%), though with minimum and maximum amounts receivable. From 2002, the State Social Security Budget Act has set these minimum and maximum amounts each year. The figures for 2003 were a minimum amount of 80 BGL 5

331 and maximum of 140 BGL. Unemployment benefit is paid monthly without any waiting period. The ratio between average unemployment benefit and the average wage has changed significantly since benefits were first introduced (Figure 3). The unemployment benefit system at the beginning of the transition was relatively generous, since it was expected that high unemployment would be a transitory phenomenon. The increase and persistence of unemployment changed the perception and led to rules becoming more restrictive. As a result, the level of unemployment benefit fell to around 30% of the average wage. In December 2001, only 15% of the unemployed received the maximum rate of benefit, and 20% received less than the maximum benefit and below the minimum wage, and 47% received minimum benefit 6. In January 2004, 22.5% were in receipt of unemployment benefit at the maximum rate and 48% received the minimum rate of benefit. Figure 3 Average unemployment benefit in relation to the average wage The number of unemployed entitled to benefit increased in the early years of the transition but has declined since (in Figure 4). Between 1990 and 1994, access to unemployment benefits rose as a result of mass layoffs due to the restructuring of enterprises and the process of privatization. After 1994, the number of unemployed eligible to unemployment benefit declined due to the restrictions on eligibility that were introduced. At present, there are also a number of active labour market programmes under which the unemployed wishing to start their own business are eligible 6 Joint Assessment of Employment Priorities in Bulgaria, October 2002, p.23 6

332 for a lump sum payment equivalent to the total amount of unemployment benefit payable and this has also contributed to reducing registered unemployment. Figure 4 Bulgaria: unemployed eligible to unemployment benefits and total unemployment Unemployed Eligible to unemp.benefits Unemployment benefit can, therefore, be regarded as a stimulus to register as unemployed for a limited period. At the same time, the declining rate of benefit has encouraged people to be active in participating in active programmes and in job search. The unemployed who participate in active measures are taken off from the unemployment register. According to the EPA and the Regulation for its implementation, the unemployed who are involved in a programme or measure are treated according to the terms and conditions of the labour contract signed between the labour office, the person unemployed, the employer and the training organization (where training is included). While participating in active programmes, the unemployed receive salaries when they are enrolled in training courses and accommodation as well as daily allowances when these take place in a different town from where the person concerned is living. Depending on the programme, there could be supplements of varying amounts on top of the normal benefit. For those unemployed, participation in active programmes is likely to be more beneficial, financially, than social benefits. The amount of social benefit in Bulgaria varies according to income. The guaranteed minimum income is 40 BGN (20 EURO). There are different social assistance schemes, with the monthly amount in 2003 varying between 36 and 60 BGN (18 and30 EUR) depending on the size of the family and between 50 and 80 BGN (25 and40 EUR) if energy support is included. It is also worth pointing out that there are other, non-monetary, benefits, such as learning to read and write, being taught basic skills or receiving vocational training, which are a key 7

333 part of the importance of ALMP for integrating people into employment. Other nonmonetary benefits include social inclusion, being able to make contact with employers, with other people unemployed and with labour office staff to obtain advice, guidance and information. These are all of importance for improving the chances of someone finding employment. In 2003, therefore, there were only some 10% of the unemployed who refused to participate in the From social benefits to employment programme (one of the main active measures), and almost 45% of these did so on the grounds of bad health. Nevertheless, despite the obvious advantages of participation in active programmes, there are still some among the unemployed who decline to participate or try to avoid participating in such measures. A major reason for this is that only those who have been registered as unemployed for six months at the Labour Office can apply for general social assistance a means-tested benefit comprising monthly cash payments together with a range of benefits in kind (free goods and services, access to health care and so) and occasional or emergency (lump sum) benefits. Eligibility is determined on the basis of the Guaranteed Minimum Income (GMI) adjusted for the size of households and the characteristics of those living there (age, health and so on). If there are dependent children, social assistance can be granted to the unemployed without the six month registration requirement. Participation in the social assistance system for families with many children is profitable because these increase the amount of social benefit receivable in addition to the child allowances (18 BGN per month per child) they give entitlement to. Although the family may be registered as poor and be in receipt of social benefits, in many cases, family members work illegally in the grey economy or earn money in other ways begging, collecting scraps, growing vegetables, selling goods in the market and so on. For these people, the payment they receive when participating in an active labour market programme may be little if any higher than they receive from social benefits and, at the same time,, they stand to lose their additional earnings and free time. They may, therefore, refuse to participate in programmes but because they do not want to lose entitlement to social assistance, they often illegally obtain certificates signifying bad health or disability. Over the past two years, the number of people with disabilities has grown substantially, and in response to this, access to social assistance has been tightened up. Loss of entitlement to benefit may occur in the following circumstances: registration at the labour office may be terminated in cases where the person concerned does not follow the recommendations included in their personal plan or the time schedule of visits to the labour office), fails to visit the labour office after receiving a written request to do so, or changes their address without notifying the labour office; 8

334 re-registration is possible after 12 months of termination, although where it is found that there are reasonable grounds for violating the above rules, re-registration occurs immediately; where someone refuses to participate in an active programme or measure, the sanction imposed can be loss of unemployment benefit, loss of eligibility to social assistance; loss of supplements on top of the normal benefit and so on. Employers as well as those registered as unemployed are liable to face penalties in cases where they submit incorrect information or violate the rules and may be asked to pay back payments received or even face criminal proceedings. In 2003, the General Labour Inspectorate General labour inspection, which is responsible for enforcing Employment Promotion Act (EPA) rules with regard to hiring the unemployed, reported a number of violations by employment agencies mediation firms in respect of the terms and conditions applying to the services they provided. In 35 cases, the activities of such firms were terminated. 5 Labour office services In accordance with the EPA (in force since 2001) labour offices are responsible for providing services to: those who are unemployed; those in employment; pensioners who would like a job; According to the EPA, services provided to people actively looking for a job by the labour offices should include: Employment services including information about available vacancies; information about existing active programmes and measures; advice on vocational courses and career guidance; vocational education and training; opportunities to participate in active programmes and measures; opportunities to obtain scholarships in cases of participation on training courses leading to vocational qualifications. Services provided to employers should include: employment services including information about the people actively looking for a job; information about current programmes and measures which are part of ALMP; opportunities to be involved in active programmes and measures; 9

335 information about grants and subsidies available when hiring the unemployed or for job creation; information about payments for participation in the provision of vocational training or apprenticeship schemes. These services have become very closely tailored to individual needs since People who become unemployed are first registered and then an individual action plan is prepared according to their needs and abilities, together with a time-schedule of regular visits to the labour office. Everyone so is obliged to follow the recommendations of the labour office, to contact their personal officer in accordance with the time-table and to undertake the activities specified in the individual action plan. The options available to the unemployed are listed in the EPA and include: programmes and measures for young people up to the age of 29 without work experience, which include the offer of a subsidized job for up to 12 months. In these cases, the unemployed are paid a monthly wage, with all employer costs being covered by the labour office, including social insurance, as and the cost of training. Where the person concerned has a disability or is an orphan, the period of employment is up to 18 months or 24 months for those under 24; measures for young people who are long-term unemployed, which also include subsidized jobs (with employers again being reimbursed for wage and social insurance costs); measures for young people who cannot read and write who are entitled to a 5-month literacy course and a scholarship amounting to 40% of the minimum wage; access to training or re-training in parallel with employment or separately; access to employment associations which provide subsidized community work in municipalities with high unemployment. Participants receive wages and have their social insurance covered for a up to 24 months; mobility promotion measures which encourage transition from passive to active programmes as well as subsidized part-time jobs. Participants are paid the minimum wage and social insurance is covered for a period up to 6 months; life-long learning which is a scheme targeted at those in employment and young people who have dropped out of secondary education and which consists of education or apprenticeship schemes. Employers are paid half the costs of hiring people who are unemployed; incentives for self-employment, which are consist of the lump-sum payment of unemployment benefits to those who draw up a business plan and have it approved; job creation schemes, which include subsidized employment for the unemployed in micro-sized firms. Employers have their social insurance contributions reimbursed for 10

336 up to 24 months for the first five unemployed people they hire, or up to 12 months if they hire fewer than five. If they take on people unemployed as part-time workers, they have the wages paid plus social insurance contributions for up to three months; equal opportunities schemes which are targeted at those disadvantaged on the labour market, including people with disabilities, for whom wages and social insurance contributions are covered for up to 12 months) or in the case of temporary, seasonal or part-time employment, for up to 6 months and single mothers, for whom wages and social insurance contributions are also covered for up to 12 months). In addition, community work is available to people unemployed in municipalities with high unemployment, to young people up to 29, to the long-term unemployed with a low level of education, to former prisoners and to other disadvantaged groups. Employment protection schemes also exist which provide additional payments to employers if they take on people unemployed without reducing the total number of employees. 6 The range of ALMP programmes As should be evident from the above, ALMP in Bulgaria includes a wide range of measures for the unemployed. A breakdown of ALMP by broad type of measure or programme shows that between 1993 and 2000, the importance of subsidized employment increased substantially, while programmes on the training and re-training of the unemployed declined (Table 2). Expenditure on administration of ALMP also declined, and spending on measures for young people remained broadly unchanged but at a relatively low level. Table 2 Broad types of ALMP as a share of total ALMP expenditure Total expenditures Active labour market policies of which: Administration services Training and retraining Youth programmes Subsidized employment Source: National Employment Service Because of significant changes in the legal basis and the system for financing ALMP, there is a break in the data series in Because of this, t comparison of the data for

337 2003 with those for the previous period in terms of expenditure and the number of participants by type of programme can only be made with caution. After 2000, the types of ALMP programmes are more closely in line with the European Employment Strategy guidelines i.e. improving employability; incentives for employment; development of conditions for life-long learning; business development and strengthening equal opportunities. The National Employment Action Plans ( ) focused on increasing employability; the development of entrepreneurship; incentives for business development and job creation and strengthening equal opportunities. The number of participants in ALMP programmes increased over the period (Table 3). Table 3 Number of participants in active programmes/measures by type of participation Total participant Participants in measures Participants in programmes Source: Employment Agency, Annual Information Bulletin. In 2003, some 82.5% of participants were involved in ALMP programmes, Most of these programmes are designed on the module principle, combining work with training. Programmes have tended to be targeted at different vulnerable groups which have changed over time in terms of priorities e.g. priority in providing access to programmes was given to young long-term unemployed, then to long-term unemployed receiving social benefits, then to single mothers, etc. The programmes and measures targeting specific vulnerable groups include subsidized employment, training and incentives to employers for hiring the unemployed. Programmes are both national and regional in scope, targeting regions with high unemployment in particular. The figures in Table 4 indicate the priorities given to different types of programme. Over the past three years, programmes have been reviewed intensively. As a result some programmes have been stopped, others where the number of participants has declined are planned to be closed down and a number of new programs have been designed and are in process of been introduced on a pilot basis. 12

338 The improved coverage of the new programmes is reflected in the increasing number of participants. These are directed mainly at increasing the employability of socially excluded groups (the long-term unemployed, people receiving benefit from the social assistance system). Education and training are also included as means for re-integration. Table 4 Number of participants by type of ALMP programme Total number of participants / Participants in ALM measures, of whom: Young people* Long-term unemployed Qualification and motivation for unemployed Territorial mobility Self-employment measures Subsidized employment Participants in ALM programmes, of whom From social benefits to employment programme Regional employment programmes Temporary employment programmes Assistance for retirement programme Increasing employability and entrepreneurship of young people project Job clubs and structural development From social assistance to employment programme ALMP in cooperation with international institutions Source: Employment Agency, Information Bulletin, 2001,2002 and A regional approach has been increasingly followed over the past three years, regional disparities being taken into account and financial resources being distributed among regions according to the rate of unemployment. Subsidized employment and training were the main priorities in More people participated in subsidized employment programmes and measures than in any of the others. This fact signifies one of the main challenges facing ALMP today how to transform subsidized jobs into stable permanent employment. The largest number of participants were enrolled in the From social assistance to employment national programme (see Box). 13

339 From social assistance to employment programme The most popular national programme, From social assistance to employment, is targeted at the long-term unemployed who are not eligible for unemployment benefit (because of being registered for 24 months or more and in receipt of benefits for 18 months) and receive social assistance, young people under 20, and single mothers not receiving social benefits. These are offered subsidized community work and paid by employers for the work they do, they are also entitled to additional payments in accordance with the Labour Code to bring their earnings up to the minimum wage and have their social and health insurance contributions covered. Everyone unemployed is eligible to participate in the programme for up to 3 years. They receive more than they would if in receipt of social benefits which are lower that the minimum wage. They also gain from participating in the programme and from the training or education, and social contacts which this entails. The National Employment Action Plans (NEAP) , outlined in the Annex show in more details the number of participants and the budgets of different programmes and measures. It shows that, in 2003, unemployed participated in programmes and measures, a unemployed were trained and were enrolled in training courses. Under the schemes included in the EPA, people started work or enrolled in training courses were enrolled in the From social assistance to employment programme, while were employed in the Beautiful Bulgaria Project and another were trained. In addition, the Jobs Opportunity through Business Support project provided work for and training for Participants in training and re-training programmes The above mentioned programs are important in view of: the large share of the unemployed with a low level of education (36% of the total in September 2003 had only compulsory education or lower). and the priority given in the European Employment Strategy to the quality employment, which cannot be achieved without extensive labour market programmes to train or re-train the unemployed. Economic restructuring has substantially changed job profiles, which has led to training and re-training becoming a priority not only to bring those who are unemployed into employment but to adapt the skills of the employed in line with new requirements and to help prevent them falling into unemployment. The share of expenditure on these types of programme and the number of participants have risen significantly in the recent past.: Between 1993 and 2000, their share of expenditure was insignificant and tended to fall (Table 2 above). The share of the unemployed, involved in training and retraining programmes was just under 2% in both 1999 and This increased to 5% in 2001 and to over 12% in 2003 (Table 4), which indicates the present importance and the priority, given to these types of programme. 14

340 There are four main reasons for training and re-training programmes: to provide professional qualifications; to provide additional professional qualifications; to update or extend qualifications; to motivate people to improve their employability and to find work; Table 5 Division of unemployed between different types of training (% of total number of participants) Years Professional qualification Additional professional qualification Update or extend qualification Motivation* * With the UPPEA, motivation was included in the activities of labour offices but was not funded as a separate measure. In 2000, 2251 people completed motivation courses, in 2001, 163 and in 2002, Training to provide additional professional qualifications increased most significantly over the period , which reflects the intensive economic restructuring process which pressed people, including the employed, to change their vocation. At the same time, vocational qualifications increased in importance. In 2003, the Employment Agency reported that vocational training courses were organized, 45% more than the previous year, and the number of individuals completing training courses was 66% higher. A persistently large proportion of people expressed interest to obtain extra training (67% of those trained), reflecting continuously growing requirements for a skilled and knowledgeable labour force on the part of employers. 7 The training of the employed has become an important preventive measure aimed at reducing unemployment and improving the adaptability of the labour force. In 2003, people in employment participated in vocational training courses, while completed vocational courses as compared with just 554 in The proportion of those obtaining employment directly after completing a vocational training course increased from 40% in 2002 to 49% in Report on the Progress Made by the Republic of Bulgaria on the Joint Assessment of Employment Priorities, May,

