Ivlevs, A., Piacentini, M. and Upward, R. (2009) The effects of the economic downturn on migration from the New EU Member States to the United Kingdom. In: COMPAS Annual Conference 2009: New Times? Economic Crisis, Geo-Political Transformation and the Emergent Migration Order, Centre on Migration, Policy and Society, University of Oxford, UK, September 2009. Available from: http://eprints.uwe.ac.uk/12487 We recommend you cite the published version. The publisher s URL is: http://eprints.uwe.ac.uk/12487/ Refereed: No (no note) Disclaimer UWE has obtained warranties from all depositors as to their title in the material deposited and as to their right to deposit such material. UWE makes no representation or warranties of commercial utility, title, or fitness for a particular purpose or any other warranty, express or implied in respect of any material deposited. UWE makes no representation that the use of the materials will not infringe any patent, copyright, trademark or other property or proprietary rights. UWE accepts no liability for any infringement of intellectual property rights in any material deposited but will remove such material from public view pending investigation in the event of an allegation of any such infringement. PLEASE SCROLL DOWN FOR TEXT.
Centre on Migration, Policy and Society, University of Oxford Annual Conference 2009 New Times? Economic Crisis, geo-political transformation and the emergent migration order The effects of the economic downturn on migration from the New EU Member States to the United Kingdom Artjoms Ivlevs Nottingham School of Economics Mario Piacentini OECD, University of Geneva Richard Upward Nottingham School of Economics COMPAS does not have a centre view and does not aim to present one. The views expressed in this document are only those of its independent author. PLEASE DO NOT QUOTE OR CITE WITHOUT PERMISSION OF THE AUTHOR. 1
Abstract The paper studies how the current economic and financial crisis affected migration flows from the new EU Member States (NMS) to the United Kingdom (UK). Using Worker Registration Scheme data, we find a strong correlation between the change in economic conditions at the country of origin and the change in migrant inflows to the UK. The NMS with the highest fall in GDP and the highest rise in unemployment have experienced the highest rise/smallest fall in migrant outflows to the UK. There is no evidence that the diversion of migrant flows has occurred: the falling migrant outflows from the largest NMS sending countries (Poland and Slovakia) to the UK were not matched by the rise in inflows to other EU States. Keywords: migration, recession, EU enlargement, diversion of migrant flows. 2
1. Introduction How do economic crises affect migration flows? Do worsening economic conditions at home make people more likely to search for better opportunities abroad? Or do uncertainties about new jobs in the destination countries make people more likely to adopt wait-and-see attitudes? This paper studies how the 2007-2008 economic and financial crisis affected migration flows from the NMS to the UK. The UK, along with Ireland and Sweden, was among the first old EU States to open its labour market to the new Europeans in May 2004, and has since then experienced a major influx of migrants from the NMS. Close to one million Eastern European citizens registered to work in the UK between 2004 and 2008 (Home Office 2009). The recession can affect migrant flows if it changes the relative balance of opportunities between the sending and receiving countries. In sending countries, lower wages and higher unemployment increase the number of potential migrants. However, in receiving countries, poorer job prospects may deter potential migrants from coming. Given that the current crisis is global, i.e. it affects both the source and the destination, the relative opportunity differential as perceived by a potential migrant may have remained unchanged, implying little change in migration flows. In this paper, we concentrate on one destination country (UK) and several sending countries (NMS). Exploiting the variation in the degree by which the economies of the NMS were affected by the economic downturn, we establish a strong correlation between economic and labour market outcomes at home and migrant outflows to the UK. In particular, we find that NMS with better GDP and employment performance experienced larger falls in the number of migrants going to the UK. In contrast, the NMS which were hit particularly badly by the crisis saw their numbers of migrant increasing. To a certain extent, the crisis allows to test the predictions of the human capital model of migration (Sjaastad 1962, Borjas 2009). Among other things, the model predicts that the deterioration of conditions at home would make the migrant more likely to emigrate. The evidence of this study supports this prediction. The remainder of the paper is organized as follows. Section 2 provides a brief theoretical discussion on how an economic crisis may affect migration flows. Section 3 reviews the 3
