1. Introduction: the role of the quality of institutions for the local development

Transcription

1 53 nd European Congress of the Regional Science Association 27 st 30 th August, 2013 Palermo Local Development and Quality of the Institutions in the EU Regions by Cristina Brasili, Alessandro Lubisco, Lucilla Spinelli Summary 1. Introduction: the role of the quality of institutions for the local development 2. The economic and social places-based scenario of the EU regions 3. Regions by productive sectors: productive specialization and districtual regions 4. A growth model for the districtual regions conditioned to the quality of the institutions Concluding remarks 1. Introduction: the role of the quality of institutions for the local development The relevance of territory for the improvement of productivity is commonly acknowledged among the scholars. The productivity in a specialized area, as a district, is normally more relevant not only for the geographical closeness of the enterprises belonging to the same productive sector, but also for the specific territorial capital which characterizes the place. A higher productivity not only depends on the economic sector but, also, on the local environment and on technical and productive characteristics jointly (Becattini, 2012, p.7). Thus, every territory has a degree of productive chorality ( coralità produttiva, Becattini 2012) not only based on the proximity of businesses but, also, on the homogeneous cultural characteristics of all families and inhabitants of that place, not necessary directly involved in the local production. The importance of the quality of institutions is, nowadays, well known in the social and economic studies. In the Nineties, Douglass North reviews the role of institutions with regards of economic development. This paper tries to answer to the question whether there is a relation between economic development and quality of the institutions, integrating value of the territory, economic and social scenario and productive specialization, particularly, in the districtual regions. Specific territorial capital and social characteristics, will be linked to the indicator of quality of the institutions (firstly proposed by Quality of Government Institute of Gothenburg) and to the economic growth of UE regions. This study takes into account the regions of France, Germany, Italy, Portugal, Spain and United Kingdom in the years from 2000 to 2012 (if available). This paper consists in the analysis of social and economic development of the regions of six EU countries (Section 2), relating to the specification of the productive characteristic of the districtual regions, following the idea of productive chorality (Section 3). The estimation of a model of economic growth, eventually taking into account, for the considered regions, the quality of the institutions and productivity in districtual regions (Section 4). Starting from the data matrix, regions by productive sectors, we would like to verify the thesis of the role of local quality of institutions in the evolution of the Industrial Districts, being true Department of Statistics, University of Bologna The authors share the responsibility for this paper. However, Cristina Brasili wrote sections 1 and 4. Alessandro Lubisco section 2. Lucilla Spinelli section 3. Concluding remarks were written jointly. Alessandro Lubisco is, also, the author of the maps. 1

2 the Becattini s hypothesis about the productive chorality instead of the merely geographical localization jointed to the productive specialization. The scarcity of territorial data prevents starting from a single cell of the matrix regions by productive sectors. For this reason we adopt the regional level (NUTS 2) for the analysis. We estimate a model of regional economic growth (Section 4) through the relationships among some relevant indicators of local development such as unemployment rate, tertiary education rate, R&D expenditure, rate of young people not in employment and not in any education training and quality of the institutions (using the regional QoG indicator, EQI 2010) in relation with the districtual regions and sectorial labour productivity. As a measure of the quality of institutions in EU regions we use data collected by the Quality of Government Institute at the University of Gothenburg (Quality of Governament Institute, 2010), combined with World Bank Global Governance Indicators data (firstly proposed in the working paper by Andres Rodriguez Pose and Josè Enrique Garcilazo, 2012). Moreover, to take into account the idea of the productive and social chorality, translation and evolution of the Marshallianan atmosphere concept, we introduce the number of the industrial districts in a region in order to weight the districtual characteristics of the regions. We know that the Nuts2 regional level is too big to focus the industrial districts, but there is no choice. If the hypothesis that the production in the Industrial Districts is a social ipso facto is correct, then the institutional quality in districtual regions could condition growth and development. The role of local institutions is more relevant in those cases where there is a local productive community which strengthen itself throughout the relationships among a great number of local institutions and cultural actors. 2. The economic and social places-based scenario of the EU regions of Italy, France, Germany, Spain, United Kingdom and Portugal. 2.1 Gross Domestic Product (GDP) Latest available data refer to In the period the GDP in EU-27, expressed in Euro per inhabitant ( /inh), increased from to in 2007 and In 2009 the level decreased to , growing again in 2010 to (graph 2.1.1). To make comparisons between the six countries about GDP, we considered the series expressed in in Purchasing Power Standard per inhabitant (PPS) and Purchasing Power Standard per inhabitant in percentage of the EU average (PPS%EU). Starting from the evaluation of the PPS%EU for each country, it appears that Italy (101%) and Spain (99%) have levels similar to the EU-27 average. This means /inh ( PPS) for Italy, and /inh ( PPS) for Spain. Germany ( /inh and PPS), UK ( /inh and PPS) and France ( /inh, higher than UK, and PPS) have levels of PPS%EU definitely higher than EU average corresponding respectively to 119%, 111% and 108% (graphs and 2.1.3). But, analysing regional data, it s possible to see that for each one of these three last countries there is one single region with a very high GDP. In particular, in Germany, Hamburg region has a level of PPS%EU double respect to EU-27 average ( /inh and PPS); in UK, Inner London reaches a level equal to 328% ( /inh and PPS); in France, Île de France records a 180% ( /inh and PPS). In all these three countries, the second region in the national rank has a value of PPS%EU clearly lower. So lower that the second region in France has a value equal to 108%, that is the national average. This means that these three regions attract heavily to higher values the national average, while levels of the three indicators in the other regions of the countries are different. It s known that Hamburg and Inner London levels are overestimated because of commuter flows. Île de France has more than 11 million inhabitants, and this is a big weight in the calculation of the national average. The geographical distribution of PPS%EU (graph 2.1.4) allows us to make some observations. First 2