341 8 Assessment of the efficiency of ALMP Over the period the total number of unemployed who found a job increased from to This represents an increase from 36.4% of the total to 46.2%. Fig 5 Number of unemployed entering employment as a % of total unemployment Total number of unemployed Number of unemployed, entering employment Number of unemployed,entering employment through ALMP An important feature which should be underlined is that both the unemployed, who entered employment through ALMP increased in both absolute and relative terms. The number rose from about 63 thousand in 2000 to over 134 thousand in 2003 and, in relative terms, from 32% of the total entering employment to 54%. ALMP has therefore become a significant means of reducing unemployment. The relative number of unemployed who entered employment after completing a training course is also an indicator of the efficiency of ALMP. (Table 6). Table 6 Number of people who entered employment after training Number of unemployed who finished training courses Number of people who entered employment after finishing training courses % of total ALMP efficiency in getting people into long-term stable employment, measured by the growth in the number of participants finding employment after completing training courses indicates a relatively high level of efficiency. The decline in the proportion concerned from 61.6% in 2001 to 48.9% in 2003 is in part due to the very sharp increase in the number of 16

342 unemployed completing training courses. For improving the balance between training and job placement, the MLSP prepared a methodology in 2004 for determining the needs of the employers for labour with specific qualifications. The proportion of people who becoming unemployed again varies across programmes and the individual characteristics of participants. In 2003, for example, only 40% of those completing the From social benefits to employment programme returned to the unemployment register afterwards, which means that 60% stayed in employment after their labour contracts under the programme came to an end. Another example is the Incentives for women to start their own business as social service providers project, which involved 226 participants, only 15 of whom re-registered as unemployed after completing the programme. Nevertheless, despite the intensive development of ALMP, there remain relatively limited opportunities for the unemployed to move out of unemployment into stable employment, while the flows from unemployment to inactivity increased during 2003 (Table 7). The flow from unemployment to inactivity might in part be a result of personal choice because of a preference for not participating in an active labour market programme, if informal employment is taken into account as mentioned above. According to some estimates, informal employment in Bulgaria is equivalent to around a third of total (formal) employment. 8 Table 7 Bulgaria: labour market flows, /1-2 Q Total Employment Unemployment Inactive Employed Unemployed Inactive /2-3Q Employed Unemployed Inactive /3-4Q Employed Unemployed Inactive Source: LFS data. 8 T. Mladenov, Control over the registration of labour contracts, in Labour Markets in Bulgaria, USAID Project, Sofia, March 2004, p

343 An evaluation of the gross effects of ALMP programmes and measures was carried out in 1997 by the National Employment Service assisted by independent national experts. In 1999, a study on the Evaluation of the net impact of the active labour market policies was undertaken by the Netherlands Economic Institute (NEI). This estimated that the net effect of programmes was to increase employment by 2.5% in the case of temporary employment programmes, by 11%; in the case of training and retraining with a guaranteed workplace, by 10%; in the case of training and retraining without a guaranteed workplace, by 39% in the case of subsidized jobs, by 11% in the case of companies set up to generate employment and 43% in the case of self-employment schemes. A new evaluation of the net effects of ALMP is planned for the end of The programmes implemented in co-operation with international organizations, e.g. the Beautiful Bulgaria project, are regularly evaluated by independent experts. The MLSP undertakes ex-ante and ex-post programme evaluation for the projects it is responsible for funding. For example, when the pilot implementation of the national From social benefits to employment programme was over, the MLSP issued calls for tender for assessing interest in the programme, its accessibility, public opinion on the appropriateness of the targeting and incentives for participation, and so on. As a result, the programme was improved by the recommendations made in the evaluations. 9 The monitoring and evaluation process The efficiency of ALMP is regularly monitored by the government. The Employment Agency undertakes monthly monitoring and ongoing supervision during the implementation of programmes on the basis of administrative statistics. The Agency collects data on the number of people included, the number who have worked before, the resources spent, the effectiveness of vocational training (the number of placements after training), the jobs created and so on. The indicators used for - monitoring are published in a monthly bulletin as well as in an annual analytical report on labour market development and labour market policy. As part of its activities for developing the monitoring and evaluation of ALMP, in 2003, the MLSP prepared a set of criteria for selecting employers who should receive subsidies for hiring and training of the unemployed. In 2004, the MLSP developed a methodology for monitoring measures, based on the main criteria, guidelines and approaches for evaluation. The implementation of this methodology will complete the system of for the ongoing supervision and follow-up of labour market policy. 18

344 10 Conclusions A number of conclusions can be drawn from the analysis of ALMP in Bulgaria. First, the importance of ALMP for reducing the gap between high labour supply and modest labour demand had increased over the years. Second, over recent years, ALMP has been subject to substantive changes in terms of financing, design, targeting, incentives for participation, and so on, all directed towards increasing the opportunities for the unemployed to re-enter employment. Third. current ALMP includes a wide range of measures and programmes designed to assist the most vulnerable groups improve their position on the labour market and to integrate into society. Fourth, funds for these activities have increased significantly over the years and now amount to around 0.7% of GDP. Fifth, the increasing role of ALMP for reducing unemployment is due not only to the doubling of financial resources and the reshaping of ALMP but also to the better provision and improved quality of employment services, as well as to their more even spread across regions. This has increased the likelihood of job placement for those registered as unemployed. The individual approach which has been implemented, accompanied by better knowledge and clearer specification of employer requirements enables the selection and recruitment of job seekers to be made more effectively. Sixth, despite the increase in financial resources, the better targeting of ALMP, the improved quality of the employment services and so on, the low level of labour demand remains a major problem for the of the unemployed to move into stable employment. Seventh, ALMP should focus more attention on the training and upgrading of the qualification of both the unemployed and those in employment in response to the pronounced structural changes in the economy and in view of the increasingly competitive climate in Bulgaria. 19

345 Annex: Table Bulgaria: National Action Plans (2003 and 2004) participants and expenditure by type of programme 2003 People employed - numbers Participants in training and qualification courses - numbers Sources of expenditure (incl. State budget, EC funds, other sources)-bgl 000 Total, included Increasing employability Development of entrepreneurship; Business development and employment Strengthening equal opportunities Total, included Active and preventive measures for unemployed and inactive Job creation and entrepreneurship Increasing adaptability and mobility in the labour market Human capital development and life-long learning Incentives for labour supply Equality between men and women Incentives for integration of vulnerable groups Actions for reducing regional disparities in employment

346 References: People are the wealth of Bulgaria, Governmental program, 2000 New Social Policy Strategy, MLSP, internet: Promotion of Employment Act, , State Gazette Employment Strategy , adopted by the Council of Ministers on November 6, 2003 Report on the Progress Made by the Republic of Bulgaria on the Joint Assessment of Employment Priorities, May 2004 Regulation for the implementation of the EPA, internet: Joint Assessment of Employment Priorities in Bulgaria, October 2002, Labour market, annual reports for 2000, 2001 and 2002, Employment Agency, Annual Information on the state of unemployment and active labour market policy, published by the Employment Agency, 2001, 2002 and 2003 Beleva, Tzanov, Labour market flexibility and employment security, Bulgaria, ILO, 2000 Labour Market in Bulgaria 2003, USAID Labour Market Project, Sofia, March

347 Economic Restructuring and Labour Markets in the Accession Countries Research Project commissioned by EU DG Employment, Social Affairs and Equal Opportunities, Contract No. VC/2003/0367 Coordinated by in cooperation with Alphametrics and The Vienna Institute for International Economic Studies Development of SMEs in Bulgaria by Iskra Beleva* July 2004 *) Institute of Economics, Bulgarian Academy of Sciences, Sofia

348 Contents 1 Introduction 2 Growth of the number of SMEs in Bulgaria during the transition period 3 The role of SMEs in employment recovery and economic growth 4 SME activities by sector 5 Regional aspects of SME development 6 State policy for supporting SME development 7 Conclusions References Appendix List of Tables and Figures Table 1 Table 2 Number of enterprises by size Growth in employment by size of enterprise (numbers employed) Table 3 Job turnover by firm size, 2000 Table 4 Share of firms in gross value added by size ( ) Table 5 Table 6 Shares of value added, firm numbers and total employment by firm size, 1996 and 2001 SMEs employment by main sectors, (shares in total sectoral employment) Table 7 Share of SMEs in employment by sectors, Table 8 SMEs contribution to value added by branches, Table A1 Table A2 Table A3 Bulgaria: number of SMEs by sectors - B16 Bulgaria: employment by economic sectors - B16 (thousands) Bulgaria: SMEs employment by economic sectors - B16 (thousands) Figure 1 Contribution of SMEs to change in employment by sector, Figure 2 Employment in the private sector by education, Figure 3 SMEs density by regions, 2002 Figure 4 Figure 5 Figure 6 SMEs by regions (numbers) SME density, GDP per head and employment by planning regions Share of SME in total employment by planning regions

349 Iskra Beleva Development of SMEs in Bulgaria 1 Introduction The purpose of this study is to outline the importance of small and medium-size enterprises (SMEs) in the process of Bulgaria s economic restructuring and development over the transition period ( ) on the basis of the available statistics and to show the relevance of the development of policy in support of these. The Small and Medium Size Enterprises Act 1 defines SMEs as firms with employed number of people from 10 to 100, grouped as micro (up to 10), small (11 to 50) and medium (from 51 to 100) sized firms. Recent economic policy links the more successful employment recovery to the expansion of SMEs. During the transition Bulgaria suffered sharp employment decline a net loss of almost 1.2 million jobs ( ). In 2003 the employment rate was 68.1% of the pretransition 1990 employment rate. Positive economic growth over the last few years (over 4% for ), and macroeconomic stability (low level of inflation, balanced budget deficit, increasing foreign investments, etc.) created a favourable environment for increasing job creation and labour demand. In 2000, SMEs were responsible for almost one third of GDP and over half of total employment (29.9% of GDP and 50.7% of employment). These figures clearly show the importance of SMEs for economic growth and employment. An understanding of this importance at present underlies government policy of increasing incentives better targeting support to the development of SMEs. 2 2 Growth of the number of SMEs in Bulgaria during the transition period The growing number of micro and small enterprises is one of the most marked feature of economic development in Bulgaria in recent years. Over the six years the total number of SMEs increased by 25%, and within this, the number of micro firms (those with 10 or fewer people employed) rose by 65%, small firms of between 11 and 50 people employed by 23% and medium-sized firms with between 51 and 100 people employed by 12%. By contrast, medium-sized enterprises with people employed and large enterprises (those with over 250 people employed) declined in number by 19-20% (Table 1) In force from 1999, Gazette N84, The policy relates not only to macroeconomic stability, but also relevant changes in the tax system; legal and institutional support; incentives for increasing employment, etc.; The term medium-sized enterprises used below refers to firms with 51 to 100 persons employed. 1

350 Table 1 Number of enterprises by size Size of enterprises All enterprises Micro enterprises Small enterprises Medium enterprises* SMEs total Enterprises with employees Large enterprises * Enterprises with persons employed Source: Report on Small and Medium-sized Enterprises and National Statistical Yearbook 2002; Micro-firms with up to 10 people employed; small firms from 11 to 50; medium-sized enterprises 51 to 100 people employed and large firms more than 250 people employed. 3 The role of SMEs in employment recovery and economic growth The growth in the number of SMEs has been associated with a substantial growth in the number of people they employ. Table 2 Growth in employment by size of enterprise (numbers of employed persons) Size of enterprises All enterprises Micro Small Medium SMEs total with employees Large Source: Report on Small and Medium-sized Enterprises and 2000; National Statistical Yearbook 2002; National Statistical Institute. In recent years, SMEs have been the only part of the economy creating employment and absorbing the labour force. In 2002, SMEs accounted for 54 % of total employment. Over the period employment in SMEs increased by 47%. Employment in micro firms and small enterprises rose by 60% and 55%, respectively, employment growth in medium sized firms went up by 11%. (Table 2). At the same time, large enterprises were still affected by job losses because of the continuing privatization process and the restructuring which has accompanied this. Employment in enterprises with people employed 2

351 also declined over this period but by less after 2000 than before. The importance of SMEs for job creation, therefore, increased over these years. While jobs have been created in small firms, they have also been lost. In 2000, for example, some 12% of those employed in micro-sized firms and around 15% of those in small firms lost their jobs (Table 3). Nevertheless, this rate of job destruction was not much different from that in larger enterprises. Table 3 Job turnover by firm size, 2000 Firm size Job creation Job destruction Job turnover Employment growth rate Micro Small Medium Large Source: Jan Rutkowski, Bulgaria, A changing poverty profile, The World Bank, 2002, p. 80. A comparison of the growth in number of SMEs with employment in SMEs shows an increase in the average size of firm over the last few years. The average number employed increased from 3.6 in 1996 to 4.3 in 2002, with the rise being concentrated in micro-sized firms (where the average number employed rose from 1.7 to 2.2). For both small- and medium-sized firms, the average employment size remained unchanged between 1998 and 2002 (at 21.4 and 70.5, respectively). The contribution of SMEs to economic growth was also significant over this period as reflected in their share of gross value added (Table 4). Table 4 Share of firms in gross value added by size ( ) Micro firms Small firms Medium firms SMEs-total Firms with employees Large firms Source: Report on SMEs, Agency for SME, 2002, p.46 and SMEs in Bulgaria, NSI,

352 Between 1996 and 2001, the share of SMEs in value added increased from around a quarter to just over a third. The average gross value added generated by firms, however, differs considerably according to their size and the sectors in which they operate. The largest firms in the public sector generate relatively high value added partly because of their monopoly position. Small firms have less opportunity to generate value added since they operate mainly in sectors with rapid turnover and low efficiency, such as distribution and various others services. Moreover, small firms tend to lack the financial resources to invest in modern technology, training, marketing, and so on. Table 5 brings together the data on the growth in the number of SMEs and on their share of employment and value-added to indicate their contribution to economic growth over the period 1996 to Table 5 Shares of value added, firm numbers and total employment by firm size, 1996 and 2001 Micro firms Small firms Medium firms 1996 Firms with employees Large firms Share in total value added Share in total number of firms Share in total employment Share in total value added Share in total number of firms Share in total employment Source: NSI. The share of micro-sized firms in value added, therefore, increased by almost 3 percentage points over these years and in employment by 10 percentage points, while their share of firm numbers declined marginally. The contribution of micro-sized firms to employment, therefore, increased significantly more than their contribution to value added over this period. The same was true of small firms with 11 to 50 people employed. These increased their share of value-added by 2.6 percentage points and of employment by 7 percentage points, while their share in firm numbers rose by 1 percentage point. Mediumsized firms, on the other hand, increased their share of value added by more than of employment by 2 percentage points as against 1 percentage point, while their share of total firm numbers declined slightly. Firms with employees also increased their share of value added, even if only slightly, while their share in total employment decreased by 3 percentage points. The share of large firms in both value added and total employment declined significantly, though more as regards the latter than the former, implying a 4