post-enlargement migration flows from the NMS to the UK. Section 4 presents data on the extent of the current crisis in the NMS and the UK. Section 5 establishes the relationships between the recent changes in economic and labour market outcomes in the NMS and the changes in the migration to the UK. Section 6 discusses whether any diversion of migrant flows from the UK to other EU countries has taken place, and section 7 concludes. 2. Migration and recession: what to expect? 1 2.1. Inflows The extent to which a crisis may affect migration flows depends largely on the underlying motivation for the migration. Economic migration where inflows and outflows are driven by wages and job opportunities is the most likely to be affected by crises which have labour market consequences. However, labour market conditions may still play a role if migration is motivated by personal (e.g. family reunification), social and political factors. Arguably, humanitarian migration is the least likely to be affected by crisis. This paper focuses on economic migration. To see how the recession may affect the number of people emigrating for economic reasons, we use a model of geographic mobility which views migration as a human capital investment (Sjaastad 1962, Borjas 2009). The model states that, at the individual level, migration decisions are determined by a comparison of the present values of lifetime earnings at home and abroad. In particular, the net gain to migration is given by: Net gain to migration = PV A PV H C, (1) where PV A is the present value of the earnings stream if the person moves abroad, PV H is the present value of the earnings stream if the person stays at home, and C are migration costs expenditure incurred in transporting the worker and his family, as well as the monetary value of the psychic costs. This simple model predicts that the worker will move if the net gain to migration is positive. 1 Parts of this section (including the evidence presented in it) draw on Papademetriou et al. (2009) and Fix et al. (2009). 4
How does an economic recession affect equation (1)? First, the economic downturn at home (lower wages, higher unemployment and higher uncertainty about one s employment prospects in the future) would reduce PV H, increasing the net gain to migration and the number of people willing to migrate. Second, the economic downturn at the destination (lower chances of finding a new job, lower wages) would reduce PV A, reducing the net gain to migration and the number of people willing to migrate. If the recession is global, i.e. if the negative economic cycles are aligned at home and abroad, the net gain to migration may stay the same and incentives to migrate may remain the same. Note that the assumption is made here that migration costs are not affected by the crisis. Considering one destination country (e.g. the UK) on the one hand (keeping PV A constant) and a group of potential migrants (or a group of migration sending countries) on the other hand, the model predicts that people (or countries) which are hit particularly hard by the recession would see the highest drops in their PV H and the highest increases in their net gain to migration. Such people (countries) would then be more likely to emigrate (or see more of their citizens emigrating). In the empirical part of this paper we test this prediction at country level: we use the variation in the change in economic conditions in different NMS to explain the changes in migrant inflows to the UK. 2.2. Outflows The recession and changing opportunity differential are likely to affect not only the inflows but also the outflows, or return flows, of foreign workers. If a migrant loses his/her job during the recession, he/she may choose to return home or move to a third country wherever job opportunities are better. However, a number of initial conditions will affect the likelihood of the return or onward move. First, was the migration move intended to be temporary or permanent in the first place? Arguably, immigrants who intended to migrate permanently are less likely to return home during a recession. Second, how strong are social and family ties in the home and host (or third) countries? This itself is related to the length of time that the migrant stayed in the host country. Longer pre-recession stays will be associated with lower return rates. Third, are the costs of return migration as high as the 5