3 of all, it s clear the large development gap between Northern and Southern Italy, as well as between Western and Eastern Germany. In the same way, the entire Portugal (excluding Lisboa region) together with Extremadura and Andalucìa in Spain have levels of PPS%EU below 75% of EU-27 average. In France, only Île de France, Rhône Alpes (108%) and Provence-Alpes-Côte d Azur (102%) have levels of PPS%EU higher than EU-27 average. Finally, in UK, the regions with low level are Western Wales regions (72%) and Cornwall (70%), while Berkshire, Buckinghamshire and Oxfordshire (143%) and its neighbours (from 109% to 115%) have high levels of PPS%EU. North Eastern (162%) and Eastern Scotland (109%) are distinguished from the other Scotland regions. This confirms the fact that, in all the countries, regions hosting national capital or big cities and the surrounding regions have higher levels of PPS%EU respect to the other areas. From 2000 up 2010 In Germany PPS%EU increased in 2010 respect to 2009 in all the regions, but respect to 2000 this happened only for 18. Also in UK, 2010 showed a sort of restarting, with 13 regions having a higher level respect to 2009, but only in Inner London and North Eastern Scotland the levels were definitely higher respect to In France, in 2009, it happened something similar to what observed in Germany in Respect to 2008, in 2009 all the French regions had higher values of PPS%EU, but this growth wasn t able to fill the gap of the previous years. For this reason, in 2010 only 3 regions had higher values respect to In Italy there was an overall decrease in PPS%EU in the period Northern regions recorded very big changes with diminutions up to 25 percentage points, but all the levels of PPS%EU remained greater than EU-27 average. In Southern regions, where the levels of PPS%EU were already lower than EU-27 average, there were reductions from 7 to 12 percentage points. In Spain, in 2007 almost all regions increased PPS%EU respect to But only Galicia in 2010 increased the value respect to 2007, while all the other regions showed lower PPS%EU values, even if 12 of them were higher respect to In Portugal, with the exception of Lisboa (114% in 2000 and 112% in 2010) and Região Autónoma da Madeira (91% and 104%), in the period all the PPS%EU values were definitely lower than EU- 27 average without great changes. Graph Euro per inhabitant ( /inh) - Years , minimum, national and maximum values (2010) 3

4 Graph Purchasing Power Standard (PPS) - Years , minimum, national and maximum values (2010) Graph Purchasing Power Standard per inhabitant in percentage of the EU average (PPS%EU) - Years , minimum, national and maximum values (2010) 4

5 Graph

6 2.2 Unemployment rate Latest available data refer to 2012 and in this year the EU-27 unemployment rate was 10.4%. Among the six countries considered the situation is very different. Portugal (25.0%) and Spain (15.7%) have unemployment rates definitely higher than EU-27 average. Italy (10.7%) and France (10.3%) are more or less near the EU-27 average. UK (7.9%) and Germany (5.6%) have the lowest unemployment rate (Graph 2.2.1). In 1999 the unemployment rate in Spain was 15.5%; it decreased until 2007 reaching 8.3% and then began to grow very quickly to the current value, due to the serious economic crisis. Portugal had the minimum value in 2000 (3.8%). During the following years the rate has continued its growth to reach the maximum in In Italy and France there was a similar behaviour of the rate that started in 1999 from a value near 12%, reaching the minimum in Italy in 2007 (6.1%) and in France in 2008 (7.8%). After that, the rate increased until In UK the rate didn t have many variations in the period from 1999 (6.0%) to 2008 (5.6%). In 2009 the rate reached 7.6% remaining on this level in all the following years. The trend of the unemployment rate in Germany was completely different from the other countries. It began in 1999 from a value of 8.9%, reached a first minimum in 2001 (7.8%) and increased until the maximum of 11.1% reached in In the following years the rate reversed the trend, reaching the actual value (5.5% in 2012) that is the lowest of the period. More or less, the unemployment rate followed the same trend of the respective country in all the regions. Anyway, there are some situations that are noteworthy (Graph 2.2.2). In Germany there are only two regions (Mecklenburg-Vorpommern with 10.8% and Berlin with 10.6%) that registered levels of unemployment rate higher respect to EU-27 average. All the other, and in particular the western German regions, have unemployment rates clearly lower. In the same way, also the great part of UK regions have low unemployment rates, excluding West Midlands (11.7%), Tees Valley and Durham (11.6%) and South Yorkshire (10.5%). The situation in France is rather uniform with most of the regions with values not too far in both directions from EU-27 average. Only two regions stand out: Limousin, with the lowest rate (7.2%), and Languedoc-Roussillon (15.7%) 1. In Italy there is the usual distinction in two parts: the first one, composed by Northern regions together with Toscana, Umbria and Marche, has unemployment rate lower than EU-27 average, with the minimum in Provincia autonoma di Bolzano/Bozen (4.1%); the second part contains Lazio, Abruzzo, Molise, Southern regions, Sicilia and Sardegna, all with unemployment rate higher than EU-27 average, with the maximum in Campania and Calabria (19.3% for both). In Portugal and Spain the situation is definitely alarming. Centro region in Portugal (12.0%) has the lowest unemployment rate in the Iberian Peninsula. While in Portugal the rates remain under 20%, in Spain more than half regions have unemployment rates higher than this level. The lowest rate in Spain is in País Vasco with 14.9%. 1 Only one little note for overseas French regions that have very high unemployment rates (from 21.0% in Martinique to 28.6% in Réunion): these high rates could not affect consistently the French average due to the small number of people involved, compared to the entire French labor force. 6

