Labor Movement and the European Neighborhood Policy: Trade in Services as a Partial Substitute for Migration?

Transcription

1 Labor Movement and the European Neighborhood Policy: Trade in Services as a Partial Substitute for Migration? Bernard Hoekman * and Çağlar Özden ** Preliminary first draft September 2006 * World Bank and CEPR. bhoekman@worldbank.org. ** Development Research Group, The World Bank. cozden@worldbank.org, Mailing address: MC3-303, 1818 H Street, NW, Washington, DC The views expressed in this paper are those of the authors and should not be attributed to the World Bank

2 1. Introduction The expansion of the European Union (EU) to include an additional 10 new member states in 2004 created a new European neighborhood, both for the EU itself and for the countries in Europe, Central Asia and the Mediterranean that are geographically proximate to the EU-25. Most of these neighboring countries have a variety of cooperation arrangements with the EU, including trade agreements that will lead to bilateral free trade in non-agricultural merchandise products in the coming decade or so. Some have the prospect of accession to the EU, but many do not. All of the southern Mediterranean countries and those in the Middle East belong to this group. One consequence of the expansion of the EU has been the development of a new approach by the EU towards its neighbors: the European Neighborhood Policy (ENP). The ENP spans a number of dimensions of the relationship between the EU and partner countries, including political (foreign policy, security, human rights, democracy), economic (policy, integration, financial assistance), and cultural cooperation. The major innovation of the ENP relative to the status quo is that it offers partners the option of a stake in the EU internal market (Hoekman, 2005). That is, the prospect is offered of integrating partner countries into specific elements of existing and evolving EU structures, law and systems on an à la carte basis. The ENP complements the treaty instruments used since 1995 by the EU and its Mediterranean partners to structure cooperation the Barcelona Process and the associated Euro-Mediterranean Partnership (EMP) agreements. The basic objectives of the EMP are to achieve reciprocal free trade between the EU and Mediterranean countries in most manufactured goods; improve access for agricultural products; establish conditions for gradual liberalization of trade in services and capital; and encourage the economic integration of Mediterranean partner countries. These objectives are pursued through binding bilateral trade agreements and various forms of development cooperation. To date, the EMP has had only limited economic impacts (FEMISE, 2005). Although it is always difficult to determine the counterfactual, on the trade front overall non-oil growth was lower in the post-1995 period than in first half of 1990s. Intra-regional (intra- Mediterranean) trade shares have increased somewhat, as has the net FDI/GDP ratio, which rose from 0.9 percent in the mid-90s to 1.2 percent in 2004.

3 Labor movements are not addressed by either the EMP or the ENP in a comprehensive manner. Indeed, a key (implicit) objective of EU member states in pursuing the EMP was to raise growth in the Mediterranean countries and reduce the incentives for migration from South to North. The economic pressures for such labor flows are generally recognized. The projected population declines in the EU are large, as current fertility levels in every country are below replacement levels. By [2050], the number of people aged in the EU25 will fall by 15 percent, while those 60+ will increase by one-third. The UN estimates some 1½ million migrants per year will be needed to maintain the current working age population in the EU-15 (to 2050). The demographic trends in recently acceded member states are also similar and their accession will do little to resolve the labor shortfall. Labor force growth has continued to exceed employment growth in Middle East and North Africa (MENA) countries, resulting in stagnant or rising unemployment levels in many of these economies. In MENA countries, many of which have been major source countries for migrants into the EU, fertility has been declining as well. However, the population aged will increase by some 4-5 million in coming years and plateau around High unemployment and limited work opportunities, including for the educated, constitute a major challenge for governments, who will need to create some 100 million jobs by As noted by Fargues (2005), the combination of a large cohort of young workers in MENA countries that confront limited opportunities at home and are relatively mobile due to limited family obligations (their parents are still relatively young and family responsibilities are divided among multiple children) create an opportunity for both sides of the Mediterranean to pursue a strategy of enhancing the possibilities for temporary migration. A major practical or rather political - challenge is how to realize the potential gains from such movements. One opportunity that is on the table is to focus on services trade. The EMP/ENP framework provides scope for negotiating specific market access commitments for services, including services provided through the movement of natural persons. In principle this may provide a vehicle through which to enhance the feasibility of workers from the Mediterranean countries to provide services in the EU. While the potential for such trade will not be large enough to address the demographic 2