initial costs of migration, C? Fourth, is it possible to migrate again when the recession is over? Finally, is the migrant eligible for welfare benefits at host country? 2.3. Different types of migrants The recession is also likely to affect different types of economic migrants low-skilled, high-skilled and illegal in different ways. The low-skilled are typically employed in sectors such as construction, manufacturing, hotels and restaurants, and are therefore more likely to suffer from the negative economic cycle. The low-skilled rely on networks to find jobs; as unemployment rises, immigrants are able to pass on less information about employment opportunities. However, despite the worsening employment opportunities, many lowskilled migrants may choose to stay in the host country. First, the prospects for the lowskilled, if they return home, may be bleak compared to the high-skilled. Second, access to welfare and unemployment benefits may encourage migrants to remain in the host country, at least for some limited period. Finally, the financial cost of returning home may be substantial, and the low-skilled may be unable to cover it. High-skilled migrants are typically young, single and without strong family ties in the host country, and hence more likely to return home if they lose their job. If the migrant is tied to the employer by the visa status, the only alternative to return may be working illegally. At the same time, if the visa constraint is not binding, the high-skilled migrants may have greater flexibility of finding another job, transferring their skills to new sectors or even downgrading temporarily. It should also be noted that many high skilled workers work in sectors which are not affected by the economic cycle. For example, more than 15 % of immigrants are employed in health and social sectors in Demark, Norway, Sweden and the United Kingdom. Illegal immigration has been found to be particularly responsive to economic cycles in host and home countries (e.g. there is close correlation between the US-Mexico wage gap and the arrests on the US-Mexico border within the current month). However, there are also disincentives which deter the return of illegal migrants in a recession. First, illegal migrants may fear that they will not be able to return to the host country after the recession. Second, illegal migrants in multi-country economic unions, such as the EU internal market, may be tempted to seek employment across the range of countries. 6
3. Post-enlargement migration flows 3.1. Migration to the EU-15 In May 2004, eight Eastern European countries the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Slovakia and Slovenia joined the EU. Most old EU Member States initially applied transition rules temporary restrictions to the movement of labour from the NMS. Only Sweden applied the full community rule, completely liberalizing its labour market and giving immediate access to the welfare benefits. The UK and Ireland opened their labour markets to workers from the NMS, but restricted access to welfare benefits for the first 12 months of employment. Various old Member States subsequently followed the UK, Ireland and Sweden, and opened their labour markets fully or partially between 2006 and 2008. Currently, only Germany and Austria have kept restrictions on access to their labour markets. However, according to the community rule, their labour markets will have to be fully liberalized by May 2011. The UK the largest of the three countries which opened their labour markets in 2004 experienced the largest immigrant inflows in absolute and relative terms (see table 1). The stock of the NMS migrants increased five times between 2004 and 2007, from 121 thousand to 609 thousand. Over the same period, Ireland saw the stock of the NMS migrants increasing four times, from 44 thousand to 179 thousand. The stock of NMS migrants in Sweden nearly doubled from 23 thousand to 42 thousand. 7
Table 1. Stock of the NMS migrants in the EU-15 Source: Brücker et al. 2009 3.2. Migration to the UK. The major data sources on NMS migration to the UK are the Worker Registration Scheme (WRS) and the Labour Force Survey (LFS). Upon taking up work in the UK, migrants from the NMS are required to register with the WRS. Although there are no penalties for non-registration, there is an important incentive to register: access to welfare benefits after 12 months of work in the UK is conditional on presenting the WRS registration certificate. The WRS data are highly disaggregated, documenting monthly inflows from each NMS to each local authority in the UK. The data are available on migrants age, gender, number of dependents aged 16 or under, dependents aged 17 or over, nationality, date of entry, date started work, hourly wage rate, hours per week, sector, occupation, and intended length of stay. WRS data has two key weaknesses. First, migrants are not required to de-register when they leave the country, and so the WRS data reflect only gross inflows. Second, self-employed workers are not required to register with the WRS. The LFS is a quarterly sample survey of around 60,000 households living at private addresses in the UK. It contains the information on respondents country of birth and year of arrival in the UK. In 2007, out of total sample of 120,000, there were 10,000 foreign-born 8