7 Graph Unemployment rate - Years , regional minimum, regional maximum and national values (2012) 7

8 Graph Unemployment rate per region (2012) 8

9 2.3 Third level education One of the five targets of Europe 2020 concerns education. In particular, this headline target aims to reduce the share of early school leavers to 10% from the current 15% and increase the percentage of the population aged having completed tertiary from 31% to at least 40%. The two variables are highly correlated and, for the purpose to estimate our model, we decided to consider only the second one. Latest available data refer to 2012 and in this year the EU-27 percentage was 35.8% (Graph 2.3.1). From year 2000, all the six countries we considered showed a growth of this indicator but with different trends, although the ranking has not changed much. In 2012 UK has the highest value with 47.1%, obtained with a continuous and steady growth started from the value of 29.0% in Also France (43.6%) and Spain (40.1%) show in 2012 high percentages, but they reached these levels respectively in 2009 and 2008, stopping in the following years the growth. In order to achieve the Europe 2020 goals, respectively 50.0% and 44.0% the two countries must start again with an increasing trend. In Germany, in 2000, the percentage was 25.7% and this value didn t change a lot until 2006 (25.8%). In the following years the percentage increased slowly reaching in 2012 the value of 31.9%, still quite far from Europe 2020 target fixed for Germany in 42.0%. Portugal started in year 2000 with the lowest value equal to 11.3%. Except in 2009, during the period the percentage increased to 27.2%, more than twice the starting value. The percentage must continue the growth to reach in 2020 the target of 40.0%. Also in Italy there was a steady growth, even if slower than the other countries. From 2000 (11.6%) the percentage of people with tertiary education in 2012 was 21.7%. Although Europe 2020 target for Italy of 27.0% is lower than the goals of the other countries, also here the growth must continue in the future years to achieve it. Regional analysis (Graph 2.3.2) shows that in Italy the percentage of people aged with tertiary education is lower than 30% in Nothern and Central regions and doesn t reach 20% in almost all Southern regions where Sardegna (15.6%) has the lowest value. Only Liguria (27.5%) has a level greater than Europa 2020 target for Italy. In Portugal, Lisboa (32.0%) has the higher percentage, while all the other regions are below 30.0% where Alentejo (21.9%) records the lowest value in the country. Likewise, in Germany, most of Northern regions are under 30.0%, with the lowest percentage in Sachsen-Anhalt (20.2%). Only Berlin (43.2%) has a percentage higher than Europe 2020 target for Germany fixed in 42.0%. In Southern regions levels are slightly better, with Obernbayern (43.6%) with the highest value. Also in Spain there s a difference between Southern and Eastern regions, with lower percentages, and the rest of the country. Ciudad Autónoma de Melilla (20.7%) and Región de Murcia (25.2%) have the lowest percentages, the only two with a value lower than 30.0%. Andalucia (31.4%), Illes Baleares (32.3%), Castilla-La Mancia (33.2%) and Extremadura (34.1%) are the other regions with levels lower than EU-27 average. The remaining regions, excluding Comunidad Valenciana (38.5%), have levels higher than Europa 2020 target for Spain (44.0%) with País Vasco (61.7%), Principado de Asturias (50.4%) and Comunidad Foral de Navarra (50.1%) with the highest values. In France, Midi-Pyrénées (54,0%) and Île de France (52.1%) have the highest percentages of persons aged with tertiary education, greater than Europa 2020 target for France (50.0%). Auvergne (31.6%) has the lowest value, but half of the regions have percentages higher than 40.0%. In UK, excluding a small number of regions like West Midlands (33.2%), Merseyside (33.4%), Kent (35.3%), Essex (35.6%), West Wales and The Valleys (35.7%) and Devon (35.7%), all the other regions have percentages higher than UE-27 average. 21 regions out of 37 have values greater than Europe 2020 target of 40.0% (no target in UK National Reform Programme), where the highest percentages are in Inner London (73.1%) and North Eastern Scotland (60.4%). 9

10 Graph Percentage of aged with tertiary education - Years , regional minimum, regional maximum and national values (2012) 10