4 fundamentals, we argue in this paper that a more active pursuit of the services trade dimension offers significant potential gains for both sides. 2. Demographic Patterns A key determinant of migration patterns is the relative wage differences across countries. The wage gaps can be startlingly high and show large variation across countries, over time and by skill level. For example, the largest relative gaps between the North and the South are for unskilled labor which implies they would be the ones to receive the most drastic welfare improvement upon migration to the North see, e.g., Winters (2003), Walmsley and Winters (2005). The wage gaps arise and persist for a wide range of supply, demand and productivity differences. The factor we would like to highlight at this point is the demographic differences between the EU and the MENA countries which are among the main determinants of the migration pressures. According to the UN population Division predictions, the population of Europe will decline to 653 million (the medium projection) from the current level of 728 million (World Population Prospects, 2004). More importantly, the age group will decline by 0.7%, while the 60+ age group will increase 0.90% annually during the same time frame. The share of Europe in the world population will decline from 11.3% to 7.2%. Finally, the dependency ratio in Europe will increase from 46.5 to 74.2 in The elderly will form 65% of the dependents in 2050 up from 50% in The population dynamics of the new accession countries are no different than the EU15 and actually many new member countries face a more dim future. This shift in the composition of the population will have significant effects on the labor markets, social programs and other government policies. Developing and least developed countries share in world population, on the other hand, will rapidly increase even though their fertility rates are also in decline. The highest population growth countries are concentrated in Africa and the Middle East. The total population of Northern Africa and Western Asia (which includes Armenia, Azerbaijan, Georgia, Cyprus but excludes Iran) will increase from 400 million in 2005 to Europe includes EU25 as well as all Eastern European, (including Russia), Balkan and Nordic countries that are not EU members. Turkey is included in Western Asia. 3

5 700 million in 2050 under the medium projection. The age group will increase from 250 million to 450 million while the 60+ age group will move from 26 million to 128 million. Even though there is a decline in the fertility rate (which started in the 1980s) and a dramatic increase in the number of the retirees, the rapid increase in the working population will offset some of the burdens on the society. The dependency ratio will decline from 60 to 52, with the old-population accounting for 40% of dependents, up from 11%. These projections are all in sharp contrast to the situation in Europe. All of this information indicates a rapid demographic transformation of the European societies and economies, resulting in major pressures on social programs and fiscal policies. One potential way to release the pressure, at least in the short term while appropriate policies are implemented, would be to admit migrant workers so that the agedistribution does not become so unbalanced. Abstracting for the moment from the negative public opinion on migration in Europe and the associated political constraints to which we return below MENA countries are a natural source, given geographic proximity, a large supply of young population and the skill match in terms of market needs in Europe. Moreover, as noted earlier, the labor market conditions in the MENA region are not very promising which creates pressures for emigration. There are limited employment opportunities for the younger workers, including the educated ones. The younger cohort is relatively mobile and without the family obligations that burdened previous generations. This is due to the relative young age of their parents and the presence of larger extended families which can share the obligations when a member emigrates. All of these factors imply that emigration will impose a relatively low burden on society but generate large potential gains. One final dimension concerns the changes in the demographic patterns and fertility behavior of migrants after migration. In a recent paper, Fargues (2006) argues that migrants adopt the behavior of their adopted countries in terms of the fertility behavior and rapidly diverge from the patterns in their home countries. As an example, he compares the Egyptian migrants to the Persian Gulf countries and North African and Turkish migrants who went to Europe. The Egyptian migrants fertility levels went up to converge with the very high levels in the Persian Gulf countries whereas North African and Turkish migrants fertility levels went down to the European levels. Furthermore, the 4

6 fertility levels in the home countries among the non-migrants also moved along similar paths. In other words, the fertility rates in the Maghreb countries and Turkey declined while they went up in Egypt. Fargues (2006) argues that migration was an important vehicle in the transmission of fertility behavior from the destination countries to the home countries of migrants via various channels. These can be economic factors, such as remittances and the related improvements in material wellbeing as well higher expenditure on education and healthcare. Or there might be transmission of cultural values, such as increased role of the women in family decisions making process. In short, given the rapid decline of the fertility rates in MENA countries, there is not likely to be a very big window of opportunity to take advantage of the demographic differences in Europe and MENA countries to ease the social and economic pressures in both regions. The current fertility patterns in Turkey, for example, indicate that after 2015 there will no longer be a large supply of young workers for the European labor markets. 3. Migration Patterns from MENA Countries A major concern from the perspective of destination countries is the flood of unskilled workers who will have a hard time to assimilate, compete for native unskilled workers and impose burdens on the social welfare policies. On the other hand, one of the main concerns in the source countries is the brain drain effects of large scale migration. Empirical research on this subject has been constrained by the scarcity of data. One exceptional source of new information on this is a dataset assembled by Frederic Docquier based on the national censuses of the receiving countries. The dataset includes bilateral migration stocks in OECD countries from 174 source countries in 1990 and 192 countries in 2000 by three education levels. A special effort was made towards ensuring homogeneity and comparability of data from different destination countries, given that national definitions of a migrant differ considerably. They also constructed a new unpublished dataset to compare the migrants educational composition to those of the native populations. These are the main sources we use in this section of the paper. Details are in Docquier and Marfouk (2006). Even though it is among the most comprehensive datasets to date, there are several shortcomings. First, it is based on the OECD countries as the destination 5

