Charles University in Prague Faculty of social sciences Institute of economic studies

Four Essays on the Economics of Migration

Doctoral Dissertation

Wadim Strielkowski

Supervisor: Prof. RNDr. Ing. František Turnovec, CSc.

Academic Year: 2007/2008

1

Title: Four Essays on the Economics of Migration Author: Wadim Strielkowski Supervisor Prof. RNDr. Ing. František Turnovec, CSc. Institution: Charles University in Prague, Faculty of social sciences, Institute of economic studies Major: Economics Academic Year: 2007/2008

2

I, hereby declare that this dissertation is my original work and that all source materials used here have been clearly identified and referenced.

Wadim Strielkowski, undersigned

…………………………………

3 ABSTRAKT

Tato diserta ční práce se skládá ze čty ř samostatných empirických studií, které spojuje spole čné téma migrace.

První studie se zabývá politicko-ekonomickými modely pracovní migrace v podmínkách jednotného trhu. R ůzné režimy mezinárodního obchodu slouží jako ilustrace k vysv ětlení spot řebitelského přebytku a tzv. „sklon k migraci“ je zaveden a používán jako nástroj k vysv ětlení nevyrovnání mezd na jednotném trhu.

Druhá studie se soust řeďuje na migra ční potenciál a „sklon k migraci“. Snaží se zejména prosadit názor, že migrace v různých zemích je podmín ěna r ůznými faktory a že každá zem ě má z řejm ě svůj unikátní migra ční potenciál.

Třetí studie se v ěnuje problematice migrací vzešlých z regionálních nebo mezinárodních integra čních proces ů. Migra ční studie často neum ějí najít odpov ěď na otázku, pro č i p řes existenci velkých rozdíl ů v ekonomickém rozvoji mezi zem ěmi z ůstávají migrace na nízké úrovni. Jako p říklad takové modelové situace je použit p řípad slovensko-české migrace po rozpadu Federace.

Čtvrtá studie stru čně p řibližuje problematiku azylových migrací slovenských Rom ů do České republiky. Testuje, mezi jiným, jaký vliv na migrace má etnická nebo národnostní p říslušnost.

4 TABLE OF CONTENTS:

INTRODUCTION ...... 7 1. PROBLEMS OF MIGRATION IN ECONOMIC THEORY...... 11 1.1. GENERAL DEFINITIONS AND TYPES OF MIGRATION ...... 11 1.2. TRENDS OF RECENT MIGRATION ...... 13 1.3. THEORETICAL APPROACH TO MIGRATION IN ECONOMIC LITERATURE ...... 16 1.4. REASONS FOR MIGRATION ...... 21 1.4.1. Comparison of migration policies in Europe ...... 23 1.4.2. Economic effects on employment opportunities of natives ...... 26 1.5. MIGRATION AND REMITTANCES ...... 31 1.6. MIGRATION AND MARKET INTEGRATION ...... 35 1.6.1. Introduction...... 35 1.6.2. Free mobility of labor in economic theory ...... 36 1.6.3. Labor migration and migration surplus ...... 39 1.7. SELECTED METHODOLOGICAL ISSUES...... 41 1.7.1. CALCULATING MARSHALLIAN SURPLUS ...... 42 1.7.2. USING BACKWARDS STEPWISE REGRESSION ...... 43 1.8. REFERENCES ...... 44 2. LABOR MIGRATION AND WELFARE EFFECTS OF FREE MOBILITY OF LABOR IN THE COMMON MARKET...... 50 2.1. INTRODUCTION ...... 50 2.2. MARSHALLIAN SURPLUS ...... 51 2.3. LABOR MOBILITY BETWEEN TWO COUNTRIES AND THE INDICATOR OF MIGRATION COST ...... 53 2.3.1. Propensity to migration and indicator of migration cost ...... 55 2.3.2. Welfare considerations...... 58 2.3.3. Illustrative example: numerical model ...... 59 2.4. EMPIRICAL MODELS ...... 61 2.4.1. Empirical model I...... 61 2.4.2. Empirical model II...... 63 2.5. CONCLUSIONS : POLICY IMPLICATIONS OF FREE MOVEMENT OF LABOR ON THE COMMON MARKET ...... 67 2.6. REFERENCES ...... 68 3. READY TO GO? MIGRATION POTENTIAL AND EU ENLARGEMENT: LESSONS FOR THE CZECH REPUBLIC IN THE CONTEXT OF IRISH MIGRATION EXPERIENCE...... 70 3.1. INTRODUCTION ...... 70 3.2. METHODOLOGY & LITERATURE REVIEW ...... 71 3.2.1. Migration patterns and economic shocks: the case of Ireland...... 73 3.2.2. Migration potential of the Czech Republic...... 75 3.3. DATA ...... 77 3.4. EMPIRICAL MODEL ...... 79 3.4.1. Dependence of international (outward) migrations on economic incentives at home...... 80 3.4.2. Dependence of regional migrations on economic incentives at home...... 86 3.5. CONCLUSIONS ...... 89 3.6. REFERENCES ...... 90 3.7. APPENDIX A...... 92 4. WHAT LIVING IS WORTH LEAVING? ACCESSING THE IMPACT OF ECONOMIC INCENTIVES ON MIGRATION FLOWS: THE CASE OF CZECH-SLOVAK LABOR MIGRATION AFTER THE SPLIT-UP OF CZECHOSLOVAKIA ...... 93 4.1. INTRODUCTION ...... 93 4.2. METHODOLOGY ...... 94

5 4.2.1. International migration and the role of economic incentives...... 94 4.2.2. Countries’ profiles: the split-up of Czechoslovakia and the successor states ...... 96 3.2.3. The implications of Czech-Slovak migrations for the European Union enlargement process...... 99 4.3. DATA ...... 102 4.4. EMPIRICAL MODEL ...... 103 4.5. CONCLUSIONS ...... 111 4.6. REFERENCES ...... 113 5. PEOPLE OF THE ROAD: THE ROLE OF ETHNIC ORIGIN IN MIGRATION DECISIONS. A STUDY OF SLOVAK ROMA ASYLUM-SEEKERS IN THE CZECH REPUBLIC IN 1998-2006 ...... 115 5.1. INTRODUCTION ...... 115 5.2. LITERATURE REVIEW : ETHNICITY AND MIGRATION ...... 117 5.3. A SHORT EXCURSE TO THE HISTORY OF ROMA IN CENTRAL AND EASTERN EUROPE ...... 119 5.3.1. Destinies of Roma population in Central and Eastern Europe ...... 119 5.3.2. Roots and destinies of the first Roma settlers in Czechoslovakia...... 121 5.3.3. Roma in the Habsburg monarchy and Austrian-Hungarian Empire...... 121 5.3.4. Roma in the independent Czechoslovakia in the after-WWI period ...... 122 5.3.5. Roma minority in Czechoslovakia during the WWII ...... 122 5.3.6. Attempts of Roma assimilation in the Communist Czechoslovakia ...... 123 5.3.7. Roma in Czechoslovakia after the collapse of the Communist regime in the 1989 ...... 124 5.3.8. The European dimension of the place of Roma minority in Central and Eastern Europe after the collapse of the Communist regimes...... 125 5.4. TYPES AND REASONS OF ROMA MIGRATION FROM SLOVAKIA TO THE CZECH REPUBLIC AFTER THE SPLIT -UP OF CZECHOSLOVAKIA ...... 127 5.5. DATA ...... 134 5.5.1. Summary statistics for the main group of asylum-seekers in the Czech Republic...... 136 5.6. EMPIRICAL MODEL ...... 137 5.7. CONCLUSIONS ...... 143 5.8. REFERENCES ...... 143 APPENDICES...... 146

6 Introduction

This doctoral dissertation deals with the economics of migration. Migration, or an inter-regional or international move of a given group of individuals from the place of their origin (often called “source” region or country in the research literature) to the new place of habitation (often related to as “target” region or country), has become an important issue worldwide at the brink of the 21 st century. In general, migration can take many forms: asylum migration, migration driven by famines and natural catastrophes, migration fueled by economic incentives, culturally-driven migration and many more. The introductory survey of this dissertation provides a brief overview of economic theory findings about migration without any ambition for completeness. Migration potential, migration incentives, propensity to migrate as well as ethnical aspects of migration are among the key topics that are highlighted here. Generally, mainly those issues that could be considered important in the context of migration in economic theory were selected. The introductory survey is followed by the four empirical essays that gave the name to dissertation. It should be noted that all four essays presented in the empirical part of this dissertation were previously published or presented as original research papers, published articles or working papers: the first essay is based on the joint conference paper with Professor Frantisek Turnovec presented at the Charles University in Prague’s International conference entitled “Central Europe in Union and Union in Central Europe” (19-21 st of September 2007), the second essay is based on the joint paper with Professor Cathal O’Donoghue and it appeared as IES Working Paper 74/2005 and was also published in Prague Economic Papers 1/2006, the third essay was created on the basis of the conference paper presented at EUROPEAUM conference at the University of Oxford in September 2002 and it was also published in Prague Economic Papers 3/2007. Finally, the fourth essay is an updated version of IES Working Paper 32/2006. The first essay is concerned with political economy models of labor migration in conditions of common market. In the first part the essay a concept of Marshallian surplus is linked with the problem of measuring welfare effects of different regimes of international trade. The second part of the essay introduces a concept of propensity to migration explaining non-equalization of wages in the common market and investigates welfare effects of free mobility of labor. In the third part a simple illustrative example of equilibrium and welfare effects on labor market is provided. The fourth part of the essay presents two empirical models based on data about migration in the EU. The basic conclusions stemming out from this study are the following: i) free mobility of labor increases total welfare in both emigration and immigration country, but has different effects for different segments of their societies, ii) contrary to the general neoclassical theory of free mobility of production factors labor migration in

7 common market does not lead to equalization of wages due to different propensities to migration in different countries, and iii) while the wage differential represents a significant factor of labor migration, it does not fully explain labor flows between countries. The second essay deals with the migration potential and analyzes the so-called “propensity to migration”. It argues that migration in different countries is likely to be influenced by completely different factors. In addition, it seems that there is a migration potential, unique for every country, that pre-determines the migration or labor mobility. We come to conclusions that migration potential of population of different countries might vary. Our findings suggest that this variation will be in favour of countries with more “mobile” population such as the Republic of Ireland (ROI). More “mobile” population measured in regional emigration and immigration responsiveness to the changing economic conditions at home might means that economic incentives abroad as well as opening of new foreign labour markets due to the political decisions (such as the EU Enlargement) will very likely to lead to larger migration of labour searching for higher wages and employment opportunities. These might explain why there was no mass emigration from the EU-8 to EU-15 after the 2004 Enlargement: migration potential of emigrants from new EU-8 was quite low (potential emigrants did not react enough to the economic changes in their countries and were not highly mobile within their countries). The third essay sets up a framework for modeling migration flows resulting from regional or international integration. Migration studies cannot explain a paradox why migrations from regions or countries with lower economic performance to regions or countries with higher economic performance remain low even though the economic incentives of emigration are high. This gives EU stakeholders solid reasons for creating serious administrative barriers and introducing “transition periods” for free movement of labor from EU-8. It is generally believed that removing barriers would cause mass labor migration. However, it might be that the problem lies elsewhere and labor migration might remain low with or without barriers. The essay analyses the pattern of Czechoslovak migrations in 1993-2004. After the split-up of Czechoslovak Federation citizens of both countries could reside and work in another country without any restrictions. This was even more simplified by the common cultural background and unique language proximity. Although in the first years of transition Slovak Republic was generally less successful in its economic growth than the Czech Republic, the analysis of the model of Czechoslovak migrations specified in this paper does not lead to the conclusion that economic differences between the two countries influence migration between Slovakia and Czech Republic. It might be that economic disparities might influence migration however it happens only after they reach some critical level.

8 Finally, the fourth essay deals with the economic testing of the origins of migrations that are distinguished by the ethnical factor. General evidence suggests that ethnicity might impersonalize strong cultural and socio-demographical characteristics that should not be omitted when accessing the volume of migration or trying to explain its flows. In particular, the case of Slovak Roma asylum migrations to the Czech Republic in the 1998-2006 is used to test an impact of ethnical factor on migration decision. Those asylum migrations of Slovak Roma became a point of controversy in the Czech Republic and triggered off a public debate about situation of minorities and Czech asylum system. The paper analysis the history and the situation of Roma in former Czechoslovakia and describes the main patterns of Slovak Roma migrations to the Czech Republic. In addition, it attempts to find out whether Slovak Roma are driven by push or pull factors and see which of them are most significant in explaining asylum migrations. The conclusion that stems from the economic analysis is that there are mainly push factors (worsening conditions at home) that might be regarded for the main determinants of Slovak Roma asylum migrations. In his book “Heaven’s Door” (1999) George Borjas was describing migration on a basis of “his own immigrant experience” (Borjas, 1999). Quite true, only an immigrant himself can fully understand the issues of migration. The author has had this opportunity, so part of the story told in this work reflects some degree of “insider’s knowledge”. This work attempts to share the new insight on the economics of migration which is based on years of research work and personal experience of the author:
By creating a novel insight of the well-known problems: although migration and economic impacts of migration have been analyzed in the research literature since the 19 th century, the field is so vast that there is always something to say. In particular, this dissertation explores some non-economic factors of migration, such as propensity to migrate (which differs for every nation), ethnicity, cultural and social ties.
By dealing with controversial subjects : parts of this dissertation analyze such sensitive issues as the role of ethnical origin in migration or free movement of labor within the EU; these issue are often suppressed in main-stream economic literature thus bringing them back to the light creates a new perspective for discussion and further research.
By analyzing the problem of migrations from different possible angles: this work does not limit itself by well-defined and well-known issues but rather looks for something that has not been already said – for instance, it deals with asylum migrations of Slovak Roma.
By assigning a great importance to empirical testing: all empirical testing used in this work was built on original data often obtained with hardships and over many constraints; the data reflects the real-life situation and makes the economic analysis and the results obtained more

9 interesting for researcher as well as for policy-makers that are to deal with the issue of migration (i.e. in defining an appropriate migration policy of the country).
By bringing a new perspective into economic knowledge: overall this dissertation brings some new and fresh ideas into economic knowledge; the topic of migration that is of a special interest nowadays due to immense globalization and integration all around the world attributes some special value to this work. In addition, it has to be noted that most of the empirical findings of this dissertation concern Central and Eastern European countries to a smaller or lesser extent. The countries most affected by this research are the Czech Republic and Slovakia. The author hopes that the unique historical experience of these countries together with the in—depth economic analysis of their past and future might be of some interest to the world of science. Not much has been said about this part of Europe in the context of migrations and economics of migrations; the author hopes that this work will trigger off further research that will draw from the rich history and experience of this part of the world. The doctoral seminar Economic Theory of Political Markets (ETPM) at the Institute of economic sciences, Charles University in Prague, which is lead by Professor František Turnovec and which the author was regularly attending, appeared to be an ideal place for shaping, articulating, communicating and discussing scientific findings that later became cornerstones of the research paper used in this dissertation. This work would never be possible without co-operation, valuable comments and never-ending support of my academic advisor, Professor František Turnovec, whose guidance I greatly appreciate. It was an honor to me to work with Professor Turnovec for all these years. I would also like to thank Professor Cathal O’Donoghue from the National University of Ireland, Galway who taught me that there are ways to overcome every trouble. I am particularly grateful to the opponents of this dissertation thesis, Professor Milan Žák, Professor Ladislav Kabát and Professor Jacek Mercik whose comments and suggestions for improving this work have been of a great value to me. In addition I thank the Institute of economic sciences, Faculty of the social sciences, Charles University in Prague, my alma mater that was a source of never-ending inspiration and a vault of wisdom that linger within its ancient walls.

10 1. Problems of migration in economic theory

1.1. General definitions and types of migration

According to Oxford English Dictionary migration is defined as the “movement of a person or people from one country, locality, place of residence, etc., to settle in another; an instance of this” (OED, 2006). In other words, migration can be perceived as the permanent change of residence by an individual or group. This change of residence occurs over such a distance that frequent face-to-face interaction with one’s family and friends in the community of origin is no longer possible (or would require substantial transaction costs). The explanation of migration boils down to the proposition that the expected gains from the move outweigh the expected costs of the move. This is, the market value of the individual must be greater in one place than in another, over and above the transactions costs. It follows from this that people are different in one country than in another (Simon, 1990). This then constitutes the basic explanation of why people decide to migrate in the first place. Most of the research literature on migration decision-making perceives it as a rational process, the one that embodies careful weighting of various alternatives by a person who seeks to maximize his or her utility or to achieve his or her aspirations (Sell and De Jong, 1978). As De Jong et al. (1983) point out: “ migration is seen as rational instrumental behavior, and decision-making is based on a “cognitive calculus” of costs and benefits that involve a subjective, anticipatory weighting of the factors in attaining certain goals…while subjective expectations about where goals can be obtained are significant determinants of the decision to move” . Whilst the process of decision-making of an individual is generally assumed to be of a rational nature and is of no interest as such, a set of main influences impacting and triggering off the migration decision is what employs the researcher. Migration decisions of individuals are thus mainly determined by a complicated set of social expectations and attitudes, which in a various possible combination might shift the weighting scales to one or the other side. Already the definition of migration itself implies some direction towards the establishment of migration typology. In general, there are criteria of separation (e.g. territorial aspect; temporal aspect decision of migration and extent of migration) according to which the kinds of migration (e.g. internal or international; constraint or permanent; free and enforced; individual, group or mass migration) can be distinguished (see table 1.1).

11 In addition to the listed criteria Lucassen and Lucassen (1999) classify migration also as subsistence and betterment migration or according to the border patrol differentiation of good or bad migrants. There is also a theory of the chain migration. Chain migration is defined as “ a process in which previous migration […] is an influence on the migration decisions of present migration ” (Wegge, 1998). Migration theory argues that once a flow of emigration to a target country has started, it encourages other potential emigrants back home to join in (see for example Massey, 1986; O’Rourke, 1994).

Table 1.1: Typology of Migration Migration criteria and aspects Types of migration Territorial aspect
Internal migration (from rural to urban areas)
International or external migration Temporal aspect
Constraint or temporary migration (e.g. seasonal workers, students, holiday takers, prisoners)
Permanent migration Decision of migration
Voluntary migration (e.g. workers scientists, students, company managers, family reunions)
Involuntary (forced) migration (e.g. war refugees, expulsion, escape due to environmental catastrophes) Extent of migration
Individual migration
Group migration
Mass migration Source: Treibel (1999)

The reasons leading to migration and the decision to move are reached through an evaluation of the incentives and obstacles to migration. Here push (unbearable or threatening conditions in the home country) and pull factors (incentives in the country of immigration) play an important role. Incentives may contain increased employment opportunities, better housing, or a more sympathetic political or cultural environment. Obstacles to migration are unfamiliarity with the new location, lack of information about distant opportunities, language barriers, transportations costs or difficulties and immigration or emigration restrictions. The pattern of migration is more or less the same everywhere in the world: people leave poorer countries in search of higher-paying jobs or achieving higher standards of life in wealthier countries. This one-way flow of migration is often interrupted by the so-called “return migration”, when certain groups of emigrants are returning back to their home country as a response to an economic boom: an example can be Ireland which has been experiencing sustained high growth rates since 1994 (and has gained a label of “Celtic Tiger” ever since) which resulted in a net inflow of immigration the majority of which were Irish emigrants from UK and USA (Barret et al., 1998).

12 The short exposé above that described basic types of migration is a beginning of a more complex insight into this problem. Further down this section trends of recent migration as well as the place of migration in economic theory will be discussed in somewhat greater detail. In addition some remarks concerning the input of the whole doctoral dissertation to the economic theory will be outlined.