353 disproportionate increase in value added per person employed as compared with smaller firms. The preliminary data for 2002 confirm the tendency for the SME contribution to gross value added and employment to rise further. Success rate of business start-ups In the middle of the last decade, SMEs in Bulgaria had a relatively high birth rate (30%), much higher than in other transition countries. Over the following two years their birth rate fell to 10.7% (in 1997), significantly below the average level in other countries. 4 There are at least three plausible explanations for this significant fall in the birth rate: the first explanation is related to the fact that this period coincided with the economic crisis and hyperinflation in the first half of In order to overcome the crisis, by mid 1996 Bulgaria had implemented a currency board regime, thus managing to maintain macroeconomic stability until now. The second reason is the relatively poor experience of business management in a market environment and unstable economic conditions. The third reason for the sharp decline in the company birth rate over the period was the loss in savings caused by the hyperinflation of 1996, which were the main source of investment for business start-ups. At the same time, banks implemented a very restrictive policy towards access to bank credits. Nevertheless, despite the difficulties in undertaking such an exercise, an estimated coefficient of survival indicates that 84% of the enterprises that started business in 1995 were still alive in By 1998, the status of the firms registered in 1995 was as follows: 63% were active; 18% had frozen their business, and 19% had failed. In 1999, the median age of micro-sized firms in Bulgaria was 5.4 years, which means that, on average, they were established in 1994 and had survived until Since 1998, the stabilized economic environment and the active policies implemented for encouraging business start-ups and supporting SMEs improved conditions and company birth rates increased. This conclusion is supported by the growing number of SMEs, increased employment, and the strategic intentions of those involved in running SMEs: The SME birth rate went on increasing until 2000, as can be seen by the growing number of micro, small and medium enterprises;. 4 5 Panorama of Enterprises in the Central and Eastern Countries in 1995, Eurostat, Phare; Development of Enterprises in the Central and Eastern Countries , Eurostat, Phare Conditions for the development of micro firms in Bulgaria, (in Bulgarian), NSI, 2000, p.61 5

354 The reduced number of micro-sized firms after 2000 is due to some of the micro firms becoming small firm (the number of small firms increased, while that of micro-sized firms fell). Other micro-sized firms failed. Consequently, although the number of microsized firms declined slightly, the number of people employed in such firms increased significantly. The growth of small firms (in terms of both numbers of firms and the number of people employed) can be attributed in part to the present economic policy which provides incentives for starting businesses (credits, labour incentives, social insurance, tax relieves, etc.). The effects of this policy and the incentives it includes are described below. The death rate of businesses declined. The strategic business objectives of SMEs for 2002 indicated that only 1.1% of the firms intended to stop business activity and the same small proportion intended to sell their business. Survival rates tended to increase. Based on the strategic intentions of SME businessmen in 2002, 56% intended to increase output; 55% - to increase profit; 52% - to increase market share. Another 29% indicated their intention to maintain the current status. 6 4 SME activities by sector Services versus manufacturing Data on SMEs by sector are available for the period. In 1997, 63% of employment in SMEs was in services, 24% in industry and 13% in agriculture. Five years later, in 2002, the SME employment in services had increased to 68%, that in industry to 25% and SME employment in agriculture had declined to 7%. In 2002, the great majority of SMEs were in services (86% of total SMEs.) Within services, most firms were micro-sized. Only 11% of SMEs were in industry and just 3% of all SMEs were in agriculture (though here most enterprises were micro-sized). As compared with 1997, SMEs had increased in services, while in industry and agriculture, the number of SMEs decreased. Over the period, employment in SMEs in services increased by 50%, that in industry by 42% and in agriculture by 29%. There are a number of plausible explanations for the lower growth of SMEs in agriculture, namely, the prolonged process of land restitution in Bulgaria; the prohibition on foreigners to buy land which limited foreign investment in the sector; the undeveloped land market and the risky character of investment in agriculture. 6 Report on small and medium-sized enterprises ; Agency for small and medium enterprises, Sofia, p. 80 The sum is over 100 as more that one answer was given. 6

355 Table 6 SMEs employment by main sector (shares in total sectoral employment) Industry Agriculture Services Source: Author s calculation based on NSI data Table 6 shows the share of SME employment by main sector. In 2002, SMEs accounted for almost 38% of employment in industry, double the share in In services, the share of SMEs in employment increased by slightly less, from 30 to 46%. In agriculture, the SME share of employment declined. Several overlapping factors underlie the higher SME growth in industry and services, namely: (i) the underdevelopment of SMEs in these sectors up to 2000; (ii) the accelerated privatization process in industry and services after 2000; (iii) the more active state policy to promote the development of SMEs in industry and services. SME employment by more detailed sector To examine the restructuring effects of the development of SMEs, sectors can be ranked according to employment growth over the period. The importance of SMEs for the development of each sector can be seen by grouping sectors by the share of SMEs in employment. Sectors are also ranked according to their share in total SME employment. The significance of SMEs for the development of employment in different sectors is shown in Figure 1. SMEs made the biggest contributions to employment growth (adding over 10% to total employment) in the following sectors: Other community, social and personal services, business activities, real estate and renting, hotels and restaurants, and wholesale and retail distribution. Employment in SMEs also increased substantially in health care, though here the data may overstate the change which actually occurred because the numbers were very small before 2000 and might not, therefore, be reliable. The increase was largely a consequence of privatization of some health services over this period (which was also the case in education). 7

356 In transport and communications, construction and manufacturing, they made a significant contribution to employment growth, adding between 5% and 10% to the total number employed. SMEs also made a positive contribution to employment over this period in education, gas, electricity and water and mining, though the numbers involved in each case were relatively small. The only sector in which employment in SMEs declined was agriculture, hunting, fishing and forestry. Figure 1 Contribution of SMEs to change in employment by sector, Other services Health Education Business activities Transport, communications Hotels, restaurants Distribution Construction Gas, electricity Manufacturing Mining Agriculture Total Change in SMEs as % of employed, Source: Author s calculations based on NSI data Over the period, therefore, the share of employment in SMEs increased in 11 of the 12 sectors distinguished, the only exception being agriculture (where the share declined from 11% to 8%). The extent of the increase, however, varied between sectors, reflecting the different stage of economic restructuring and the different pace of privatization. There were four sectors in which SMEs accounted for half or more of total employment in Wholesale and retail trade (84%), Hotels and restaurants (69%), Business activity, real estate and renting (57%) and Construction (50%). In all of these, their share of employment increased significantly over the preceding 5 years. 8

357 In three other sectors, Manufacturing (38% in 2002), Other community, social and personal services (25%) and Transport and communications (24%), the share of SMEs in total employment expanded significantly over the period. This was also the case in Health care, though there is some uncertainty about the precise extent of the increase. In Mining and quarrying and Gas, electricity and water, the share of SMEs in employment also increased but it remains small. This was also the case in Education. Table 7 Share of SMEs in employment by sector, Agriculture, hunting and forestry, fishing Mining and quarrying Manufacturing Electricity, gas and water supply Construction Wholesale and retail trade; sale and repair of motor-vehicles Hotels and restaurants Transport, storage and communication Real estate, renting and business activity Education Health and social works Other community, social, personal services The expansion of SME employment in Manufacturing and Transport and communications indicates their potential for job creation after privatization. There is likely to be an important niche for SMEs in both sectors and the possibility of them to increase their share of employment further. The concentration of SME employment in services and in Business activity, real estate and renting in particular is a reflection of the growth of more modern services, which had been undeveloped or underdeveloped in Bulgaria before the transition. There are also examples of SMEs developing in more traditional sectors, such as Hotels and restaurants, and, within Manufacturing, in wood processing, essential oils and wine production, for instance. The ranking of the sectors according to their share of total SME employment in 2002 confirms the importance Wholesale and retail trade, which accounted for 34% of the total employed in SMEs. Manufacturing accounted for 25%, Business activities and real estate for just over 8%, and Construction and Hotels and restaurants as well as Agriculture for almost 7% in each case. 9

358 A similar picture is evident for value-added, with Wholesale and retail trades accounting for 38% of total value added generated by SMEs, slightly more than their share of employment and Manufacturing for 20%, less than its contribution to employment in SMEs, implying that labour productivity in SMEs in manufacturing was lower than in other sectors (Table 8). The same was true of Hotels and restaurants, while, by contrast both Business activity and real estate and Construction made a larger contribution to the value-added generated by SMEs than to employment. Table 8 Share of SMEs in value added by sector, Agriculture, hunting and forestry, fishing Mining Manufacturing Electricity, gas and water supply Construction Wholesale and retail trade; repair of motor-vehicles Hotels and restaurants Transport, Storage and communication Real estate, renting and business activity The contribution of Wholesale and retail trades to the value added produced by SMEs increased between 1999 and 2001, while that of Manufacturing, Construction Business activities and real estate and Hotels and restaurants declined. In terms of firms of different size, micro-sized enterprises made an increasing contribution to value added in three sectors over these two years Agriculture, hunting, forestry and fishing ; Business activities and real estate and Wholesale and retail trade. Small firms made an increasing contribution to value added in 5 sectors, including - Business activities and real estate ; Wholesale and retail trade and Hotels and restaurants. Medium-sized firms were responsible for a growing share of value added in three sectors, Business activities and real estate, Hotels and restaurants and Electricity, gas and water, though in the last, their share remain small. Innovative versus traditional sectors The development of innovative as opposed to more traditional sectors of activity needs to be balanced. It depends on national priorities and the interests of foreign investors, 10

359 integration into the EU market, and existing national advantages for successful competition in international markets. At the end of 2001, in the EU-Phare Project Capacity Building for the Accelerated Growth of the SME sector in Bulgaria, a cluster approach was adopted to identifying priorities. The approach is designed to encourage the setting up of groups of economically related companies within a specific geographical area with the aim of achieving a more effective transfer of resources from national and regional authorities to improve the competitiveness of SMEs. The study identified the following sectors as priority areas for the economic development of Bulgaria and for attracting foreign direct investment: mechanical and electrical engineering; chemicals; electronics; R&D; metals; construction and food, drink and tobacco. The role of foreign enterprises in the development of SMEs in Bulgaria 7 Bulgarian SMEs have limited opportunities to produce and sell goods competitively in areas where patents, trademarks or industrial design are important. In some sectors, moreover, Bulgarian SMEs are closely linked to foreign direct investment and are highly dependent on international markets because of very limited domestic demand. The Report on Small and Medium-sized enterprises presented a study of the activities of 843 industrial enterprises in Bulgaria at the beginning of According to its findings, the prevailing relationship of Bulgarian enterprises with their foreign partners were mainly based on export activities. In almost 60% of cases, the Bulgarian enterprises were suppliers (sub-contractors) to their foreign partners. 8 In 4% of cases, enterprises established joint production arrangements or joint companies. A very small proportion of Bulgarian SMEs have any relationship with foreign partners based on a strategy of jointly producing high tech goods through establishing patents, prestigious trademarks or the use of the latest technology. Innovation activities of Bulgarian SMEs The most recent information on innovative activity of SMEs in Bulgaria comes from the 2004 study on SMEs, undertaken by the National Statistical Institute. 9 The study shows the relatively high innovation activity of Bulgarian SMEs over the period despite the still imperfect innovative structure in the country. This conclusion is supported by the following facts: For more details see Report on SMEs , Agency for SMEs, Sofia, p.101 Report on SMEs , Agency for SMEs, Sofia, p.103 SMEs in Bulgaria, Statistical analysis, NSI, Sofia, 2004 p The analysis follows the Eurostat definition for innovative activity incl. introduction of innovations; product innovation and process innovation. 11

360 one in every 10 enterprises was involved in organizing and implementing innovative activities; 28% of small firms introduced innovations; of all SMEs that declared they were involved in innovative activity, 39% had introduced innovative products, 19% had introduced innovative processes and 42% had introduced both innovative products and innovative processes; SMEs undertaking innovative activities accounted for 29% of all SMEs in Manufacturing ; 25% of those in Electricity, gas and water ; 19% in Business activities and real estate, 19% in Mining and 15% in Construction ; innovative micro-sized firms accounted for 69% of innovative SMEs, less than their share of the total number of SMEs (88%); innovative activity (the introduction of innovative products or innovative processes) increased the competitiveness of firms in international markets (those in the neighbouring countries and the EU). Some 10% of innovative firms declared this to be the case, whereas only 3.5% of non-innovative firms reported that their competitive position was stable in international markets; SMEs reported innovation-related benefits in the form of higher turnover and a wider range of goods; greater flexibility and capacity of production; labour costs and energy expenditure; better working conditions and the increased health of staff; the share of SMEs involved in joint innovation activity over the period was about 26% (27.1% for the firms introducing innovative process; 25.9% for the firms introducing new products); the joint innovation activity concerned mainly took the form of contracts with suppliers and clients, as well as with universities, research centres and business consultants; the joint innovation partners in the case of research institutes were mainly national; EC\EACT innovative partners were mainly suppliers of technology or software; customers or business consultants; innovation partners in central and eastern European countries tended to suppliers of technology, materials, components and so on. A number of companies also had partnerships with US and Japanese companies; whether or not SMEs are at the forefront of innovative efforts is a question which is difficult to answer because of the lack of data, but a number of pieces of evidence suggest that large firms are more advanced in terms of innovating. SMEs in traditional activities 10 The study mentioned above on business clusters in Bulgaria indicates a possible interaction between Bulgarian SMEs and foreign companies in a number of areas. These 10 This examples are from the Report on small and medium enterprises , Agency for SMEs, Sofia,