7 countries and, as such, ignores several major destination labor markets such as the Persian Gulf, Singapore, South Africa etc. Especially the Persian Gulf countries are important destinations for the purposes for this paper, but until these countries make their data public, there is not much can be done. Nevertheless, Docquier and Marfouk dataset captures over 90% of global brain drain. Second, the dataset only includes employed migrants and overlooks current students. Current students, especially in post-graduate programs, are a major component of brain drain and they are not present in the analysis. Middle East Migration to OECD Countries 10,000,000 1,000, ,000 10,000 1, Algeria Bahrain Djibouti Egypt Iran Iraq Israel Jordan Kuwait Lebanon Libya Morocco Oman Qatar S. Arabia Syria Tunisia Turkey UAE Yemen Figure 1 The aggregate migration numbers from the Middle East and North Africa to the OECD countries is dominated by several countries. Figure 1 presents the number of migrants in OECD labor markets from the countries in the region. The largest source countries are Turkey (close to 2m workers in the OECD labor markets as of 2000), Morocco (slightly over 1m), Algeria, Iran (both around 500 thousand), Egypt, Iraq, Lebanon and Tunisia (each around 250 thousand). It should be emphasized that the figure is on a logarithmic scale otherwise these large source countries would visually dwarf the smaller countries. Figure 1 has data for both 1990 and 2000 and all countries have experienced increased emigration during the decade. The overall number of migrants has 6

8 increased by 40% - but this naturally includes the children of the migrants who were born in the destination countries and who are considered migrants in most of the EU countries. Smaller countries have experienced more rapid increases in their migration. Iraq is an exception among the larger source countries, reflecting political instability. The impact of the post-2000 conflicts in the region is not yet reflected in the data. As expected, countries with larger populations have a large number of migrants. Another measure of migration is the migration level as a ratio of the native labor force (Figure 2). On this measure, Lebanon and Morocco stand out. Lebanese migration is linked to decades of political instability as well as the historical integration of the native population with economic centers of Europe. Moroccan migration is also linked to the relatively low cost of migration to France and Spain due to historical links and geographic proximity. The next group of high emigration countries is composed of Turkey, Tunisia and Algeria countries with also high absolute level of migration. Thus, migration is an economic issue for the large, oil-poor countries with rapidly growing populations. It should once again be noted that, there is significant migration to the oilrich Persian Gulf countries from other countries such as Egypt and Jordan which is, unfortunately, not captured in these two figures due to data limitations explained earlier. Migration as a Share of Labor Force 20% 18% 16% 14% 12% 10% 8% 6% 4% 2% 0% Algeria Bahrain Djibouti Egypt Iran Iraq Israel Jordan Kuwait Lebanon Libya Morocco Oman Qatar S. Arabia Syria Tunisia Turkey UAE Yemen Figure 2 7

9 There is significant variation in the destinations of MENA migrants (Figure 3), partially reflecting geography and partially historical linkages. For example, vast majority of migrants from Algeria, Morocco and Tunisia go to France while the Turkish migrants choose Germany, the Netherlands and Belgium. Migrants from ex-british colonies such as Egypt, Iraq and Jordan migrate to the UK as well as the other English speaking new world countries such as the US, Canada and Australia. There is also significant migration to Spain from Morocco and to the UK from the oil-rich countries Gulf countries. Finally, there are a significant number of migrants from Iraq, Iran and Syria in Turkey - this is the missing portions in the figure. This migration is due to geographic proximity and as a transit stop en route to Europe. Middle East Migration to OECD by destionation, % 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Algeria Bahrain Djibouti Egypt Iran Iraq Israel Jordan Kuwait Lebanon Libya Morocco Oman Qatar S. Arabia Syria USA / CAN / AUS FRA UK Germany / Netherlands Spain Tunisia Turkey UAE Yemen Figure 3 The geographic destination choices for the educated migrants from the region are similar to those of overall migration (Figure 4). For example, majority of educated Turkish migrants are in Germany and North African migrants are in France. The most important distinction is that the share of US/Canada/Australia is much higher among tertiary educated migrants for every country. For example, while only 5% of Turks and Moroccan migrants go to US/Canada/Australia, close 30% of educated Turks and 25% of 8