1.2. Trends of recent migration

International organizations, such as United Nations often proclaimed that international migration will be “one of the greatest challenges of the 21 st century”(UN, 2006). Quite true, nowadays we are facing various sorts and kinds of migration: as the world gradually turns into one global village, more and more people are on the move. Migration has played crucial role in the economics of many countries over time and is gradually becoming even more important issue today, with the opening up of international borders and integration and globalization processes throughout the modern world. Migration that affects not only the country of origin but also the host country is a global issue that has a double-folded impact on both sides. Today millions of people in the world live in a country other than the country of their origin (IOM, 2007; Friedberg and Hunt, 1995). The most common destination for migration is the United States of America, Canada and Australia. These countries have been created by immigrants and with the help of migration flows that provided labor and settled in unoccupied territories. The long-term strategy of these “traditionally immigration countries” has always been to draw more immigrants from Europe and later from the third-world countries. Perhaps, the best country to study immigration and its implications is the United States of America (US). Most of the studies dealing with immigration chose their example. This is done for two main reasons: (i) ever since immigration figures have been available it is clear that the US has been the most common destination; (ii) most of the economic literature that has been written about the economic impact of immigration relates to the US. Immigration from poorer countries to the US grew after the immigration act of 1965 which opened doors for immigrants from non-European countries. Between 1951 and 1960 a lot of immigrants came from the developed countries such as Canada, Germany, Italy and the UK, and indeed the only developing country among the top five was Mexico (Freeman, 1993). According to Friedberg and Hunt (1995) immigration has became an increasingly important component of population growth in the US rising from 13 per cent of growth in the 1960s to 19 per cent in the 1970s, and increasing to grow over 25 per cent in the 1990s (Friedberg and Hunt, 1995). The immigration act of 1965 abolished the National Origins Quota system of the 1920s, so that the country’s composition of immigration changed greatly since 1965. Over half of all immigrants in the 1950s came from Europe; in 1993 however only 18 per cent of immigrants were European, almost 40

13 per cent of recent immigrants came from Asia and a roughly equal amount originated in Mexico, the Caribbean and Latin America (Friedberg and Hunt, 1995). Most of immigrants therefore are from countries that are poorer and less educated than the United States. On the other hand, European migration is often under-estimated but it still remains an important factor in the population structure of many European countries. For example, 11 per cent of the population of France, 17 per cent of Switzerland, and 9 per cent of the United Kingdom are foreign- born (Borjas, 1994). As a result of this the impact of immigration is being more likely to be more debated in Europe. This issue is becoming increasingly important since the creation of the European Union (EU) and its consequent Enlargements. The question of the free movement of labor in Europe that would eventually lead to the creation of the EU Single Market is one of the most debated today, as far as contrary to the principle of Single Market, restrictions and barriers to the free movement of workers are applied. The disparities in the life-level and income in the West and East of Europe are not new. There have always been considerable differences between the former Communist block and the Western Europe, EU respectively. With regard to this also political migrations from the East to West should be mentioned: due to political repressions and oppression many open-minded people were forced to leave the Socialist bloc and settle down in the West. There were also labor migrations, of course, but their rate was negligible. Only after the fall of the Communist regimes in the CEE countries and opening to the trade and migration with the EU Member States the movement from the poorer countries to the richer ones in a search of the higher income became a reality. Until the income gap is there, the pressure to migrate will remain. The recent history of European migration is quite a long one. One of the great movements of population took place in Europe in the 1950s and 1960s and included emigrants from Southern Europe. This was a period of high economic growth and low unemployment in Western and Northern Europe. Some of the migration was sponsored by employers but some by the workers themselves. Germany recruited employers in Turkey and Yugoslavia with the assistance of German Federal Labour Institute. The Italians (mainly represented by the emigrants from the agricultural and low- developed regions of Southern Italy) were moving into Switzerland and France, and Spaniards and Portuguese poured into France. The total net migration northward from these Mediterranean countries between 1950 and 1970 was about 5 million people (Layard et al., 1992). However, everything has changed within Europe since those days. European unemployment rose rapidly, which caused the fall in the northward movement of population between countries. According to the UN World Population Monitoring in 1989, the number of migrants from southern European countries to Northern countries (defined as France, Germany, Benelux, Switzerland and Sweden)

14 decreased by 30% from 1970 till 1985. The fall in the number of Spaniards living and working in the North of Europe was more than 40%; for the Yugoslavs it was a little bit less – 23%. The stream of migrants started to dry out (UN World Population Monitoring, 1989). One of the reasons was that from the 1974 onward all European countries stopped issuing further working permits to primary immigrants, though relatives continued to come in. The reason for the change of policy was partly unemployment and partly pre-1974 build-up of communal tensions, which was then fueled by rising unemployment. Since 1974 the number of Southern European workers in the North has fallen. According to Layard et al. (1992) two provisional conclusions can be drawn from this chronicle:
Very large migrations can occur for purely economic reasons, despite vast cultural and linguistic differences. They tend to peter out, if jobs become hard to find.
Previous channels of migration help, but if employers want labor new channels can develop rapidly (Layard et al., 1992). In addition to that, one more remark can be added to the conclusions above: the so-called “economic reasons” that Layard et al. (1992) points out have to be very large in order to overweight cultural and linguistic differences. There may be some mass migrations, especially in a search of employment or better conditions of life, however in order to happen they have to be triggered out by some major disaster such as the war conflict, natural disaster or devastating economy downfall. The tendency in most of the EU candidate countries nowadays does not reveal any of these signs. Thus cultural and language differences, together with various restrictions on obtaining a job without permission may be of a significant role in preventing people to move abroad. No spectacular labor migration has materialized within the current EU area together with the joining states represented by the countries of the Eastern and Central Europe. The European Union’s efforts to reduce further any remaining barriers to mobility helped some individuals, but on the whole they have hardly created a new momentum for migration. This does not mean, however, that movement between countries will come to a standstill. Indeed, in some fields of employment worker migration might actually increase:
Among skilled and highly qualified labor. Specialists, managers, technicians, personnel with special qualifications and executives are in demand everywhere. This includes intra- company transfers, such as when multinational business sends specialists and executives to another member state for a fixed term.
In border regions of EU countries, working in the neighboring country while living in the home country – or vice-versa – is being more readily taken for granted.
Short periods of residence for work experience, business trips, consultations, training etc. are on the increase.

15 Nevertheless, these trends cannot be described as the full proof of the mass labor migrations in the future. Overall, the research evidence suggests that the size and direction of migration flows are strongly determined by international differences in economic opportunities, and that the migration flows from the acceding countries to the current Member States of the EU will likely be relatively small.

1.3. Theoretical approach to migration in economic literature

a) Neoclassical economics approach

The first references to migration can be found in the works of Adam Smith (1776) and Ernest George Ravenstein (1889). The basic assumption of these authors was that individuals were maximizing their utility subject to a budget constraint. According to the neoclassical approach real wages in the source and the target country are the main reason for migration. Migration mainly occurs because of geographical differences in the demand and supply of labor markets. Regions with a shortage of labor relative to capital are characterized by a high equilibrium wage, whereas regions with a large supply of labor relative to capital are faced with low equilibrium wages. This wage differential causes a migration flow from low wage to high wage regions. In response to this migration flow, the supply of labor in the high wage region increases; subsequently, the wage in this region falls. Similarly, due to migration, the supply of labor in the low wage region decreases and the wages in this region rise. The migration flow ends as soon as the wage differential between the two regions reflects the costs of movement from the low wage to the high wage region. As a result, the model argues, labor migration emerges from actual wage differentials between regions, i.e. the larger the wage differential the larger the migration flow (Bauer and Zimmermann, 1999). This framework was later developed and extended by Todaro (1968, 1969) and Harris and Todaro (1970). In order to explain rural-urban migration in less developed countries they dropped the neoclassical assumption of full employment in the sending and the receiving region and included consideration of the probability of employment in the destination region by migrants. With this extension it is possible to explain the observed large migration flows from rural to urban regions, although the urban regions are often characterized by a scarcity of jobs. The prospect of finding a high-paying job in urban regions causes labor migration out of rural areas, even though this migration could lead to unemployment. Contrary to the pure neoclassical theory, migration in this extended model is determined by expected, rather than actual earnings differentials. The key variable for

16 migration is earnings weighted by the probability of finding employment in the destination region. Several modifications of the basic Harris-Todaro model have been developed to make it more realistic. However, these modifications do not change the basic findings of the original model. In economic theory there are two hypotheses regarding the movement of workers within integrated markets. According to the integration theory, the creation of a common market leads to increased economic welfare when workers are allowed to move to an area where their productivity is greatest and the wages they can earn are hence also highest. This assumes a movement away from the less productive to the more productive workplaces in the integration area. This process last until marginal productivity and, at the same time, wages (for the same jobs) have reached equilibrium throughout the area. The underlying assumptions are, of course, that the workforce is mobile, that information about available employment is easily accessible to all, and that there are no obstacles in the way of migration such as legal barriers (work permits, residence permits), non-recognition of qualifications, or cultural and language differences. By contrast, classical foreign trade theory starts from the premise that a workforce does not move between countries. Given that theory, different endowments of production factors (natural resources, capital, the state of technology, labor) a balance between countries and an increase in prosperity is achieved through trade between them. Each country concentrates on producing the goods that give it a comparative advantage in that it can produce them more cost-effectively than other countries (Hecksher-Ohlin theorem). Foreign trade then brings about a division of labor which corresponds to these comparative cost advantages. The countries engaging in this trade grow more prosperous as a result. According to this theory labor migration is not always necessary. Foreign trade is seen as a substitute for migration. Quite apart from this, capital is generally seen to be more mobile than workers. Generally, the neoclassical theory can be used to predict the exchangeability of labor migration and free movement of goods. Suppose that there are two countries or two regions. One of them is labor abundant and the other is capital abundant and there exists free trade between the two. Further, suppose that producers in both countries (or regions) have the same technology. The disparity in relative scarcity of labor and capital in the two countries causes different prices of input factors and thus also different prices of labor-intensive and capital-intensive goods. The region with relatively scarce capital will have comparative advantage in labor-intensive production and will import capital- intensive goods and specialize in production of labor-intensive goods. Trade between the two regions will decrease wages in labor-scarce region and will also diminish rate of return of capital in capital- scarce area. Wages and rate of return of capital will equalize in the long run. In this situation the

17 temptation to migrate will disappear. Therefore labor migration and trade flows may be seen as substitutes.

b) Push- and pull- migrations

Reasons for migration can be divided into so-called pull-factors and push-factors. Pull-factors start working in a situation when the level of income that can be earned in the potential country of immigration is higher than in the home country and, in addition, there is low employment (or low earnings) in the home country. When both sets of factors exist in two countries, there is in principle a corresponding migratory “push” or “pull” between them. For migration to actually take place the situation must be fully transparent for the workers concerned, the immigration itself must be legally permissible (leaving aside the illegal migration) and there must not be any other barriers such as language, cultural differences, or a great geographical distance. Demand-pull migration and supply-push migration defined in line with shifts in the aggregate demand and supply curves of the receiving economy are outlined in Zimmermann (1995a) and Bauer and Zimmermann (1999). Figure 1.1 represents a standard price-output diagram with an upward- sloping supply curve. If aggregate demand increases, output and prices (or wage) rise. With rising wages, it seems beneficial to allow immigration in order to avoid inflation and to obtain a further increase in output. Hence, the supply curve shifts downward, and the distance AB in Figure 1.1 is pull migration. Conversely, an inflow of migrants without a change in demand shifts the supply curve downwards; prices fall, while output rises. Hence, the distance AC in Figure 1.1 is push migration. A different case of push migration occurs, if, due to a supply shock, (a reduction of native labor supply, for instance), the supply curve shifts upwards (say from equilibrium point C to A in Figure 1.1). This is compensated by immigration, so that the equilibrium moves again down the aggregate demand curve (Bauer and Zimmermann, 1999). Push migration arises from various sources: positive economic conditions in the target countries relative to the sending regions as measured by variables such as unemployment, wages, working conditions, social security benefits, and the structure of economy. Demographic determinants such as size and age distribution of the working population as well as family migration and inflow of asylum seekers and refugees are also considered to be push migration (Bauer and Zimmermann, 1999). It is, however, a question how strongly the different factors operate in affecting the net desire to migrate. International migration almost never reflects exactly the desire to migrate. In the twentieth century there have almost always been controls, and the strength of the controls has constantly varied, making econometric estimation virtually impossible (Bhagwati, 1976).

18 Figure 1.1: Push and pull migration and the economy

Source: Bauer and Zimmermann (1999)

For further integration within countries this problem does not arise, and there are many studies that confirm the influence of comparative wage levels (actual and expected), comparative unemployment rates and unemployment benefits, the availability of housing, the costs of migration, travel costs and the social costs of adaptation, thus the factors influencing on migration (Grossman, 1982).

c) Network migration and formation of immigrants’ clusters

Migration theory stresses the importance of immigrants’ networks. Networks pass information about the country to which people wish to migrate. Such knowledge can come from previous experience with the country, or from migrants who already live there and who can make it easier for their compatriots to immigrate. Most importantly, the existence of networks is determined by the direction of migratory flows, but can also intensify existing movements, such as those of Algerians to France (colonial history) or Turks to Germany (joining one’s family). Cultural and geographical proximity are likewise important. For Germany’s neighbor Poland (although this can be scarcely applied to the Czech Republic, which is Germany’s another close neighbor) such considerations are certainly significant: in 1998 there were 69,000 Poles working in Germany. This accounts for almost two-thirds of all Polish workers in the EU (Honekopp and Werner, 2000). The theory of immigrant’s networks is closely interconnected with the migration clusters’ approach. Immigrant’s clusters formation has long been of a specific interest among economists and

19 social scientists. Clustering, or geographical concentration of migrants from the same source region who are distinguished by the same ethnic origin, background or language is frequently observed in many countries. For instance, there is a large immigrants’ community of Tamils in Switzerland, Turks in Germany, Italians in Argentina, Albanians and Bosnians in Italy, Ukrainians in the Czech Republic or Mexicans in the U.S. While in some cases the formation of immigrants’ clusters is influenced by the geographic proximity of the host and target countries (i.e. Mexico and U.S. or Albania and Italy), in other cases we have a reason to believe that there are some other reasons standing behind the whole process of immigrants’ clustering. It seems that in a large number of cases not only ethnic and language similarities do matter but also spatial characteristics create basis for such clusters’ formation: for instance, there are Macedonians from Skopje who constitute large immigrants’ community in Gothenburg, Sweden. The story of these clusters is quite simple: a person from a small town or a rural community moves to the foreign country. After a considerable period of time necessary for settling down this person is followed by the closest family and relatives. Those people are then followed by their relatives, friends and neighbors and so on. In such a case local knowledge and “knowing-your-neighbor” which is mostly developed in dense communities where people know each other well appear to be of greater importance than the ethnic or cultural ties. Generally, in the existing migration literature there are two basic approaches to be distinguished that concern the theory of formation of migration clusters: i) network externalities and ii) so-called “herd behavior” (see for example Winters et al., 2001 or Bauer et al., 2002). Beneficial network externalities appear when a network of established immigrants reduces the costs of relocation to a foreign culture by providing assistance in finding a job as well as food and shelter to the recent immigrants. Thus, by joining the ethnic community the immigrant increases his positive utility (see for example Gottlieb, 1987; Grossman, 1989; Stark, 1991 and Church and King, 1993). On the contrary, negative network externalities arise when high productivity immigrants do not want low-productive immigrants (Stark, 1991; Stark, 1995) or if growing number of foreigners in a specific location increases competition for jobs and lowers wages. In this case immigrant will not be able to profit from the presence of his compatriots in the given location. Herd behavior encourages migrants to discount private information. According to this hypothesis an emigrant will follow the flow of other immigrants assuming that they have the information that is not obvious to him (Bauer et al., 2002). This clearly shows that there are three particular aspects to economically motivated migration: trading relations, income differentials and labor market situation.

20 1.4. Reasons for migration

The research literature on migration suggests that there exists a very complex set of inter- relationships between social, psychological background, social structural and educational factors, as well as certain constraints on individual’s migration decisions. Individual’s motives or reason for migrating require explanation. As De Jong (1983) suggests, they are mainly explainable in terms of individual’s locations in the economic and social structure, and level of educational background: because everything to what the individuals are aspiring to is given and pre-determined by their educational level and social-economic or socio-cultural characteristics. However, individuals’ believes or perceptions about how satisfactory local or foreign communities may be, also depends on the nature of the local economy or labor market (mainly on such factors as the rurality of the community and its unemployment level). Thence, a direct linkage exists between the nature of the home community’s economic structure and satisfying one’s economic aspirations locally. The constraining factors are thought to affect migration decisions directly, as well as to affect the probability that the presence of motivating factors will lead to a decision to migrate and, in addition to that, the probability that once a decision has been taken to migrate that it will proceeded with. It appears that there are three main types of social psychological factors predicting migration decisions. First, there are economic and social status, or social mobility, motives. These are primarily beliefs about the extent to which one’s occupational status and income aspirations can be achieved in one’s present location or in some alternative community (Hannan, 1970; Richardson, 1974). Hannan (1970) found that “ beliefs about one’s ability to fulfill occupational and income aspirations…were the most closely associated with migration intentions ” (Hannan, 1970). While Walsh (1984) in a larger replication of Hannan’s study found that such motivating factors were still as influential in the later 1970s as they had been in the 1960s (Walsh, 1984). As Walsh (1984) shows in his paper published in the same year, two different aspects of such aspirations are relevant: the level of aspiration – primarily the occupational status and income level aspired to; as well as the beliefs that something cannot be achieved in the current community but can be done in the alternative one. These variables may act independently – in the same way as both the number and the level of occupational opportunities varies. As Walsh (1984) points out, since most immigrants are young and are at the beginning of their working life migration can be generally looked at as a very important part of their strategy to maximize mobility within the individual “career cycle” (Walsh, 1984). Besides income and occupational aspirations some other important social psychological factors are of relevance, particularly an individual’s attitude toward his/her community or culture: high levels of

21 alienation from it or high attachment to it significantly influences migration decisions. Hannan (1970) found these “solidarity” motives to be of some importance and together with “family obligations” to stay locally, they are the main “solidarity motives” in his analysis. Walsh (1984) also found these to be important – though like Hannan (1970), much less so than economic and social mobility motives. Although some researches argue that it is mainly such emotional or attitude factor that influence migration decisions, most of the research evidence available at that point indicates that such emotional factors do not out-weight purely rational evaluations of the balance of likely advantage to goal achievement of staying locally or migrating. The most close, social psychological, factors affecting migrating decisions, therefore, are thought to be: (i) beliefs about the attainment of important goals or aspirations locally – primarily occupational and income goals for young individuals; (ii) solidarity motives – attitudes which either attract or repel people away from their own locality or from local institutions; (iii) certain individual or personal characteristics which are likely to affect people’s willingness and ability to migrate – particularly their own extent of self-confidence and sense of personal competency De Jong et al. (1983). The existing literature on migration focuses almost exclusively on the question as to why people migrate, even though the vast majority does not (Fischer, Reiner and Straubhaar, 1997). It is only recently that increased attention has been paid to why it can also pay to stay put. The following reasons are often highlighted:
Some knowledge, skills, and experience are not transferable, as they only apply at their place of origin. Examples of this are work-related preferences (business philosophy, a company- specific range of products or production processes, knowledge of customer behavior etc.) or leisure-related advantages (social environment, friends, shopping facilities, housing market etc.)
Most people are risk-average. If they move to another country, they expose themselves to incalculable risks, as they do not posses all the relevant information they need and cannot assess their own ability to adapt.
There is a danger of covert or overt discrimination. As a rule the danger is greater the more the immigrant differs from the current population – by language, appearance, qualifications or income.
It can be worth to choose the option value of waiting, if by doing so uncertainties are alleviated and information gaps are reduced. In the meantime the situation in the home country may improve. The decision to migrate is postponed and ultimately abandoned.

22
There are some further reasons that should be given here, although they are no longer relevant in the context of the EU freedom of movement’ regulation: border controls, legal barriers (residence and work permits), and differences in the social benefits systems. Migratory movements in present-day Europe follow the economic and political changes and upheavals on the continent. Generally, they can be broken down into three major groups: a) political changes in Europe, b) labor shortages during the economic development of the 1950s and 1960s, c) large economic differences that motivated illegal immigrants from Eastern Europe, but also Africa and Asia. After World War II, Europe experienced four major migration movements. The first one lasted from 1945 until 1960 due to war displacements (for instance, larges-scale deportations of people of German origin from Poland and Czechoslovakia or repatriations of Poles to the “newly acquired” lands that used to belong to Germany). The second one is dated by the mid 1950s and 1973 when there was a strong labor demand resulting from the strong economic development of Western Europe. The third one started in 1973 and was mainly family migration. The fourth major migration movement which started in the 1980s was initiated by the liberalization of Soviet Union policy. Labor immigrations contributed strongly to Europe’s post-war economic development. The European migration differed from those of the U.S. or Australia because it was intended to be temporary. Permanents migration was not envisaged, because the immigration was only occurring to deal with short-run shortages in labor. This situation seems in theory at least, to be beneficial to everyone: the host country solves a temporary disequilibrium problem in the labor market, and the migrants’ home country benefits from the experience and knowledge that the workers gain abroad. Today labor market conditions are different: unemployment is becoming a major problem and there is not excess demand for labor. There are however some sectors of the economy which still depend on foreign labor, simply because some jobs are not attractive or well-paid enough for native workers.