361 include the production of essential oils, wood-processing, textile manufacture, tourism, and wine making. Production of essential oils Production of essential oils in Bulgaria is a typical traditional sector, favoured by the specific climate conditions and flora. Up until 1989, the state company Pharmachim was the main producer at the national level, producing high-quality raw material for the perfume, cosmetic and pharmaceutical industry. Currently, there are more than 50 producers in this sector and private ownership exceeds 95%. These businesses are located mainly in the regions of Sredna Gora, the Rose Valley, and in and around 13 towns or cities, including Plovdiv, Stara Zagora and Chirpan, where refineries for the distillation of the raw material are located. Production of essential oils employs about people, while some people are involved in the cultivation of the raw material in 25 different locations. 95% of the companies are SMEs, producing essential oils from roses, lavender and so on. 98% of the production is exported to France, the US, the Netherlands, UK, Switzerland, Germany, Japan, Italy and so on. There is a Research Institute of oleaginous rose and medical herbs (established in 1934), that provides support to companies belonging to the National Association of Essential Oil Manufacturers. Recently, productivity has tended to be depressed by plants not being renewed, the lack of a new laboratory to apply French standards and the use of modern small agricultural machines. A study of wood processing in the Mesta river region The study covered producers (112 wood companies operating in the region, 102 of which were micro-sized firms) in four municipalities, where about 64% of the total area is woodland. Wood processing accounts for 14% of employment in the region. The production amounts to 11% of the national total. The recently established Mesta Business Centre provides training as well as support in finding foreign markets and partners. Mesta Mebel can be regarded as a very successful company it employed 27 people and through its trade contacts as a supplier for a Dutch business, it established a joint venture in Bulgaria Hamefa-Mesta Mebel. As a result of closer manufacturing relations, the joint venture succeeded in increased productivity by some 60%. At present, it has 98 employees and 34 local suppliers. The enterprise exports to Belgium, France, Germany, Norway, Switzerland and the US. SMEs in innovative sectors Good opportunities for the development of SMEs exist in microelectronics, a sector where there good possibilities for attracting foreign investment and creating high-tech jobs. One important advantage of Bulgaria is the considerable number of well-trained specialists. Some Bulgarian SMEs, especially in computer software, have succeeded in achieving rapid growth in the local market and have opened representative offices abroad. These include Sirma AI EOOD, established in 1992 by a group of 5 young Bulgarian specialists who have seen their Sirma trademark adopted by a multinational group of IT companies located in Canada and US as well as in Bulgaria. It comprises 3 companies - EndView System Corporation, specializing in CAD/CAM products and engineering solutions in 13

362 packaging design and video-measurement systems; WorkLogic.com Corporation, specializing in internet-based software for corporate business management, and Sirma AI EOOD, which is the Group Research Centre. At present 70 highly qualified programmers work in the Sirma Group in Bulgaria and 15 work in the marketing and sales divisions of the Canada-based companies. They also include Netage Solutions Inc, specializing in the development of complete internet applications, especially for the financial sector, with a staff of 40 and offices in Boston and San Francisco as well as Sofia. 11 Microelectronics firms are concentrated in Sofia, in the former centre of the sector, Botevgrad, and in Plovdiv. Most of the firms are located in Sofia (including Sillway Semicondutions Inc. with 234 employees, Hybrid integral schemes, with 96 employees; Expect Ltd, with 150 employees) and TC-IME Inc. with 58 employees. The case of Mikroelectronika, a company based in Botevgrad (around 100 km from Sofia, the Bulgarian Silicon Valley before the transition). The company was declared insolvent in 1996 and sold to TC-IME jointly with some newly established companies in this area. New SMEs emerged to fill the various technological niches. In the process of market restructuring, these companies were divided into several groups: Small and medium export oriented innovative enterprises (established with Bulgarian, German, Belgian, Dutch and French capital innovative activity includes over 130 original circuit board projects annually and 24 patents registered in France, Belgium, the US and elsewhere) micro and small sized innovative enterprises with service functions; technological centre; finishing and accompanying back-end activities in microelectronics; product restructuring in microelectronics is directed mainly towards their use for special purposes, cars; industry and automation, entertainment and household applications, children s toys and telecommunications. The above examples demonstrate that: there are good examples of the development of SMEs in both basic (e.g. essential oils, wood-processing, textiles, wine-making and tourism) and innovative sectors and activities (e.g. microelectronics, software and business management); the development of SMEs depends greatly on foreign direct investors, who tend to be interested mostly in investing in more traditional activities, such as food, drink and tobacco, textiles, wood and non-metallic mineral products; cases of Bulgarian SMEs with links with foreign partners, based on high or modern technology, patent protection or prestigious trademarks are relatively limited; Bulgarian SMEs involved in more innovative activities are usually contractors or subcontractors of foreign companies. Their production tends to be exported and the extent of technological transfer to the national economy is limited. Moreover, it is questionable 11 More examples can be seen in the Report on Small and Medium-sized enterprises , Agency for SMEs, Bulgaria, Sofia. 14

363 whether the products or services they produce are transferable to the national economy at the present level of technological development. Restructuring and changes in the skill levels of those employed in SMEs 12 The effects of restructuring on the quality of the labour force at the beginning of the transition were negative, since many highly qualified people lost their jobs. Over the years many of these have emigrated and others have changed what they do according to the jobs available (which usually require lower qualifications). Many people with high qualifications went into the private sector. Thus, in September 2000, around 25% of those with higher education and in employment worked in the private sector. By September 2003, this share had increased to 36%. Over the period, therefore, the number of people with higher education employed in the private sector increased by 68%, while the number with upper secondary education rose by 31.5%. By contrast, the number of employees with lower secondary education only hardly increased at all and the number with primary or lower education declined. Figure 2 Employment in private sector by education, higher upper secondary low er secondary primary or low er Sep-00 Sep-01 Dec-02 Sep-03 Although the share of people with higher education employed in the private sector has grown (Figure 2), private employers still mainly hire people with upper secondary education, especially those with technical and vocational qualifications. In sectors with expanding employment in SMEs, those with upper secondary level technical and vocational education represented nearly half of the total employed (46% in both Wholesale and retail trade and Hotels and restaurants) in September In 12 Educational structure of employment by branches is available from LFS since

364 Business activities and real estate, 55% of the people employed had higher education and 41% had upper secondary level with only 4% having less than this. Over the period , increasing numbers of people employed in Wholesale and retail trade had higher education, while those with lower secondary education declined to 9%. This tendency is also evident in other sectors with high employment in SMEs employment - e.g. Hotels and restaurants (those with lower secondary education down to 11% in September 2003) and Business activities and real estate (down to 4%, as noted above). The share of people in employment with lower secondary education in sectors where SMEs were less important in September 2003 was higher 58% in Agriculture, hunting, forestry and fishing, 20% in Manufacturing, 19% in Mining and quarrying. In 2003, 15% of the self-employed in Bulgaria, who account for a large proportion of microsized firms, had higher education, 41% had upper secondary education, 35% had lower secondary education and 8% had a level lower than this. Comparing these proportions with the educational structure of the self-employed in 2000 (22% with higher education; 56% with upper secondary; 20% with lower secondary and 3% with primary or lower) shows that self-employment increased disproportionately among people with relatively low levels of education. This reflects in part the more active involvement of long-term unemployed in labour market programmes for subsidizing employment in agriculture in particular. Over the years, along with the economic restructuring and the growth of SMEs, the demand of private sector employers for qualified worker, in terms of educational levels, has increased. At the same time, an increasing number of private sector employers have become interested in labour market programmes for training and retraining. 5 Regional aspects of SME development 13 The distribution of the SMEs across the 6 planning regions is relatively uneven. SME density calculated per 1000 inhabitants by planning regions in 2002 shows an average density of 27.9 SMEs per 1000 inhabitants for the country as a whole, but this varying from 31.6 SMEs per 1000 inhabitants in the South-West region to only 19.9 per 1000 inhabitants in the North-West region. Elsewhere, the North-East region had 30.6 SMEs per 1000 inhabitants, South-Central, 29.5, South-East, 25.8 and North-Central, 25.4 (Figure 3). 13 The planning regions in Bulgaria were established in 1999 as regional structures to respond to governmental priorities related to the implementation of targeted regional policy, as well as to meet EU requirements for Bulgarian participation in the Structural Funds. 16

365 Figure 3 SMEs density by regions in 2002 Average for the country South-West South-Central South-East North-East North-Central North-West In 2002, SMEs were, therefore, disproportionately concentrated in the South-West region, where the capital city is situated, while the North-West region had the lowest concentration). Regional disparities were more significant in 2002 than in Over these two years, the number of SMEs increased only in one region South-West, where the density was already highest while the number declined in South-Central and remained broadly unchanged elsewhere. (Figure 4). Figure 4 SMEs by regions ( numbers) North- West North- Central North- East Sout h- East Sout h- Central Sout h- West

366 Although there is some relationship between the density of SMEs and GDP per head across regions, the correlation is relatively weak. There is, however, a much closer association between the density of SMEs and the employment rates (the total employed relative to population of working age) (Figure 5). Nevertheless, it should be noted that differences in employment rates between regions are relatively small, with the exception of the South-West (57% of population aged 15 to 64 in 2002) and the North-West (44%). Figure 5 SME density, GDP per head and employment rates by planning region in Bulgaria SME density, 2002 GDP per head, 2001 Employment rate, 2002 % of national average South-W est North-East South- Central South-East North- Central North-W est Source: Statistical Yearbook, NSI, 2000, 2001 and The role of SMEs in regional employment The contribution of SMEs to employment is relatively more balanced between planning regions than the numbers of SMEs (Figure 6). SMEs play a significant role for employment in all regions since they account for over 50% of the total in all six cases. The North-East and the South-East regions have the highest share (61.3%) and (60.6%), respectively. In the other four planning regions the shares varies between 59% in South-Central to 55% in South-West, with North-Central and North- West in between (56% in both cases).. 18

367 Figure 6 Share of SME in total employment by planning region Total North-West North- Central North-East South-East South- Central South-West Interestingly, there is not a very close association between the share of SMEs in employment and the overall employment rate, which possibly suggests that the presence of SMEs in a region tends to boost employment in all parts of the economy, or, alternatively, that the development of SMEs is boosted by a relatively high level of employment, which is likely to reflect a high level of economic activity. In the case of micro-sized firms, the highest employment growth rate over the period was registered in the South-West region (an increase of 7%). Only in the South - Central region was there a decline in employment in micro-sized firms and then only marginally. In the case of small firms with 11 to 50 people employed, employment increased in all regions. The North-West region had the largest increase (of 12%), and North-Central, the smallest (5%). Employment in firms with 51 to 100 employees increased on average by 4%, the largest rise being registered in the South-West region (of 8%), while employment declined by 2% in South-Central. The data on changes in employment by company size by regions show that small firms were the main job suppliers across regions, followed by firms with employees and those with 10 or fewer people employed. Small firms contributed 33% of total employment and 38% of value-added in the country as a whole, the largest effect being in the South-West, South-Central and South-East regions. 19

368 6 State policy for supporting SME development The present state policy for encouraging and supporting SME growth combines two complementary approaches: 14 maintaining a favourable business environment specific support for starting and developing SMEs The first approach includes multiple activities relating to the legal, institutional, financial, technological and entrepreneurial environment for SMEs. In this regard, the main measures included in state policy consist of a reduction in trade barriers, privatization, curbing monopoly power, improvements in the taxation regime and the development of an entrepreneurial culture. The second approach is aimed at developing various programmes providing support for business start-ups, finance and investment, company growth, quality management, export activities, training and personnel development, which are implemented through the funds and programmes of various ministries. It includes a package of employment, social insurance, regional, tax and social policy incentives for SMEs. The legal basis of SMEs has been subject to change over the last few years. In 1999, Bulgaria adopted the Law on Small and Medium-sized Enterprises (LSME). This was amended in 2001 through the enactment of the Law on Craft Trades. As a result, the scope of application of the LSME was widened, since those involved in craft activities were recognized as enterprises. The Law of Crafts establishes the principle of obligatory membership of regional crafts chambers. The law provides for the setting up of a National Crafts Chamber, which implements the overall policy for protecting the interests of craft trade people and their organizations. The National Crafts Chamber participates in the development of projects and prepares opinions on regulatory legislation relating to crafts and relevant training programmes; establish professional rules; organize training and coordinate the activities of the regional crafts chambers. The Law for Privatization and Post-Privatization Control provides specific rules for the participation of SMEs in the privatization of state- and municipality-owned enterprises. 14 The recent Governmental policy is based on the National Plan for Economic Development and a number of laws: The Small and Medium-sized Enterprises Act (SMEA), amended in 2001; The Crafts Act, in force since May 2001; The Employment Promotion Act (in force since January 2002; The Privatization and Post-privatization Control Act; on The National Strategy for Promotion of SMEs ( ) adopted in 2001 and by the Action Plan for its implementation and on a number of rules, decrees, etc. regulating the implementation of policy, its effectiveness and control. 20

369 The Employment Promotion Act (in force since 1 January 2002) assigns to the Employment Agency responsibility for providing training to employees of SMEs to enable them to acquire or upgrade their professional qualifications. The first National Strategy for Promotion of SMEs covered the period This was followed by a second programme for the period along the same lines, further developing the direction, the main goals and priorities of the strategy. The Action Plans present in more detail the concrete activities as well as specifying the time periods for providing support. The institutional framework for SMEs in recent years has also been subject to development and improvement. According to the LSME, the Agency for SMEs (ASME) is in charge of the implementation and co-ordination of state policy towards these firms. An advisory Council works under the President of the ASME with the purpose of ensuring and guaranteeing effective communication between the state authorities, businesses and nongovernmental organizations. The ASME provides a wide range of services to SMEs, starting with voluntary registration; information on the procedures for obtaining licences and permits and meeting other requirements, including through its internet site. Simplifying administrative and regulatory requirements for SMEs is one of the priorities of state policy in this area. The measures undertaken in this area have included the review and optimization of the licensing, permit and registration systems. In 2002, a specially established inter-ministerial group for reviewing these systems proposed the repeal of 73 systems out of 361 and the simplification of another 118. In April 2000, as part of the negotiations for Bulgaria s accession to the EU, negotiations on Chapter 16 Small and Medium sized Enterprises was conditionally concluded, one of the first chapters to be so. As part of active labour market policy, the Ministry of Labour and Social Policy operates a wide range of programmes and measures targeted at helping the unemployed people find employment by assisting them to start their own business or encouraging employers to take them on. There are both national and regional programmes, as well as programmes supported by international organizations National programmes According to the Employment Promotion Act (EPA), any unemployed person willing to start a business, alone or with a partner, to produce goods or services, is entitled to financial 21

370 support once their business plan has been approved by the regional department of the National Employment Service (NES). If the business plan provides employment for another member of the family, the person concerned is also entitled to financial support. The scheme also provides for vocational training and free advice on how to write a business plan, specialized literature on small company management and so on. Lump-sum unemployment benefits 15 are available in the event of liquidation or financial rehabilitation of commercial companies (a lump-sum financial compensation plus BGN 1,000 (around EUR 510) in case of starting own business) and layoffs by state limitedliability companies in coal mining (BGN 2000 plus BGN 1000 in the case of starting up a business). Regional programmes The Guarantee Fund for Micro-Crediting project started at the end of 2001 in 9 regions and covered 18 pilot municipalities with high unemployment. The initial funding came from the state budget and amounted to 20 million BGN. The project seeks to create new jobs through facilitating access to funds people willing to establish their own business and to free resources for business development and expansion. In 2002, 1473 were granted credits amounting to BGN 14 million, or an average of BGN (about EUR 5 100) per credit, and new jobs were created. In 2003, were granted credits amounting to BGN 27 million, and the number of new jobs created was The number of people, who were consulted was The distribution of credits by economic sector was as follows: agriculture, 34%, trade, 29%, services, 18.8%, industry, 16% and tourism, 1.3%. 16 There are numerous regional programmes providing encouragement, support, and training to the unemployed to encourage them to start their own business, as well as creating relevant supporting structures (Business Incubators). They include: Start in business and the Small business development micro-project (Sofia region); Successful business (Kjustendil region); Start-up Business Management (Russe region); Start your own business (Gabrovo region); Business Incubator (Montana region, ); Selfemployment my choice (Veliko Tarnovo region); and Development of small and medium-sized farms (Sliven region). Programmes, supported by international organizations The Bulgarian Ministry of Labour and Social Policy and National Employment Service have co-operated with international institutions to create structures supporting SMEs, including: Council of Ministers Decree N100/ ; N141/2000; N100/1998 Agency for SMEs 22