10 educated Moroccans migrate to these countries. There are various reasons for this distinction the most important being that the educated migrants might perform better in these labor markets compared to the European countries. Also, it might be easier US/Canada/Australia to be legally and socially accepted. Middle East Skilled Migration to OECD by destionation, % 80% 60% 40% 20% 0% Algeria Bahrain Djibouti Egypt Iran Iraq Israel Jordan Kuwait Lebanon Libya Morocco Oman Qatar S. Arabia Syria Tunisia Turkey UAE Yemen USA/CAN/AUS FRA UK Germany/Netherlands Spain Figure 4 In the brain drain debate, an important statistic is the ratio of the educated migrants within the migrant population. The educational distribution of migrants is reported in Figure 5, where the missing portions of each bar are the migrants with only primary education. One striking fact is that the tertiary educated migrants are a small portion of the overall migrant population especially in large migrant sending countries such as Turkey, Algeria, Morocco and Tunisia. For example, between 10-15% of migrants have tertiary education in Turkey, Tunisia, Morocco and Algeria as of year On the other hand, more than 50% of migrants to OCED countries from other countries, including Egypt, Iran, Iraq, Lebanon and Jordan have tertiary education. There are several reasons for this difference. First, migrants with low levels of education from many of the smaller and wealthier countries, such as Kuwait, have no reason to migrate since there are plenty of employment opportunities in their own countries. Second, 9

11 unskilled migrants from other larger countries (Egypt, Lebanon, Jordan) migrate to the Persian Gulf countries. As a result, the figures capture high levels of educated migrants from some countries to the OECD (Persian Gulf countries, as well as Egypt, Jordan, Lebanon) and relatively high levels of unskilled migration from the other countries (Turkey and North Africa). Finally, the figures reveal that the share of educated migrants increased from 1990 to Middle East migration to OECD, by education, % 80% 60% 40% 20% 0% Algeria Bahrain Djibouti Egypt Iran Iraq Israel Jordan Kuwait Lebanon Libya Morocco Oman Qatar S. Arabia Syria Tunisia Turkey UAE Yemen secondary tertiary Middle East migration to OECD, by education % 80% 60% 40% 20% 0% Algeria Bahrain Djibouti Egypt Iran Iraq Israel Jordan Kuwait Lebanon Libya Morocco Oman Qatar S. Arabia Syria Tunisia Turkey UAE Yemen secondary tertiary Figure 5 10

12 A major reason why the portion of the highly educated migrants from Maghreb countries and Turkey is so low in the overall migrant population is the rather low level of education among the native population. Figure 6 presents the portion of the migrants who have tertiary education in vertical axis and portion of the native population with tertiary education in the horizontal axis for all countries in the database. Countries above the 45 degree line have migrants who are more educated than the natives. Figure 6: The Selection Effect Migrants from almost every developing country are more educated than the native population. There are many reasons for this selection effect the migration policies of the receiving countries are biased towards educated migrants, they face fewer constraints in terms of financial and social costs when they migrate, the returns to migration are higher etc. The selection effect is stronger for some countries compared to others. For example, among the countries in the region, Turkish migrants are the most similar to native population in terms of education profile around 10% of natives and migrants have tertiary education. Similarly, the bias is relatively small for North African countries such as Morocco and Tunisia. On the other hand, the bias is largest for the Persian Gulf 11

13 countries and Egypt over 60% of migrants have tertiary education as opposed to 15% of the native population. Share of tertiary educated migrants in tertiary educated population 50% 40% 30% 20% 10% 0% Algeria Bahrain Djibouti Egypt Iran Iraq Israel Jordan Kuwait Lebanon Libya Morocco Oman Qatar S. Arabia Syria Tunisia Turkey UAE Yemen India Mexico Poland Caribbean SS Africa Figure 7 A key question in the brain drain debate is what percent of the educated population migrate. It is possible that educated migrants form a large portion of the migrant population. However, if the underlying native population is also highly educated, then the economic impact of migration is not likely to be negative on the sending country. For example, this is the case in India and China. Even though majority of Indian migrants, especially to the US, have college degrees, they still present a small portion of the educated labor force in India. Figure 7 presents data on the tertiary educated migrants share in the overall tertiary educated labor force from the sending country. There is wide variation among the countries in the region. A significant portion of the college educated migrants from poorer countries have migrated. Lebanon and Iran are at the extreme, especially as of 1990, mainly due to political instability. Educated migrants have better prospects in the destination countries, have more resources and face fewer constraints to migrate in times of chaos. The next group is Morocco, Tunisia, Iraq and Algeria with around 10-15% of educated workforce migrating. The middle range includes Jordan, Syria, Turkey and Egypt with 5-8% migration rate among the educated. The migration 12