respondents, including 1,200 from the NMS. The LFS allows one to track the evolution of the stock of NMS migrants, providing estimates for net inflows (inflows minus outflows). However, the LFS estimates may be imprecise for smaller countries as they are represented by a small number of respondents in the sample. Table 2 reports the WRS and LFS data on the NMS migrant inflows and stocks in the UK. The WRS data suggest a lower number of initial applications (migrants taking up a new job) in 2008, compared to 2007 and 2006. The LFS data also show a much smaller net inflow of migrants in 2008 than in 2007 and 2006, which can be attributed both to lower inflows and higher outflows. Figure 1 shows the quarterly WRS migrant inflows from the eight NMS to the UK. A clear downward trend is observed, starting in the 4 th quarter of 2007 and continuing till the 1 st quarter of 2009. The beginning of this period corresponds to relatively high growth rates, both in the UK and the NMS. By the end of the period, all countries, except Poland, were in a recession. The most recent data suggest that the inflows stopped falling in the second quarter of 2009; however, it is difficult to say in which direction the future flows will evolve. Table 2. Inflows and stocks and flows of the NMS migrants in the UK Source: Home Office (2009), the British Labour Force Survey Figure 1. The number of initial WRS applications, by quarter, 2004-2009 9
70000 60000 50000 40000 30000 20000 10000 0 May/June Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 2004 2005 2006 2007 2008 2009 Source: Home Office (2009) Concerning the characteristics of the NMS migrants, Poland has been the most important sending country (66% of total inflows), followed by Slovakia (10%) and Lithuania (9%). The typical Eastern European migrants in the UK are young 81% are 18-34 years old and are employed primarily in the low-skilled occupations. The male-female ratio of the NMS migration is 56:44; only 8% of migrants reported a dependent living with them in the UK; 97% work at least 16 hours/week, and 86% work at least 35 hours/week; 74% earn 4.50 5.99 per hour and 19% earn 6.00 7.99 per hour. Upon arrival, 60% of NMS migrants intend to stay for less than 3 months in the UK (Home Office 2009). Thus, migrants from the NMS to the UK can be characterized as young, working in low-qualified jobs, and coming with an intention of temporary stay in the UK. 10
4. The extent of the crisis This section briefly reviews the extent of the economic recession in the NMS and the UK, focusing on two variables the GDP growth rate (table 3 and Figure 2) and the unemployment rate (table 4 and Figure 3). By the fourth quarter of 2008, all New Member States except Poland and Slovakia had registered a fall in their GDP. The Baltic States seem to have suffered the most. Latvia and Estonia were the first to enter the recession in the 2 nd quarter of 2008, and their economies continued to fall at more than 15% through the first half of 2009. Lithuania has outperformed its Baltic neighbours with the GDP fall of 23% in the second quarter of 2009. The dramatic GDP fall in the Baltic States contrasts with their above-average Baltic Tiger economic performance in 2006 and 2007. Interestingly, Poland has managed to maintain a positive GDP growth throughout the period. The remaining NMS (the Czech Republic, Slovakia, Slovenia and Hungary) and the UK have seen their GDP falling by 5 10% in the second quarter of 2009. To a large extent, the unemployment rate has followed the evolution of GDP. The sharpest increases in unemployment were observed in the Baltic States: the unemployment rate in Estonia, Latvia and Lithuania increased from 4 5% at the end of 2007 to 14 17% in the second quarter of 2009. A Eurobarometer survey conducted in June 2009 reveals that 20% of Latvians, 18% of Lithuanians and 15% of Estonians have lost their jobs as a result of the economic crisis the highest proportions among the NMS. At the same time, there has been only a slight increase in unemployment in Poland and Slovakia, after a long period of sustained decrease in the unemployment rate. At 5 6%, the unemployment rates in the Czech Republic and Slovenia were almost unaffected by the crisis; the unemployment rate in Hungary has increased from 7% in 2007 to 10% in the second quarter of 2009. Thus, the data reveal an important variation in the extent to which the economies of the NMS were affected by the crisis. In the next section we examine whether this variation explains the changes in migration flows to the UK. 11