371 business centres; business incubators; regional development centres; informal local development structures, local strategies and programmes for economic development. A good example is the Job Opportunities through Business Support (JOBS) project, which is planned to open 24 agribusiness centres in 24 towns and cities in 2001, permanent and temporary jobs were created and people were trained under this project. The level of taxation on labour can have a negative impact on job creation and can depress demand for labour in the formal economy while encouraging the growth of the shadow economy. Since the beginning of the transition period there has been a significant problem in the form of many enterprises, especially SMEs, avoiding paying social insurance taxes of all types, thus keeping down labour costs. In January 2000, the Government introduced an Obligatory Social Insurance Code (OSIC) aimed at improving the acceptance of entrepreneurs and self-employed of social insurance. An attempt was also made to ensure better coverage of employees by re-registering labour contracts. Thus in 2003 about employees emerged from the grey economy into legitimate employment. As part of the ongoing reform in the health system, the health insurance system has also been revised recently. As a result more and more people, employed in SMEs are being covered by the social insurance system. Social insurance contributions (total rate 42.7%) are considered by private sector employers as a barrier to business. 17 In 2000 and 2001, the ratio of employers contributions to employees was 80:20. In 2002 the ratio was changed to 75:25 and the Government s objective is to reduce social contributions further and progressively to achieve a 50:50 split between employers and employees by In 2002 with the amendments to OSIC 18, a number of changes were made aimed at further reducing the insurance burden on employers/contractors by both reducing overall payments and changing the employer/employee ratio; changing the minimum and maximum thresholds of insured income; providing opportunities to count additional insured years of service upon retirement, which is regarded as an additional incentive for people to pay social insurance. Some simplifications have been introduced for the self-insured with regard to providing data for the Pension Registers. SMEs still claim to have some problems relating to social insurance, in particular, the frequent introduction of new requirements concerning the terms of payment of contributions; frequently changing and unclear legislation and so on Capital Weekly, March, 2002, a Survey carried by Alfa Research Renamed in 2003 to Public Social Insurance Code. 23

372 Development of SMEs is a key element of regional policy. 19 The Action Plan for 2001 and the National Programme for Regional Development (NPRD) included an independent programme for Small and Medium-sized Enterprises. The main objectives of the programme are to support the development of SMEs and to create a favourable financing and competitive environment for entrepreneurial development to make best use of local resources. The understanding of what constitutes a more favourable business environment includes reducing administrative barriers (introduction of one-stop-shop), diminishing corruption and a stable legal framework. State policy towards improvement of the business infrastructure includes provision of information, promotion of SME business abroad, assisting SMEs to expand capacity and supporting the development of business incubators. Specific measures include: Action plan 2001 covering 18 regional projects with a budget of BGN million. The distribution of the projects across regions (South-West - 1; North-West and North- Central - 2; South-East - 3; North-East - 4 and South-Central - 6) is related to the capacity for the development of SMEs, including making optimum use of local resources for economic recovery, as well as improving the effectiveness and competitiveness of SMEs. The Business Infrastructure programme including 29 infrastructure projects and a budget of BGN million. Projects are divided among regions as follows: (South- West - 9; North-West -3; North-Central - 1; South-East - 5; North-East - 3 and South- Central - 8). The projects consist of the construction of new business premises and the reconstruction of existing ones and the creation of a business environment according with European standards and international requirements and are aimed at reducing unemployment through the provision of new jobs. Technological Parks Programme, comprising the establishment of a business park in Montana, a high-tech park in Gabrovo and the Iskar-Sever Technological Park in Sofia. The total budget was BGN million. Projects implemented under the Phare Economic and Social Cohesion 2000 Programmes in the North-West and South-Central planning regions, including Introduction of Quality Management Systems, Investment in Business Incubators in Regions in Industrial Decline, High-Tech Business incubators, and Services to Small and Medium sized Business and Technology Grants. A Regional Innovative Strategy for the South-Central region. 19 The part is based on the programs and projects, pointed in the Report on small and medium-sized enterprises

373 Availability of financing has been recognized as one of the critical factors for successful development of SMEs. 20 Access to financing was ranked as the second most important obstacle to SME growth after insufficient demand. As a result, recent policy in this area has focused on: establishing and developing the general environment for financing enterprises, including SMEs (indirect element) ensuring direct financial assistance to SMEs (direct element) The indirect element is aimed at improving financial conditions, especially tackling high interest rates, the requirement of collateral exceeding 150% of the amount of the loan; the long period for processing loan applications and the lack of alternative sources for financing of SMEs. A 2001 survey of the business environment for SME development indicated that the main sources of funding for starting up business were personnel savings (79.5%); bank loans (19%); leasing (3%) and personnel loans (1%). According to the same survey, around 23% of SMEs had not relied on any external financial support at all. In , the funding sources and instruments for the direct financing of SMEs expanded and improved substantially due to the greater availability of bank finance and the more stable banking system. The main sources for financing SMEs were bank loans, leasing, equities (or investment funds), guaranteed loans and grants. There has been increasing direct financial support for the development of SMEs, especially from a substantial growth in foreign financial assistance to banks and SMEs themselves. (In 1999, foreign sources of financial support for SMEs amounted to US$ 16.2 million and increased to US$ 23.4 million in 2000.) A number of banks have specialized in projects and programmes for financing the SMEs. In , the Encouragement Bank granted 100 loans amounting to BGN million, including 29 for long-term investment, 7 for export financing, 37 for investment financing; 24 for micro-investment and 2 for improving competitiveness; 21 DSK Bank retained its leading position in consumer credit and lending to micro-sized and small enterprises, as well as to specialized markets (e.g. general practitioners in health services). For the period , the bank made loans amounting to BGN The Report on small and medium-sized enterprises refers to several national and international studies, carried out in , outlining the importance of financing to SME development. Report on small and medium-sized enterprises , Sofia, p

374 Leasing also became quite common over the period its main advantage over bank financing being that no collateral is required especially for purchasing vehicles, office equipment and industrial and agricultural machinery Guarantee schemes of the Bulgarian Government for micro-loans were set up at the end of 2001 as a follow-up to the Regional Initiatives Fund that provided finance of BGN 28.6 million for 201 micro-projects for infrastructure improvement over the period The scheme provided a further BGN 20 million for Its main aim was to create jobs in regions most seriously affected by unemployment. Under the scheme, loans were provided for both investment (of up to BGN in each case) and working capital (of up to BGN ). Support of EU Programmes for SME development in Bulgaria Under the Third EU Multiannual SMEs Programme, a network of 8 EURO Info Centres were created to provide information and consultancy to SMEs so as to widen their opportunities for participation in EU programmes. Bulgarian participate in the Second Horizontal Programme Promotion of Innovation and Support to SMEs under the EU Fifth Framework RDT Programme. The country was successful in 31% of craft project proposals submitted for funding and qualified for another 4 research projects for the period April 1999-April A number of investment projects were also included in the EU-Phare Programme for supporting Bulgarian SMEs, such as Grant scheme for adoption of system for quality management, Investment in business incubators for SMEs in regions of industrial decline, Services for SMEs and technology grant scheme and High technology business incubators. 7 Conclusions The development of SMEs in Bulgaria over recent years has been quite strong, in part because of government policy to maintain macroeconomic stability and to support SMEs by improving the legal and institutional framework, by better coordination of the specific measures to promote SMEs, by labour market and regional policies and by carefully targeted funding sources. SMEs have increased in number and their contribution to employment growth has become significant. Their contribution to value-added has also increased as they have become more competitive in the new business environment. SMEs in Bulgaria have had a relatively high survival rate in the past few years. In 2002, 56.5% of those managing SMEs declared an intention to increase output and 55.4% to 26

375 increase profit. These intentions reflect current policy of supporting and encouraging the start-up of businesses by easing access to finance, and by providing guidance, advice, education and training to employers and employees in SMEs. Growth of employment in SMEs occurred in Wholesale and retail trades, Hotels and restaurants, Business activities and real estate and Other community and personal services, in particular, though also in Manufacturing and Construction. There are examples of SMEs development in both traditional and innovative sectors and activities. However, the development of SMEs in Bulgaria depends greatly on foreign direct investment, which is usually directed towards more traditional activities, such as food, drink and tobacco, textiles, the wood industry and non-metallic mineral production. The relationships of Bulgarian SMEs with foreign partners which are based on high technologies and protected by patents, prestigious trademarks or modern technology are still relatively limited. The innovative activities of these firms ere not easily transferable to the national economy at present because of the low level of technology which typically applies. The development of SMEs requires an increasingly qualified labour force, though much of the growth in employment which has occurred in recent years has been among those with relatively low levels of educational attainment. Current government policy towards SMEs consists of two main elements: maintaining a favourable business environment and providing specific support for starting-up and developing SMEs. Both elements are important. The first includes multiple activities relating the legal, institutional, financial, technological and entrepreneurial environment in which firms operate. The main measures of policy in this regard include a reduction in trade and administrative barriers, privatization and reducing monopoly position, improvements in the tax regime and the development of an entrepreneurial culture as well as of information systems. The second element of policy is aimed at developing various programmes providing support for business start-ups, finance and investment, company growth, quality management, export activities, training and personnel development. A number of different ministries are involved in this. Measures include a package of incentives in respect of employment, social insurance, taxation, regional development and social policy. Many problem areas and specific obstacles remain, but the success of the development of SMEs in Bulgaria stems to a significant extent from an effective policy approach and a widespread understanding of the importance of SMEs for economic prosperity. 27

376 References Report on small and medium-sized enterprises (2002), Agency for small and medium-sized enterprises, Sofia. Report on small and medium-sized enterprises (2000), Agency for small and medium-sized enterprises, Sofia. Bulgaria. A Changing Poverty Profile, Poverty Assessment (2002), The World Bank, p. 81. Competitiveness of the Bulgarian economy, Annual report (2000), Centre for economic development, December, pp (in Bulgarian). Conditions for the development of micro firms in Bulgaria (2000), National Statistical Institute (in Bulgarian). Small and medium-size enterprises in the Republic of Bulgaria (2004), National Statistical Institute (in Bulgarian). 28

377 Appendix Table A1 Bulgaria: number of SMEs by sectors - B16 Sector Total Agriculture, forestry Mining and quarrying Manufacturing Electricity Construction Wholesale and retail trade Hotels and restaurants Transport & Communication Real estate, renting Education Health Other community Source: National Statistical Institute, Statistical Reference Book for the relevant years. Table A2 Bulgaria: employment by economic sectors - B16 (thousands) Sector Total Agriculture, forestry Mining and quarrying Manufacturing Electricity Construction Wholesale and retail trade Hotels and restaurants Transport & Communication Real estate Education Health Other community Source: National Statistical Institute, Statistical Reference Book for the relevant years. 29

378 Table A3 Bulgaria: SMEs employment by economic sectors - B16 (thousands) Sector Total Agriculture, forestry Mining and quarrying Manufacturing Electricity Construction Wholesale and retail trade Hotels and restaurants Transport & Communication Real estate Education Health Other community Source: National Statistical Institute, Statistical Reference Book for the relevant years. 30

379 Economic Restructuring and Labour Markets in the Accession Countries Research Project commissioned by EU DG Employment, Social Affairs and Equal Opportunities, Contract No. VC/2003/0367 Coordinated by in cooperation with Alphametrics and The Vienna Institute for International Economic Studies Developments in education and training in the Czech Republic directed by Filip Zeman August 2004 *) EuroProfis Ltd., Prague

380 Contents Introduction...1 A Participation of young people in education and training...2 B Changes in recent years...8 C The adaptation of education and training system to the new requirements on the labour market...10 D Develompents in continuing training...16 E Government policy towards education and training...18

381 Filip Zeman Developments in education and training in the Czech Republic Introduction During the early 1990s, the Czech Republic succeeded in maintaining a relatively stable macro-economic climate. Between 1994 and 2000, the activity rate declined slightly, especially for women, from 52.6% to 51.6%. There was a much more significant decline in the rate for those under 20 (from 35% to 15%) partly as a result of them staying longer in school and high participation rates in education. In 2000, the participation rate in education for people aged 16 was over 95%, the highest among the new Member States. Resources for education and training are limited. Public expenditure on education in the Czech Republic was only slightly above 4% of GDP during the period Expenditures of the state budget for education were about 3.75% of GDP in 2002, which is 11.5% of the state budget (without other public expenditure). 6 Expenditure on education (in % on GDP) in % EU 15 Czech Republic Spending on education as a proportion of total public expenditure declined from 10.5% in 1994 to 8.9% in As a consequence of public administration reform, however, the funding mechanism was modified and finance for initial vocational education is now channelled through the new regional authorities, which means that expenditures spent on education can implicitly vary in each region. Vocational education and training has always been an important part of the education system in the Czech Republic given the high proportion of those completing basic education going on to enter upper secondary education and vocational training (about 81.5% in 2000), among the highest in Europe. 1 Eurostat 1

382 Over the past decade there has been a shift towards educational programmes leading to higher qualifications. The reform of post-compulsory education and initial vocational training was mainly determined by a bottom-up approach, encouraged by the liberalization of regulations in the early 1990s. As part of the reform of public administration and the establishment of regional self-governing bodies, the responsibility for upper secondary education was transferred to the latter. Expenditure on different levels of education (in %, 2000) 100% 80% 60% 40% 20% 0% Primary Secondary Tertiary 2

383 A Participation of young people in education and training Student participation The rate of participation in education in the Czech Republic is high among young people of between 15 and 19 years of age. In 2000, about 90% of the population of 16- to 17-yearolds, 58% of 18-year-olds and 41% of 19-year-olds participated in education. These rates have tended to increase slightly in the younger age groups and more rapidly in the older ones among those aged The participation rate, therefore, rose by 27.6 percentage points among 18-year-olds and by 14 percentage points among 19-year-olds between 1990 and The main reasons for these increases include the introduction of an extra year of basic schooling (to 9 years) and the higher numbers in longer education and training programmes (four years, leading to the Maturita leaving school examination and report). The rise in participation of 19-year-olds in education also reflects the growing numbers going on to university (or into tertiary education), though as noted below, the proportion still remains low. Tertiary education in the Czech Republic ( ) Tertiary education - TOTAL Tertiary programmes with academic orientation Tertiary programmes with occupation orientation Advanced tertiary education Source: Eurostat In terms of international comparisons, the Czech Republic scores high in participation of young people in upper-secondary education, exceeding the EU average. However, as the average duration of upper secondary studies is shorter than the average elsewhere in the EU and young people tend to enter the labour market earlier, at the age of 19, participation in education falls below the EU average. Participation falls even further below the EU average for those aged 20 and above, as the Czech Republic lags behind EU15 countries in the proportion of young people in tertiary education. In the Czech Republic, only 11% of young people aged are enrolled in tertiary education while in the EU15, the figure is 18% MoEYS Eurostat 3