14 rate of the educated is lower in the Gulf countries reflecting the high demand for their services in their native labor markets. The data reveal that large migrant countries with the exception of Turkey - and politically unstable countries are the main victims of brain drain in the Middle East and North Africa. The main reason of brain drain in the first group is the low levels of education, rather than specific migration of the educated as the vast majority of the migrants are unskilled. Poorer countries, such as the ones in North Africa, also have lower shares of college educated workers in the labor force. A very large portion of their population tends to migrate if given the chance since the economic opportunities seem to be rather limited for everyone, not just for the highly educated. As a result, highly educated migrants make up a smaller portion of the migrant population but they form a large portion of the overall educated labor force. On the other hand, for the wealthier countries, the total number of migrants relative to the population is much smaller and a smaller portion of the highly educated chooses to migrate hence the lower levels in the figure for these countries. But the highly educated form a large portion of the overall migration as in the earlier figures. A second observation is that in many of these victims, the situation has improved in the 1990s as the level of brain drain is lower in 2000 compared to Finally, several large migrant countries and regions are included as comparison at the end of the figure on the right. The brain drain levels are around 15% for Mexico and Poland and 6% for India. On the other hand, it is staggering 43% for the Caribbean and 15% for sub- Saharan Africa (however, the rate is above 50% for many countries in the region). These rates indicate that the problem is not much worse for the Middle East and North Africa compared to other regions of the world. An overlooked issue is where the migrants obtain their education. In the previous section, we established that, in many countries, a large portion of the college educated people born there have migrated to the United States. The key fact we are going to present in the section is that a large portion of those people have actually received or completed their education in the destination countries since they migrated either as children with their families or migrated as young adults to complete their education and stayed to work. 13

15 Figure 8 presents the age-of-arrival distribution of educated migrants to the New World countries the United States, Canada, Australia and New Zealand for Majority of the educated migrants to these countries migrate as adults - after age 22. For example, 87% of Algerians, 73% of Egyptians, 68% of Moroccans and 60% of Turks migrated after completing their college education in their home countries. It is possible that some of these migrants obtained post-graduate degrees in their destination countries, but nevertheless, they can be considered brain drain for their home countries. On average, another 10% of the migrants arrived between the ages 18-22, indicating they migrated to complete their college education. The countries that defy this pattern are the small Persian Gulf countries where majority of educated migrants have arrived as children. Given that these countries send a large number of students abroad to complete their education, these data imply that these students return home upon graduation. Middle East Skilled Migration to New World, by Age % 80% 60% 40% 20% 0% Algeria Bahrain Djibouti Egypt Iran Iraq Israel Jordan Kuwait Lebanon Libya Malta Morocco Oman Qatar S. Arabia Syria Tunisia Turkey UAE Yemen less than 11 yrs. old betw een 12 and 17 yrs. old betw een 18 and 21 yrs. old above 22 yrs. Old Figure 8 Parallel data for France reveals a different pattern, especially for large migrantsending countries in North Africa (Figure 9). For example, only around 30% of Moroccan, Tunisian and Algerian educated migrants arrived after the age of 22. Actually, around 60% have arrived France before age 18 which indicates they came as children and grew up there. Similarly 40% of Turkish and Lebanese migrants the other two large 14

16 groups also arrived as children. There are several reasons for this pattern. Migration to France is a much older phenomenon and the migrant communities from MENA are more established. Thus, the majority of educated migrants in France are actually the children of the previous generation, not recent migrants who consumed the resources of their native countries for their education. 100% Middle East Skilled Migration to France, by Age % 60% 40% 20% 0% Algeria Bahrain Djibouti Egypt Iran Iraq Israel Jordan Kuwait Lebanon Libya Malta Morocco Oman Qatar S. Arabia Syria Tunisia Turkey UAE Yemen Mali less than 11 yrs. old betw een 12 and 17 yrs. old betw een 18 and 21 yrs. old above 22 yrs. Old Figure 9 4. Temporary Entry: Incentive Problems and Design Challenges The data presented above are likely to have strong implications for the debate on brain drain and migration policies. Several stylized facts emerge clearly: (i) There is already a large stock of migrants already in the EU. They are relatively unskilled, especially compared to native populations, but a large portion entered the EU as children and completed their education there. One result is the rising resistance to migration in current public opinion surveys and political debate in Europe. (ii) There is a significant brain drain/waste problem for the MENA countries, but much of this is non-eu in that the brain drain from the region, in relative terms, seems to be flowing towards the New World the US, Canada and Australia. This is partially reflecting proximity effects (migration costs for less skilled workers to enter the EU are lower than they are for North 15