11 Graph Percentage of aged with tertiary education - (2012) 11

12 2.4 NEET Latest available data refer to 2012 and in this year the EU-27 percentage of young people aged not in employment and not in any education training (NEET) over people aged was 17.0% (Graph 2.4.1). In 2000 the value was 16.9% and reached the minimum of 13.9% in In the following years the EU-27 percentage of NEET increased, reaching the actual level. Germany has the lowest level of NEET (9.8%), value reached in 2012 after a period of growth from 2000 (11.2%) to 2005 (15.6%), followed by a continuous decrement. In France, the percentage of NEET was more or less the same in the period , with values around 13.5%. In 2009, the percentage raised to 16.1%, remaining at this level until 2012 (16.2%). In 2000, UK percentage of NEET was 13.9% and it reached a minimum in 2004 (10.1%). In 2007 (14.9%) the percentage started to increase. It became greater than EU-27 average and reached in 2012 the value of 18.1%. In Portugal the percentage of NEET in 2000 (10.3%) was the lowest of the six countries considered, but in the following years the level started growing, reaching in 2012 the maximum of 18.7%, for the first time greater than EU-27 average. In Spain, from 2000 to 2007, the percentage of NEET was between 13.6% and 15.0%, always under EU-27 value. In a couple of years, the percentage increased rapidly reaching in %, the highest among the six countries. In 2012, the percentage of NEET in Spain was 23.8%. During the entire period , with the only exception of 2009, the percentage of NEET in Italy was the highest among the six countries analyzed. Starting from 23.1%, it reached a minimum in 2007 (20.1%), but after that, the percentage increased reaching in 2012 the highest level ever reached before (27.0%). Regional analysis shows that all the regions in Germany have low percentage of NEET compared to the other countries (Graph 2.4.2). In particular, Southern regions have values lower than 10.0% like Oberbayern (5.9%), Tübingen (6.2%) and Freiburg (6.3%). The highest percentages are in Mecklenburg-Vorpommern (13.6%), Brandenburg (13.5%) and Berlin (13.5%). In France, only three Northern regions, Picardie (22.8%), Nord Pas-de-Calais (22.3%) and Champagne-Ardenne (21.2%), together with Languedoc-Roussillon (22.7%) have percentages of NEET above 20%. Eleven other regions have values lower than EU-27 with the lowest in Bretagne (10.9%), Île de France (12.5%) and Midi-Pyrénées (12.8%). In UK, only Tees Valley and Durham (27.1%) overcome 25% of NEET. Also East Yorkshire and Northern Lincolnshire (24.7%) and West Wales and The Valleys (24.4%) have high levels, while, on the opposite side, Dorset and Somerset (10.3%), Bedfordshire and Hertfordshire (10.6%) and Berkshire, Buckinghamshire and Oxfordshire (11.5%) are UK regions with the lowest percentages of NEET. In Portugal, Região Autónoma da Madeira (29.4%) and Região Autónoma dos Açores (29.2%) have the highest percentages of NEET of the country, where the lowest level is in Norte (16.3%), the only one region under EU-27 average. In Spain, Ciudad Autónoma de Ceuta (36.5%), Canarias (30.2%) and Ciudad Autónoma de Melilla (30.2%) have the highest values, but also in other regions like Andalucía (28.7%), Extremadura (27.2%) and Illes Balears (27.1%) the percentage of NEET is very high. Only Comunidad Foral de Navarra (13.7%) and País Vasco (15.3%) are under EU-27 average. In Italy, all Southern regions have percentages of NEET greater than 30%: Sicilia (39.9%), Campania (38.0%), Calabria (37.8%), Puglia (33.2%), Sardegna (33.0%) and Basilicata (31.3%).Conversely, Marche (19.2%) and some Northern regions are under 20%: Liguria (19.9%), Emilia-Romagna (19.7%) and Lombardia (19.1%). Only Provincia Autonoma di Bolzano/Bozen (10.6%) and Provincia Autonoma di Trento (16.4%) record percentages lower than EU-27 average. 12