384 Drop-outs There are no specific data available in the Czech Republic monitoring the progress of students through the education system. Neither are there statistics on students who leave school before properly completing their studies. Nevertheless, the proportion dropping out of the education system before obtaining qualifications, as indicated by data on education attainment levels, is relatively low. It is estimated that some 4% 5% of young people students leave basic compulsory education before the ninth year, so failing to complete basic education and being unable to continue to further education..such people entering the labour market without qualifications face serious problems. Proportion of general education in study programmes Problems can also be significant for those who drop out of upper secondary vocational schools. The extent of the problem may be inferred from the fact that there are 7% 8% of people with only basic education in the age group, which is significantly below the EU average. Drop-outs from gymnasiums and secondary technical schools do not normally constitute a departure from the education system. Students who fail in this type of school tend to transfer to another, easier type of education programme i.e. gymnasium students usually move to secondary technical schools and secondary technical students to secondary vocational schools. There is a higher drop-out rate from tertiary education, from higher professional schools and universities (though precise figures are not available). Although failure to complete studies at this level is less serious than at lower levels, it may 4

385 still cause problems for students graduating from gymnasiums without the professional qualifications which employers may require. Young men are more likely to drop out than young women, particularly from programmes leading to the Maturita and in tertiary education. On the other hand, drop-out among young women at secondary vocational schools and schools without Maturita are higher than for young men. Although the education system provides opportunities to complete basic and secondary vocational and technical education, which most young people benefit from, a number still have problems There has been some criticism of the rigid education programmes which make it impossible for students to interrupt their studies when undertaking individual study units and which can deter these people from returning to school if they have to leave, since it would mean them having to start virtually all over again. The education system, moreover, lacks embedded mechanisms for the early identification of students at risk and is not sufficiently inclusive to provide individualized assistance and reintegration into studies. The work of educational counsellors at basic and secondary schools is held to be formal and inefficient in this respect. 4 Access to higher levels of education and employment The majority of young people in the Czech Republic start their working life after finishing upper secondary school and only a small proportion continue their studies further. After completing their upper secondary education with the Maturita, students have two possible ways of entering tertiary education: they go into either higher professional schools or universities. The participation rate in tertiary education, on the latest data available (for 2001), is around 22% of those aged 19, 28% of those aged 20(EU15 average around 32%), 26% of those aged 21 (EU15 average 30%) and 21% of 22-year olds (EU15 average 25%) National Observatory of the Czech Republic, VET in the CR, 2001 National Observatory of the Czech Republic, VET in the CR, Eurostat figures 5

386 Proportion of students of 1st year of secondary schools 50 % Gymnasium STS with "maturita" STS without "maturita" SVS (apprenticeship) SVS with "maturita" Source: Vývoj vzdělanostní a oborové struktury žáků ve střením a vyšším vzdělávání, Vojtěch Jiří, NÚOV, Praha, 2002/3 The enrolment at secondary schools has to be expressed only in %, as there were demographic changes during the second half of 1990 s, which would affect total figures 67. Some 31-32% 8 of those graduating from secondary schools go on to tertiary education. Of these, about 24% enrolled in universities and 9% in higher professional schools. However, it can be inferred from analysis of admission procedures that many students, particularly those from gymnasiums, enrol in tertiary institutions some time after completing their secondary education. Students from gymnasium tend to be much more successful in gaining admission to university than those from secondary technical schools who are more likely to go to higher professional schools Development of educational and field structure of secondary students in Czech republic and regions, National VET Institute, 2002/3 SVS = apprenticeship National Observatory of the Czech Republic, VET in the CR, 2001 Figures only for 1999: There were 53% gymnasium graduates, 33% from secondary technical schools and 14% from secondary vocational schools (including follow-up courses) of students enrolled in universities. At higher professional schools, secondary technical graduates prevailed (63%), with a much smaller proportion of gymnasium graduates (28%) and of secondary vocational graduates (9%). 6

387 Development in number of students in upper secondary education 7

388 The limited capacity of tertiary education in relation to demand remains a problem despite some lessening of pressure after the introduction of higher professional schools. The number wishing to enter higher education is steadily growing, and although universities are also expanding their capacity, it is not keeping pace with the rate of applications. Even in 2000 which was not typical, since as a result of an increase in the number of years of compulsory schooling, a whole year of students completing secondary education was missing 10, reducing the number of applicants significantly universities rejected around half of applicants on average, though the proportion varied markedly across fields of study. Students at universities tertiary education Tertiary education in the Czech Republic by field ( ) Unknown or not specified Services Health and social services Agriculture Architecture and building Manufacturing and processing Engineering and engineering trades Computing Mathematics and statistics Physical sciences Life sciences Law Business and administration Journalism and information Social and behavioural sciences Humanities and arts Educational sciences It is evident from the above that some two-thirds of those completing upper secondary education enter the labour market immediately after doing so, though some of them only temporarily because they simply postpone seeking admission to the tertiary sector. It can be estimated very approximately that those entering the labour market consist of some Four years ago a nine-year basic school system was launched so the admission of students from basic school to secondary education was postponed a year. 8

389 20%of all those completing gymnasium education, 85-90% of those graduating from vocational school without Maturita and 70-80% of those completing vocational-school courses with Maturita. Those entering the labour market face varying sized problems of finding a job depending on their field of study, the labour market situation in the place they live and their personal qualities. Tertiary education by field and participation of men and women (2001) M F TOTAL Educational sciences Humanities and arts Social and behav. sciences Journalism and information Business and administration Law Life sciences Physical sciences Mathematics and statistics Computing Engineering Manufacturing and processing Architecture and building Agriculture Health and social services Services Unknown or not specified Source: Eurostat 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% B Changes in recent years Demographic changes have a continuing effect on employment developments and rates of participation in education. The slowdown in population growth over recent years, especially among young people, is a common trend in many CEECs (except Slovakia and Slovenia) and it has affected the numbers in education and entering the labour market. The change in the population aged 15 to 24 varies significantly across countries (see graph). The number in this age group in the Czech Republic has declined markedly in recent years, the extent of the fall between 1999 and 2003 being more than in any other EU country. 9

390 % Changes in population aged ,0 5,0 0,0-5,0-10,0-15,0 Czech Republic Belgium Germany Netherlands Hungary Poland Slovakia EU15 EU C The adaptation of the education and training system to the new labour market requirements Curricular policy In terms of its main features, curricula in the Czech Republic are still largely determined centrally, since all documents, with the exception of textbooks, are subject to approval by MoEYS. On the other hand, the Czech Republic is one of the countries where, particularly in the first half of the 1990s, strong liberalizing trends occurred which resulted in a relatively large degree of autonomy of schools. This autonomy has given rise to a certain lack of transparency in vocational training in that a great number of educational programmes were approved without due attention being paid to their quality. This situation has not as yet been fully addressed, although a Standard of Secondary Vocational Education was adopted in 1998 which sets out a framework for MoEYS requirements concerning the aims and content of VET curricula. The Standard of Secondary Vocational Education, originally developed as part of the Phare VET Reform project, envisaged the implementation of a bipolar curricular policy. This means that schools themselves were intended to draw up their own educational programmes based on framework requirements defined in the standard. Although this model was tested in pilot schools and the standard was then adopted as a binding document for the entire VET system, the authority to approve educational programmes prepared by schools has remained with central government. The reason for this can be seen not only in the traditionally centralized education policy, but also in the lack of preparedness of school directors and teachers in vocational schools to take on new tasks without sufficient notice. 10

391 At present, efforts are being made to introduce a new balance in the distribution of responsibilities between ministry, regions and schools in the context of the new administrative set-up and, consequently, changes in the management of education. The draft of the Long-Term Development Plan is aimed at making changes in curricular policy, with the ministry determining framework objectives and the nature of vocational education, as well as Framework Educational Programmes (FEPs). The FEPs will form the basis of school-based curricula (a second part of the curricular document), which will be developed by individual schools taking account of the needs of the regional labour market. In order to facilitate the transition of young people from school to employment, practical training needs to be part of education programmes. While in secondary vocational schools the proportion of practical training undertaken as part of vocational programmes is relatively large (although it often takes place only in school and, therefore, may pose problems of quality), practical training (in the workplace) is negligible in programmes leading to the Maturita. This is particularly true of business and engineering programmes (agriculture and health care being exceptions, with a traditionally high proportion of practical training). The reason is related to the organization of placements and practical training in companies. There is often a lack of interest on the part of employers, who therefore, offer limited opportunities for workplace training. Accordingly, the important question remains of how to motivate companies to take part in vocational education. In consequence, even vocational training without Maturita is of a school-based nature without employers being involved. One area in which curricula have been modernized and improved is in foreign language teaching (see Graph), although the quality of tuition remains low. In line with the requirements of lifelong learning, the Standard of Secondary Vocational Education specifies an obligation to introduce a foreign language into all training programmes. The introduction of a foreign language as an obligatory subject in some training programmes without Maturita, however, is still under discussion, since the achievement of the students concerned has not been satisfactory. Although the implementation of IT courses in curricula in secondary schools has risen (see Graph, which shows the average proportion containing IT courses together with the trend), problems still remain as regards the use of new information technology. It is disturbing that some students still do not have access to computers in their secondary vocational education. It is estimated that on average only 68% of new programmes developed after 1990 include use of computer technology.. 11

392 Students of foreign languages The development of IT courses in curricula Improvements, however, can be expected as a result of a massive ICT support scheme in education which has been launched by MoEYS and for which the ministry has earmarked to receive substantial financial resources (e.g. more than CZK 4.4 billion EUR 140 million for the project Internet into schools which consists mainly of investment in hardware and software for thousands of schools but also training). Curricula still mostly take the form of, linear programmes lasting two, three or four years. Although a modular structure of teaching was tested as part of the Phare VET Reform experiment, the widespread introduction of modules is still to occur. Part of the reason is that the advantages of this kind of structure are limited by the persistent efforts of MoEYS to modify the modular system to comply with existing legislation. (Under the Educational Code, curricula are determined centrally and only a limited part of curriculum for each programme up to 10% - can be modified by the school itself). This therefore, significantly 12

393 reduces the benefits of adopting a modular system and so only isolated programmes of this kind are being introduced. The standard vocational curriculum applying at present does not allow for a modular approach. Even greater rigidity is introduced by the fact that for the most part it is not possible for students to modify their study programmes. While a limited number of optional subjects can be taken, most of what they are taught is fixed from the beginning to the end of their studies. The Czech Republic lacks, in addition, a well-developed evaluation system and, consequently, there is no assessment of the quality of education programmes and the outcome of vocational education. Evaluation in general (and this also applies to vocational education) is limited to achievements in general subjects. Evaluation of the professional and vocational abilities of students is carried out by individual schools with little coordination of this evaluation. The absence of the social partners (most importantly, representatives of employers) in the evaluation process makes it difficult if not impossible to assess the relevance of vocational education to the needs of employers. Public expenditure Share of GDP on public expenditure in education (Czech Republic) 11 Public expenditure by type of school 11 Modernization of VET in transition countries National Report Czech Republic, Czech National Observatory, Prague,

394 Labour market developments challenges for the education system 12 The labour market in the Czech Republic has been undergoing rapid change linked to the restructuring of the economy as a whole and individual companies within it, to technological advance and demographic trends. This has led to considerable changes in requirements in relation to qualifications, professional and vocational skills and the ways of acquiring them. Shifts of the labour force between companies and activities The maintenance of a high level of employment is a major objective of policy and will continue to be so. It is likely, however, to be increasingly difficult for people with low qualifications to find jobs. At the same time, new investors are reporting a shortage of qualified workers with the skills they are looking for. Foreign companies are beginning to change their focus from cheap labour to workers with high education levels, particularly university degrees. There is a need, therefore, to introduce measures to prepare employees for occupational change, which entails the development of career guidance services and an increase in participation in continuing education and training. The use of high technology A major challenge associated with labour market developments consists of keeping pace with current trends in globalization and of improving labour force skills. The proportion of sectors of activity with a high skill intensity is progressively growing in the Czech Republic. Exports of electrical engineering, electronics and telecommunication products are rising and employment in services associated with the use of modern technologies is growing, as are business services generally. In addition to the necessary expansion of initial training courses, there is a need for the systematic development of continuing training and re-training to enable people to acquire new skills and qualifications. Increase in employment in small and medium-sized companies Small and medium-sized enterprises (SMEs), which already employ almost 60% of the workforce in the economy, have in many cases specific skill requirements for the workers they take on. At the same time, SMEs tend to have a lower level of human resource development only half the proportion of small companies with up to 20 staff provide training for their employees as of large companies. 12 Modernization of VET in transition countries National Report Czech Republic, Czech National Observatory, Prague,

395 There is, therefore, a need to improve the provision of training by adopting new methods, particularly distance learning, and to adjust courses to the specific needs of SMEs. Employers in SMEs need also be encouraged to improve their approach to human resource development and to develop information channels and networks for the exchange of experience and good practice. Improving and extending entrepreneurial skills Czech entrepreneurs often lack management skills. It is important, therefore, to support those starting new businesses through the provision of relevant counselling and advice as well as access to the -training required, which itself needs to be improved in terms of quality. Growing competition in the labour market Competition in the labour market is increasing and the position of those with disadvantages (particularly those with low qualifications, those with disabilities, members of the Roma community and women with small children) is worsening. There is a need to adjust the forms of continuing training provided to meet the differing needs of the various groups among these and to give stronger incentives for companies to pay more attention to lowskilled employees and to facilitate their participation in training. The position of early school leavers in the labour market The situation of early school leavers is more difficult than that of employees with work experience. Although the number of young people who leave school with only basic education is not large, these individuals have little chance of finding a job and still less one of reasonable quality. Training provided by labour offices for such people does not seem to address this problem, since those participating in the courses which exist do not obtain a certified qualification on completion. The need is to minimize the drop-out rate from schools and to provide more opportunities for those who do to re-enter the education system. The recognition of qualifications obtained outside the formal education system is also important in this respect. Differences in the development of regional labour markets The disparities between regional labour markets caused by their uneven development potential are widening. There is an increasingly important role for regional development strategies and the involvement of key actors in policy on employment and education. 15