17 America, and thus a larger portion of the unskilled emigrate there ) and partially reflecting the possible skill premium in the North American labor market ; and (iii) brain drain effects are attenuated insofar as many migrants obtain their education once in the EU. Temporary movement of workers from MENA countries, as opposed to longerterm migration would appear to an obvious solution to the excess demand for labor in the EU and the (temporary) excess supply in MENA. *** Such a policy response would both alleviate some of the brain drain problems identified above and address political concerns in the EU. First, educated migrants are likely to be more sensitive to their legal status and job security, relative to undocumented unskilled migrants. Thus, a legal migration framework could attract a portion of the educated migrants from MENA countries to the EU who might otherwise migrate to North America. This would benefit MENA countries as well since the extent of the brain drain would be diminished insofar as educated migrants would be returning home after their contract ends. They would also be geographically closer to their home countries, enhancing the prospects for economic and cultural spillovers to materialize. Among these spillovers are knowledge and technology transfers as well tighter integration of the national economies into the EU markets. Another benefit would be to reduce the extent of brain waste. Migrants would have proper documentation and would be placed in well-specified and job categories that are commensurate with their skill levels. Thus, migrants would not be seeking and placed in relatively unskilled job categories that do not require legal documentation. Given that a large number of migrants complete their education in the destination countries, they obtain valuable human capita and they, are more successful in the labor markets and less of a burden on the education systems of their home countries. Furthermore, there are tremendous benefits to MENA countries if they were to return. One possibility this suggests is to couple a temporary migration program with a Hatton and Williamson (2004) analyze the selection effects that may be associated with migration costs. It should be noted that the brain drain data is based on education levels and hides some of the important skills which are in high demand in service sectors in the European labor markets. That is probably the reason behind the large number of highly educated European workers migrating to North America as well. *** Winters (2003) and Walmsley and Winters (2005) analyze the potential global welfare gains associated with a relaxation of barriers to temporary movement of workers. 16

18 development assistance program that would provide education to students from MENA countries in the EU and allow them to be employed for a certain period of time. This would benefit MENA countries, especially in terms of human capital accumulation of their citizens and knowledge transfer upon their return. Another issue is the importance of other skills, such as language ability, that is not necessarily captured by pure education measurements. For example, a significant portion of the population in many MENA countries speaks a foreign language mostly French and English due to colonial legacies reflected in secondary education systems. These are valuable skills that are not necessarily rewarded in the local labor markets but would be valuable in the EU. For example, many workers might be employed in the (seasonal) tourism sector, say in Greece, to serve English or French speaking tourists. Targeted temporary visa schemes that allow such services to be provided through temporary movement of workers would have significant welfare improving outcomes for all parties. Despite the importance of brain drain issues in temporary migration policies, the main benefits to both EU and the MENA countries are likely to arise from an increased migration of semi-skilled workers employed in service sectors such as home, hospitality and healthcare services. These are the areas where there is much demand in EU markets and the wage gaps are the largest. Unfortunately, these are also the areas where there is much public opposition to increased migration. Numerous surveys have shown the increased negative perception in many European countries towards migration and migrants. Cultural as well as economic characteristics of the natives influence these opinions and perceptions. Evidence provided by Mayda (2005) shows that cultural assimilation and security/crime concerns are relatively more important in shaping attitudes than labor market variables (such as skill composition). The opinion surveys reveal seemingly conflicting attitudes toward migration and the economic interests of natives. For example, older people tend to be more anti-immigration even though they are the ones likely to benefit the most from migrants whether via increased supply of domestic and health service providers or lower dependency ratios. Given that no policy reform can occur without public support, the prevailing public opinion and attitudes suggest that an advantage of exploring the potential for Among these are the World Values Survey, international Social Survey Program and others. 17

19 temporary migration policies is that this may be perceived to be less threatening. By choosing portions of the labor market where there is less competition with native workers or significant shortages in terms of supply, temporary migration programs can be implemented with relatively lower political opposition compared to a general market access program. Furthermore, such segments of the labor market will also be the areas where the wages are high due to supply constraints. Thus, targeted temporary migration schemes will not introduce large economic distortions. We are not aware of much survey data on attitudes towards temporary entry as opposed to migration more generally. However, in a recent EU consultation on development policy the question was asked what should be the objective of EU migration policy. Some 56% of respondents were of the view that migration should benefit source countries, 38% worried about negative effects of brain drain; and 27% thought that temporary movement possibilities needed to be enhanced. While probably not representative of the EU population as the questionnaire pertained to economic development policies this does suggest it may be possible to find support for effective temporary labor movement schemes. The challenge of course is to ensure that temporary movement is indeed temporary. A common presumption of many EU voters and policymakers based on past experience is that nothing is more permanent than a temporary guest worker. This preconception is a major factor impeding a serious effort to negotiate and put in place effective temporary movement schemes. A problem is that the preconception is indeed an accurate one, in that incentives for both employers and temporary entrants to ensure temporariness may be weak. However, if truly temporary, labor inflows will have little impact on demography of EU, thus attenuating political costs/resistance in the EU, while addressing labor force decline and increasing the access of EU citizens to labor services. And, if truly temporary, labor flows will also generate less in the way of brain drain losses, while generating remittance flows and returnees that could help raise productivity in source countries. EU Public Consultation Report, 2005: 18