13 Graph Percentage of NEET over people aged Years , regional minimum, regional maximum and national values (2012) 13

14 Graph Percentage of NEET over people aged per region (2012) 14

15 2.5 Intramural R&D expenditure (GERD) Last national available data refer to Regional data, instead, are available for 2010 (Italy, Portugal and Spain) or 2009 (France, Germany and UK). In 2011, in EU-27 the percentage of GERD on GDP was 2.03% (Graph 2.5.1). This corresponds to euro per inhabitant of expenditure for research and development. In years from 1995 to 2005, in EU-27 the percentage of GERD on GDP was about 1.8/1.85%. In the following years the percentage increased overtaking 2.00%. Germany has the highest value among the six countries considered, both in percentage of GERD (2.84%) and in euro per inhabitant (901.4). Germany reached these levels with a steady increase from 1995 (2.20% and /inhabitant), even if in 1991 the percentage was already 2.47%. Also in France, in the last 15 years, the level of GERD was always greater than EU-27 average. Considering available data, values varied from 1.90% in 1981 to 2.27% in 1996 and In 2011 the percentage of GERD in France was 2.25%, corresponding to per inhabitant. In UK, the percentage of GERD on GDP, in 1981 was 2.35%. In the entire following period, the percentage decreased, reaching a minimum of 1.69% in 2004 and settling to 1.77% in 2011 (495.9 /inhabitant). In the mid-90s Portugal had the lowest percentage of GERD (0.52%), value that slowly increased until 2005 (0.78%). In the following years, the value raised more rapidly reaching the maximum of 1.64% in 2009 and returning to 1.49% in The value of /inhabitant is the lowest among the six countries considered. In Spain, from 1981 (0.41%) the percentage of GERD on GDP increased slowly but constantly reaching in 2009 the maximum of 1.39% and decreasing in the two following years to the value of 1.33% in 2011 (307.3 /inhabitant). In 1990, the first with available data, in Italy the percentage was 1.25%. After a minimum of 0.97% in 1995, the percentage slowly increased reaching in 2009 the same level of twenty years before (1.27%). In 2011, the percentage of GERD on GDP in Italy was 1.26%, the lowest among the six countries considered, even if /inhabitant is greater than Portugal and Spain values. Regional analysis shows that, in Germany, Braunschweig (8%) has the highest percentage of GERD on GDP (Graph 2.5.2). With per inhabitant of expenditure for research and development, this is the most R&D intensive area in Europe. Half of the German regions have values higher than EU-27 average of 2.03%, and most of them is located in the south, like Stuttgart (6.34%), Oberbayern (4.66%) and Tübingen (4.58%). On the other hand, in some regions the percentage is very low like in Niederbayern (0.38%) and Koblenz (0.73%) together with some Northern regions: Weser-Ems (0.58%) and Brandenburg (0.74%). In France, Midi-Pyrénées (4.40%) and Île de France (3.02%) are the only two regions with value higher than 3%. Corse in particular (0.30% that means 73.6 per inhabitant), Champagne-Ardenne (0.80%) and Nord - Pas-de-Calais (0.82%) have the lowest percentages. In UK there is great variability. Seven regions have percentage of GERD on GDP greater than 3%, with Cheshire (6.51% corresponding to per inhabitant) and East Anglia (5.57%) at the top of the ranking. Conversely, more than ten regions are under 1%, where Cornwall and Isles of Scilly (0.19% meaning 31.4 per inhabitant) and Lincolnshire (0.26%) have the lowest values. In Portugal, only Lisboa (2.32% corresponding to per inhabitant) has a percentage of GERD on GDP greater than EU-27 average. All the other regions have low percentages with the two autonomous regions of Madeira (0.32%) and Algarve (0.38%) at the minimum, but also with Alentejo and Algarve both under 0.50%. In Spain, only Comunidad de Madrid (2.05% that is per inhabitant), Comunidad Foral de Navarra (2.04% or per inhabitant) and País Vasco (2.03% meaning per inhabitant, the highest of the country) have percentages greater than EU-27 average. Excluding the two autonomous cities of Ceuta (0.09%) and Melilla (0.16%) and the islands Baleares (0.43%) and Canarias (0.63%) that have very small percentages, but low weight, Castilla-la Mancha (0.63%), Extremadura(0.88%) and Región de Murcia (0.92%) have the lowest values in the country. In Italy, only Provincia autonoma di Trento (2.13% that means per inhabitant) has a percentage of GERD on GDP greater than EU-27 average, followed by Piemonte (1.83%) and Lazio (1.80%). The other Central and Southern regions like Calabria (0.46%) and Molise (0.50%), together with Valle d'aosta/vallée d'aoste (0.59%) and Provincia Autonoma Bolzano/Bozen (0.54%) have percentages lower than 1%. Campania (1.22% 15

16 meaning 198 per inhabitant) is the only Southern region that overcomes this level. Graph Percentage of GERD on GDP - Years , regional minimum (2009/2010), regional maximum (2009/2010) and national values (2011) 16

17 Graph Percentage of GERD on GDP per region (Italy, Portugal and Spain: 2010; France, Germany and UK: 2009) 17

18 3. Regions by productive sectors: productive specialization and districtual regions. 3.1 The productive chorality : regions by productive sectors matrix There are significant disparities between European Union (EU) regions in terms of the importance of different productive activities. The reasons for regional specialization are varied and include the availability of natural resources, infrastructure, legislation, climatic and topographic conditions, the proximity to markets and, more broadly, the degree of "productive chorality" of the territory. The productive chorality is a broader concept, which also includes the homogeneous cultural characteristics of all the inhabitants of that place, not necessary directly involved in the local production. The starting point of this analysis was the construction of a matrix of data, regions by productive sectors, in which a measure of the importance of all business sectors is provided for each region, as suggested by Becattini (2012). This matrix allows to study the importance of each sector in each region and at the same time, the geographical distribution of each sector. As a reference measure has been taken the labour productivity, defined as the value added per employed person 2. The calculation of the indicators has been limited by the availability of data at the regional level (NUTS 2), and with a sectoral detail. Considering the large number of missing data, we had to maintain a high level of aggregation of sectors of economic activity. The sectors that are part of the matrix are therefore as follows: - Agriculture, forestry and fishing (A); - Industry, except construction (B-E); - Construction (F); - Wholesale and retail trade, transport, accommodation and food service activities, information and communication (G-J); - Financial and insurance activities, real estate activities, professional, scientific and technical activities, administrative and support service activities (K-N); - Public administration and defence, compulsory social security, education, human health and social work activities, arts, entertainment and recreation, repair of household goods and other services (O-U). As shown in Graph 3.1.1, in 2010, latest available data, the business sector that generated the highest productivity was financial, real estate, renting and business activities (NACE section K- N). The high degree of dispersion of productivity in this sector is due to the particularly high value of the region Inner London ( euro). The highest value, after the maximum recorded, is in fact much lower, and belongs to the region Valle d Aosta ( euro). To the opposite side we find another region of the UK, Outer London, which records the minimum value (50.04 euro). Above average values are recorded in all the Italian regions, in most German regions (25 of 38), in eleven regions of the United Kingdom and in the Região Autónoma da Madeira (PT). Considerations analogous can be made for the G-J sector, in which the maximum value, always recorded by the Inner London region ( euro), deviates from the next higher value, in the UK s region Berkshire, Buckinghamshire and Oxfordshire (71.09 euro). In this case the minimum value is recorded by a region of Portugal, the Norte (26.83 euro). Above average values are recorded from all Italian regions and some regions of Germany (7 of 38), Spain (13 of 19) and the UK (15 of 37), all regions of Portugal show lower values. As regards the industrial sector (NACE B-E), we note that the upper end lies the region Inner London ( euro) while in the lower the Portuguese region Norte (26.21 euro). Few regions show a below-average productivity, only 15 distributed as follows: two in the United Kingdom, five in Portugal and eight in Italy. The construction sector (F) is the one with the highest average 2 The data used in this analysis were collected by EUROSTAT database, in the Branch account section of the Regional economic accounts ESA95. 18