1.4.1. Comparison of migration policies in Europe

What follows is the comparison of two varying models of migration policies in Europe (represented by Switzerland and the Federal Republic of Germany). The countries that have been selected for comparison apply different approach to labor migration. An important aspect of any migration analysis is to examine the migration policies and how successful they are. The pros and cons of both approaches is discussed and analyzed in greater detail

23 Swiss migration policy as an example of restrictive migration policy

With the economic boom of the 1950s and 1960s the demand for labor in the Swiss economy induced a strong inflow of migrants and pre 1963 migration policy was aimed solely to satisfy the expanding demand for labor. Then, starting from 1963, measures were taken to reduce the percentage of foreign workers in Switzerland. This changed Swiss migration policy and increased the role of the state in regulation of migration. The state intervened directly in the employment decisions of Swiss firms. Switzerland distinguished between four categories of foreigners: a) Foreigners with a yearly work and residence permit. The permit may be reduced based on labor market requirements, such as no Swiss worker being available for the job b) Seasonal workers, where workers in a job with a seasonal nature may stay in the country up to nine months c) Commuters (“pendlers”) who have no permission to live in Switzerland, but are granted work permits to work in the country d) Foreigners with a permanent work and residence permit, who after 5-10 years on uninterrupted stay are granted the same rights in the labor market as the Swiss themselves. Thus, it can be seen that the Swiss policy follows a rotational principle and the new permits are almost all temporary. This scheme has proved to be relatively successful. According to Dhima (1991) who had been analyzing the behavior of 1981 migration cohort for the period of ten years, 63 per cent of immigrants returned home between 1981 and 1989.

German migration policy as an example of liberal migration policy

After 1955 the German economy experienced a strong upward swing, which was accompanied by the increase in the number of job vacancies. Bilateral recruitment agreements between Germany and Spain, Greece, Turkey, Portugal and Yugoslavia which were signed in the 1960s lead to an increase in foreign workers in the country to 1.2 million in 1965 (a 5.5 per cent of the total labor force). The peak of foreign workers in the country was reached in 1973 with 12 per cent of the labor force. German recruitment policy was based on the assumption that foreign workers immigrated on a short term basis. This, however, did not happen and became especially evident after 1973 when many immigrants brought in their families. In the 1980s another increase accompanied by economic decline in Germany changed the government’s attitude in favor of an active return policy. In 1983 three basic instruments were brought in to motivate return: financial assistance, reductions in return barriers and consultation of potential returnees. The aim was to encourage between 220.00-250.000 foreign workers to leave.

24 The main difference of German migration policy was that Germany followed a more laissez-faire approach, where migrants themselves basically decided their length of stay. Return programs when they came were based on financial incentives (Dustman, 1992).

Comparison of both policies

It is obvious that the Swiss migration policy is by far the most beneficial. In times of labor shortages a required number of temporary work permits can be granted to foreign workers. In case the labor shortage has been solved, the work permits can then easily be repealed. The German policy was beneficial for a long time but after the contraction of the economy in the 1973 the problems with laissez-faire policies became evident. When the economy slowed, unemployment became a problem. Germany could not lay off immigrants that easy or substitute them by native workers in the periods of economic decline because of union laws. That is why it had to come up with financial incentives to encourage immigrants to return home. Laws which control the numbers of immigrants are then seen to be the most beneficial in the host country in the long-run. Such “traditional” immigrant host countries like the U.S. did not really have this problem because of the numerous immigration acts which controlled the number of visas allocated. The only problem for the U.S. were illegal immigrants, and these only held low-paid jobs anyway which natives might be reluctant to accept. The main goal in a return migration policy was to satisfy the needs of an expanding labor market. Whether a migrant is temporary of permanent will solve some bearing on how the migrants will perform in the host country. A migrant is permanent if she remains until retirement and temporary if she returns before then. A migrant that returns before retirement will condition her behavior not only about the future economic situation in the host country, but also on expectations of the future in the home country. Permanent migrants on the other hand form their expectations solely on the situation in the host country. Let us consider two identical workers in a West European manufacturing company to show how the migrant’s status (temporary or permanent) affects their performance in the host country. Worker 1 is a native and therefore intends to spend her entire future in the home country. Worker 2 is a migrant from a low-wage country, who has to return home after a limited contract has expired. Her job will yield a far lower wage if she were to do it in her home country. Consequently, for worker 2, leisure time is relatively more expensive in the host country than in the home country: while for worker 1 leisure time costs the same in all time periods. Worker 2 will therefore choose to work harder in the host country, where working time has a higher value, and work less in the home country in the future as work time has a relatively low value. Worker 1 will work equally hard in all time periods.

25 The optimum policy for the policy-makers to enforce given this analysis would probably be one of temporary migration with the government deciding the length of stay. This would be the most successful way to ensure that immigration will improve welfare in the receiving country. Therefore, policy-makers must use effective policy tools to reconcile the demand for cheaper labor with the desire to limit the social and political impact of migrations.

1.4.2. Economic effects on employment opportunities of natives

According to Friedberg and Hunt (1995), the theoretical predictions of the impact of immigration on the wages of natives depend on the employed model. In a model of a closed economy, immigrants will lower the price of factors with which they are imperfect substitutes and raise the price of factors with which they are compliments (Friedberg and Hunt, 1995). For example, consider an economy where production takes place using capital and skilled labor, which are complimentary and unskilled labor which is a substitute for the two other factors. If immigration of unskilled workers occurs, the wage of unskilled workers will consequently fall. The effect on the return to capital and the skilled wage will be ambiguous, because a fall in the unskilled wage will induce employers to substitute away from capital and skilled labor to unskilled labor. However, since the greater supply of unskilled labor means the optimal output is higher, this scale effect will induce employers to use more of all inputs. Immigration will lower the wage more if immigrants are willing to work for less than natives, which is plausible in the case of illegal immigrants. An influx of such immigrants not only shifts labor supply, but also makes it more elastic. In the most commonly employed open economy model; the Heckscher-Ohlin model, the results are quite different. If the country is a large one, the increase in output of poor intensive goods spurred by immigration will reduce the world price of these goods. This may reduce the wage. Any change in the wages of native groups will be accompanied by changes in native employment of hours worked. If we take the example of an influx of unskilled workers into a closed economy, although total employment of unskilled workers will increase, the fall in the unskilled wage will cause some natives to leave the labor force and the employment rate of natives may fall. These models do not directly predict that unemployment will result from immigration, yet they predict movement of factors between production factors of different intensity (Friedberg and Hunt, 1995).

Labor market assimilation

26 An alternative argument on the effects of immigration on earnings and employment opportunities of natives is made by Nixon (1993) who examined the effects of immigration by developing the argument into a Chiswick versus rivalry, simply by the respective use of cross-sectional versus cohort analysis between the two. Chiswick (1978) examines the effects of foreign birth and length of time in the United States on the earnings of foreign-born males arguing that differences such as foreign males lower level of education, fewer weeks worked, higher age and smaller likelihood of getting married can in some cases be expected to lower immigrants earnings, while in other cases can be expected to increase the earnings of the foreign labor (Chiswick, 1978). Chiswick’s hypothesis is one of the so- called “labor market assimilation” where recent immigrants are likely to have fewer of the characteristics associated with the higher earnings of the natives. Due to the fact that they are recent arrivals, they also have less “insider’s” knowledge of the customs and the language relevant to US jobs, job opportunities and less firm specific training. As time passes however, immigrants begin to gain knowledge which narrows the deficiency in the wage gap. Also the initial earnings deficiency in the wage gap would be smaller, the greater the similarity between the country of origin and the country of immigration (represented here by the US). Chiswick (1978) argues that as the earnings rise with time in the US, the earnings of foreign-born immigrants may exceed that of natives (Chiswick, 1978). He predicts this by implying that for the same schooling, age and other factors immigrants to the US have more innate ability and motivation relevant to the labor market than the natives do. He states that: “ Other things the same, the earnings of the foreign-born are 9.5 per cent lower than natives after five years in the country, equal after thirteen years and 6.4 per cent greater after 20 years” (Nixon, 1993). Orcutt and Regrets (1989) support these claims: by constructing a model of human capital investment they come up with some interesting conclusions. In particular, they find that the immigrants will have faster earnings growth than the natives due to both greater human capital investment and greater value over time of source country skills. It was found in 1988 the median annual wage for immigrants was 2.3 per cent higher than that of natives and that 13 per cent of immigrants aged 25-34 were enrolled in school, roughly double the corresponding percentage for natives (Orcutt and Regrets, 1989). Simon (1990) agrees with Chiswick on this issue, he states that within a few years in the US the average contemporary immigrant male or head of family comes to earn more than does the average native head of family. Much of this superior earnings performance and the relatively high educational level is due to the fact that immigrants immigrate when they are young, and educational levels have been rising all over the world (Simon, 1990). Alongside with this Borjas (1990) points out that only highly-productive individuals migrate because: “only persons with exceptional ability, drive and motivation would pick up everything they own, leave family behind and move to another country”

27 (Borjas, 1990). Borjas (1990) argues against Chiswick’s theory by asking why these highly-motivated individuals do not remain at home and become successful there, although fails to consider some important factors such as personal climate, taxation and other regulations immigrants used to face in their country of origin (i.e. neglects the “push factors” of migration). According to Borjas (1990) the analysis of immigration cannot be that simple which means that Chiswick’s study might bear some empirical fault: he states that the notion that immigrants earnings catch up to and then surpass those of natives is preposterous (as many as 33 per cent of immigrants leave the US within a decade, most of whom had held low-paid jobs, which clearly represents labor market failures). On the contrary, recent immigrants in the US earn 10-25 per cent less than natives, but after 10-20 years reach parity, and remain at parity until retirement (Borjas, 1990). The comparison of results of Borjas (1990) and Chiswick (1978) can be seen at figure 1.2 below: Borjas’s findings are represented in Panel B, as compared with Chiswick’s findings in Panel A). Borjas (1990) carries his argument further in attempt to find out whether immigrants increase the unemployment of the natives (in his model he assumes that immigrants are substitutes in the labor market). The main findings can be described as the following:
Women and all male inputs are substitutes
Immigrant men are compliments with native men but substitutes for women
Black men and women are substitutes
Immigrant men are compliments with black man

Figure 1.2: Comparison of Chiswick (1978) and Borjas’s (1990) results on earnings of immigrants in the US

Source: Nixon (1993)

28 In fact, Borjas (1990) says that immigrant men in the US narrow black/white wage differentials. Statistically, Borjas finds that the impact of non-Hispanic immigrants on the earnings of native males is always positive and usually significant. The impact of Hispanic immigrants on native men’s earnings is neither positive nor negative. In all, immigrants tend to be weak compliments for some groups and weak substitutes for others (Nixon, 1993). Simon (1990) agrees with Borjas to some extent on the effect of immigration on unemployment. He concludes that the correlation is relatively low and taking everything into consideration (i.e. immigration levels, national employment levels and the like) immigrants do not have a great effect on unemployment levels. Simon goes further and looks at the effect of illegal immigrants on employment in the host country. Loss of jobs by natives and reductions in natives wages are blamed on immigrants and on illegals in particular. American workers are hurt by being forced into competition with illegal immigrants who tend to work hard and do not complain about unsafe working conditions and do not create trade unions. Claims are being made that it is illegal immigration that is preventing unemployment in the US from being brought under 4 per cent (Simon, 1990). In a report published in 1996 Borjas states that a very optimistic picture of immigration had been painted in the early 1980s. This was one of the immigrants arriving at an economic disadvantage but opportunities improving over time. Within a decade or two their earnings would overtake those of natives, they did not harm native employment, and were less likely to receive welfare assistance than natives. Borjas disagrees with this optimistic opinion and believes that the story has altered radically in the last ten years. New research has established a number of points:
The relative skills of successive immigrant waves have declined. In 1970 latest immigrant arrivals on average had 0.4 fewer years of schooling and earned 17 per cent less than natives. By 1990 immigrants had 1.3 fewer years of schooling and earned 32 per cent less than natives,
The rate of assimilation is not very rapid, so the earnings of the newest arrivals may never reach parity with the earnings of natives. Recent arrivals will probably earn up to 20 per cent less than natives throughout much of their working lives,
The large-scale migration of less skilled workers has done harm to economic opportunities of less skilled workers in the US. Immigration may account for up to 1/3 of the recent decline in the relative wages of less skilled native workers in the US,
New immigrants are more likely to receive welfare assistance than earlier waves. About 21 per cent of immigrant households participate in some social-assistance program.
The increasing welfare dependency may create a substantial fiscal burden on the most affected areas,

29
There exists a strong correlation between the skills of immigrants and of their American-born children. In effect, immigration has set the stage for sizeable ethnic differences and socio- economic outcomes for the years to come (Borjas, 1996). It seems that no matter how much immigration policy of the US changes in the future, circumstances have already been set in motion that will alter the economic prospects of native workers, and the costs of social insurance programs not only in the present generation in the US but also in the generations to come. Simon (1995) is much more optimistic on the economic consequences of immigration and draws on a substantial body of support for his ideas. Fix and Passel (1994) review on the effects on aggregate earnings of natives across industries and come with the following findings: “Immigration has no discernible effect on wages overall. Wage growth and decline appear to be un-related to immigration – a finding that holds for both skilled and unskilled workers” (Fix and Passel, 1994). Butcher and Card (1991) tracked wages in 24 major cities across the US during the period from 1979 to 1989 using census and population data. They conclude: “ the evidence confirms earlier conclusions, that there is little indication of an adverse effect of immigration, either cross-sectionally or within cities over time” (Butcher and Card, 1991). La Londe and Topel (1991) have also come up with some interesting conclusions after intensive research in this area. They studied immigrants and natives earnings and education in the US censuses and found that although immigration has small effects on equilibrium wages, virtually all of the burden falls on the immigrants themselves. Labor markets effects for non- immigrants are non-negligible. These results imply that any adverse effects of current immigration on the US labor market and on native welfare will be small (La Londe and Topel, 1991). For instance, a 10 per cent increase in the size of the immigrant population reduces native earnings by 0.2 per cent. The overwhelming consensus seems to be that native and immigrant workers are, on average, weak substitutes in production. Many argue that blacks are the one group whose economic progress is most likely to be hampered by immigrants (see for example Melvin and Glick, 1982; Briggs, 2004). The recent econometric evidence finds no proof to support this claim For instance, according to Borjas (1990) recent econometric evidence cannot establish any instance in which an increase in the supply of immigrants had a significant adverse impact on the earnings of natives (Borjas, 1990). Fix and Passel (1994) study that has been already mentioned also deals with the effects of immigration on the wages of minorities and the poor and also sheds some light on immigration effects on African-Americans. According to them immigration has no negative impacts for black workers taken as a whole according to the evidence. But less skilled black workers in high immigration areas with stagnant economies might be adversely affected. Generally, native African-Americans in areas of high immigration fared better than

30 native African-Americans in low immigration areas, in terms of wage growth. In high immigration areas, however, native African-American wages do not keep pace with the rising wage trend that immigration brings for Anglo’s and Hispanics. Immigrants increase the labor market opportunities of African-Americans in strong local economies but reduce them where demand is weak. Thus, increased immigration may hurt African-Americans in period of recessions and help them in periods of economic growth. Immigration increases the percentage of the overall labor force but reduces the weekly earnings of less-skilled African-American men and women (Fix and Passel, 1994).

1.5. Migration and remittances

The discussion of the labor migrations would be incomplete without the analysis of the effects of migration on the labor market of the source country through the remittances. Although many studies focus on the effects of migration on the economic prospects of natives (for instance Borjas, 1994 and 1995 discussed further in this chapter), however, the impact of migration on the labor market in the home country has received far less attention. According to Drinkwater et al. (2003) the most obvious effect of migration on the labor market in the source economy is that, in itself, migration should lower the unemployment rate by reducing the supply of labor. However, it seems that this relationship might be far more complex. Due to the fact that those who migrate are usually the most skilled workers, the total effect on the labor market of the source country might be negative. But it does not always happen like this. For instance Stark et al. (1998) show how a positive employment probability in the host country provides an incentive to increase human capital formation because of higher returns to human capital overseas. Beine et al. (2001) outline the conditions under which a brain drain can be beneficial to the source economy. Domingues Dos Santos et al. (2000) also suggest that migration can improve the efficiency of the home country's labor market since migrants can diffuse back the knowledge they have obtained overseas and return migrants will bring both physical and human capital back with them. However, it can be argued that one of the most important effects of migration on the labor market of the source country is due to remittances or remittance transfer. According to Biller (2007) a “remittance transfer” represents “the transfer of money from an individual, usually a person who has emigrated from her city or country of origin, to another individual, usually a relative who remains at home” . Remittances to developing countries officially recorded by the World Bank through Remittance Service Provider (RSP) that includes transfers via banks, credit unions, and wire-transfer

31 services constituted about $199 billion worldwide in 2006, while total remittances in 2006 (including funds transferred via informal channels) were $268 billion (World Bank, 2006). According to Drinkwater et al. (2003) evidence from 20 lower and middle income countries indicates that 3.46 per cent of their GDP is equal to remittances. Apart from official forms of transferring funds other informal forms of remittances exist – the most notoriously known include Hawala and Hundi. These remittances that are based on mutual trust between the remittee and the broker are carried out without being recorded by any official financial system. Undoubtly, remittances represent an important channel of leveling unemployment and increasing welfare. For example Djajic (1986) shows that those who stay in the source country may also benefit from migration if migrants send a sufficiently large amount of remittances back home. In general, remittances can have two opposing effects on the labor market of the source country. First, remittances might actually increase the incomes of the unemployed individuals in the source country. Second, remittances might be invested (by the firms in developing countries that are financially constrained and lack capital). In this case the net effect of remittances is far from obvious and they can actually decrease the unemployment rate in the source country. There are a number of studies that are dealing with the effects of remittances in Less Developed Countries (LDS). For instance Funkhauser (1992) studied the effect of remittances on the participation decisions in the source country. According to him the receipt of remittances from emigrant workers abroad could reduce the participation rates of their household members (in particular females). He used cross-section data from El Salvador to estimate labor force participation. The findings suggest, among all, that having a household member abroad had a negative and significant effect on the participation decision of males and a negative but insignificant effect on the participation decision of females. In addition, remittances have a negative and significant influence on the labor force participation of both males and females. Therefore, he concludes that for males, the negative income effect from remittances dominates all other effects but for females the positive but small effect of the local labor market is enough to outweigh the negative remittance effect (Funkhauser, 1992). Zachariah et al. (2001) provided evidence that the effect of remittances can be very similar to those of welfare payments. They report the results of a case-study from Kerala, India which shows that the worker-population ratio was 55 per cent amongst non-migrant households but only 31.6 per cent in migrant households. In addition, unemployment rate was 20.8 per cent and 8.1 per cent for emigrant and non-emigrant households respectively. A suggestion that comes from these findings is the following: employment seekers from emigrant households are more selective with regards to their job match because they have the financial support of the emigrant members (Zachariah et al., 2001).