20 A recent paper by Amin and Mattoo (2005) analyzes the incentives for host and source countries to prefer temporary over long-term or permanent migration. They show that unilateral migration policies are globally inefficient in that they lead to too much permanent migration relative to temporary movement, and too little overall movement of workers. They develop a model in which the perceived net social/political cost of permanent migration in a host country exceeds that associated with temporary entry, and that the latter generates turnover costs for host country firms. Source countries prefer temporary movement as it reduces the brain drain losses associated with permanent migration, but migrants have incentives to stay in the host as their incomes there will be higher than if they return. If host country firms prefer permanence ex post given turnover costs, temporary entry will generate permanence ex post. This is not optimal for either the host (because of the assumed social costs/preferences) or the source (brain drain). However, both host and source countries prefer permanent migration to no migration. Thus, there is a time consistency problem if host firms confront significant costs of worker turnover and migrants have financial incentives to stay. This analysis suggests that the problem is to design self-enforcing cooperation agreements that make temporary entry feasible ex post. This could be done by increasing the cost to host firms to retain workers by taxing them; requiring them to deposit lumpsum payments into an escrow account that is refunded on return of workers; or by increasing worker s incentives to return by refunding social security taxes; or deferring a share of wages (see e.g., Martin, 2006). The implication is that negotiating better access to the EU for workers say through specific commitments pertaining to mode 4 is necessary, but not sufficient for temporariness to be credible/feasible. In addition to time inconsistency problems in host countries, an added complication concerns the presence of undocumented workers in the EU (Martin, 2003). Here actions will be needed on the part of both host and source countries. Amin and Mattoo (2006) show that guest worker programs alone cannot be an efficient tool to combat illegal migration. While beneficial to source countries, the negotiation of additional access to a source country will not by itself be enough to induce adequate action to control illegal movement. Additional incentives will need to be provided for source (and transit) countries to control such flows, e.g., financial or other transfers. 19

21 5. Services Trade as a Partial Solution? Debates on and policies towards migration often are driven by domestic labor market, cultural, security, and justice concerns. Although the powerful demographic forces are recognized, the approach taken towards migration policy is only partially driven by economic considerations. From the perspective of source countries, a problem with migration policy is that it tends to be determined unilaterally. Insofar as there are bilateral discussions and efforts to cooperate, these tend to be very asymmetric source countries are regarded as demandeurs, and access is treated as a concession. The focus tends to be on control of net inflows, not just at the border but also through repatriation and readmission agreements, and through efforts to help source countries realize and sustain higher levels of economic growth with a view to attenuate outward migration pressures. The latter types of efforts are primarily of two types: trade and investment agreements, and development cooperation and assistance. Despite rather extensive development cooperation programs and the implementation of bilateral free trade agreements, economic growth in many MENA countries has been insufficient to create jobs on the scale that is required. Indeed, even if Chinese style growth rates could be achieved in the coming decade, this would be insufficient to address the short/medium term demographics in MENA countries. Given the demography of the EU, it makes little sense in any event not to allow greater entry of workers into the EU. An explicit focus on expanding the opportunities for services trade could both allow greater MENA employment in EU service sectors and improve economic policies and activity in source nations through quid pro quo reforms in these countries. In 2005, a framework protocol for negotiations on services was agreed in the EMP context (Hoekman, 2005). This envisages employing a so-called positive list approach to determining the sectoral and substantive coverage of market access commitments, allowing flexibility in terms of what sectors to include and what types of commitments to make. The protocol also calls for progressive alignment with the acquis in the area of services. 20

22 Because services often cannot be exchanged in a manner where the provider is located in one country and the demander in another, international trade may require either provider or demander to move to the other s location. An implication is that liberalization of trade in services involves movement of people, be managers of affiliates of multinational firms that establish in a market to supply services locally, consultants who move temporarily to provide a service, or consumers who travel on holiday. Because of the non-storable and non-transportable nature of many services, the WTO defines trade in services as spanning not just cross-border exchange through telecom networks and the Internet, but also services where exchange involves movement of providers/consumers, including the movement of natural persons. The latter mode of supply is the salient one from a labor movement perspective. All these different modes of trading services are on the table in EU-MENA partner country negotiations to extend the existing EMP trade agreements to cover services. The European Commission has the mandate to negotiate liberalization on behalf of the EU as a whole, although in practice the EU is not (yet) a customs union when it comes to services trade policies. A key feature of the protocol is that it envisages reciprocal liberalization on a (regional) most-favored-nation (MFN) basis. This standard dimension of trade agreements differs importantly from migration discussions where there is much less, if any scope, for reciprocity and may serve to overcome some of the resistance in the EU to enhance access to its service markets. The more there are specific groups in the EU that see an interest in lobbying for additional access for MENA service suppliers the more likely this becomes feasible. Such groups are likely to involve services firms with an interest in investing in the MENA countries e.g., financial firms, distributors, etc. who need to have a physical presence in order to sell services. Another advantage of pursuing labor mobility through a trade rather than migration framework is that access can be conditioned and managed by tying it to specific sectors or activities, which can include both more skilled and less skilled parts of the labor spectrum. For example, one implication of the demographic changes that are ongoing in Europe is that there is/will be large excess demand for workers in the healthcare and home or personal care sectors. Employers that provide such services could be the focus of specific quotas (visas) for workers from MENA countries who would be 21