19 productivity after that indicated by K-N. Regions with above average values are numerous and include all those of Spain, 22 of France, ten of Italy, eight of the UK and five of Germany. The O-U sector is one in which there is less dispersion of the values of productivity, but only ten regions show values above the average. Among these are Inner London, Valle d'aosta and eight French regions: Île de France, Haute-Normandie, Alsace, Aquitaine, Rhône-Alpes, Languedoc- Roussillon, Provence-Alpes-Côte d'azur and Corse. Finally, the agricultural sector is the one with the lowest values of productivity, with a minimum value of about 3,700 euro in the in the Centro region of Portugal. Graph Sectoral productivity at regional level- Minimum, Maximum and Mean values, 2010 GVA per person employed (KEUR) Inner London (UK) Inner London (UK) Inner London (UK) Essex (UK) Inner London (UK) Inner London (UK) Centro (PT) Norte (PT) Outer London R. A. Açores (PT) Norte (PT) (UK) Lincolnshire (UK) A B-E F G-J* K-N* O-U max min mean Berkshire, Buckinghamshire and Oxfordshire (UK) Valle d'aosta (IT) * The data of French regions are missing Looking at the ranking of the regions based on the value of the productivity of the sectors (Tab ) we see that the first place, as was imaginable, is occupied by Inner London region, with a very high value of productivity, amounting to about three times the average value. Follow the Île de France and Hamburg regions: the first show high levels of productivity in all sectors, although there are some missing values, while in the second only the agricultural sector has a lower productivity. Then we find a group of less-spaced composed of two German regions (Darmstadt and Oberbayern), the region of North Eastern Scotland, two Italian regions (Lombardia and P.A. di Bolzano) and the French Provence-Alpes-Côte d'azur. At the bottom of the ranking (Tab ) we find almost all the regions of Portugal, and four regions of the United Kingdom. Among these we distinguish the region of Lincolnshire, which presents values of above-average productivity in agriculture and in the construction sector 3. The last section of this section takes a look at how the sectors have evolved in terms of productivity, GVA and employment in the three time periods: , and The complete matrix of the regions by sector is in Appendix A 19

20 Tab The ranking of regions for the value of productivity: the first 10 Country Region TOTAL All NACE activities GVA per per person employed Mean=100 A B-E F G-J K-N O-U UK Inner London France Île de France Germany Hamburg Germany Darmstadt Germany Oberbayern UK North Eastern Scotland Italy Lombardia Italy Valle d'aosta Italy P. A. di Bolzano France Provence-Alpes-Côte d'azur Tab The ranking of regions for the value of productivity: the last 10 Country Region TOTAL All NACE activities GVA per person employed Mean=100 A B-E F G-J K-N O-U UK Shropshire and Staffordshire Portugal Lisboa UK West Wales and The Valleys UK Lincolnshire UK Cornwall and Isles of Scilly Portugal Alentejo Portugal Algarve Portugal R. A. dos Açores Portugal Norte Portugal Centro

21 Lets first look at the changes that occurred between 2000 and 2007 (Graph 3.1.2). The average annual rate of total productivity increased of 2.63% in the period The G-J sector (trade, transport, accommodation and food service activities, information) has generated the highest productivity growths (5.37% a year), but due to a growth of GVA (3.86% a year) and a reduction of employees (-1.43% a year). The same dynamic occurred in the industrial sector (B-E), but with a decrease of 1.07% a year of the employees, and in the K-N sector, with a very little decrease of the employees, 0.54% a year. The sector with the biggest increase in GVA, however, is construction, with an average annual increase of 6.04%. The agriculture sector, however, saw its total employment shrink between 2000 and 2007, at an average of 1.65% a year, and GVA decreased by 1.13% a year in the same period. Graph Changes in Productivity, GVA and employment in 6 sectors, 2000 to 2007, Annual average % change TOTAL A B-E F G-J K-N O-U Productivity Employed GVA 21

22 As shown in Graph 3.1.3, in the period between 2009 and 2008 are clearly manifested signs of the economic crisis, with reductions of indicators (productivity, employment and GVA) in almost all sectors. The construction sector is the only one in which productivity has not recorded negative value (+0.19%), but it is also one in which the number of employees is reduced to a greater degree (-7.86%). the only two sectors that have favored an increase in employment were the K-N (+1.51%) and the O-U (+2.07%). In agriculture and in industry sector, there has been a significant reduction in added value, by 12.71% and 11.89% respectively. In the industrial sector has also verified the greatest reduction in the number of employees (-6.09%), after that of the construction. Graph Changes in Productivity, GVA and employment in 6 sectors, 2008 to 2009, Annual average % change TOTAL A B-E F G-J K-N O-U Productivity Employed GVA 22