32 Galasi and Kollo (2002) studied the literature on remittances in the framework of total effects of unemployment benefits on the incidence and duration of unemployment. According to them, the outcome of remittances can be similar to the effect of increasing unemployment benefits and hence the individual's replacement rate and their reservation wages. They argue that an increase in benefits does not necessarily cause the diminishing interest for employment. This happens due to the following reasons: i) longer job search period can produce a better job by allowing individuals to find their most suitable (Burtless, 1990), ii) an increase in the reservation wage as a result of more generous benefits also increases the number of job offers an individual rejects then this will increase the number of vacancies for other job seekers (Atkinson, 1981). Another important aspect is the use of remittances in the source country. Generally, remittances can be either spent on consumption or re-invested. With regard to this remittances may increase the dependency and have negative impact on the income inequality in the home country. In addition to that, in case most of the remittances are spent on goods and services this could cause inflation which leads to excessive wage claims. On the other hand, remittances can be equality enhancing and have a positive impact on the development of poor areas, especially if they are invested in productive activities (Drinkwater et al, 2003). The majority of remittances studies find that remittances are generally spent on consumer goods (such as food and clothing or housing). For instance, Durand et al. (1996) report that 10 per cent of their sample of Mexican migrants to the United States who reported that they sent remittances or brought savings back with them spent at least some of the remitted money (i.e. “migradollars”) productively. In addition to that, 14 per cent reported that they spent some of their migradollars on housing and the remaining 76 per cent reported that they spent the migradollars only on consumption (Durand et al., 1996). Studying the remittances in Greece in 1971 Glytsos (1993) estimates that only 4 per cent of all migrant remittances sent to Greece was invested in machinery and another 4 per cent was invested in small shops, whilst 63 per cent was spent on consumption, 22 per cent on housing and 7 per cent on land. This matched the amounts spent on housing and consumption for the similar study on Pakistan done by Gilani et al. (1981). Therefore, it seems that the majority of remittances in LDCs are spent on everyday consumption. However, there is another use of remittances. According to Glytos (1993), Adams (1998) and Rozelle et al. (1999) even though a small proportion of remittances may be invested directly by migrants or their families, remittances can be channeled into productive use by the banking system. For instance, Mesnard (1999) and Kule et al. (2000) found that the majority of entrepreneurial projects started by return migrants were financed by their own savings. Leon-Ledesma and Piracha (2001) studied the impact of migration and remittances on employment performance of Central and Eastern European

33 countries (CEECs). They adopt a positive view of the relationship between migration and development and claim that most of the CEECs migration is temporary and that the main sources of the migrant' savings from overseas are used productively in the home country (Leon-Ledesma and Piracha, 2001). The case of remittances and return migration is also explored by McCormick and Wahba (2001) study of remittances in Egypt. They come to a conclusion that return migrants invest in projects and enterprise in their own country and that the duration of stay overseas increase the probability of opening an enterprise. According to some estimates one out of every ten people participates in some form of remittance transfer. Of the estimated US$ 167 billion of formal remittances that transpired in 2005, almost US$ 50 billion went to Latin American and Caribbean countries (LAC) with nearly US$ 22 billion bound for Mexico alone. The LAC represents the highest volume, and most rapidly growing, market for international remittances and is therefore of special focus throughout this brief (Biller, 2007). Apart from the United States among the largest net remitters are also Germany and Saudi Arabia. On the other hand, net receivers of remittances are such countries as Mexico and India. According to IMF Balance of Payments Statistics remittances that go to developing nations exceed the amount of official aid (Official Development Assistance, ODA) and are equivalent to 90% of foreign direct investment (FDI) received by those same countries (IMF BOPS, 2006). Some words should be said about the remittances in countries which do not belong to LDSs. Truly, not only the developing countries are among the recipients of remittances. According to Migration News (2007), the country receiving the most worker remittances from those abroad more than 12 months in 1995 was Portugal, with $3.8 billion, followed by Mexico ($3.7 billion) and Turkey ($3.3 billion). In addition to that, it can be shown that amongst top ten countries in remittances in 1994 was France, which was followed by Mexico, Portugal, Egypt, Greece, Turkey, Italy and Brazil. Although remittances have been practiced for generations and it is well-known that they represent an important channel of the redistribution of wealth and helping economic development of LDSs, a large share of remittances passes by unnoticed. Due to the importance of monitoring and studying remittances (especially due to the recent development in the world and the rise of international terrorism) new ways of recording them should be found. This might happen due to new regulations proposed in many Western countries. It is believed that tighter control over remittances transfers would allow increasing global security and studying international migration in a more effective way.

34 1.6. Migration and market integration

1.6.1. Introduction

According to Molle (1994) economic integration has two major aspects: the first concerns markets of goods, services and production factors, and the second aspect concerns policy (Molle, 1994). While the policy aspect which describes the main stages of integration of economic policy (economic union, monetary union and economic and monetary union) has limited relevance to the topic of this dissertation, the market integration aspect is concerned with its main scope and therefore should be described in a greater detail. Generally, market integration is proceeding in several stages. According to Mirus and Rylska (2002) three progressively higher levels of market integration can be distinguished:
“Modest” integration which means an agreement to apply symmetric preferential treatment of imports and assign supporting functions and instruments to jointly operated institutions. An example of this stage can be NAFTA agreement which includes commitment to eliminate tariffs among its members, dispute settlement provisions and the various working groups and committees that serve to facilitate trade and investment among the three partners.
Harmonization of instruments over which the parties retain control (different national approaches): migration of workers, competition policy, and production standards. An example of this stage might be the Single European Act of 1987. Among other provisions this act applied the “principle of mutual recognition” to product standards. More co-operation and supranational institutions, such as a joint tribunal on competition policy, are also characteristic of this second level.
Coordination of national policies and the creation of further supranational bodies which entail not only economic but increasingly political integration. An example of this stage can be the story of European integration with the creation of a common currency and central bank, and even a supranational parliament (Mirus and Rylska, 2002). Table 1.2 shows the main stages of market economic integration. It can be shown that each successive stage includes implications from the previous stage added by some additional provisions. As it stems from the table 1.2, the stage of a common market (CM) is reached when the obstacles to free movement for not only goods and services, but also production factors among partners are removed. More superior stages to the common market are Economic Union (EU) and Economic and Monetary Union (EMU).

35 Table 1.2: Stages of market integration Basic Elements of the Stages of Economic Integration Free Trade Agreement (FTA) Zero tariffs between member countries and reduced non-tariff barriers Customs Union (CU) FTA + common external tariff Common Market (CM) CU + free movement of capital and labor, some policy harmonization Economic Union (EU) CM + common economic policies and institutions Economic and Monetary EU + common currency or linking of national currencies through the Union (EMU) fixed exchanged rates Source: Holden (2003)

1.6.2. Free mobility of labor in economic theory

According to Mundell (1957) perfect mobility of factors results in a factor-price equalization (Mundell, 1957). The general analysis of the effects of labor mobility can be analyzed in the framework presented in Krugman and Obstfeld (2005). Suppose that there is a static model with two countries: country P and country H. Both countries have the same technology but different overall land-labor ratios. If country P is a labor-abundant country, workers in P will earn less than those in the country H, while land in the country P earns more than in the country H. This obviously creates incentives for the production factors to move. The workers from the country P will be willing to move to the country H. Suppose that the borders between the two countries are open and workers are able to choose the country where they can settle down. Obviously, the workers from the country P will move to the country H. This movement will reduce the labor force in the country P and will raise the real wage in this country. On the contrary, the labor force in the country H will increase and the real wage will decline. According the predictions of the economic theory, this process will continue until the marginal product of labor is the same in the two countries. Figure 1.3 illustrates this process: the horizontal axis represents the total world labor force. The workers employed in the country P are shown on the left, the workers employed in the country H are shown on the right; the left vertical axis shows the marginal product of labor in the country P and the right vertical axis shows the marginal product of labor in the country H. It is assumed that there are OL 1 workers in the country P and L 1O* workers in the country H. The real wage rate is lower in country P (point C) that in the country H (point B). If workers can move freely to whichever country offers that higher real wage, they will move from the country H to the country F until the real wage rates are equalized (Krugman and Obstfeld, 2005).

36 Figure 1.3: Causes and effects of international labor mobility

MPL MPL*

B A

C

MPL MPL*

O 2 1 O* L L Home employment Foreign employment

Migration of labor from Home to Foreign

Total world labor force Source: Krugman and Obstfeld (2005).

However, in real life nothing is as simple as in the model. An issue of wage equalization in the common market is a controversial topic. Although some authors provide the supporting evidence for the real wage convergence (especially using the data from the so-called “The Age of Mass migrations” 1 (Williamson, 1995), others consider this assumption to be too much simplistic (see for example Caves, 1960; Young, 1991; Rassekh and Thompshon, 1998). For instance Caves (1960) called the assumption of the equalization of wages a “supreme example of non-operational theorizing” (Caves, 1960). Economic theory offers justification for decrease as well as increase in wages among countries (Molle and van Mourik, 1989b). Generally, there are two schools: the convergence school (represented for example by Meyer and Willgerodt, 1956; Fisher, 1966; Butler, 1967 or van Mourik, 1993) and the divergence school (represented by Giersch, 1949; Seers et al., 1980 or Tovias, 1982). The convergence school assumes that free labor mobility will lead to the elimination of any differences in wages. The divergence school says that the conditions for equalization of wages will be never reached. Free movements of labor are most likely to be hampered by the economic and social

1 The period identified in the literature as the end of the 19 th century and the beginning of the 20 th century. This period was remarkable for the large movements of population: specifically from Europe to the United States of America, Canada, Australia, New Zealand and some countries of Latin America.

37 factors such as: costs of moving and settling down in the new country, cultural distance, language barriers, social ties, etc. The divergence schools advocates that even in a customs union or common market free movement of production factors can be impeded by transport costs, collusive practices or multinational firms monopolizing new technologies. The technological advance of certain countries implies that they will always select the new product with a high value-added as soon as they come to market, abandoning products as soon as their value-added drops and no longer sustains higher wages. Therefore, the wage gap that accompanies the technology gap is not only perpetuated but even accentuated (Molle, 1994). It appears that the rigorous testing of both alternative views can be very difficult because of insufficient data (Tovias, 1982; Gremmen, 1985, van Mourik, 1987; Molle, 1994). However, in practice one can rather observe the equalization of interest rates on the common market rather than the equalization of wages. Economic literature comes with several explanations for that. One of the most widely mentioned explanations is the imperfect factor mobility (especially labor mobility) among countries. As Faini et al. (1999) point out international migration is the “great absentee” in the liberalization global factor markets (Faini et al., 1999). Net annual emigration rates in the developing countries represent less than 0.1 per cent of total populations, although the GDP per capita in these countries is less than 1/10 of that in the developed world (World Bank, 2005). Even in case of poor countries and regions that are situated very close to substantially wealthier regions (and therefore the costs of emigration are very low) such as Eastern and South-Eastern Europe, Northern Africa and Central America the net annual emigration rates are about 0.15 per cent (Brücker and Schroder, 2006). The introduction of free movement in the context of the EU Southern Enlargement did not increase immigration from the South, and opening the labor markets to citizens of the New Member States (NMS) in 1/3 of the Member States of the EU and the European Economic Area (EEA) after May 2004 has resulted in a net migration of less than 0.2 per cent of the population of the NMS, despite the GDP per capita in the NMS being approximately 25% of the average of the old EU Member States (Brücker and Schroder, 2006). As Molle (1994) points out, the lack of international mobility is not an international phenomenon but an ubiquitous issue that appears on the national level among regions, sectors of activity and occupational groups (Molle, 1994). This happens due to many reasons, the most relevant of which is some sort of “migration costs”. It becomes apparent that there is some cost of migration to which the large strata of the population are subjected to. Even in the case of apparent advantages of migration for the better job and living standards (that can be found on the common market), many individuals from the less economically developed countries are simply not ready to bear the economic and socio- psychological costs of migration.

38 It become apparent that migration costs can become a serious obstacle for the labor mobility on the common market as well as the reason of incomplete convergence of real wages among member states. As far as there has never been an attempt to measure this migration aversion factor (or index) in the literature, it seems to be quite an interesting and challenging task.

1.6.3. Labor migration and migration surplus

When studying labor mobility in economic theory with the special stress on market integration, it would not be excessive to look at the issue of migration surplus. The framework for this sub-chapter has been adapted from Borjas (1994, 1995) with some expansion of Borjas’ results for the case of the small Central European economy (represented hereinafter by the Czech Republic). Following Borjas’ specification of production technology in the host country, we would also assume that it can be summarized using the aggregate production function with two inputs: capital ( K) and labor ( L), so that the total output is equal to Q = f(K,L) . The workforce is composed of N native and M immigrant workers. Initially, some important assumptions should be made. First of all, it is assumed that all capital is owned by natives (to ignore the possibility that immigrants might augment the country’s capital stock). Second, it is assumed that all workers are perfect substitutes in production (no skill differentials). Third, it is assumed that both supplies of capital and foreign and native labor are perfectly inelastic (relaxing the assumption of inelastic supply curves would make the calculations cumbersome). The aggregate production function defined above exhibits constant returns to scale. As a result all output is redistributed to the owners of capital and to workers. The equilibrium in the economy (prior to the admission of M of foreign labor) required that each factor price is equals to the respective value of the marginal product. Let us suppose that the price of capital is initially r 0 and the price of labor is w0. The price of the output is the numeraire (input prices are measured in units of output). Before the admission of M immigrants, the national income that goes to natives QN, is the price of capital times the quantity used, plus the price of labor times the number of workers hired, or, in another words: QN,

= r 0K + w 0N

When the immigrants enter the country, the supply curve shifts and the market wage falls to w 1. The increase of the national income accruing to the natives which is created as a result of this process, is called “ immigration surplus ” (Borjas, 1995). It can be shown that the immigration surplus is equal to: ½*(w 0 – w 1)*M. This formula can be manipulated and re-written to give the following expression: ∆Q 1 N = − sem 2 Q 2

39 where s is labor’s share of national income; e is the elasticity of factor price for labor (the percentage change in the wage resulting from a 1 per cent change in the size of the labor force); and m is the fraction of the workforce that is foreign-born ( m=M/L ). According to Borjas’ estimations the immigration surplus for the United States is on the order of 0.1 per cent of GDP, making it a mere $7 billion per year in a $7 trillion economy 2 (Borjas, 1995). Further, it would be useful to estimate the net change in the native workers and capitalists expressed as a fraction of GDP. According to Borjas (1995) this is equal to: Change in Native Labor Earnings = sem 1( − m) Q

Change in Incomes of Capitalists 1 = −sem 1( − m) Q 2 If the elasticity of factor price is -0.3, native workers lose about 1.9 per cent of GDP ($133 billion in a $7 trillion economy); native capital gains about 2 per cent of GDP ($140 billion). The small immigration surplus of $7 billion disguises a sizable redistribution of wealth from workers to the users of capital. Let us try to re-calculate Borjas’ results for the small Central European Economy. It is expected that the results will be different: first, due to the smaller size of the economy and therefore smaller size of GDP and, second, due to the smaller size of foreign-born labor force. First of all, let us calculate the immigration surplus. Using the data from IMF (2007) it can be seen that the labor’s share of national income in most of the EU countries around 2005 was on the order of 65 per cent. According to Hamermesh (1993) it can be accepted that the elasticity of factor price for labor is equal to (1 – s) 2/η, where η is the output-constant elasticity of labor (there is some consensus that η is about -0.3). Because the share of labor income is 0.65, the elasticity of factor price is equal to -0.4. Now, let us look at the fraction of the workforce that is foreign-born. Due to the Czech Statistical Office (2007), the number of officially registered foreigners who were working in the Czech Republic in 2005 was on the order of 200.000 persons. However, it is quite obvious that these estimates do not include the illegal labor migrants (i.e. from Ukraine). Although the data on illegal migrations is not recorded officially, it is estimated that another 200.000 illegal workers can be present in the Czech Republic (Ministry of Internal Affairs of the Czech Republic, 2006). Given that in 2005 there were 4.764.000 employed people in the Czech Republic, the fraction of the workforce that is foreign-born

2 Borjas (1995) accepts that the share of labor income in the US in the 1995 was on the order of 70 per cent, the fraction of immigrants in the US workforce was slightly less than 10 per cent. In addition, in the model specified above (only two factors of production), the elasticity is about -0.3.

40 (depending whether we are considering legal labor migrants or also illegal ones) might range from 5 to 10 per cent. Given this data it is easy now to estimate the immigration surplus. It equals to the 0.03 percent of GDP (0.1 percent of GDP taking into account illegal labor migrants). Czech GDP in 2005 was around 2970 billion CZK, so that economic gains from labor immigration in the Czech Republic would be around 950 million CZK (or 3.8 billion CZK in case illegal immigrants are also taken into account). The net change in incomes of native workers and capitalists calculated for the Czech Republic yields the following results: assuming that the elasticity of factor price is -0.3, native workers would lose about 1.23 percent of GDP (about 36 billion CZK) in case just legal working migrants are taken into account or around 2.3 per cent of GDP (69 billion CZK) in case illegal migrants are also concerned. The native capital would gain about 1.2 per cent of GDP (37 billion CZK) in case only legal working migrants are concerned and 2.4 per cent of GDP (71 billion CZK) in case the approximate estimate of illegal migrants is also taken into account. The small immigration surplus of 950 million CZK (or about 3.8 billion CZK in case we take into account estimates of illegal migrants) thus disguises the redistribution of wealth from workers to the users of immigrant labor in the Czech Republic. The results can be compared with those of Borjas (1995). His estimate of immigration surplus in a $7 trillion economy (USA) is $7 billion. Using the actual CZK/USD exchange rate it can be easily computed that the immigration surplus for the US economy is 167 times larger of those for the Czech Republic (with the Czech economy being just about 50 times smaller than those of the USA expressed in US dollars). This might mean that labor immigration to the Czech Republic is very modest in comparison with the US immigration. However, this situation is quite understandable: the Czech Republic has just developed into a market economy and is not yet regarded by potential migrants as a highly attractive place for working and living. Therefore, there are still gaps to bridge in the Czech Republic with regard to attracting more migrants and increasing the level of immigration to the country.

1.7. Selected methodological issues

The following two sub-sections describe selected methodology tools that have been used in this dissertation: calculation of Marshallian surplus and using backwards stepwise regression. While the former issue is related to the estimation of welfare gains in the elaboration on labor migration and welfare

41 effects of free mobility of labor in the common market, the latter one is used throughout the dissertation as a tool for selecting the appropriate variables for the models. Due to the fact that selection of methodological tools might be put under doubt an attempt has been made to explain their use and draw the rational for selecting them out of the range of other available tools.

1.7.1. Calculating Marshallian surplus

In this sub-section the way to calculate Marshallian surplus is explained in somewhat greater detail. Marshallian surplus can be expressed explicitly in terms of simple market equilibrium models. The calculus provides us with the following geometrical meaning of definite integral: let y = f(x) be continuously differentiable function over an interval (a, b) and its end points, then the value of definite integral b ∫ f (x)dx a provides the volume of the shaded area in Figure 1.4.

Figure 1.4: Geometrical meaning of definite integral

y y=f(x)

a b x

Source: own drawing

Let u(x) is a continuously differentiable function such that du (x) = f (x) dx then

42 du (x) ∫f (x)dx = ∫ dx = u(x) + const dx (integral of a function is a function such that its derivative is equal to the function we are integrating) and b ∫ f (x)dx = u(b) − u(a) a From it follows algorithm for calculation consumers’ surplus: Find function e(x) such that de (q) = D(q) dq where D(q) is an inverse demand function, then consumers surplus is equal to q* CS = ∫ D(q)dq − p *q* = e(q*) − e )0( − p *q * 0 In the same way we obtain producers’ surplus: Find function r(q) such that dr (q) = S(q) dq where S(q) is an inverse supply function, then producers’ surplus is equal to q* PS = p *q *− ∫ S(q)dq = p *q *−r(q*) + r )0( 0 Marshallian surplus: MS = CS + PS = e(q*) − r(q*) + r )0( − e )0(

1.7.2. Using backwards stepwise regression

In this sub-section the use of backwards stepwise regression is described in detail. Backward stepwise regression is used in most of the essays that constitute the following elaboration in order to explain and defend the use of candidate variables in regression models which test different aspects of migration. Generally, in statistics the method called hereinafter “stepwise regression” means regression models in which the choice of predictive variables is carried out by an automatic procedure (see for

43 example Hocking, 1976; or Draper and Smith, 1981). The procedure is run with the use of a sequence of F-test, t-tests, adjusted R-square, Bayesian information criterion, Mallow’s Cp or false discovery rate, just to name a few. There are three main approaches to the stepwise regression: i) forward selection (starting with no variables in the model and adding them one by one); ii) backward selection (starting with all candidate variables in the model and testing them for significance), iii) combination of the previous two. In spite of the fact that stepwise regression methods are considered to be controversial by a number of critics (see for example, Rencher and Pun, 1980, or Copas, 1983), they are extensively used in data mining (extracting of implicit, previously unknown, and useful information from data sets and databases") (see for example, Hand et al, 2001). For this dissertation, the backward stepwise regression was used in most of the cases. Backwards stepwise regression represents the approach that often applied in multilinear regression. Generally it represents a common task to determine the "best" set of independent variables to use in the fit. Backward step regression used here starts with the initial model that contains all the independent variables. Then one variable is deleted at each stage (at first the variable that causes the smallest drop in adjusted R-squared is dropped) until the best model is reached (see for example Menken (2005). The justification for using backwards stepwise regression throughout this elaboration is threefold:
In most of the models presented here a set of candidate variables was defined on the basis of economic theory and extensive literature review; it seemed important to extract insignificant variables from the models in order to use the results for explanations of migration-related issues and deriving relevant policy implications
Backwards stepwise regression provided the possibility to rule out the “suspicious” variables by testing the models with and without them in turns.
The use of backwards stepwise regression created solid grounds for elaborating on the outcomes of regression models; the results obtained throughout the regression analysis provide reasonable illustrations for the migration debate in single essays.

1.8. References

Adams, R. H. (1998), “Remittances, investment, and rural asset accumulation in Pakistan", Economic Development and Cultural Change , 47, 155-173.
Barrett, A., Trace, F. (1998), “Who is Coming Back? The Educational Profile of Returning Migrants in the 1990s”. Irish Banking Review.