23 permitted to enter the EU on a time-limited basis. As discussed previously, incentives could easily be put in place to ensure that these temporary access quotas do not result in permanence by making the firms responsible for the return of workers, by tying the visas to employers, and through forced saving schemes that provide workers with a substantial lump sum payment upon their return. Sector-specific access commitments can also be designed so as to have a minimal impact on domestic labor markets. For example, Kremer and Watt (2006) argue that temporary entry programs that target home work childcare and household/personal services would free up native home workers desiring to do so to enter the domestic labor market. Foreign workers who provide such services would not crowd out domestic workers, would earn incomes that exceed what they would earn in their home countries generating remittance income flows that benefit source countries and may actually increase unskilled wages in the host country. 6. Concluding Remarks Movement of people generates gains for both the EU and MENA countries. Both have good reasons partly political, partly economic to prefer more temporary movement over permanent migration. The economic rationales are particularly compelling from the perspective of MENA countries, given brain drain losses and the temporary nature of the demographic bulge of young workers that are entering the labor market. The problem is that the policy debate tends to be focused primarily on permanent migration and not temporary movement. However, the institutional framework to begin negotiating additional temporary entry for workers exists in the shape of trade agreements. While these have inherent limitations in that access must be restricted to services, much of the demand for workers is in fact most likely to be in service activities. Negotiating additional temporary access is not a panacea. To make temporariness credible, specific measures need to be put in place that address the incentives of EU firms and MENA workers to resist the turnover costs associated with regular mobility. To make temporariness feasible is likely to require transfers as well as more access to induce the South to manage illegal migration pressures. The trade approach to temporary movement of service providers has the major advantage of having to function in a context 22

24 of reciprocity. A services-centric negotiating agenda will include deep domestic policy and regulatory reform commitments that are critical for job creation and competitiveness in MENA, as well as improved access to MENA markets that is of interest to EU firms. Even though the scope to harness reciprocity forces to leverage additional concessions on mode 4 access is likely to be limited given the small size of MENA markets, and recent EU debates and developments around the Services Directive reveal that progress will be difficult, negotiating better temporary access to specific sectors where there is clearly excess demand for workers should be pursued. References Amin, Mohammad and Aaditya Mattoo Does Temporary Migration Have to be Permanent?, World Bank, mimeo. Amin, Mohammad and Aaditya Mattoo Can Guest Worker Schemes Reduce illegal Migration?, World Bank, mimeo. Chaudhuri, Sumanta, A. Mattoo and R. Self Moving people to deliver services: how can the WTO help?, World Bank Working Paper Fargues, Philippe Temporary Migration: Matching Demand in the EU with Supply from the MENA, Analytic and Synthetic Notes 2005/11, Euro- Mediterranean Consortium for Applied Research in International Migration (CARIM). Fargues, Philippe The Demographic Benefit of International Migration: Hypothesis and Application to the Middle East and North African Contexts, mimeo Fargues, Philippe, Jean-Pierre Cassarino and Abdelkader Latreche Mediterranean migration: an overview, in P. Fargues (ed.) Mediterranean Migration: 2005 report. Florence: European University Institute. FEMISE The Euro-Mediterranean Partnership 10 years after Barcelona: achievements and perspectives Marseille: Institut de la Méditerrané. Hatton, T. and Jeffrey Williamson International Migration in the Long Run: Positive Selection, Negative Selection and Policy, NBER Working Paper Hoekman, Bernard From Euro-Med Partnership to European Neighborhood: Deeper Integration à la Carte and Economic Development, ECES Working Paper No. 103, Cairo. Kremer, Michael and Stanley Watt The Globalization of Household production, Harvard University, mimeo. Martin, Philip Managing Labor Migration : Temporary Worker Programs for the 21 st Century, International Institute for Labour Studies, Geneva, mimeo. 23

25 Martin, Philip Population and Migration, in Bjorn Lomborg (ed.), How to Spend $50 Billion to Make the World a Better Place. Cambridge: Cambridge University Press. Mayda, Ana Maria Who is Against Migration? A Cross Country Investigation of Individual Attitudes Toward Immigrants, mimeo, Georgetown University Walmsley, Terrie and L. Alan Winters Relaxing the Restrictions on the Temporary Movements of Natural Persons: A simulation Analysis, Journal of Economic Integration, 20(4): Winters, L. Alan The Economic Implications of Liberalizing Mode 4 Trade, in A. Mattoo and A. Carzaniga (eds.), Moving People to Deliver Services. Washington D.C.: Oxford University Press and World Bank. 24