Table 4. Quarterly GDP growth rates, in %, with respect to the same quarter of the previous year. Source: Eurostat Figure 2. Quarterly GDP growth rates, in %, with respect to the same quarter of the previous year. Source: Eurostat 12
Table 5. Quarterly unemployment rates, in % Source: Eurostat Figure 3. Quarterly unemployment rates, in % Source: Eurostat 13
5. Crisis and migration from the NMS to the UK Figure 4 shows the aggregate quarterly NMS migrant inflows (WRS data) after the 2004 enlargement. There is a clear downward trend in the applications from Poland and Slovakia, starting at the end of 2007. Given that these two countries are the biggest migrant senders to the UK, the fall in the number of migrants from there essentially explains the fall in the total number of NMS migrants to the UK (Figure 1). The downward trend, however, is less evident for other countries. Lithuania, a major migrant sender in the first months after the enlargement, has seen the number of its UK migrants oscillating around 3,000 per quarter during the 24 months to June 2009. Since the fourth quarter 2007 Latvia has exhibited an upward trend in the number of migrants, with a significant increase in the second quarter of 2009. Over the whole post-enlargement period, an upward trend is also observed for Hungary and a downward trend for the Czech Republic. Figure 5 shows the WRS migrant applications relative to home country population. Now the Baltic States Lithuania and Latvia appear to the major migrants sending countries, especially at the beginning and at the end of the post-enlargement period. Latvia, which experienced a major fall in GDP in the beginning of 2009, in the 2 nd quarter of 2009 sent to the UK the same proportion of its population as it did just after the enlargement. The rates for Poland and Slovakia are below the two Baltic States, although the positions of the latter were challenged in 2006 2008. The Czech Republic, Estonia, Hungary and Slovenia have been sending significantly fewer migrants relative to their populations over the whole period. Can the recent changes in the NMS migrant inflows be explained by the relative recent economic performance of the NMS? Figure 6 depicts, for the four quarters between Q3 2008 and Q2 2009, the correlation between fall in GDP and the change in WRS migrant inflows of the same quarter. In the third quarter of 2008, when the majority of the NMS were still experiencing economic growth, the correlation between the growth of GDP and the growth in migrant inflows was negative but weak (R 2 = 0.21). However, as the recession spread to almost all NMS in Q4 2008, the correlation between the two variables became more pronounced (R 2 = 0.60 0.80). From Q4 2008 onwards, a 1 percentage point lower GDP growth was associated with 4 5 percentage points higher migrant outflows to the UK. 14
Figure 4. Aggregate migrant inflows from the NMS to the UK, by quarter. Source: Home Office (2009) Figure 5. Migrant inflows from the NMS to the UK, as % of home country population, by quarter. Source: Home Office (2009) and Eurostat 15
Figure 6. The same quarter year-on-year growth in GDP and migrants inflows to the UK. Source: Home Office (2009), Eurostat 16
A similar picture emerges if the growth of migrant inflows is regressed on the previous quarter growth of GDP (Figure 7), and the year-on-year unemployment differential (Figure 8). For the latter, a 1 percentage point increase in the unemployment rate between Q1 2008 and Q1 2009 was associated with 10 percentage points higher growth of migrant outflows to the UK in Q2 2009 relative to Q2 2008. Figure 7. Growth in WRS migrant inflows and GDP growth of the previous quarter. Source: Home Office (2009), Eurostat Figure 8. Growth in WRS migrant inflows and a 12-months increase in unemployment rate of the previous quarter. Source: Home Office (2009), Eurostat 17
The results of a recent (June 2009) Eurobarometer survey confirm a strong correlation between economic conditions in source countries and migrant outflows. A 1 percentage point increase in the number of respondents who reported having lost their job due to the economic crisis was associated with 9 percentage points higher migrant outflows to the UK in the same quarter (Figure 9). Similarly, a 1 percentage point increase in the number of respondents expecting the worst of the economic crisis yet to come was associated with 4 points higher migrant outflows to the UK (Figure 10). Figure 9. Figure 10. 18