44
Bauer T., G. S. Epstein and I. Gang (2002): “Herd Effects or Migration Networks? The location Choice of Mexican Immigrants in the US”, CEPR Discussion Paper Series , n. 3505, August 2002
Bauer, T., Zimmermann, K. (1999), “Assessment of Possible Migration Pressure and its Labour Market Impact Following EU Enlargement to Central and Eastern Europe“ , IZA Bonn Research Report No. 3, July.
Bhagwati, J.(1976), ”The International Brain Drain and Taxation”, North Holland
Biller, S. (2007), “UICIFD Briefing No. 3: Remittances”, The University of Iowa Center for International Finance and Development
Borjas G. J. (1995): “Ethnicity, Neighborhoods, and Human-Capital Externalities”, American Economic Review , Vol. 85 (3), pp. 365-390.
Borjas, G. (1990), “Friends or Strangers”, Basic Books, Inc, New York, NY
Borjas, G. (1994), “The Economics of Immigration”, Journal of Economic Literature, Vol. XXXII, pp. 1667-1714.
Borjas, G. (1995), “The economic benefit from immigration", Journal of Economic Perspectives , 9(2), 3-22.
Borjas, G. (1996), “The New Economics of Immigration”, The Atlantic Monthly
Borjas, G. (1999), “Heaven’s Door”, Princeton University Press.
Briggs, V. (2004), “The Economic well-being of black Americans: the overarching influence of U.S. immigration policies”, The Review of Black Political Economy , Vol. 31, No. 1-2, pp. 15-42.
Brücker, H. and Shroder,P. (2006), “International Migration with Heterogeneous agents: Theory and Evidence”, Paper presented on the Annual Labor Economics Association conference in Prague, September 2006
Brücker, H., Siliverstovs, B. (2004), “The Macro-Determinants of International Migration in Europe: Evidence from Germany”, Paper presented at the conference “Immigration Issues in EU-Turkish Relations”, Bogazici University, October 8-9, 2004
Butcher, K. C., Card, D. (1991), “Immigration and Wages: Evidence from the 1980s”. American Economic Review, Papers and Proceedings, 81 (2), 292-296.
Butler, A. (1967), “Labor Costs in the Common Market”, Industrial Economics, vol.6, no.2, pp. 166-83
Caves, R. (1960), “Trade and Economic Structure”, Cambridge, M.A: Harvard University Press;
Chiswick, B. R. (1978), “The Effect of Americanization on the Earnings of Foreign born Men”, Journal of Political Economy, October, Vol. 86, No. 3, pp. 897-921.
De Jong, G. et al. (1986): “Migration Intentions and Behavior. Decision-making in a Rural Philippines parish”, in J.T. Fawcet (ed.), “Population and Environment”, Vol.8, No.1
De Jong, G., et al. (1983): “International and Internal Migration Decision-making”. International Migration Review, vol. 17, No. 3.
Dhima, G. (1991), “Politische Ökonomie des Schweizerischen Ausländerregelung”, Chur/Zurich, Rüegger.
Djajic S. (1986), “International Migration, Remittances and Welfare in a Dependent Economy", Journal of Development Economics , 21, 229-234.

45
Domingues Dos Santos, M. and Postel-Vinay, F. (2003), “Migration as a source of growth: The perspective of a developing country", Journal of Population Economics .
Drinkwater, S., Levine, P., Lotti, E. (2003), “The Labor market effect of remittances”, Flowenla Discussion Paper No.6 , Hamburg Institute of International Economics
Dustmann, C. (1996), “Return Migration: the European Experience”, Journal of Political Economics, April (22), pp. 213-234.
Faini et al. (1999), “Migration. The Controversies and the Evidence”, Cambridge: Cambridge University Press
Faini, R., Venturini, A. (1994), “Migration and Growth: The Experience of Southern Europe“, CEPR Discussion Paper, No. 964, London, May, www.cepr.org/pubs/dps/DP964.asp
Fischer, P. A., Reiner M., and Straubhaar, T. (1997), “Should I Stay or Should I Go?”. In T. Hammar, G. Brochman, K. Tamas, and T. Faist (eds.), “International Migration, Immobility and Development: Multidisciplinary Perspectives”. Oxford and New York: Berg.
Fisher, S. (1966), “Wage Discrimination in an Integrating Europe”, Sythoff, London
Fix, M., Passel, J. S. (1994), “Immigration and immigrants: Setting the record straight”. Washington, DC: The Urban Institute.
Freeman, R. (1993), “Immigration From Poor to Wealthy Countries, Experience of the US”, European Economic Review (37), pp. 443-452.
Friedberg, R., Hunt, J. (1995): “The Impact of Immigration on a Host Countries Wages, Employment and Growth”, Journal of Economic Perspectives, Spring, vol. 9, No.2;
Funkhauser, E. (1992), “Mass emigration, remittances and economic adjustment: The case of El Salvador in the 1980s", in G. Borjas and R. Freeman (eds), Migration and the Workforce: Economic Consequences for the United States , University of Chicago Press, Chicago, IL.
Galasi, P. and Kollo, J. (2002), “The disincentive and re-employment effects of unemployment benefits" in K. Fazekas and J. Koltay (eds), The Hungarian Labour Market 2002, Institute of Economics, Hungarian Academy of Science.
Giersch, H. (1949), “Economic Union Between Nations and the Location of Industries”, Review of Economic Studies , vol. 17, pp. 87-97
Giliani, I., Khan, M. F. and Iqbal, M., (1981), “Labour Migration From Pakistan to the Middle East and its Impact on the Domestic Economy", Final Report, Research Project on Export of Manpower to the Middle East, World Bank , Washington DC.
Glytsos, N. P. (1993), “Measuring the income effects of migrant remittances: A methodological approach applied to Greece", Economic Development and Cultural Change , 42, 131- 168.
Gremmen, H. (1985), “Testing Factor Price Equalization in EC: An alternative approach”, Journal of Common Market Studies , vol. 23, no.3, pp. 277-86
Grossman, J. (1982), ”The Substitutability of Natives and Immigrants in Production”, Review of Economic and Statistics, 64. Nr.4.

46
Hamermesh, D. (1993), “Labor demand”, Princeton, N.J., Princeton University Press.
Hannan, D.F. (1970): “Rural exodus” Geoffrey Chapman, London
Hannan, D.F. (1986): “Schooling and the Labour market”. Shannon Cirruculum Development Centre
Harris, J.R. and Todaro, M.P. (1970), “Migration, Unemployment and Development: A Two Sector Analysis,” American Economic Review 60 (1), 120-142.
Harris, N. (1995): “The New Untouchables. Immigration and the New World”, Taurus publishers, London.
Harris, R., Schmitt, N. (2003), “The Consequences of Increased Labour Mobility within an Integrating North America”, in Harris, R. (ed.), “ North American linkages : opportunities and challenges for Canada, Calgary: Calgary University Press;
Holden, M. (2003), “Stages of Economic Integration: From Autarchy to Economic Union”, Parliamentary Research Branch, PRB 02-49B, 13 th of February.
Honekopp, E., Werner, H. (2000), “Eastward Enlargement of the European Union: a Wave of Migration?”, Bundesansalt für Arbeit, IAB Labor Market Research Topics, No. 40.
IMF Balance of Payments Statistics (2006), www.imf.org
International Organization for Migration (2006), “Global Estimates and Trends”.
Krugman, P., Obstfeld, M. (2005), “International Economics, Theory and Policy”, 7 th Edition, Pearson. Addison Wesley.
Kule D., Mancellari A., Papapanagos H., Qirici S. and Sanfey P. (2000), “The causes and consequences of Albanian emigration during transition: evidence from micro-data", University of Kent, Discussion Paer No.98/18.
Layard et. al (1992) :”East-West migration”, MIT Press, London.
Leon-Ledesma, M.L. and Piracha M. (2001), “International Migration and the Role of Remittances in Eastern Europe", Department of Economics Discussion paper 01/13, University of Kent at Canterbury.
Lindert, P. (1986), “International Economics”, 8 th edition, Irwin, Homewood, Illinois.
Lucassen, J., Lucassen, L. (1999), “Migration, migration history, history: old paradigms and new perspectives”, Bern; New York: Peter Lang.
Massey, D. (1986), “The Settlement Process of Mexican Migrants to the United States”, American Sociological Review 51, No. 5 (1986), pp. 670-84.
Mc Cormick B. and Wahba J. (2001), “Overseas Work Experience, Savings and Entrepreneurship amongst Return Migrants to LDCc", Scottish Journal of Political Economy, 48(2) .
Melvin, L., Glick, M. (1982), “An Analysis of the New Orthodoxy on Black Mobility”. Social Problems , Vol. 29, No. 5 (Jun., 1982), pp. 511-523.
Mesnard A. (2001) “Temporary Migration and Capital Market Imperfections", University of Toulouse working paper .
Meyer, F. and Willgerodt, H. (1956), “Der Wirtschaftpolitische Aussegewert Internationaler Lohnvergleiche”, in Internationale Lohngefaelle, Wirtschaftpolitische Folgerungen, und Statistische Problematik , Deutsche Bundesverlag, Bonn

47
Migration News (2007), www.migration.ucdavis.edu
Mirus, R., Rylska, N. (2002), "Economic Integration: Free Trade Areas vs. Customs Unions." In Smith, P. and Chambers, E. (eds.), “NAFTA and the new millennium” , U of A Press.
Molle, W. (1994), “The Economics of European Integration, Theory, Practice, Policy”. Dartmouth, Aldershot;
Molle, W., van Mourik, A. (1989b) (eds.), “Wage Structures in the European Community, Cobvergence of Divergence?”, Gower, Aldershot.
Mourik, A. van (1987), “Testing Factor Price Equalization in EC: An alternative approach: A Comment”, Journal of Common Market Studies , vol. 26, no. 1, pp. 178-197
Mourik, A. van (1993), “Wages and European Integration”, PhD, RL, Maastricht
Mundell, R. (1957), “International Trade and Factor Mobility”, The American Economic Review , Vol. 47, No. 3 (June), pp. 321-335;
Mundell, R. (1968), “A Theory of Optimum Currency Areas”, in “ International Economics ”, New York: Macmillan, pp. 177-186;
Myrdal, G. (1956), “Trade as a Mechanism of International Inequality”, in Meier, G. (ed.), “ Leading Issues in Economic Development” , 5 th edition, New York: Oxford University Press, 1989;
Nixon, F. (1993), “Friends or Strangers? Examining the Empirical Evidence of Immigration”, International Journal of Social Economics, Vol. 20, No. 10, pp. 46-54.
O’Rourke, K. (1994), “The Economic Impact of the Famine in the Short Run and Long Run”, American Economic Review Papers and Proceedings 84, no.2 (1994), pp. 309-13.
Orcutt, H., Regrets, M. (1989), “Immigrants and Human Capital Investment”, American Economic Review, Vol. 89, Issue 2, pp. 186-190.
Oxford English Dictionary, 2006 (on-line version, www.oed.com ).
Rassekh, F., Thompson, H. (1998), “Micro convergence and macro convergence: factor price equalization and per capita income”, Pacific Economic Review , 3(1), pp. 3-11;
Richardson, A. (1974): “British Immigrants and Australia: A Psychio-Social Inquiry”, Australia National University Press, Canberra.
Rozelle, S., Taylor, E. and de Braw, A. (1999), “Migration, remittances and agricultural productivity in China", American Economic Review, 89(2), 287-291.
Seers et al. (1980), “Integration and Unequal Development, The Experience of EC”, St. Martin’s Press, New York
Sell, R. and De Jong, G. (1978), “Towards a Motivational Theory of Migration Decision Making”, Journal of Population, Vol.1, 315-335.
Simon, J. (1990), “The Economic Consequences of Immigration”, Basil Blackwell.
Simon, J. (1995), “Immigration: the Demographic and Economic Facts”, Cato Institute and the National Immigration forum ( http://www.cato.org/pubs/policy_report/pr-immig.html ).

48
Stark, O., Helmenstein, C. and Prskawetz, A. (1997), “A brain gain with a brain drain", Economics Letters , 55, 227-234.
Todaro, M. P. (1969), “A Model of Labor Migration and Urban Unemployment in Less Developed Countries.” American Economic Review 59 (1), 138-148.
Todaro, Michael P. (1969): “A Model of Labor Migration and Urban Unemployment in Less Developed Countries.” American Economic Review 59 (1), 138-148.
Tovias, A. (1982), “Testing Factor Price Equalization in the EEC”, Journal of Common Market Studies , vol. 20, pp. 165-81
Treibel, A. (1999), “Migration in Modernen Gesellschaften. Soziale Folgen von Einwanderung, Gastarbeit und Flucht”, Weinheim: Juventa, 2. Aufl. 1999
United Nations (2006), “International Migration One of the Greatest Challenges of the 21 st Century”, Commission on Population and Development, 39 th session, 7 th meeting report.
Walsh, B.M. (1974): “Expectations, Information, and Human Migration: Specifying an Econometric Model of Irish Migration to Britain”, Journal of Regional Science, 14, 107-120.
Walsh, B.M. (1987): “The Impact of Demographic Variables on Unemployment”, in “The Challenge of Unemployment”, Administration, vol. xxxv, No. 3, Dublin.
Walsh, B.M. (1987): “Why is Unemployment so high in Ireland?” in “Perspectives if Economic policy”, Centre for Economic research, UCD, Dublin
Walsh, B.M. (1989): “Testing for the Existence of Macro Economic Feedback from Large-scale Migration”. The Economic and Social Review, Dublin.
Wegge, S. (1998), “Chain Migration and Information Networks: Evidence from Nineteenth-Century Hesse-Cassel”, The Journal of Economic History , Vol. 58, No.4 (Dec., 1998), pp.957-986.
Williamson, J. (1995), “The Evolution of Global Market Since 1830: Background Evidence and Hypothesis”, Explorations in Economic History, 32 (1995), pp. 141-196;
Winters, P., A. de Janvry, and E. Sadoulet (2001): “Family and Community Networks in Mexico-U.S. Migration,” Journal of Human Resources , 36, 159-184.
World Bank (2005), “World Development Indicators”, Word Bank, Washington, D.C.
Young, A. (1991), “Learning by Doing and the Dynamic Effects of International Trade”, Quarterly Journal of Economics , 106, pp. 369-405
Zachariah, K. C., Mathew, E. T. and Irudaya Rajan, S. (2001), “Social, economic and demographic consequences of migration on Kerala", International Migration , 39(2), 43-68.
Zimmermann, K. (1995), “European Migration: Push and Pull”, Proceedings of the World Bank Annual Conference on Development Economics Review and The World Bank Research Observer, March, pp. 313-342.

49 2. Labor migration and welfare effects of free mobility of labor in the common market

2.1. Introduction

According to Holden (2003) common market (CM) is reached when the obstacles to free movement for not only goods and services, but also production factors among partners are removed. Economic Union (EU) and Economic and Monetary Union (EMU) are the most superior levels of integration. Whilst Economic Union means a common market with a high degree of coordination and unification of the most important areas of economic policies, Economic and Monetary Union includes linking the currencies of the member states through the irrevocably fixed exchange rates or creating a single currency (as well as the high degree of coordination of macroeconomic and budget policies) (Molle, 1994). The common market pre-supposes efficient allocation: according to expectations capital and labor would move from activities with a low marginal product of labor to those with a higher one. Whilst there are almost no problems with the free movement of capital on the common market, labor movement is a different issue. Generally, there is a free international movement of labor on the common market created by two or more parties if: (i) nationals of one party may unrestrictedly look for and accept a job in another party, and (ii) self-employed people from one party are free to settle in another party to exert their profession or activity (Molle, 1994). In practice, however, this might take various forms: even when all restrictions among partner parties are abolished, some positive integration is needed as well (for instance, coordination of labor market and employment policies, coordination of social policies and taxes). This essay is concerned with the political economy models of labor migration: specifically it focuses on the labor migration on the common market, propensity to migration and welfare effects of migration that might help to define winners and losers of market integration. The paper begins with the introduction of the concept of “welfare”, discussion of the so-called “Marshallian surplus” and listing problems of measuring welfare effects of different regimes of international trade. Consequently the paper provides the explanation of the effects of free mobility of factors in neoclassical static two- country model. It draws the conclusions of the model for the free movement of labor and equalization of real wages. Then it comes with implications and explanations of why the real wages are not equalized in practice (e.g. in real life). It explains why the migration within the common market is hampered and comes with a measure of migration cost (i.e. the cost of labor migration that can be expressed in both economic and psychological terms.). Further, the paper analysis the welfare effects

50 of the free movement of labor after the creation of the common market taking into account self- defined “indicator of migration cost” (denoted by α) and draws implications for the target and source countries of labor migration. Finally, the paper draws an empirical model of average wages and migration flows in the EU-27 that aims to point out that the indicator of migration cost and thus the propensity of migration are always country-specific and varying amongst countries. In the conclusion the paper discusses the policy implications and consequences for the main actors involved in labor migrations on the common market. The main value-added of this essay is an attempt to draw the political economy model of labor migration. The paper is trying to explore the well-known framework and bring new insights into the migration model. It draws several important conclusions about the total welfare in both emigration and immigration countries and its effects for different segments of their societies. It confronts the general neoclassical theory of free mobility of production factors and labor migration in it by taking into account propensity to migration. Finally, the essay puts into doubt a well-known paradigm of the wage differential in labor migration as the factors that fully explain labor flows between the countries.

2.2. Marshallian surplus

The idea to measure economic welfare that is based on the concepts of so-called consumers´ and producers´ surplus was introduced by Alfred Marshall. Marshall originally applied this idea to the analysis of effects of taxes and price shifts on market equilibrium. 3 That is why an aggregate of two components of welfare, consumers´ and producers´ surplus, is often called “Marshallian surplus”. An individual consumer’s surplus is the difference between the price a consumer is ready to pay for a product and the actual price of the product. If a consumer is willing to pay more than the actual price of the product, her benefit from the eventual transaction is how much she saved when she did not pay that price but the price prevailing on the market. For example, a person is willing to pay a tremendous amount for water which represents a basic measure of survival, however since there are competing suppliers of water on the market the water is available for less than a consumer is willing to pay. The difference between the price a consumer is willing to pay and the price she pays is the consumer’s surplus. Consequently, the aggregate consumers' surplus is the sum of all individual consumer’s surpluses. An individual producer‘s surplus is used to measure the welfare of a group of firms selling a particular product at a particular price. An individual producer‘s surplus is defined as the difference between what the producer actually receives for selling her product and the amount she would be

3 See in Marshall (1920).

51 willing to accept for the product unit. The aggregate producers' surplus is the sum of all individual producers’ surpluses. Normally, consumers’ willingness to pay for a product and producers' willingness to accept payments for the product can be derived from a market demand curve and market supply curve for a product. The market supply curve shows the quantity of good that firms would supply at each and every price that might prevail on the market. Read the other way, the supply curve tells us the minimum price that producers would be willing to accept for any quantity demanded by the market. Let us consider the market with just one commodity with inverse demand function p = D(q) and inverse supply function p = S(q), where q is a quantity demanded or supplied and p is the price. Then, for the closed market economy, q* such that D(q*) = S(q*) is the equilibrium quantity and p*= D(q*) = S(q*) is the equilibrium price, see Figure 2.1. By p C we denote the zero demand price, by pP the zero supply price, and by E position of market equilibrium. Clearly, the area 0q*Ep C expresses the amount of money consumers are ready to pay and the area 0q*Ep* expresses the money consumers have actually to pay in market equilibrium. The difference CS between area 0q*Ep C and area 0q*Ep* 0q*Ep* is nothing else but consumers’ surplus in market equilibrium E.

Figure 2.1: Market equilibrium and Marshallian surplus p pC Marshallian surplus = CS+PS

S CS

E p* PS

D pP

0 q* q

Source: own drawing

Comparing two different market situations (equilibrium positions) it becomes apparent that the market equilibrium with higher consumers' surplus is better for consumers (it provides more welfare) than the market equilibrium with lower consumers’ surplus. On the other hand, the market equilibrium with higher

52 producers' surplus is better for producers (provides more welfare) than equilibrium with lower producers' surplus. It is clear that equilibrium with higher Marshallian surplus is preferable to equilibrium with lower Marshallian surplus. In this sense Marshallian surplus can be used as a reliable measure of welfare. It is apparent that changes in regimes of international trade and international economic cooperation, including different forms of economic integration, affect the equilibrium situations at the markets. In order to evaluate economic implications and desirability of these changes the Mashallian surplus can be used as a measure of welfare. It can also be used for eventual implications about welfare effects. Marshall’s original motivation was to study of welfare effects of taxation. Some authors (see for example Viner, 1950) used concepts of consumers’ and producers’ surplus to analyse welfare effects of customs unions. Moreover, in the 1950s and 1960s Marshallian surplus became one of the instruments of economics of international trade and economics of integration (Johnson, 1965; Molle, 1994; Pelkmans, 1997; Hansen and Nielsen, 1997; Svendsed, 2003; Turnovec, 2003). This essay continues to apply this concept to the study of political economy of the free mobility of labour in common market.