23 In the last period, between 2009 and 2010, productivity has returned to growth in almost all sectors, mainly due to the increase in GVA (Graph 3.1.4). The only exception is the construction sector (F), where GVA and employees dropped significantly. The other two sectors where continued the reduction of employees have been the industrial (-2.31%) and, to a lesser extent, the G-J (- 0.84%). As in the previous two years, the K-N and O-U sectors have continued to support employment, with annual increases of 0.87% and 1.15% respectively. Graph Changes in Productivity, GVA and employment in 6 sectors, 2009 to 2010, Annual average % change TOTAL A B-E F G-J K-N O-U Productivity Employed GVA Then we analyze the growth of these sectors in the three periods previously analyzed, at the regional level. The agricultural sector is not reported because its low level of productivity has recorded very high fluctuations. As we previously saw, the industrial sector in the period showed a positive average annual change, and only in four regions of the United Kingdom there have been decreases (Greater Manchester, Inner London, Outer London, Highlands and Islands). In the recessionary period ( ) many regions remained in a growth phase: Oberfranken in Germany, twelve regions of Spain, four in France, six in Portugal and sixteen in the UK. Past two years the only regions that are not returned to growth were the following: six in the UK, two in France, Ciudad Autónoma de Ceuta (ES), Sardegna (IT) and Região Autónoma da Madeira (PT). Regarding the construction sector, over the period the only region that recorded a negative annual average change in productivity was the North Yorkshire. In the period, the situation has definitely changed and there were decreases in 79 regions. In the last period that 23

24 C. F. de Navarra (ES) R. A. da Madeira (PT) North Yorkshire (UK) number fell to 71 and the decreases were less intense, the lowest value is increased from -30.7% ( ) to -9.7% ( ). G-J sector in the period between 2000 and 2007 showed average change rates very close between the various regions. In the following period the situation changed and regions continued to show positive changes were: six in Italy and in the UK, all the regions of Spain and Portugal. In the last period ( ) the rates of change were again approached and only nine regions have registered negative values: seven in Italy, Centro (PT) and Bedfordshire and Hertfordshire (UK). Similar considerations can be made for the O-U sector. In the number of regions that remained on positive values was high (79), and these were concentrated in Portugal (all), Italy (17), Spain (all) and Germany (36). Between 2009 and 2010 this number increased, reaching 140 regions and including many regions of France and the UK. Finally, as regards the KN sector, over the period only six regions of Germany (Köln, Münster, Arnsberg, Sachsen-Anhalt) have recorded decreases. Between 2008 and 2009, the financial crisis has changed the situation, and only 51 regions maintained a positive productivity: 19 in Germany, 18 in Spain and 13 in Italy. In the last period most regions of the United Kingdom and Portugal began to grow again, and their number increased to 82. Graph Changes in Productivity in 5 sectors, , and , Average annual change (%) Inner Inner London (UK) London (UK) C. A. de Ceuta (ES) Outer London (UK) Cumbria (UK) N. Yorkshire (UK) Lincolnshire (UK) Saarland (DE) C. F. de Navarra (ES) Cheshire (UK) Cumbria (UK) Castilla y León (ES) max min mean Hampshire and Is. of Wight (UK) Berlin (DE) Cheshire (UK) Devon (UK) East Anglia (UK) Puglia (IT) Münster (DE) Calabria (IT) Devon (UK) Cornwall and Com. de Is. of Scilly Madrid (ES) N. East. Scotland (UK) N. East. Scotland (UK) Kent (UK) Lincolnsh. (UK) B-E F G-J K-N O-U 24

25 3.2 The Districtual regions in France, Germany, Italy, Spain, United Kingdom and Portugal Local clusters and industrial districts have been studied intensively since the 1990s. The process of identification of the Districtual regions had as starting point the search for empirical studies on the countries subject of the analysis. In Italy there is a vast literature on the qualitative and quantitative economic analysis of industrial districts (for necessity of shortness we don t give references about this theme) and several methodologies for the identification of industrial districts have been developed (several methodologies has been employed in Italy to identify Marshallian industrial districts getting different results). The most commonly accepted of these methodologies is the Sforzi-ISTAT, an algorithm which departing from local labour markets and activity data, provide a first operative approximation to mapping industrial districts. Sforzi (1987 and 1990) provided the first rigorous attempts to find criteria to identify industrial districts, revised and updated by the Italian Institute of Statistics (ISTAT) in 1996, 1997 and This methodology consists of two phases: in the first phase, the Local Labour Systems (LLS) are delimited, and then a battery of nested coefficients of specialization is used for identifying those LLS showing characteristics of Marshallian industrial district. The application for the year 2001 identified 156 districts in Italy (ISTAT 2006). In the others considered EU countries (United Kingdom, France, Spain, Portugal and German) we don t have the same methodologies but we have tried to compare the identification methods to have the number of industrial districts by region. The Sforzi-ISTAT methodology was applied by De Propris (2005) to the United Kingdom to identify the various forms of the local system and industrial districts, in particular. For the spatial diagnostic analysis four criteria need to be considered: the share of employment in the manufacturing sector, the size of firms, industrial specialization and a combination of industrial specialization and firms size. The application showed the existence of 47 industrial districts in UK. We used the results of this analysis, summarized in a table containing for each district the local system of reference, region (NUTS 1 level) and production sector. In a study for the Délégation à l'aménagement du Territoire et à l'action Régionale (Datar), aimed at promoting local production systems (LPS), Courlet has attempted to identify industrial districts, specialized "cluster" of the French economy. The study established the existence of fifty "industrial districts" of which 25 were selected, of very different sizes and a variety of activities. The list of identified districts, that we use, is shown in a work of Bernard Guesnier (2004). In a research of Boxand Galletto (2006) an identification of Local Labour Markets in Spain is performed using the ISTAT s methodology. The application of the adapted methodology produced 806 local labour systems in the year 2001, and 205 of them would show characteristics of Marshallian industrial district. Industrial districts are concentrated in the centre and the east of Spain. The autonomous regions with a greater number of districts are Valencia, Castilla-La Mancha, Catalonia and Andalusia. They account for 68.6% of the industrial districts in Spain. In the results of the analysis shows the number of districts at the regional level, and these have been used by us. Information on industrial districts of Portugal were derived from work carried out by Cerejeira da Silva (2002). In that work they intend to carry out the segmentation of the entire country into concelho 4 groups, slightly homogeneous, regarding certain pre-defined characteristics, in order to identify those that might be regarded as industrial districts. They use as methodology, multivariate statistics proceedings and, as classification technique they use the cluster analysis with an appropriate variant for the spatial analysis. The result shows the concelhos that are characterized by a significant number of industrial plants, spatially agglomerated, and the intraindustry specialization of each concelho (the first and second most dominant industrial sectors). To identify the presence of industrial districts in Germany regions was based on the Brenner paper (2006), in which he presents a method that allows local industrial clusters to be identified. A 4 In the Portuguese language, concelho is the word municipality, when referring to the territory. 25