6. Diversion of migrant flows? Despite the strong correlation between economic and labour market outcomes on the one hand and the migrant outflows to the UK on the other hand, the decrease of migrant inflows from some NMS (e.g. Poland and Slovakia) in 2008 could be due to other factors, such as the diversion of migrant flows. After May 2006, most old EU Member States liberalised their labour markets for the new Europeans (the restrictions are currently maintained only by Germany and Austria, and they will have to be lifted by the end of the transition period in May 2011). Could it be that the Polish and the Slovaks simply preferred migrating to other countries instead of migrating to the UK? One alternative to the UK is Ireland, which, along with the UK and Sweden, was the first to open its labour market to the citizens of the NMS. However, Figure 11 shows that since 2006 the total NMS migration to Ireland has been on a downward trend. This applies in particular to the largest migrant sending countries Poland, Lithuania and Slovakia which in 2008 sent 37,000, 4,000 and 3,000 fewer migrants to Ireland than in 2007 (the similar figures for the UK are 51,000, 3,000 and 5,000). Clearly, there has not been a diversion of the NMS migrant flows from the UK to Ireland. Interestingly, the correlation of the NMS growth of GDP and migrant outflows to Ireland is similar to that observed in the UK: higher falls in GDP were associated with lower falls in migrant inflows (Figure 12). Table 6 summarises the evolution of migrant flows in 2008 from the NMS to selected Western European countries (for which the 2008 migration data were available). It appears that for the Czech Republic, Lithuania, Slovakia and Poland, the 2008 decline in migrant inflows to the UK (and Ireland) was not matched by the increase in migration to other European countries. Latvia saw an increase in the number of migrants to all countries, except Ireland, Spain and Sweden, and Hungary sent fewer migrants only to Ireland and Spain. Only for Estonia could the diversion argument possibly hold water: the out-migration from this country to the UK and Ireland decreased by 186 persons in 2008 compared to 2007, while the out-migration to other European countries increased by 313 persons. 19
Figure 11. Ireland: Issued Personal Public Service Numbers to the nationals of the NMS. Source: Department Of Social and Family Affairs Figure 12. Ireland: Growth of GDP and the issued PPSN in Ireland, in %. Source: Department Of Social and Family Affairs and Eurostat 20
Table 6. The absolute change in migrant flows from the NMS to selected EU countries, 2007-2008. UK Ireland Spain Denmark Norway Finland Sweden Nether- lands All, except the UK and Ireland Czech R. -1270-1076 -614 79 1-18 118-434 Estonia -110-76 24 16 135 170-32 313 Hungary 1515-484 -720 140 49 104 242-185 Latvia 160-947 -39 64 162-15 71 243 Lithuania -3375-4285 -1317 222 501 25-3 -572 Slovakia -4830-3381 -660 152 86 28 35-359 Slovenia 5 24-51 22 30 10 11 22 Poland -51010-37262 -8591 2982 277 133-555 3863-1891 Source: Statistical offices of the respective countries 7. Conclusion. Total migrant inflows from the new EU Member States to the UK between July 2008 and July 2009 were 42% lower than inflows over the same period one year earlier. However, the drop was not uniform: migrant inflows from some NMS actually increased. This paper explains the recent change in migrant inflows to the UK by the change in economic and labour market conditions in the NMS. We find that the change in migrant inflows was negatively correlated with the GDP growth and positively correlated with the change in unemployment rates. The Baltic States appear to be affected by the crisis particularly badly, and to have experienced the highest relative increases of migrant flows to the UK. At the same time, there is no evidence that migrant flows have been diverted from the UK to other European countries. 21
References Borjas, George (2009) Labor Economics, McGraw-Hill Brücker H., T. Baas, I. Beleva, S. Bertoli, T. Boeri, A. Damelang, L. Duval, A. Hauptmann, A. Fihel, P. Huber, A. Iara, A. Ivlevs, E. J. Jahn, P. Kaczmarczyk, M. E. Landesmann, J. Mackiewicz-Lyziak, M. Makovec, P. Monti, K. Nowotny, M. Okólski, S. Richter, R. Upward, H. Vidovic, K. Wolf, N. Wolfeil, P. Wright, K. Zaiga, A. Zylicz (2009), Labour mobility within the EU in the context of enlargement and the functioning of the transitional arrangements, Report Prepared for DG Employment. Fix, Michael, Demetrios G. Papademetriou, Jeanne Batalova, Aaron Terrazas, Serena Yi-Ying Lin, and Michelle Mittelstadt (2009) Migration and the Global Recession, A report commissioned by the BBC World Service, Migration Policy Institute, Washington, DC. Home Office, Department for Work and Pensions, HM Revenue & Customs and Communities and Local Government (2008). Accession Monitoring Report May 2004 March 2009, published May 2009, available at http://www.ukba.homeoffice.gov.uk/sitecontent/documents/aboutus/reports/accession_moni toring_report/ Papademetriou, Demetrios G., Madeleine Sumption, and Will Somerville (2009) Migration and the Economic Downturn: What to Expect in the European Union, Migration Policy Institute, Washington, DC. Sjaastad, Larry A. (1962) The Costs and Returns of Human Migration. Journal of Political Economy 70: 80-93. 22