2.3. Labor mobility between two countries and the indicator of migration cost

Let us present a simple neo-classical model of labor mobility between two countries: country H and country P. Let us suppose that these countries created a Common Market (CM). This means that: (i) an internal market with fully free internal movement of products (goods and services) and of production factors (labor and capital) has been set; and (ii) common external regulations for both products and production factors have been established. Although it is unlikely to observe a fully free internal movement of production factors between members of the CM (especially when it comes to the free movement of labor) we will use this assumption in our model. Now suppose that all obstacles for mobility of production factors such as capital and labor in both countries have been removed. Let = = LH SH (w) and LP SP (w) be supply functions of labor and = = QH DH (w) and QP DP (w) be demand functions of labor in country H and country P respectively. Let w be a price of labor (wage) partial equilibrium analysis assuming that the quantities of other factors used as primary inputs are fixed.

53 Figure 2.3 represents the situation when the labor markets in countries H and P are separate and there is no migration. Figure 3 is a modification of a neo-classical static two-country diagram.

Let w H , w P be the equilibrium wages in country H and country P respectively. Let LH L, P be the equilibrium supply of labor (and demand of labor) in country H and country P respectively. Equilibrium wages and quantities are solutions of equations = = SH (w) DH (w) and SP (w) DP (w)

The welfare effects in the situation without migration are represented by the “consumers’ surplus” (in this case employers’ welfare) represented by the blue areas as well as “producers’ surplus” (in this case workers’ welfare) represented by the red areas.

Figure 2.3: Labor mobility: situation without migration (single markets, labor markers in H and P are separate)

country H country P w w D S H H S E P H D w P H

EP w P

LH L LP L

Source: own drawing

In our case (labor markets in countries H and P are separate and there is no migration) we have > w H w P (wage level in the country H is higher than the wage level in the country P). As it has been stated in section 1, labor is a production factor and the common market entails the free mobility of this factor. According the expectations of general economic theory it can be envisaged that workers from the country P (the lower wage country) would migrate to the country H (the higher wage country) until the wages become equal (equalization of factor prices assumption). But in reality we do not observe wage equalization in common market.

54 2.3.1. Propensity to migration and indicator of migration cost

Let us define the concept of propensity to migration in the framework of the two countries model as  α 1− for α ≤ w − w Π α = − H P (wH , wP , )  wH wP  α > − 0 for wH wP

α where w H , w P are the equilibrium wages in country H and country P respectively and is the indicator of migration transaction cost. The migration transaction cost deserves further elaboration. It is clear that any migration induces a cost. It might be a financial cost of moving to another place and settling down there, but it also can be the cost of leaving family or friends, adjusting to a new environment, cultural differences, language difficulties, which is hard to represent in financial terms. Therefore, as far as the cost of migration includes both tangible and intangible aspects, α can be expressed in terms of the following two aspects that seem to be most relevant for the purpose of this analysis:
Monetary (financial) cost of migration (keeping two homes, bearing travel expenses, various traveling and administrational adjustments and re-settling, etc.),
Cost of (intangible) psychological factors (habits, language barriers, breaking of social ties in the country of origin, deprivation related to migration, etc.). For simplification of further consideration we assume that migration transaction cost can be fully expressed in financial terms. In order to analyze the implications of migration transaction cost for labor market equilibrium, the situation with free mobility of labor between countries H and P that have created a common market is illustrated in Figure 2.4. In conditions of transition to free mobility of labor between P and H the original equilibrium situation with equilibrium prices w P and w H and equilibrium labor supply L P and L H in countries P and H we assume migration transaction cost from country P to country H is equal a. If wage differential 2 1 wH-wP > a, as in our case, labor starts to flow from P to H. Assume that quantity of labor L P -LP will 2 1 flow from P to H. The inflow of L P -LP migrants from country P to country H pushes the wages in 1 2 country H down from w H to w H and increases demand for labor in country H from LH to L H . In 1 country P the supply of labor falls down from L P to L P , what leads to increase of wages from w P to 1 1 wP and decrease of demand for labor from L P to L P . The new wage differential wH1-wP1 still exceeds migration transaction cost a. The flow of labor from country P to country H will continue until

55 difference between wages in H and P is equal to the migration transaction cost (as in Figure 2.5), what provides the common market equilibrium on labor market.

Figure 2.4: Labor mobility: migration from P to H and transaction cost

country H country P w w D S H H S E P H 1 D w E P H H w1 H E1 a P w1 wP E P P

1 2 1 2 LH LH LH L LP L P LP L

Source: own drawing

There are two major implications that have to be pointed out:
If the difference between wages in country H and country P is less than the indicator of migration cost, there is no mobility motivation for the workers to move from country P to country H. It means that economic and psychological costs are greater than benefits of wage differential. In another words, it can be said that the propensity to migration between two countries equals to zero.
If the difference between wages in country H and country P is greater than indicator or migration cost, workers from country P will move to country H. In another words, it can be said that the α propensity to migration between two countries equals to 1− . This means that, for a − wH wP given α the greater is the difference between wages, the higher is the propensity to migration (from lower wage country to the higher wage country). In the situation of the common market and the free mobility of labor between countries H and P, the equilibrium (zero propensity to migration) is reached when the difference between wages in country H and country P is equals to α. The equilibrium can be derived from equation

56 + − α = + − α DH (w) DP (w ) SH (w) SP (w ) Let w* be the solution for this equation, representing the equilibrium wage after the common market is created and the obstacles to free migration of labor are removed. The equilibrium wages that CM = * CM = * −α are set in countries H and P would be: w H w and w P w for the country H and country P H = CM respectively. The domestic supply of labor in country H is expressed as LH SH (w H ) , the total supply CM = CM of labor in country H (including migration of workers from country P) is expressed as LH DH (w H ) . The size of “imported” labor (workers that migrated from country P to country H) is expressed CM − H as LH LH .

Figure 2.5: Labor mobility: Common market equilibrium

country H country P w w D SH H S E P H D w CM P H E H wC H EC a P wC P E w P P

1 2 1 2 LH LH LH L LP L P LP L

Source: own drawing

CM = CM −α The domestic supply of labor in country P is expressed as LP SP (w H ) and the domestic P = CM −α demand for labor in country P is expressed as LP DP (w H ). The size of “exported” labor (workers that left country P in the search of better employment opportunities in country H) is CM − P expressed as LP LP .

57 2.3.2. Welfare considerations

Analyzing the situation of the free movement of labor after the establishment of the common market, it is obvious that for both countries (H and P) there has been an increase in total welfare. This welfare increase can be distinguished for each of two countries separately:
For country P (country with lower factor price, the labor): the welfare increase has happened due to the increasing revenues of factor owners (workers) who are exporting this factor (their labor) to the country with higher factor price,
For country H (country with higher factor price): the welfare increase has happened due to the increase of the product Taking into account specific behavior of labor markets it is still not clear whether the establishment of common market and free mobility of labor increases welfare of all involved parties. The situation is depicted in Figure 2.6.

Figure 2.6: Welfare considerations of the free mobility of labor in common market

country H country P w w D H SH S P EH CM w E D H H P wCM H CM a E P wCM wP P E P

H CM P CM LH LH LH L LP L P LP L

Source: own drawing

In country H (the country with higher wages) wages are falling down compared to starting situation (single market), employment of workers from country H decreases, although total employment of workers from country H and migrants from country P is increasing. Employees from country H are loosing their surplus (“producers’ surplus” represented by the red area), although

58 employers from country H are gaining. In total it is obvious from Figure 2.3 that the employers’ gain is greater than domestic workers’ loss. Therefore, total welfare effect in country H is positive. On the contrary, in country P (the country with lower wages) wages are increasing compared to the starting situation (single market). Total employment of employees in countries P and H is growing. Employees from country P (both those employed in country P and country H) are gaining, whilst employers from country P are loosing. In total, gain of employees from country P is greater than loss of employers. Therefore, the total welfare effect in country P is positive. However, the situation is not that simple if we consider winners and losers of free movement of labor after the establishment of the common market. In country H (the country with higher wages) the employees are loosing whilst the capital owners are gaining. On the contrary, in the country P (the country with lower wages) the capital owners are loosing and the employees are gaining. If we consider the political pressure that can be induced by workers and employees in country H (say, using the most obvious tool of the political pressure nowadays – trade unions), it is clear that it will try to revert the migration and even impose restrictions (or at least transitional periods) on migration flows. Therefore, it seems logical to assume that in country H it will be trade unions who will try to lobby for the introduction of some ban to the free movement of labor. Projecting this situation to the European Union, it becomes obvious who was affecting the governments of EU-15 for keeping the transition periods for the free movement of labor from EU-8. On the other hand, capital-owners and entrepreneurs in country H gain from market integration. Their interest groups can evolve other political pressure on the government aiming to abolish the restrictions to the free movement of labor. It is entrepreneurs and capital-owners in EU-15 who are interested in opening the labor markets to citizens of all New Member States.

2.3.3. Illustrative example: numerical model

The theoretical implications of welfare considerations can be illustrated using the numerical = − example. Let the demand for labor in country H be equal to DH (w) 70 5w and the demand for labor = − in country P be equal to DP (w) 40 4w . = − + Consequently, let the supply of labor in country H be equal to SH (w) 20 4w and the supply of = − + labor in country P be equal to SP (w) 9 3w . = The “no mobility” equilibrium in country H would then be equal to: SH (w) DH (w) , which is equal to:

59 = = 9w 90 , w H 10 , = = SH (w H ) DH (w H ) 20

= On the other hand, the “no mobility” equilibrium in country P would be equal to SP (w) DP (w), which is in its turn equal to: = = 7w 49 , w P 7, = = Sp (w p ) Dp (w p ) 12 > In addition, it holds true that w P w H . It is obvious from the example that if the free mobility of labor is allowed (under the provisions of common market), the direction of labor flow is going to be from country P to country H.

This is where we introduce the indicator of migration cost, α. As it has been shown above, the value of this indicator is crucial for assessing propensity to migration. For instance, if in the case of the numerical example shown above α ≥ 3, propensity to migration will be equal to zero. Let us assume that α < 3. The free mobility equilibrium will be expressed in the form of the following equation: + −α = + −α DH (w) DP (w ) SH (w) SP (w ) Attributing the numbers from the numerical example that is shown above this is going to yield: 70 −5w + 40 − (4 w −α) = −20 + 4w −9+ (3 w −α) which gives: 139 +7α 139 +9α 16 w =139 + 7α, wCM = , wCM = H 16 P 16

Putting α = 2yields: 153 121 wCM = = .9 5625 , wCM = = .7 5625 H 16 P 16 Substituting this into demand and supply functions gives the following results:

CM = − CM = For demand for labor in country H: DH 70 5w H 22 .1875

CM = − CM = For demand for labor in country P: DP 40 4w P 75.9

CM = − + CM = For supply of labor in country H: SH 20 4w H 18 25.

60 CM = − + CM = For supply of labor in country P: SP 9 3w P 13 .6875 To summarize, in country H the domestic labor supply is 18.25 and the labor supply from country P is 3.9375. In country P the employment is equal to 9.75 and the migration to H is 3.9375. The welfare effects in country H and country P are presented below. Welfare gain for country H (welfare gain for employers) is represented by the following equation: 1 1 (migration to H (*) w − w CM ) = .3( 9375 10(* − .9 5625 )) = .0 86133 . Welfare gain for country P: 2 H H 2 1 1 (migration to H (*) wCM − w ) = .3( 9375 .7(* 5625 − ))7 = .1 10742 2 P P 2

2.4. Empirical models

In this section we present two empirical models describing the relationship between inter-EU migrations and real wages differentials.

2.4.1. Empirical model I

As shown in the previous section, full labor migration is unlikely to materialize in the situation of two countries, one of which is a country with higher wages and the other one being the country with lower wages. This happens due to the lower propensity to migration than the one that is needed for reaching the full migration of labor. The propensity to migration is, in its turn, pre-supposed by the existence of migration cost which prevents from the complete equalization of factors (e.g. real wages) to happen in real life. In the model employed in this sub-section uses the data from the selected EU countries (Denmark, Germany, Greece, Luxembourg, Portugal, Spain, Sweden and the United Kingdom) for the period of 1996 to 2001. The data has been purchased from the Eurostat office in London on a special request. The model aims to show that there exists a cost of migration that influences the propensity to migration and therefore plays a very important role in preventing the full labor migration flows to equalize the differences between the countries. In addition, this migration cost is country-specific, as far as the propensity to migration is likely to be different for every country. The data used in this model present numbers of immigrants in selected EU countries by the countries of origin (data from EU-15 countries are used in the model), as well as differentials in hourly labor costs between the same EU countries. Each variable has 372 observations which gives the model

61 a good predictive power and makes it statistically reasonable. The formal model can be presented in the following form:

= α +β − + ε ≠ = = Mij 1(HLCi HLCj) , i i,j 1,.., j;n 1,.., .k (2.1)

where M is a number of immigrants residing in selected EU countries at 1996-2001 (by their EU countries of residence), HLC is hourly labor costs in selected EU countries in 1996-2001 (always corresponding to the previous selection of immigrants for the integration are variable) and ε is an error term. The index i indicates country of immigration and j indicates the country of emigration. The differential (HLCi − HLCj) denotes the difference in hourly labor costs between the country of immigration and the country of emigration. The will is trying to estimate, whether employment incentives represented by a higher wages will lead to higher migration and whether high a difference in wages will be significant for determining labor migration. For estimations of the model presented in (1) was used STATA 7® statistical software. The results can be presented in the following form:

= + − + ε Mij 3701 .921 94 45. (* HLCi HLCj) , [320.057] [35.021]

Both the constant and the hourly labor costs difference constant are significant at 1% level. Adjusted R-squared (0.016) is very low. The results of estimation are presented in table 2.1.

Table 2.1: Results of the estimation of the empirical model (M, number of immigrants) Migration (M), Std. errors (in brackets) 94.45549*** Hourly labor costs difference [35.02118] 3701.921*** Constant [320.057] N 372 R-squared 0.019 Adjusted R-squared 0.016 Source: own estimations; Note: * significant at 10%; ** significant at 5%; *** significant at 1%

The coefficient of the dependent variable has a positive sign and is quite large. Therefore, it might suggest that the larger is the difference in hourly labor costs in immigration and emigration countries used in our model, the larger would be migration. In addition, the p-value of the independent variable

62 is very low, which might suggest that the hourly labor costs variable should be very significant (at the 1% level of significance). However, the coefficient of determination, adjusted R-squared, is very low (0.016) which might show that a great deal of variances in the model still remains unexplained. Therefore, the model provides a very weak explanation of how labor migration between the two countries can be induced by the differences in wages. Due to a large number of observations used in the model it might provide a hint that in the case of selected EU countries labor migration has never reached its massive scale in spite of wage differences. This might be attributed to the existence of the indicator of migration costs that is country-specific and determines different propensity to migration in different countries. Indicator of migration cost is difficult to estimate, especially when models as the one presented above are used. What can be shown using such models as the one above is the fact that migration is not mainly influenced by the differences in wages and unlikely to equalize wages between the countries. This migration is influenced by other intangible factors which enter the propensity to migration. The existence of propensity to migration means that there is still room for people to make independent decisions, for example, disregarding the incentives provided by the wage differences between the countries rather than following them blindly. Economic incentives expressed in wages differentials do not always play the most important role in people’s migration decision. That is why the existence of transitional periods for the free movement of labor in the EU-27 imposed and preserved by some EU-15 countries seems to be merely a political move undertaken on the pressure of trade unions in the respective EU-15 countries.

2.4.2. Empirical model II

In this sub-section an empirical model describing the cross-tabulations of migrations between selected EU-15 and EU-12 countries and real wages differentials is presented. As outlined in the previous section, full labor migration is unlikely to materialize in the situation of two countries, one of which is a country with higher wages and the other one being the country with lower wages. This happens due to the different propensities to migration when the urge to migrate in a country with lower level of economic performance might be very low. The propensity to migration is in its turn influenced by the existence of migration cost which prevents from the complete equalization of factors (e.g. real wages) to happen in real life. In our model we are using the data from EU-27 (EU-25 plus Bulgaria and Romania). The data has been obtained from the national statistical offices of the respective countries, Eurostat (2007), Finfacts

63 Ireland (2007) and European Commission report on remuneration of researchers in the public and private commercial sectors (European Commission, 2007). The data in this model consists of the observations of differences in monthly gross statutory minimum wage rates (full-time adult employees, aged 23 and over) in selected EU-27 countries as well as number of immigrants from selected EU-12 (EU-10 + Romania and Bulgaria) to EU-15 in 2006. In order to make a cross-tabulation of the available data on immigrants from selected EU-12 in the selected EU-15 countries (captured by the country of their citizenship) with monthly minimum wage differentials, wages should be first transformed and expressed in the terms of standardized PPS according to the following formulae: w w = EUR *100 (2.2) PPS δ where w PPS is the salary in terms of standardized Purchasing Power Standards (PPS), w EUR is the salary in EUR and δ is the corrective coefficient 4. The corrected monthly minimum wages seem to be a more precise measure for comparison of migration incentives in various countries as far as the PPPs are also concerned and embedded into the analysis. This approach seems to be consistent with the ones used in similar studies (see for example European Commission report on the remuneration (European Commission, 2007)). Table 2.2 depicts the wages, corrective coefficients and corrected wages for selected EU-27 countries.

4 The corrective coefficients are PPP as defined by Eurostat PPS is the original Purchasing Power Parity (PPP) transformed into a standardized form using the EU25=100 as a base. This artificial unit (currency) is named PPS (Purchasing Power Standard). In order to convert those figures into adimensional figures, they have been divided by the annual exchange rate (Local currency/Euro) in 2006.

64 Table 2.2: Monthly gross statuary minimum wages (in EUR) in selected EU-27 countries, corrective coefficients and corrected wages in EUR in 2006 European Wages in EUR Corrective coefficients* Corrected wages in EUR Countries Belgium 1234.21 104.4 1182.1 Bulgaria 81.80 36.4 224.7 Czech Republic 263.93 53.1 497 Estonia 191.93 55.8 343 France 1217.88 107 1138.2 Greece 658 83.3 790 Hungary 247.26 57.1 480.3 Ireland 1326 122.3 1084.2 Latvia 128.06 48.6 263.5 Lithuania 159.29 46.7 341 Luxembourg 1503.42 113.5 1324.5 Malta 584 69.6 839 Netherlands 1272.60 104.2 1221 Poland 233.01 54 431 Portugal 385.90 87 443.5 Romania 97.07 46.6 208.3 Slovakia 183.14 50.2 364.8 Slovenia 511.87 73.1 700.2 Spain 540.90 89.8 602.3 U.K. 1272.28 106.2 1198 * Adapted from the study by the European Commission (2007); Source: European Commission (2007); Finfacts Ireland (2007); Eurostat (2007).

Now that we have similar grounds for comparison, we can make the cross-tabulation of the numbers of immigrants from the selected New Member States to the EU-15. Due to the incomparability and unavailability of the data obtained from the Eurostat and local national statistical offices, we are only able to make a comparison between the following EU-12 states: Bulgaria, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Slovakia and Slovenia, and the following EU-15 selected countries: Ireland, Spain, the Netherlands and the United Kingdom. The results of the cross-tabulation are presented in Table 2.3.

65 Table 2.3: Cross-country tabulations with the number of immigrants in selected EU-15 countries (displayed vertically) by the country of citizenship (displayed horizontally) and wage differentials (in brackets) between the target and the source countries in 2006. Source/target BG CZ EST HU LV LT PL RO SK SLO country

IRL n.a. 4447 2260 2693 9207 18064 40973 n.a. 7190 85 [587.2] [741.2] [603.9] [820.7] [734.2] [653.2] [719.4] [384] ES 83418 3782 563 2298 1246 11389 32843 287087 3188 426 [377.6] [105.3] [259.3] [122] [338.8] [261.3] [171.3] [394] [237.5] [97.9] NL 1924 1776 284 2029 361 970 10968 3020 1359 256 [996.3] [724] [878] [740.7] [957.5] [880] [790] [1012.7] [856.2] [520.8] U.K. n.a. 8340 1475 7055 9485 17055 162390 n.a. 21725 185 [701] [855] [717.7] [934.5] [857] [767] [833.2] [497.8] Source: Eurostat (2007); Office for National Statistics (2007); Niessen at al. (2006)

From the Table 2.3 it becomes apparent that larger wage differentials of monthly gross statutory minimum wage rates (selected as the basic measure of wages, as far as most of the unskilled immigrants are likely to start earning these wages throughout the first months or years of their stay in the country with the higher level of economic development) does not mean larger number of immigrants. In quite a number of cases (e.g. Polish or Slovak migrants in the U.K. or Polish immigrants in the Netherlands) the number of migrants is relatively big in spite of smaller wage differentials than in case of other migration countries (e.g. differences between U.K. and Latvian or Estonian adjusted wages or Dutch and Lithuanian or Estonian adjusted wages). The data depicted in the table shows that wage differentials themselves cannot be the only determinants of migration. It appears that there is a unique propensity of migration for every country that is pre-supposed by the index of migration cost. Therefore, the index of migration cost which is related to the wage differentials might actually be a better explanation of migration processes. Economic incentives expressed in wages differentials do not always play the most important role in people’s migration decision. That is why the existence of transitional periods for the free movement of labor in the EU-27 imposed and preserved by some EU-15 countries seems to be merely a political move undertaken on the pressure of trade unions in the respective EU-15 countries.