396 The ageing of the work force The proportion of older people in the population is growing. Existing training opportunities only partially meet the needs of adult learners and it is necessary to develop innovative approaches (in terms of methods, content and demands on time) which would better suit these. Participation of adults in education and training is also conditional on better information and guidance being provided and on the development of an appropriate system of certification of knowledge and skills, including those acquired informally. A number of policy proposals have been made on the development of initial and continuing vocational education and training (IVET and CVET) strategies to meet the needs of the labour market and society as a whole. These proposals have not as yet been passed into legislation, nor have they been incorporated into a system of specific measures for implementation. As a consequence, IVET reform is based on ad hoc measures, which may be consistent with strategic plans, but which are not sufficiently coordinated either between the various ministries responsible (particularly the MoEYS and MoLSA), or between different levels of government. The implementation of coherent policies is further complicated by the reform of public administration and the devolution of responsibility for important parts of IVET to regional authorities. The way that strategic development plans are to be coordinated and how differences in plans between regions are to be reconciled is not yet clear. D Developments in continuing training Various surveys among training providers suggest that, in general, the supply of vocational training courses on offer substantially exceeds the demand for these across the country. However, the problem lies not so much with excessive supply as will deficient demand, which tends to be depressed by low real income levels, poor motivation for training, the limited time available to pursue courses and so on. In this respect, the Czech Republic still has some way to go to catch up with the EU15 countries considered as a whole, as indicated by the lower participation of active population in CVET as compared with the EU15 average. CVET programmes vary widely, in terms of both content and duration. The longest programmes lead to the acquisition of a recognized educational qualification and may be the same duration as full-time programmes for young people, or perhaps one year longer (i.e. 3-5 years at secondary schools, most frequently 3 years at higher professional schools and normally 5 years at universities). The content of these programmes is analogous to that of vocational or technical programmes in secondary schools, higher professional schools and universities. 16

397 Number of secondary schools and schools providing part-time courses for adults CVET programmes for job seekers (i.e. retraining courses) are also diverse in terms of content. Surveys into this type of CVET indicate that the content of courses is tending to expand, focusing on retraining for manual vocations as well as technical and administrative ones. In the past five years, the number of vocations towards which this type of retraining is directed has more than doubled to almost 300, while the length of the most frequently provided retraining courses ranges from 600 to 4,000 hours. According to the CVTS survey carried out in 1999, around half of companies in the Czech Republic offer continuing training to employees in their own internal training units, while the other half use external providers. The most frequently used providers of staff training are private organizations, while secondary schools and universities rank among the least used (accounting for only 1.4% and 3.7% of training, respectively). The main players in the training of those in employment are the employers for which people work and these largely determine the extent of access of people to continuing training and the type of training they receive. The links between employers and the education system are, however, very weak. There are means in place for ensuring the quality of CVET only for some parts of the system. The quality of CVET provided by schools, leading to the acquisition of a particular educational qualification, is monitored (as in the case of IVET) by the Czech School Inspectorate. A Commission for Accreditation of Retraining Programmes was set up at MoEYS as early as the first half of the 1990s in order to maintain the quality of retraining courses,. When organizing retraining for job seekers, labour offices give preference to programmes which have been accredited. There are, however, no universally valid means for assuring the quality of in-service training in companies, though there are requirements for the content and quality of CVET programmes in certain professions (physicians, teachers, accountants) fixed by the institution concerned (ministry, association). 17

398 Structure of re-qualification course providers accredited by the Ministry of Education, Youth and Sport Innovative educational methods (such as distance learning and modular systems) have also been applied to CVET, but only to a very limited extent. Traditional teaching methods predominate, although surveys show a greater use of computers in CVET in companies than in schools or colleges. Curricular development in CVET In most areas, curricula have been developed specifically for CVET. These specific curricula are used in companies, in retraining courses for job seekers, in compulsory CVET for certain professions and in adult education courses. Most CVET providers, however, develop their own individual curricula and since there are no educational standards to meet, these do not need to conform with any particular requirements. Nevertheless, CVET providers in the public sector have to use curricula approved by MoEYS or by another State body appointed to do so. Private CVET providers, companies for the most part, need not have their curricula approved by any independent body, except in the case of courses leading to the acquisition of a formal educational qualification, where MoEYS needs to give approval. On the one hand, therefore, the lack of controls facilitates the rapid development of new training programmes, on the other, the quality of these varies significantly as a consequence. E Government policy towards education and training After a period of relatively spontaneous development since 1999, intensive work on legislative and policy proposals has recently got under way. This should create the conditions for more systematic reform of further education. Recent publications in this area include the National Programme for the Development of Education (White Paper) which was approved by the government in early 2001, and the Human Resource Development 18

399 Strategy, National Employment Plan, and National Employment Action Plan. Implementation of these programmes still requires a major effort. The structure of the secondary education system is still relatively inflexible, does not enable students to combine programmes, and makes it difficult to transfer from one type of school to another without falling behind, losing credits and, given the length of studies, a lot of time because of schools not accepting study results achieved elsewhere. If students do not complete the whole of a programme and pass the final exams, they do not receive credit for the part they have completed. In addition, students cannot extend their specialization and obtain an apprentice certificate once they have obtained their Maturita from a secondary technical or a secondary vocational school. Learning in schools is still focused on building up encyclopaedic knowledge and lacks provision for developing desirable attitudes, skills and knowledge. The introduction of ICT in schools and, particularly, its use in teaching, still lags behind despite the adoption of the State Information Policy in Education in The prevailing preference for specialization within secondary technical and vocational education results in only a small proportion of students (18-19% of the total) following general secondary education programmes (in gymnasiums). Steps to address this imbalance were initiated in 1998 through the introduction of a smaller number of broad fields of study and a Standard of Secondary Vocational Education. The standard provides for a relatively high proportion of general education (between 30% and 45%), puts emphasis on key competencies and, within the vocational part of the curriculum, defines a general-vocational common core for a group of courses. Separate educational pathways with limited possibilities of transfer between them highlight the importance of the initial choice made. The system and arrangements available to inform people about labour market requirements, educational pathways and the quality of schools and to provide career guidance need a major overhaul and to be better integrated. A broader range of learning programmes and growing school autonomy make it hard for job seekers and employers to find their way in the system. The system also impedes comparisons of outcomes and the quality of schools. Although a number of individual evaluations have been conducted, there is no comprehensive system of assessing educational programmes based on defined learning objectives and on the use of appropriate criteria and methods. The current evaluation methods rely mainly on pedagogical criteria rather than the placement on the labour market of those completing education and their initial training. Schools do not gather systematic information on the success of their students in their working careers. 19

400 The inadequacy of links with the labour market is a major flaw of the Czech school system. There are no links between educational standards of vocational schools and professional standards; while industry is insufficiently involved in setting goals and determining the content of education and training programmes and does not regularly participate in quality control, final exams or the introduction of new methods of teaching. Contacts between schools and companies are so ineffectual and random that most young people enter the labour market with little knowledge of the work environment. Lack of practical experience among teachers is a further reason for the separation of the worlds of education and work. In-service training of teachers does not eliminate the problem. However, greater involvement of the social partners in curricular development, evaluation of education results and final exams is envisaged in the newly prepared Long-Term Development Plan. The main bottleneck of the Czech education system is tertiary education. Although the number of universities has increased as private institutions have been set up, these add very little to study places and the number of applicants is still twice as high as those admitted. This is despite the considerable demand from the labour market for specialists with university degrees. As part if the reform of public administration from 2000, administration of education is gradually being decentralized. All secondary and higher professional schools are now administered by regional self-governing authorities which will also allocate financial and other resources to schools. There is no systematic basis for continuing education and training in the Czech Republic, although this area developed rapidly in the 1990s. Despite the extensive supply of programmes, the quality of continuing education is insufficient, training companies are often small and ineffective and participation of universities and vocational schools in the provision of CVET is limited. In the policy proposals currently under preparation, most of the above-mentioned weaknesses and problems of the education system would gradually be eliminated. However, the pace of reform, the right conditions, and the active approach of education institutions and others involved are critical for their successful implementation. One promising reform at present is the establishment of more active cooperation between the ministry, the regions and the schools, which could accelerate positive changes in education. Major changes in the education and training system The most important steps (including government acts) in education and vocational training over the transition are as follows: 20

401 1990 Deregulation of curricular policy: schools were granted the right to modify up to 10% of their study plans and 30% of the relevant curricula and to design and put forward new education programmes to the MoEYS for approval. The establishment of private and church schools was made possible. Labour offices were set up in all districts of the country The establishment of vocational training centres was made possible. The centres integrated initial and continuing vocational education and training Discussion initiated concerning a new model of so-called state maturita (standardized tests set and assessed at national level are proposed in general education) Publication of the document Further Steps in Transformation proposing specific measures for a further development of VET on the basis of the outcomes of the Phare VET programme The Strategy for Human Resource Development prepared by the National Training Fund was made public Since the beginning of 2001, the education system has been undergoing decentralization as part of the reform of public administration. Self-governing regional authorities (at NUTS III level) have been taking over responsibilities for the setting up and administration of all upper secondary schools and higher professional schools. The Government approved the final version of the National Programme for the Development of Education in the Czech Republic (White Paper). The document identified ambitious targets in line with the Lisbon strategy The MoEYS developed and presented to the Government a Long-Term Development Plan of Education and the Education System in the Czech Republic (based on the ideas in the White Paper). The plan aims to ensure a uniform approach to the development of education and implementation of state education policy in a context of decentralization. It provides a methodological framework for regional plans. It recommends the establishment of a system of strategic management for human resource development at the regional 21

402 level as at the national level. The regional council for HRD, including political, professional and social partner representatives, should have a key role and should be able to initiate plans and activities in this area. Another important step is the building up of centres of excellence, intellectual support for entrepreneurship and for new technologies and methods The Government approved a strategy of Human Resource Development for the Czech Republic. This reflects global and national challenges which affect human resources and defines basic common knowledge, skills and values in this area. It puts forward objectives and makes recommendations in wide range of areas, such as lifelong learning, education of young people, tertiary education and science, continuing education and training, management and entrepreneurial skills, company strategies, vocations for the free-market economy and HR in the public sector. It suggests strategic ways for realizing this strategy through national and regional HRD management, information and communication networking, quality management, partnership and cooperation and funding. MoLSA in co-operation with MoEYS and other ministries develops and presents to the European Commission the Operational Programme for Human Resource Development and the Single Programming Document for Objective 3 for Prague. The programmes will be funded from the European Social Fund in the period and include, inter alia, priorities and measures for the development of the initial and continuing education and training systems. 22

403 Economic Restructuring and Labour Markets in the Accession Countries Research Project commissioned by EU DG Employment, Social Affairs and Equal Opportunities, Contract No. VC/2003/0367 Coordinated by in cooperation with Alphametrics and The Vienna Institute for International Economic Studies Industrial policy in the Czech Republic by Filip Zeman* August 2004 *) EuroProfis Ltd., Prague

404 Contents Introduction A Developments in government policy towards industry and services 1 Overview of main industrial features of the Czech economy 2 The transition from public companies to private enterprises 3 Developments in government policy towards industry and services 4 Policies adopted for the creation of new businesses and their operation B The form and extent of government support 1 Finance (subsidies, grants, cheap loans) for businesses 2 Advice and assistance (advice centres, support for marketing, exporting and research and development) 1 3 Low-cost business premises: the establishment of industry and business parks and science and technology parks 2 C Policy towards foreign investors 1 Investment incentives 2 Other support measures D Relative effectiveness of the policy 1 2 Support of entrepreneurship in the Czech Republic, MIT, 2004 Czechinvest

405 Filip Zeman Industrial policy in the Czech Republic Introduction Attempts to review industrial policy in the Czech Republic and its effects have to contend with the following problems: (1) Although there exists a concept of industrial policy, the link between conceptual measures and the concrete tools deployed is indirect. (2) Support for SMEs is not consolidated but managed by different bodies (Ministry of Industry and Trade, CzechInvest, Ministry for Regional Development, and others). (3) There are at best only vague attempts to measure and monitor the effectiveness of policy and the concrete measures deployed. The accession of the Czech Republic to the EU and harmonization of political approaches is likely to encourage more to be done in this regard. Monitoring of outcomes and evaluation of the effectiveness of each measure is an essential part of the implementation of the Structural Funds and this also applies to spending financed from national sources. A Developments in government policy towards industry and services Industrial policy in the Czech Republic is largely non-interventionist and in line with the Government s market-oriented concept of industrial development. Industrial policy starts from analysis of the situation across industry and the economy as a whole and of on the challenges arising from EU accession and entry into the single market. The Czech Republic has historically been one of the more advanced industrial economies. During the period of economic transformation and privatization, industry has undergone extensive changes in ownership and has had to face the advent of competition in a market environment. The Czech economy has opened up, so that it is now affected by forces associated with the globalization of the world economy. Support has been given in the recent past to the development of small and medium enterprises (SMEs) with the aim of enhancing industrial competitiveness and, to a certain extent, of expanding exports and research and development. In addition, the Czech economy has also been affected by the pre-accession steps taken to prepare for at entry into the EU. 1

406 1 Overview of main industrial features of the Czech economy The share of industry in the generation of GDP (including mineral extraction and energy) is still relatively high and is declining only very slowly. In 2003 industrial value-added amounted to nearly 31.4% of GDP 3, which was significantly above the EU15 average of around 25%. The structure of the Czech economy has changed markedly over the 1990s, the share of the tertiary sector (services) has increased while that of the primary and secondary sectors (agriculture and industry) has declined. Since 2000, the share of each sector has been relatively stable (see Figure 1). Figure 1 Share of the sector in GDP (in %) in basic prices of gross value added in % Czech Republic primary sector agriculture + mining Czech Republic secondary sector industry (manufacuring, electricity, gas and water suply), construction Czech Republic secondary sector manufacturing only Labour productivity in industry in the Czech Republic is only around half of the EU average (adjusted for purchasing power parity). 4 The growth of labour productivity is faster in businesses with a higher share of foreign capital. Industrial employment in the Czech Republic (30% of the total) is 12 percentage points above the EU average (18%) which reflects the lower level of labour productivity. 3 4 Survey of the Czech Economy and MIT Sectors in 2003, MIT, 2004 Eurostat,

407 Industrial assets owned by the state have gradually been privatized. A fully liberalized market economy has been created, excessive production capacity has to a large extent been phased out and fundamental environmental issues inherited from the past have been addressed. In industry and the distributive trades, state assets worth about EUR 23 billion were initially earmarked for privatization. 70% of assets and over 74% of entities were privatized by the end of In 1999 the private sector accounted for 85% of total industrial production and proceeds from industrial activities. Small and medium enterprises (SMEs) in aggregate employ over 40 % of the industrial workforce. SMEs in industry have proved to be more dynamic in terms of the generation of value-added. However, they are weaker economically and the banks regard them as highrisk partners. They account for 36% of all national exports and 52% of all imports (figures in 2002), which is a slightly lower than in Figure 2 Development of output and value added of SMEs in the Czech Republic, Output (mil. CZK) Value added (mil. CZK) 5 Statistics of SMEs, MIT,