26 complete list of all local clusters that existed in Germany in 2001 and are identified by this method is given. This paper provides a methodology to identify empirically the threshold in the number of firms that separates those regions containing a local cluster in a certain industry from those that do not. The local industrial clusters that have been identified are listed above according to the industry to which they belong. A list of all identified local clusters is given, and contains, for each industry studied, the names of all the administrative districts that satisfy the above condition. Therefore our work was to be identified the regions that contain clusters. Bringing together the results achieved in the work above mentioned we obtained a variable that indicates the number of districts in the regions showing the dark blue regions as the Districtual regions (Graph 3.2.1). Graph

27 4. A growth model for the districtual regions conditioned to the quality of the institutions. 4.1 The data on quality of institutions To analyse a growth model for the districtual and non-districtual regions conditioned to the quality of the government we chose the concept of good governance, given in the Nineties, by the most important international organizations conceptualisation (World Bank, UN, IMF). The European Commission, also, in the Fifth Report on the social and territorial cohesion (2010) shows the centrality relevance the quality of institutions saying: The economists are ever more sure the quality of the institutions would be important effects on the economic growth and development. Poor institutions can, in particular, hinder the effectiveness of regional development strategies (EU, 2010, p. 65). Several social scientists, from geographers (Amin and Thrift, 19949) and political scientists (Putnam, 1993, 2000), up to economists (North, 1990; Acemoglu et al., 2001), have underlined how the quality of the institutions determines the long-term economic growth more than others economic factors (Rodrick et al, 2004) and affects the effectiveness and returns of public policies (Rodríguez- Pose e Garcilazo, 2012). Douglass North in his contribute Institutions, Institutional Change and Economic Performance (1990) gives to the institutions a central role in the development, together with the constraints derived from those and with political and economical organizations working in the frame determined by the same institutions. North defines institutions as the basic rules of the road in an economy. This includes formal systems, such as constitutions, laws, taxation, insurance, and market regulations, as well as informal norms of behavior, such as habits, customs, and ideologies. Starting from this definition, we have to adopt a measure of the quality of the institutions. The quality of the institutions has been measured throughout a lot of indicators, all of them subjected to some warnings because they are only based on perceptions. The World Governance Indicators (WGI) given by the World Bank, identify six dimensions of the governance, each measured by an index: Voice and Accountability, Political Stability and Absence of Violence/Terrorism, Government Effectiveness, Regulatory Quality, Rule of Law, Control of Corruption. The values of the six indicators show how EU countries have a good quality of governance in average respect to the other countries. Northern European countries have the best governance level. The scholars of Quality of Government Institute (QoG Institute) of Gothemburg considered four of the six WB indicators, defining fairness, efficiency, transparency, lack of corruption. With these elements they gave a measure of the quality of institutions in EU regions. The regional level of the indicator on the quality of institutions has been captured by the interviews in the EU regions to study fairness, efficiency, transparency and lack of corruption in three public services which jurisdiction is, generally, at regional level: education, health and rule of law 5. The map on the QoG index, relative to European countries in 2009, shows a European Union divided in two parts: the Northern one has high values of the index, while, they are lower in the Mediterranean area (graph 4.1.1). The variability of the QoG index inside European Union is quite relevant, while in some countries, at regional level, is much higher (graph 4.1.2). Sweden is characterized by a low variability of QoG index. Italy shows a very high variability from Northern regions (Bolzano has the highest value, comparable with the best values of QoG index in the other northern UE regions) to Southern regions (Campania has the lowest value of the QoG index among UE regions). Only Mediterranean countries show values close to the EU average, equal to 0 (Spain and Portugal) or lower than the EU average (Greece and Italy). 5 To delve the methodology of the quality of institutions indicator to see RegiosS,

28 Graph The European Quality of Government Index, 2009 Source: Charron, Lapuente, Dijkstra (2012) Graph Regional variations of the European Quality of Government Index, 2009 Source: RegiosS (2013) elaboration on the Quality of Government Institute data 28