66 2.5. Conclusions: policy implications of free movement of labor on the common market

The essay analyzed the political economy model of migration. Specifically it focused on the shift from single market to the superior form of integration – common market on the example of two countries, one of which is capital-abundant country with a higher real wage rate and the other is labor- abundant country with a lower real wage rate. It appears that on the common market the real wage rate for member countries is unlikely to equalize. The main reason for this is the lack of international mobility which can be expressed by the low propensity to migration. Propensity to migration depends on the indicator of migration cost, α, which consists of tangible and intangible components. Amongst the tangible components range the cost of resettling or adjustment in the new country, the intangible components include, for example, psychological costs of migration (such as breaking the social ties, leaving family or friends or languages barriers). The indicator of migration cost which is a country-specific variable might be used for assessing the volume of international migration: it can be shown that if this indicator is greater than the difference between wages of member states in the CM, there is no motivation for the labor to move (propensity to migration equals to zero). In such a case economic and psychological costs of migration are larger than benefits of wage differentials. Although analyzing the welfare considerations of market integration it can be shown that the total welfare effect is positive, one should distinguish between those who win and those who lose from that process. The index of migration cost can shed some light on the situation of winners and losers of market integration and free movement of labor. Specifically, it appears that usually it is entrepreneurs in the country with higher wages and employers in the countries with lower wages that profit from the market integration and free movement of labor. On the other hand, employees in the country with higher wages and entrepreneurs in the countries with lower wages lose from market integration. Therefore, it is employees in more economically developed countries that are trying to prevent from market integration and especially from opening the labor market to happen. They have a large variety of instruments to achieve their goals with the most effective of them being the lobbying power of the trade unions. This can be easily traced in the recent history of the EU Enlargement – those EU- 15 countries where the trade unions are traditionally strong (Austria, France and Germany) are the ones that are still restraining the transition periods for the labor force from the EU-8, Bulgaria and Romania. On the other hand, countries in which entrepreneurial associations have strong position (United Kingdom, Ireland, Denmark) have opened their labor markets to citizens of the New Members States that joined the EU after May 2004.

67 2.6. References

Brücker, H. and Shroder,P. (2006), “International Migration with Heterogeneous agents: Theory and Evidence”, Paper presented on the Annual Labor Economics Association conference in Prague, September 2006
Brücker, H., Siliverstovs, B. (2004), “The Macro-Determinants of International Migration in Europe: Evidence from Germany”, Paper presented at the conference “Immigration Issues in EU-Turkish Relations”, Bogazici University, October 8-9, 2004
Butler, A. (1967), “Labor Costs in the Common Market”, Industrial Economics, vol.6, no.2, pp. 166-83
Caves, R. (1960), “Trade and Economic Structure”, Cambridge, M.A: Harvard University Press;
European Commission (2007), “Study of Remuneration of researchers in the Public and Private Commercial Sector”, Final Report, Research Directorate General, Directorate D-Human Resources, April.
Eurostat (2007), data retrieved from: http://epp.eurostat.ec.europa.eu
Faini et al. (1999), “Migration. The Controversies and the Evidence”, Cambridge: Cambridge University Press
Faini, R., Venturini, A. (1994), “Migration and Growth: The Experience of Southern Europe“, CEPR Discussion Paper, No. 964, London, May, www.cepr.org/pubs/dps/DP964.asp
Filer, R., Hamermesh, D., Rees, A. (1996), “The Economics of Work and Pay”, HarperCollins College Publishers.
Fisher, S. (1966), “Wage Discrimination in an Integrating Europe”, Sythoff, London
Giersch, H. (1949), “Economic Union Between Nations and the Location of Industries”, Review of Economic Studies , vol. 17, pp. 87-97
Gremmen, H. (1985), “Testing Factor Price Equalization in EC: An alternative approach”, Journal of Common Market Studies , vol. 23, no.3, pp. 277-86
Hansen, J. D. and J. U-M. Nielsen (1997), “An Economic Analysis of the EU”. McGrawhill, London.
Harris, R., Schmitt, N. (2003), “The Consequences of Increased Labour Mobility within an Integrating North America”, in Harris, R. (ed.), “ North American linkages : opportunities and challenges for Canada, Calgary: Calgary University Press;
Holden, M. (2003), “Stages of Economic Integration: From Autarchy to Economic Union”, Parliamentary Research Branch, PRB 02-49B, 13 th of February.
Johnson, H. G. (1965), “An Economic Theory of Protectionism, Tariff Bargaining and the Formation of Customs Unions”. Journal of Political Economy, 73, 256-283.
Krugman, P., Obstfeld, M. (2005), “International Economics, Theory and Policy”, 7 th Edition, Pearson. Addison Wesley.
Lindert, P. (1986), “International Economics”, 8 th edition, Irwin, Homewood, Illinois.
Marshall, A. (1920), “Principles of Economics” (8 th edition, first edition in 1920). Macmillan and Co., London.
Meyer, F. and Willgerodt, H. (1956), “Der Wirtschaftpolitische Aussegewert Internationaler Lohnvergleiche”, in Internationale Lohngefaelle, Wirtschaftpolitische Folgerungen, und Statistische Problematik , Deutsche Bundesverlag, Bonn
Mirus, R., Rylska, N. (2002), "Economic Integration: Free Trade Areas vs. Customs Unions." In Smith, P. and Chambers, E. (eds.), “NAFTA and the new millennium” , U of A Press.
Molle, W. (1994), “The Economics of European Integration, Theory, Practice, Policy”. Dartmouth, Aldershot;
Molle, W. (1994), “The Economics of European Integration, Theory, Practiced, Policy”. Darmouth Publ. Comp., Aldershot.
Molle, W., van Mourik, A. (1989b) (eds.), “Wage Structures in the European Community, Cobvergence of Divergence?”, Gower, Aldershot.
Mourik, A. van (1987), “Testing Factor Price Equalization in EC: An alternative approach: A Comment”, Journal of Common Market Studies , vol. 26, no. 1, pp. 178-197
Mourik, A. van (1993), “Wages and European Integration”, PhD, RL, Maastricht
Mundell, R. (1957), “International Trade and Factor Mobility”, The American Economic Review , Vol. 47, No. 3 (June), pp. 321-335;

68
Mundell, R. (1968), “A Theory of Optimum Currency Areas”, in “ International Economics ”, New York: Macmillan, pp. 177-186;
Myrdal, G. (1956), “Trade as a Mechanism of International Inequality”, in Meier, G. (ed.), “ Leading Issues in Economic Development” , 5 th edition, New York: Oxford University Press, 1989;
Niessen, J., Schibel, Y., Thompson, C. (eds.) (2006), “Current Immigration Debates in Europe: A Publication of the European Migration Dialogue”, Ireland, Piaras Mac Einri for NCCRI.
Office for National Statistics (2007), United Kingdom, data retrieved from: www.statistics.gov.uk
Pelkmans, J. (1997), “European Integgration, Methods and Economic Analysis”. Addison Wesley, New York.
Rassekh, F., Thompson, H. (1998), “Micro convergence and macro convergence: factor price equalization and per capita income”, Pacific Economic Review , 3(1), pp. 3-11;
Seers et al. (1980), “Integration and Unequal Development, The Experience of EC”, St. Martin’s Press, New York
Svendsen, G. T. (2003), “The Political Economy of the European Union”. Woodward Elgar, Cheltenham.
Tovias, A. (1982), “Testing Factor Price Equalization in the EEC”, Journal of Common Market Studies , vol. 20, pp. 165-81
Turnovec, F. (2003), “Political Economy of European Integration”. The Karolinum Press, Prague.
Viner, J. (1950), “The Customs Union Case”, Stevens and Sons, London.
Williamson, J. (1995), “The Evolution of Global Market Since 1830: Background Evidence and Hypothesis”, Explorations in Economic History, 32 (1995), pp. 141-196;
World Bank (2005), “World Development Indicators”, Word Bank, Washington, D.C.
Young, A. (1991), “Learning by Doing and the Dynamic Effects of International Trade”, Quarterly Journal of Economics , 106, pp. 369-405

69 3. Ready to go? Migration potential and EU Enlargement: lessons for the Czech Republic in the context of Irish migration experience. 5

3.1. Introduction

Among the most frequently named catalyzes of migration from a poor country or region to the wealthier one are wage differentials, economic disparities between regions, differences in GDP per capita and unemployment differentials. The works of Hannan (1970), Todaro (1969), Harris and Todaro (1970) and Walsh (1984) usually focus on factors that are solely economic in their nature. Few of the migration theories look at the EU accession, other than the institutional changes that allow for easier work permits, as the major factor of inducing migration moves. This is why it was surprising to witness the discussions around the transitional period introduced for the labor force from the EU-86 Member States after the EU Enlargement on the 1 st of May 2004. While the majority of experts estimated migration flows from the EU-8 Member States as rather modest ones, this view has not been accepted by the policy-makers and public opinion in the EU-15 Member States. The EU Enlargement was consciously and unconsciously tied up to the triggering off the mass inflow of workers from the “new” countries7 (Bauer and Zimmermann, 1999; Boeri and Bruecker, 2000). The recent experience shows that three EU-15 countries – Ireland, Sweden and the United Kingdom – which did not follow other EU Members and opened their labor markets to the newcomers do not experience any adverse effects associated with mass migrations or over-flooding of their labor markets with cheap labor force. The freedom of movement (of persons) is, alongside with the freedoms of movement of capital, goods and services, one of the basic rights determining European Union itself. In theory, the creation of a Single Market pre-supposes the creation of many additional employment and earnings opportunities for the workers in various Member States of the EU. In addition, unrestricted labor flows should substantially reduce regional differences in economic opportunities within the EU.

5 We are grateful to the Irish Social Science Data Archive for providing the data. 6 Many authors use terms “core” EU countries and “new” EU countries referring to the EU members before and after the Enlargement on the 1 st of May 2004 (specifically the CEE countries without Cyprus and Malta). We will call them EU-15 and EU-8 respectively for the sake of simplicity. 7 In this context those countries are mainly represented by the Central and Eastern European countries and the Baltic States which have joined the EU on the 1.5.2004.

70 The basic focus of this paper is to advocate the idea the scope and size of migrations might differ for different countries. In another words, they strongly depend on the migration potential of population of a given country. For the population of countries with high migration potential it is more likely to react to the problems at home by increased emigration to the countries with higher income and lower unemployment (Fidrmuc, 2002). Some countries, however, do not posses that migration potential and thus might not react to wage and unemployment incentives abroad. This might be the case of the EU-8 Member States represented by the CEECs. We will test these hypotheses basing on the experience of two countries – Ireland, as a representative of the classical emigration country in the 1970s and 1980s, but belonging to the EU-15; and the Czech Republic, which is a newcomer and a country which has got rid of the heritage of communist regime and socialist economy relatively recently

3.2. Methodology & literature review

The reasons leading to migration and the decision to move are reached through an evaluation of the incentives and obstacles to migration. Here push (unbearable or threatening conditions in the home country) and pull factors (incentives in the countries of immigration) play an important role (see for example Ravenstein 1876, 1885 and 1889; or Dorigo and Tobler, 1983). Incentives may contain increased employment opportunities, better housing, or a more sympathetic political or cultural environment. Obstacles to migration are unfamiliarity with the new location, lack of information about distant opportunities, language barriers, transportations costs or difficulties and immigration or emigration restrictions. The pattern of migration is more or less the same everywhere: people leave poorer states in search of higher-paying jobs in richer states. The same patterns hold true within the European Union. The research literature on migration suggests that a very complex set of inter-relationships exist between social, psychological, background social structural and educational factors, as well as certain constraints on individual’s migration decisions. Individual’s motives or reason for migrating require explanation. As De Jong (1983) suggests they are mainly explainable in terms of individual’s locations in the economic and social structure, and level of educational background: because everything to what the individuals are aspiring to is given and pre-determined by their educational level and social-economic or socio-cultural characteristics. However peoples’ beliefs (or perceptions) of how satisfactory local or foreign communities may be also depend on the nature of the local economy or labor markets (mainly on such factors as the rurality

71 of the community and its unemployment level). Thence a direct linkage exists between the nature of the home community’s economic structure and satisfying one’s economic aspirations locally. For its rather short history of its existence, the EU has experienced five enlargements already. The story with each enlargement was different, and only four of them, including recent Enlargement, can encounter as those having an impact on the international migration and the issue of inter-EU free movement of labor. EU membership, as it appears, by no means necessarily induces uncontrolled immigration into the core EU Member States. On the contrary, in the aftermath of their EU accession net emigration from Greece, Spain and Portugal in the first half of the 1980s has substantially declined. In spite of that, recent EU Eastern Enlargement, encountering mostly post-Communist countries of the Eastern and Central Europe aroused many doubts to what will be the effect of introducing the free movement to labor onto the countries with their GDP per capita hardly reaching the EU-15 average. There are various reports and papers on that issue (see Bauer and Zimmermann, 1999, Boeri and Bruecker, 2000; Zimmermann, 2004) assessing the post-enlargement migration, however none of them contains a catastrophic scenario. Even now, more than three years after the Enlargement, the issue remains open (and the free mobility of labor chapter closed) so to speak - all this in spite of the fact that there has never been registered a mass migration basing on the data from those EU countries who opened their labor markets to the workers from the EU8 Member States as the first, namely Ireland, Sweden and the United Kingdom. There are some methodological issues to be discussed here in connection with what has been said previously. First, even though the migration from the EU-8 Member States to the EU-15 countries might be generally caused by the economic incentives, it does not necessarily holds true that the "response ratio" will be the same as in the case of traditionally emigration countries, such as Ireland or Portugal. Second, it might be relevant to measure the migration potential of the population in the EU-8 Member States. Various reports (see for example Fidrmuc, 2002) suggest that the labor mobility in the EU-8 Member states has been low and falling, this happening even in spite of increasing wages and unemployment disparities across regions. In fact, it seems that only the prosperous regions are dealing with migration, moreover this migration covers high-skilled workers. If this is true, then the EU accession will have the adverse effect on the EU-8 Members.

72 3.2.1. Migration patterns and economic shocks: the case of Ireland

The topic of emigration, including labor migration in the first place, has been of considerable importance for Ireland since the early part of the last century. For much of the last century emigration has remained high and the population decline continued until 1961. But even in the 1960s emigration continued and after a decade of unprecedented inflows in the 1970s, net outflows resumed in the 1980s. Part of the reason for the deeply-felt Irish sensitivity to the emigration problem (and the reason why it becomes a focus of social, economic and political debate) stems from its long-standing historical nature. While the populations of most European countries have substantially increased since the middle of the 19 th century, Ireland, on the other hand, has been unique in recording a population decline for the most of the period. Most of the more or less industrialised Northern countries were net importers of people during the economically prosperous 1960s. This period is represented by the era of the “guest worker”, as millions migrated from the countries of Mediterranean basin to find employment in the industrial centres of Germany, France and other northern European countries. This pattern continued on into the early 1970s but the scene changed in middle of that decade when economics conditions deteriorated following the first oil price shock in late 1973. Not surprisingly, migrant workers, who were generally unskilled and held less secure jobs in the economies concerned, suffered disproportionately through unemployment in the following economic fallout. Many returned to their home countries; indeed, in some countries those workers, who originated from outside the EU were forced to do so by various means. Ireland also experienced a reversal of its long-standing migration pattern during the 1970s, but the underlying reasons were different from those indicated above for the other EC countries. Most of the influx derived from improvement in employment opportunities and enhanced social conditions at home, rather than from a desire to escape from the recession abroad. Indeed, even though the effects of the mid decade recession in Ireland were as great as they were in other countries, the net inflow of population continued throughout its duration. Emigration, when viewed in a wider European context over the post-war period, is not a feature which is unique to Ireland. One must, therefore, when considering the wider context of a deregulated EU labour market, exploit the possibility of continuing intra-European movements, with a likely gravitation towards these centres where economic growth will tend to be concentrated. Differences in demographic structure between the EU countries are also likely to contribute to such movements.

73 Waves of Irish migration

In the decades since political independence of Republic of Ireland there has been a substantial and almost never-ceasing net outflow of emigrants.

Chart 3.1: Gross migration flows: Republic of Ireland (1987-2003)

80.0 70.0 60.0 50.0 40.0 30.0 20.0 10.0 0.0 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

Emigrants Immigrants

Source: Aidan Punch and Helen Cahill, CSO (2004)

As the NESC report of 1991 states this peaked in the decade 1951-1961. After the decline in the net outflow (1961-1971) and the subsequent net inflow (1971-1981), the net outflow figure rose again rapidly from 1981 (NESC, 1991). This process has however been stopped in the beginning of the 1990s and for the last decade of the 20 th century Ireland has been the net importer of the labour force. Some authors (Mac Laughlin, 1994 and Hazelkorn, 1990) show that for the most part of the 20 th century Ireland remained highly-dependent upon Britain both as a market for her exports and as a host of her “surplus” labour. The population shifts from Ireland to the U.K. were merely a part of a broader general process of urban-rural drift. The situation of the WWII when the neutral Ireland was an automatic supplement of labour for Britain continued to play its role in the 1950s and 1960s. Labour emigration in this context was clearly caused by the Ireland’s geographical proximity to Britain. The location of Ireland caused its position of an emigrant nursery for the U.K. as well contributed to the process of peripheralisation in Ireland (Coulter, 1994). While the improving conditions at home made many Irish migrants to return back, the beginning of 1980s was again marked by the high numbers of net outflows. As Mac Laughlin (1994) points out since 1980s, emigration is particularly affecting Irish teenagers and represents a mixture of urban-rural migration. Results from most recent surveys show that this “new wave” emigration is still largely a

74 survival strategy for working-class and small-farming families and primarily affects middle-class families. Furthermore, the traditional view of an Irish migrant looked upon as an unskilled rudimentary worker should be revised – more and more emigrants posses second- and third-level qualifications (NESC, 1991). The EC accession seemed to have no substantial influence on the Irish labour migration – in fact the period of 1971-1981 was marked by the decrease in emigration outflow. Britain and the English-speaking countries absorbed the largest part of Irish excessive labour flows (Niall Farrell, 1991). It is easy to understand why the rate of emigration from Ireland has been so high over the past two centuries. Few other countries face the combination of circumstances that have led to such a high long- term rate of emigration from Ireland: a high birth rate and the pressure on the labour force, low average income levels and a long-term rate of economic growth that has not been sufficient to close the gap in living standards relative to the other EU members, the widespread use of English and free access to Britain.

3.2.2. Migration potential of the Czech Republic

Unlike Ireland presented in the previous case, the Czech Republic is a country with low labour mobility. This concerns both inter-regional migrations and outward international migrations.