408 Figure 3 Development of foreign trade of SMEs in the CR ( ) Export (mil. CZK) Import (mil. CZK)

409 Share of SMEs in the Czech economy (% of total) Indicator Number of enterprises Number of employees Output Value-added Wage costs Investment 1) 1) 1) 1) Exports Imports GDP 1) 1) * GDP: CZSO data, other indicators: calculations of MIT. 1) unavailable Output and value-added generated in SMEs in Sector Output Value added EUR million share in sector EUR million share in sector Industry Construction Retailing Hotels and restaurants Transport Financial services Other services Agriculture Total The adverse effect on the environment impact of industrial production was significantly reduced during the 1990s. The implementation of stringent environmental legislation is comparable to that in EU15 countries in many respects and took effect as early as the late 1990s. Foreign trade has shifted to other markets as traditional markets, which accounted for over 55 % of Czech exports before 1990, collapsed. There was a transitional increase in the share of exports of commodities with low added value. Since 1996, exports of higher value-added goods have been gradually increasing and the overall volume of exports has been rising continuously. Exports were redirected towards the EU and other competitive markets. Mechanical engineering and chemicals continue to be the most important exporting industries, while light industry is regarded as a potential source of export growth. 6 MIT CR,

410 Investment in the Czech Republic is about a third higher than the EU average in relation to GDP. The construction industry has been restructured and privatized and employs a relatively large work force (accounting for around 9.5 % of total employment in 2003 as against an EU15 average of 8%). Its value-added as a share of GDP is also higher than the EU15 average. Housing accounts for a very small share of total investment. In the distributive trades, large multinational companies are growing in importance and the former rural trading network has practically collapsed. Legal awareness and consumer protection as well as adherence to industrial rights and copyright are below EU standards. Regions have retained their industrial character established over the historical past. Industrial microregions are key components of the manufacturing base in the Czech Republic and the state of their industries is crucial for their social and economic well-being. Nearly two thirds of all industrial products were produced in 31 of the 76 districts and in Prague, the capital; i.e. in 40% of the land area of the Czech Republic. The industrial significance of Prague and Brno has declined due to the growing share of services in these cities. Ostrava and Kladno have also lost some of their importance as their coal and steel production has been partly phased out. 2 The transition from public companies to private enterprises From the beginning of the 1990s, state assets were gradually privatized with the aim of creating a fully liberalized market environment, of changing property relations, of phasing out excessive production capacity (in coal mining and steel production especially) and of tackling environmental problems inherited from the past (through desulphurization of coal power stations, closing down inefficient and antiquated production plants and so on). There were 1721 enterprises earmarked for privatization in industry and services together. The total value of the state share in these companies amounted to EUR 24 billion. By the end of 2000, nearly 75 % of all entities and 70 % of all assets were privatized. a number of large companies, often of strategic importance, where the state owns a majority of shares, remain, though the state also remains a shareholder in many companies which are not of great importance in the market. Adjustment to the new economic circumstances tends to be slower in large industrial organizations. Extensive restructuring of property rights, production, technology, 6

411 organization and personnel policy have been introduced but not completed. The gradual shift from public to private ownership during the 1990s is shown in the table below. 7 Indicator Unit Non-state sector share of GDP (constant prices) Non-state sector share of industrial production % * % * - estimated 3 Developments in government policy towards industry and services For the period 2001 to 2006 four policies priorities were established to achieve the strategic aim of increasing production and the competitiveness of industry: Support for entrepreneurial activity in industry and industrial services Restructuring the industrial manufacturing base Increasing the competitiveness of industrial production Developing human resources in industry These specific aims are intended to be achieved through: 8 The development and transfer of technology, innovation and new processes Investment in industrial infrastructure Reducing requirements and risk of the production process Improving trade relations Promoting SMEs in industry Investing in human resources in industry Despite the strategy set out and the positive shifts towards more effective concentration and coordination of measures, practical realization of the policy is still limited in particular by the division of responsibilities among the various bodies involved, including the regions. 7 8 Industry - Discussion paper, MIT, 2001 These aims are defined by the industry policy approved by the government as an initial basis for negotiation with the EU during accession period. Some of the areas are covered in programmes under the industry Operational Programme prepared for

412 4 Policies adopted for the creation of new businesses and their operation The creation of new businesses, mainly SMEs, is supported by a number of programmes operated by the Ministry of industry and trade (MIT). The merger of three agencies (CzechInvest, CzechIndustry, Agency for Business for support of investments and entrepreneurial activities) into one (CzechInvest) serves to consolidate this support. Legislation approving the support for SMEs was introduced in 2002 (new amendment in 2004). The Czech Government approved the terms and conditions for assisting SMEs in New businesses are in practice supported by both the national programmes of the Ministry of industry and Trade and the Operational Programme for Industry co-financed by the Structural Funds. As part of these programmes, new entrepreneurs receive support from the START scheme (bank loan for setting up new businesses), PORADENSTVÍ (Consultations - subsidies for consultancy for new entrepreneurs) and KREDIT (bank loans). Assistance for other SMEs is provided from the following schemes: ZÁRUKA (Guarantee), TRH (Market), SPECIAL, KOOPERACE (Co-operation), MARKETING, MALÉ PŮJČKY (Small Loans), DESIGN, REGION, VESNICE (Community), PREFERENCE, PROVOZ (Operation), REGENERACE (Regeneration) and REGIOZÁRUKA (Regional Guarantee). Českomoravská záruční a rozvojová banka, a.s. (the Czech-Moravian Guarantee and Development Bank hereafter CMZRB Bank) has been entrusted with the implementation and management of these programmes, except MARKETING (implemented by the Czech Trade Promotion Agency Czech Trade), MALÉ PŮJČKY/Small loans and PORADENSTVÍ/Consultations (managed by CzechInvest) and DESIGN (managed by the Czech Republic Design Centre). Other support is available under the PHARE 2003 programme, the Human Resource Development operational programme (with priority given to Adaptability and Entrepreneurship, including the training of entrepreneurs as well as their employees). In certain regions, Joint Regional Operational Programme measures can be used for support of SMEs as well as for tourist businesses. 8

413 B The form and extent of government support 1 Finance (subsidies, grants, cheap loans) for businesses Support for SMEs All projects must take place in the Czech Republic and must involve industrial production, construction, handcraft production or services to be eligible for support, which is provided by the Czech-Moravian Guarantee and Development Bank, CzechInvest, CzechTrade and Design centre Czech Republic. Projects supported by guarantee, loan or grant for interest ( ) 9 Indicator Number of projects Total costs (EUR million) Applicants Natural bodies 10 (%) Applicants Legal bodies (%) Programmes GUARANTEES (Záruka /Guarantee programme) This programme provides support to SMEs in the form of a guarantee for bank loans, operating credit, leasing, capital contribution, offers in business tenders. The Guarantee provided ranges from CZK to CZK 10 million. Guarantees provided in by sector (% of new guarantees) Sector Industry Of which: food industry Transport Building industry Retailing Accommodation services Services to the public Health care Other sectors Total Czech-Moravian Guarantee and Development Bank Natural body = single entrepreneur, Legal body = company established according Business Code 9

414 Guarantees provided and guaranteed loans in Number of guarantees Guarantees (EUR million) Guaranteed loans (EUR million) Average volume of guarantees (%total) Figures for programmes GUARANTEE, REGIONAL GUARANTEE and RECONSTRUCTION: Czech-Moravian Guarantee and Development Bank. PREFERENTIAL LOANS (programme KREDIT, START) These include investment loans with a preferential or zero interest rate with maturity of up to six years. The KREDIT programme offers loans of up to CZK 7 million, while the START programme offers zero interest rate loans up to CZK 1 million (also for entrepreneurs setting up new businesses). Preferential interest rate loans in Number of guarantees Guarantees (EUR million) Guaranteed loans (EUR million) Figures for programmes MARKET, REGION, COMMUNITY, REGENERATION, REGION 2, RECONSTRUCTION: Czech- Moravian Guarantee and Development Bank Guarantees provided in by sector (% of new guarantees) 11 Sector Industry of which: food industry Transport Building industry Retailing Accommodation services Services to the public Health care Other sectors Total Czech-Moravian Guarantee and Development Bank 10

415 CONTRIBUTIONS TO INTEREST PAYMENTS (VESNICE, REGENERACE programme) The contributions amounting to 3 % (REGENERATION) or 5 % (COMMUNITY) a year differ according to the programme and are available for investment loans of up to 4 years. Businesses supported are those: with less than 10 employees, in municipalities with less then 2999 inhabitants in town landmark areas 12, landmark reservations and others, preferably with less than 50 employees OTHER CONTRIBUTIONS (SPECIAL, TRH, KOOPERACE, ZÁRUKA programmes) Grants are available for increasing employment (SPECIAL) of up to CZK 4,000 per month (EUR 125) for up to 4 years if those who are socially excluded are taken on A grant of 50% (up to a in maximum of CZK 300,000 EUR 9,400) for acquiring a ISO 9001 or ISO from the TRH scheme or for implementing the EMAS programme 13. A contribution (TEST) is also available of up to 50% of costs, CZK 30,000 per product (EUR 940) and CZK 200,0000 (EUR 6,300) per recipient for he acquisition of a standard ČSN compliance mark (signifying compliance with Czech standards). A contribution to help cover capital costs (ZÁRUKA) enables investors to cover the costs of external services. The maximum limit is 3% of the total cost up to CZK 250,000 (EUR 7,850). A contribution to cover the costs of association activities, of up to 50% with a limit of CZK 1.5 million (EUR 47,000) (KOOPERACE) is also available. Other contributions provided in Number of contributions Amount of contributions (EUR million) Areas with registered and protected landmarks (UNESCO, etc.) EMAS (Eco-Management and Audit Scheme) - a tool for implementation of EMS (Environmental management systems) in the companies. Czech-Moravian Guarantee and Development Bank 11

416 2 Advice and assistance (advice centres, support for marketing, exporting and research and development) 15 The centres which exist to provide business advice are as follows: CzechInvest An agency of the Ministry of industry and trade for investment and support of entrepreneurs, providing detailed information about the support available to entrepreneurs, investment incentives, the support for industry/business zones and parks and the programmes which exist for Czech sub contractors. Czech Republic Design Centre A centre providing information, consultation and training services on design in production, trade and services; services are provided under the MIT DESIGN programme. Business Info A web portal providing information about all the support available in the Czech Republic organized by topic (legislation, finance and taxes, exporting, the EU, consumer protection, entrepreneurial support, analysis, etc.) instead of by provider. National Training Fund An agency of the Ministry of Labour and Social Affairs providing support for human resource development (lifelong learning, quality and effectiveness of human resource development, employability and social development). Economic Chamber of the Czech Republic and Czech Agrarian Chamber Offices providing services to members in respect of legal advice, support for SMEs, customs, foreign contacts, information services and so on. Czech Confederation of Industry An association providing its members with services in respect of foreign contacts, economic policy, employment relations and social policy, research and development, legislation, communications, and so on. Czech Confederation of Commerce and Trade An association representing members in commerce and trade working to promote equal business conditions, a competitive environment and market conditions and providing 15 Support of entrepreneurship in the Czech Republic, MIT,

417 information about the single European market, legislative changes affecting the sectors of activity of their members as well as training in business negotiation. Association of small and medium enterprises and traders of the Czech Republic An association undertaking activities for improving the entrepreneurial environment, access to finance, the dissemination of information, relations with the Government and Parliament and providing other services for members. CzechTrade (Czech Trade Promotion Agency) An agency of the Ministry of Industry and Trade providing individual support abroad, training in exporting and information about current export conditions. Amount of support and number of active participants in programmes EUR million Participants Subsidies given by company size, % 2% 13% 36% do 10 do 50 do 100 do 200 do % More than 80% of supported companies have up to 100 employees Czech Export Bank A financial institution providing direct export credits, refinance export credits, loans for exporting, export guarantees and interest rates subsidies. Export Guarantee and Insurance Corporation (EGAP) An agency providing commercial guarantee and insurance as well as services with state support. 13

418 Results of EGAP s activities in the year 2003 indicate resumption of the growth trend in the volume of insured credits, with an overall improvement in the insurance loss ratio. The multiplicative effect of the concluded insurance on the aggregate volume of export contracts enabled by this insurance was several times higher. The year 2003 was characterized by a significant decrease in the volume of claims paid of around 64% in relation to 2002 (to CZK 374 million EUR 12 million), mainly due to insurance with state support. 14

419 Aggregate export support through insurance of EGAP in (CZK million) 16 Note: Aggregate values for individual years include values of export contracts supported by insurance with state support and insured values of receivables insured against marketable risks 3 Low-cost business premises: the establishment of industry and business parks and science and technology parks 17 Since 1995 CzechInvest has maintained the country s most comprehensive database of industrial properties in the Czech Republic. The database is available online and contains details of more than 200 fully serviced industrial zones or sites available for strategic services and technology centres. Foreign investors can also find details of industrial premises tailor made for them.. All of the properties are in good condition and ready to start production. There are four main programmes for supporting the development of industrial zones, aimed at: preparing zones renovating zones building new premises or reconstructing existing ones accrediting zones Exchange rate is around CZK 32 to EUR 1. Czechinvest 15

420 The above map indicates the strategic sites which obtained a subsidy for construction or for land transfer at a preferential price under the Programme for Support of Industrial Zones Development. State investment in industrial zones Year Investment Area supported Details (EUR million) (ha) Karviná, Bystřice and Pernštejnem Induced investment of 235,4 EUR million, 2900 new jobs Philips Strategic industry zones By the end of 2002, there were 71 industrial zones in the Czech Republic with a total area of 2,149 ha, with around 51% utilization and providing potentially 85,000 jobs. Since 2002, a new project has been launched called brownfields, which has focused on old industrial areas and premises in need of renovation. 16

421 Since 2003, the focus has been on involving the private sector in the preparation of industrial areas. Science- technology centres A new Framework Programme for Support of Technology Centres and Centres of Business Support Services was launched in February, This merged together two previous programmes introduced in June 2002 the Framework Programme for the Support of the Establishment and Expansion of Technology Centres and the Framework Programme for the Support of Strategic Services. Investment incentives for technology centres and business support services Support is provided to technology centres and centres for business support services. Technology centres are intended to stimulate innovation and are closely linked to the production of high-tech products and the use of new technology in production. The sectors of activity concerned are: Aerospace Computers and office machines Electronics and microelectronics Telecommunications Pharmaceuticals The Automotive industry Support is also available for Customer contact centres of all kinds; shared-service centres, handling a range of different services for companies, such as finance, accounting, marketing and training; ICT expert solution centres for implementing computer and telecommunication systems,; software development centres, aimed especially at foreign companies; and high-tech repair centres for computers, electronic devices, precision instruments and so on. The maximum amount of support in the form of subsidies differs between regions as shown in the following map (which is valid from 2002 to 2006). Economically poor regions and those affected by structural change are preferred. 17

422 The forms of support consist of a subsidy for business activity and a subsidy for training and retraining. C Policy towards foreign investors 1 Investment incentives The incentives available for foreign investors in Czech manufacturing include: corporate tax relief full tax relief for 10 years (newly-established companies) partial tax relief for 10 years (expanding companies) job-creation grants training and re-training grants transfer of infrastructure and land at a discount transfer of land owned by the Czech State at a discount The centrepiece of the Czech Government s incentive package is corporate tax relief for manufacturing investors for up to 10 years. The incentives package also contains two employment-related benefits: 18