Chart 3.2: Gross migration flows: Czech Republic, 1991 to 2003

70000 60000 50000 40000 30000 20000 thous. people thous. 10000 0 1990 1992 1994 1996 1998 2000 2002 2004 year

Emigrants Immigrants

Source: Czech Statistical Office (2005)

Similarly to Ireland with the UK as the market for its “surplus” labour, the Czech Republic has had Germany and Austria with only one difference: political and economic barriers did not turn these

75 migrations into the large-scale flows. Legal employment of the Czech citizens after the collapse of the Iron curtain was processed in several forms: seasonal works (up to three months in a year), employment on the basis of agreements, bilateral employments with educational purposes and so- called “pendling” (VUPSV, 2001). “Pendling” or simply working in one country without residing there (a worker usually returns home (commutes) every night and spends most of his earnings in the home country) existed on a large-scale in the first half of the 1990s (12000 workers annually), however by the end of the decade the number of Czech commuters has decreased by 60 per cent (VUPSV). One of the reasons for that was, on one had the decrease in issuing of working permits for Germany and Austria and on the other hand loss of interest in low-paid and low-qualified jobs. According to Austrian statistics, prior to the EU enlargement (and thus gaining some special privileges for the Czech workers) employees from the Czech Republic were merely 10 per cent of the largest group of employees from the former Yugoslavia and yielded 10.8 thousand people in the 2002 (Austrian labour statistics, 2003). The second-largest group were the Turks and later followed Hungary and Poland. The same situation could have been observed in Germany: in the 2002 Czech employees constituted 4 per cent (13.2 thousand persons) of the largest group of foreign employees from Turkey (German labour statistics, 2003). This comparisons show that under the equal treatment and labour law, only a small fragment of the Czech labour force was seeking employment abroad, even though the geographical proximity of the Czech Republic is considerably higher than in the case of Turkey of Yugoslavia. One more important fact that we mentioned when describing Irish migrations, was the knowledge of language. One of the most crucial determinants of the Irish migration is the fact that the Irish speak English and thus can easily take up employment in the U.K. or the USA. With the increasing popularity of English as lingua franca more and more young people from the Czech Republic are likely to look for employment opportunities in the English-speaking countries, which can also be used as an explanation of loosing interest in Austria and Germany. When it comes to regional migrations, the most important factor of accessing the country’s migration potential, it can be stated that like any other post-communist country (perhaps with the exception of Poland) the Czech Republic has very low regional migration turnover in comparison with the EU-15 Member States. The reasons to the low regional mobility are mostly the housing-stock market and the cultural specifics (mostly granted employment in the Communist times and thus tightness to the place of work and residence). Some authors (for example Vavrejnova, 2004) suggest that up to now the decisive factor of the low regional migrations has been badly-functioning housing-stock market. Privatization of state-

76 owned apartments, especially selling apartments in the regions with high unemployment has contributed to “chaining” local labor force in these regions and helping to spreading the speculations with those apartments (this, in its turn, lead to decreasing of available housing stock for the social cases). In spite of privatization, private apartments still constitute just 46 per cent of the housing stock in the Czech Republic (Czech Statistical office, 2004).

An interesting fact is that change of the workplace and moving closer to the workplace are the most irrelevant factors of the regional migrations according to the Czech Statistical office. In 1999, when the unemployment rate in the Czech Republic was about 8.5 per cent, these two reasons made 6 per cent of the total regional migration turnover and remained on the same scale even when migration had risen to 10 per cent in 2002 (Czech Statistical Office, 2004).

3.3. Data

We have elaborated a comparative study of two countries: Ireland and the Czech Republic in order to demonstrate the dependence of migration flow on the economic factors and analyse the migration potential. We analyze inward, outward and net (immigration minus emigration) migrations in both countries. Another part of the analysis covers the migration potential measured by the number of migrants from and to the administrative regions of the two selected countries as a reaction to the economic incentives such as expected income and employment rate within each state. To our mind Ireland is distinguished by the high migration potential and Irish labour force reacts quickly to the economic changes at home and incentives elsewhere. Irish population has been marked by a high rate of emigration throughout the last century subject to the worsening conditions at home. The Czech Republic, on the other hand is a typical example of the post-Communist country, which has undergone the change from zero unemployment to the problems of the capitalist economies. Some evidence let us suspect that the migration potential here is low and thus nothing can verify alleged mass migrations to the West. This is given by local specifics and, to the great extent, by the heritage of the Communist economy. The data are covering the periods of several years and the scope of data differs somehow. To make the data, and thus the analysis of both countries, comparable we had to adjust the data from Irish Central Statistical Office to those from the Czech Statistical Office. The problem with the data was mainly the fact that Irish migration and economic data are gathered in whole-national Censuses taking place every 5 years. We have focused on the Censuses of 1981, 1986, 1991, 1996 and 2002 and used data on emigration, immigration (both without distinguishing the regions of origin or destination of

77 migrants), disposable income 8 and employment by micro-regions (County or DD1 level) in Ireland to generate the year-by-year values. While no emigration data is explicitly collected in the Irish Census, it has been possible to impute estimates of emigration flows. Utilising information on population change, deaths and births at the county level, we were able to impute the number of net migrants between Census points (generally 5 years). This together with information on the number of immigrants in the previous year allows us, using interpolation methods to calculate the number of emigrants. We have also performed an exercise of clearing our data on Ireland off the people aged 65 and over (as those not being relevant for inter-regional migrations as far as many pensioners are retiring and returning back to Ireland from Scotland and USA). In this way the data on international migration from and to Ireland (1980-2002) as well as on Irish regional migrations (1986-2002) has been gathered. The data for the Czech Republic covers the period of 1991-2006 in case of international migrations (1993-2003 in case of regional migrations) and represents the annual observations of disposable income (average monthly salary in Czech Crowns) by regions and employment structure (officially recorded % of people in productive age employed and unemployed in the main sectors of economy). We have also included GDP per capita for each region and perform the analysis with and without it. Most of the data has been obtained from Eurostat, Irish Statistical Office and the Czech Statistical Office. The regional data from the Czech Republic and Ireland are divided by regions (called “kraje” in Czech, 14 regions in whole) in the Czech Republic and counties (34 counties) in Ireland. The data report overall immigration and emigration per region, without distinguishing the regions of origin or destination of migrants. The fact that the data report migration of populations rather than labor migration may cause problems when interpreting the results, but this is a problem of all migration studies and is subjected to the availability of the data. The comparative study will show the patterns for both countries: one with high migration potential and the other with the low one and will help us to draw the results and conclusions for the inter-EU migrations.

8 We utilise time invariant disposable income as our repressor as far as time series of the county level - disposable income is not sufficiently long. However as there is a high level of temporal correlation, this is not a major issue.

78 3.4. Empirical model

In this section of our paper we will draw an econometric model estimating the dependence of migration on economic factors and migration potential of the population. When it comes to estimating the interdependence between the migration and various economic factors, the usual procedure adopted in migration studies (see for example Walsh, 1974 or Geary and O’Gráda, 1989) is to relate the annual rate of net migration, M (net emigration is treated as negative migration) with wage ratio of emigration and immigration country (W emig /W imm) and unemployment rate ratio (U emig /U imm) . The aim of such studies is to try to explain the year-to-year fluctuations in the level of migration. It is valuable to combine the findings of these studies with those of studies of the factors that increase the likelihood that individuals will emigrate. Most of the migration studies depict evidence that these variables affect migration in the expected manner. This seems to be consistent with the hypothesis that fluctuations in the rate of emigration reflect changes in labour market conditions (wage levels and employment opportunities) in country of emigration relative those to the country of immigration. In spite of that, a number of unresolved issues arise from the econometric evidence. The problem is, of course in the data sources. While the data on aggregate net flows are available the allocation of this total among individual years could be subject to error. Besides, the annual net migration data relate to the whole population and not just those in the labour force or of working age. These limitations reduce the reliability of econometric models (Keenan, 1981). In their study Geary and O’Gráda (1989) do the migration model of Irish migrations to the UK for the period of 1953-1982. They advocate the use of expected income (Y) and relative expected income (RELY- ratio of expected income in Britain to that in Ireland) as key determinants of migration, where Y is defined as follows: Υ = uB + 1( − u)W 1( − t) , where u = unemployment rate (as a proportion), B = the level of social welfare payments to the unemployed, W = gross wages and t = income tax rate (including social security contributions). Expected income is therefore a weighted average of the income received by the employed and unemployed labour force, using the proportions of the labour force unemployed and employed as weights. RELY variable combines the influence of both relative unemployment rates and relative wages in a single variable. Geary and O’Gráda report the following results for the Irish migrations for the period of 1953- 1982: M = 67.88 – .758*RELY + .195*M-1, R 2 = 0.852,

79 where M = net Irish migration of 15-64 year olds expressed in thousands and RELY = Y irl /Y uk expressed as an index to base 1980 = 100 (Geary and O’Gráda, 1989). The result shows that the higher is expected income in Britain relative to the Republic of Ireland, the higher the rate of emigration.

3.4.1. Dependence of international (outward) migrations on economic incentives at home

In this section of our paper we will attempt to construct an econometric model referring to the study of Geary and O’Gráda (1989) in order to estimate the impact of economic factors in the emigration country on the outward migration. We argue that the econometric model can be tested in the linear form for two basic reasons: i) the model has been tested for linearity and found linear 9; ii) similar migration models in similar studies were also linear. The model can be presented in the following way:

= β +β +β +β + = Mi 0 1Y 2W 3U ui , i 2,1 ,.... n (3.1) where M i s the net emigration from the country, Y is GDP per capita, W is the average wage in the country and U is the unemployment rate in the country. Unlike the previous studies quoted, because of the multitude of potential destinations of migration, we will not estimate the dependence of emigration on the difference in economic factors of the country of origin and countries of immigration, but rather focus on the push factors (worsening the economic conditions at home) as an explanatory variable of emigration. This decision pre-determines the choice of the independent variables for our model. The core of this approach is the fact that we are focusing on the migration potential of the population and therefore the first step will be to access the reaction of the population to the internal shocks: for this purpose we run the same regression for the Czech Republic and Ireland.

9 The testing has been done in two ways: by checking the relationships of dependent and independent variables (plotting them into a graph) and by plotting a graph of residuals of independent variables. Both methods lead to the conclusion that for this particular model and this particular data set linearity of the model can be accepted.

80 The case of the Czech Republic

Table 3.1 represents the results of the regression for the net migration (dependent variable) from the Czech Republic for the years of 1991-2006 and the independent variables (GDP per capita, nominal wage and unemployment). Limited number of observations is due to the data availability, as far as only statistical data from 1991 onwards can be used as satisfactory measures of GDP per capita, nominal wage and employment (there is a question, however, to which extent was the data gathered during the Communist regime manipulated by the authorities).

Table 3.1: Determinants of net migration: Czech Republic (1991-2006), net migration (NM) Net Migration (NM), Std. Errors -232.9193* GDP per capita [114.1499] .0041139** Nominal wage [.0017487] -.0045457 Unemployment [.0060188] -572.4722*** Constant [174.4089] N 16 R-squared 0.74 Adjusted R-squared 0.64 Source: own estimations. Note: Standard errors in brackets; * significant at 10%; ** significant at 5%; *** significant at 1%

Overall, the model seems to have a good predictive power with the adjusted R-squared that equals to 0.64. However, the unemployment variable came through as an insignificant one. The insignificant coefficient on the unemployment variable in the model might be a proof of the assumption that the factors influencing migration in different countries vary. For instance, emigrants from the Czech Republic might not consider the unemployment in their country when making their migration decisions. In order to test this we run the backwards stepwise regression 10 in order to make sure whether unemployment should belong to the model specified in (3.1). The results of the estimation are presented in table 3.1.1 that follows. It can be seen that if the backward stepwise regression is performed, the only remaining significant independent variable is GDP per capita which becomes significant even at 1% interval (table 3.1.1).

10 For the explanations of backwards stepwise regression see section 1.7 of this dissertation.

81 Table 3.1.1: Determinants of net migration: Czech Republic (1991-2006), net migration (NM), backwards stepwise regression Net Migration (NM), Std. Errors -19.1074*** GDP per capita [5.588155] -2.64183** Constant [1.1314] N 16 R-squared 0.65 Adjusted R-squared 0.61 Source: own estimations. Note: Standard errors in brackets; * significant at 10%; ** significant at 5%; *** significant at 1%

Due to this fact it seems justified to test the model defined in (3.1) without the unemployment variable. The “improved” model can be written as the following:

= β +β +β + = Mi 0 1Y 2W ui , i 2,1 ,.... n (3.2) where M i s the net emigration from the country (the Czech Republic), Y is GDP per capita, W is the average wage in the country. The results of this estimation are presented in table 3.1.2. It can be noted that after the unemployment variable is taken out of the model, the significance of the model improves slightly (the adjusted R-squared increases by 8% and the signs of the independent variables increase) with the significance of the other two variables (GDP per capita and nominal wage) remaining the same. The sign of the nominal wage variable is as expected: it suggests that the increase in the nominal wage would lead to the increase in net migration and decrease respectively. An interesting fact is the behavior of GDP per capita variable that comes through all three modifications of the model. It seems that it has a negative influence on the net migration, thus an increase in GDP would lead to decrease of net migration, indicating a rise in emigration (as far as net migration is, by definition, the difference between immigration and emigration). The possible explanation for this can be the following: the majority of emigrants from the Czech Republic are well-qualified 11 and thus, in times of recession, it is the unskilled workers who suffer the most making migration less sensitive to changes in the disposable income they are provided with (expressed by the GDP per capita).

11 We speak here about “white-collar” migrations: i.e. technical and IT specialists, managers, etc. In the case of the Czech Republic, we do measure the emigration and immigration from and to the country without considering seasonal and random workers.

82 Table 3.1.2: Determinants of net migration: Czech Republic (1991-2006), Net migration (NM) Net Migration (NM), Std. Errors -245.0262* GDP per capita [109.6472] .0047865** Nominal wage [.0016635] -585.872*** Constant [170.3701] N 16 R-squared 0.79 Adjusted R-squared 0.72 Source: own estimations. Note: Standard errors in brackets. * significant at 10%; ** significant at 5%; *** significant at 1%

The results of the model specified in (3.2) can be therefore reported in the following way: = − − + + Mi 585 .872 245 .026 *Y .0047 * W u i (3.3)

These results show that the unemployment variable might be of negligible significance in the model used for explaining determinants of net migration for the Czech Republic. Instead, it seems that GDP per capita and nominal wage represent more decisive factors that explain the net migration for this particular model and the data set.

The case of Ireland

Now, let us do the same exercise for the Irish emigration. This time we will use more extensive database going back to the 1980 in order to see the development of the Irish migration rate over time and in the periods of extensive outward migrations. Table 3.2 below presents the results of the regression for the net migration (dependent variable) from the Republic of Ireland for the years of 1980-2002 on the independent variables (GDP per capita, nominal wage and unemployment). The data is of a time-series nature and includes periods of large scale emigration in the late 1980’s and immigration in the late 1990’s.

83 Table 3.2: Determinants of net migration: Ireland (1980-2002), Net migration (NM) Net Migration (NM), Std. Errors .0000364 GDP per capita [.000719] -.0007804 Nominal wage [.0027722] -1.527573* Unemployment [.864728] -94.04865 Constant [228.4288] N 23 R-squared 0.37 Adjusted R-squared 0.23 Source: own estimations. Note: Standard errors in brackets; * significant at 10%; ** significant at 5%; *** significant at 1%

Looking at the table 3.2 one can notice that the R-squared values are lower than expected, although this is quite common for similar studies that involve the analysis of such social and demographic phenomenon as migration. It is accepted in the research literature that for social studies coefficients of determinations of about 0.2 and more are acceptable (Visek, 2006). This is also due to the fact that coefficients of determination are only one of the set of indicators reporting the “quality” of the model. Contrary to the results of the model (3.2) that uses the data from the Czech Republic, unemployment variable seems to be the only significant determinant in the case of Irish net migrations. We run the stepwise regression in this case, trying to see whether the model defined in (3.1) should be adjusted in the case of Ireland as it has adjusted in the case of the Czech Republic. The results of this estimation are presented in table 3.2.1.

Table 3.2.1: Determinants of net migration: Ireland (1980-2002), Net migration (NM), backwards stepwise regression Net Migration (NM), Std. Errors -1.958388*** Unemployment [.5810984] 18.31967** Constant [7.333411] N 23 R-squared 0.35 Adjusted R-squared 0.32 Source: own estimations. Note: Standard errors in brackets; * significant at 10%; ** significant at 5%; *** significant at 1%

84 The results of the stepwise regression show that the GDP per capita and nominal wage variables are neither significant in the “full” model defined in (3.1), nor come through as significant in the stepwise regression. Based on these results, it might seem justified to apply the model of net migration without the GDP per capita and nominal wage variables. The “improved” model can be written as the following:

= β + β + = Mi 0 1U u i , i 2,1 ,.... n (3.4) where M i s the net emigration from the country (Republic of Ireland) and U is the unemployment rate in the country. The results of estimating model can be also reported in the following way:

= − + Mi 18 .319 .1 958 * U u i (3.5)

It is apparent that is the case of Ireland, unemployment might be the only decisive factor that influences the immigration and emigration in the country. It is also apparent that when the two independent variables were taken out of the model, the significance of the remaining variable and the overall predictive power of the model improved slightly.

Comparison of both countries: determining net migration

If we compare the results for both countries, the Czech Republic and Ireland, we can clearly see the difference between the two. It is clear that net migration in each country is influenced by different factors. It appears that the significance of the estimates are opposite in each country with unemployment being the main decisive “push” factors for the Irish migrants. Comparing the push factor mode with the Geary and O’Gráda (1989) model highlights the important role pull-factors play, indicating the fact that the UK and Irish labor markets operate similar to a regional market rather than international labor markets. It also highlights that for much of the period in question; unemployment and wealth followed a different pattern, with high unemployment levels continuing, except for the end of the period, even during periods of higher growth. Thus differential labor demand between Ireland and the UK was an important driver. We can clearly see that in spite of methodological differences, the outward migrations from Ireland were dependent on the internal economic factors to the less extent that those in the case of the Czech Republic due to the importance of the pull factors experienced by Ireland from the UK.

85 3.4.2. Dependence of regional migrations on economic incentives at home

According to Fidrmuc (2002) migration can be one of the principal mechanisms for absorbing asymmetric shocks. A region hit by, say, negative demand shock will face increasing unemployment and decreasing wages. There are several ways to absorb this shock: for instance, residents of the regions can move to the region with higher wages and lower unemployment; or lower wages can attract potential investors. In this part of our model we will compare the emigration potentials of Ireland and the Czech Republic by measuring the responsiveness of regional flows to respective regional economic characteristics. The dependent variables are gross and net migration flows. The data record the total number of migrants arriving to or leaving the region in a given year without any special regard to their destination. We assume that in order for the country’s population to have strong migration potential, net immigration should be positively related to average wages and negatively to unemployment, while gross emigration should be positively related to unemployment and negatively to wages. The vice versa situation suggests the existence of immobile population which does not react to asymmetric shocks. The formal model can be represented by the following three equations:

= β +β +β + = NM h 0 1DISP _Yh 2EMPRATE _Yh ui i 2,1 ,.... n (3.6) = β +β +β + = EM h 0 1DISP _ Yh 2EMPRATE _ Yh ui i 2,1 ,.... n (3.7) = β +β +β + = IM h 0 1DISP _ Yh 2EMPRATE _ Yh ui i 2,1 ,.... n (3.8) where NM is the net migration flow, EM is the emigration flow (total numbers of migrants leaving a district in a given time period without identifying the destination) normalized by population and IM is the immigration flow (total numbers of migrants arriving to a district in a given time period without identifying the destination) normalized by population. DISP is the dependent variable representing disposable income per capita in the region and EMPRATE is the dependent variable that represents the employment rate in the region. We expect both employment and disposable income to have a disproportional relationship with respect to emigration. In order for migration to be effective as a channel of regional adjustment, gross

86 (and net) immigration should be positively related to both disposable income and employment, while emigration should be negatively related to employment and wages. It does not seem reasonable to apply stepwise regressions in this case due to the fact that we only operate with two independent variables in each of the models.

Table 3.3: Determinants of regional migration: Republic of Ireland (1986-2002), Immigration flow (IM), Emigration flow (EM), Net migration flow (NM) Immigration (IM) Emigration (EM) Net migration (NM) Std. errors Std. errors Std. errors Disposable income -3.21e-06 6.00e-06 -.0002455 [.0000998] [.0001228] [.0002856] Employment rate .1001516*** .1546487*** .1739801*** [.0204264] [.0251237] [.0584371] Constant -.0318659*** -.0609712*** -.0665878** [.0109108] [.0134199] [.0312143] N 136 R-squared 0.16 0.24 0.06 Adjusted R-squared 0.15 0.23 0.04 Source: own estimations. Note: Standard errors in brackets; * significant at 10%; ** significant at 5%; *** significant at 1%

Table 3.3 summarizes the determinants of regional migrations for the Republic of Ireland 12 computed using the models’ equations described in (3.6)-(3.8). We can see, as expected, that the Irish population is highly mobile reacting to the changes in economic incentives within region. In each case, we see that county income differentials are not significant, while the employment differentials are highly significant and thus important in driving migrations. While this may be due to the fact that there is no inter-temporal variation in this variable (merely inter-county variation), it may indicate that individuals migrate more in search of work than in search of a higher standard of living. To compare the situation in both countries subjected to our analysis, the output for both should be comparable and presented in the similar way. In order to achieve that we have compared the results for multiply regression model of Ireland with those of the Czech Republic. The only modification we had to introduce was that we reversed the relationship for employment rate. Due to the data availability we have used the data for unemployment rate instead of employment rate, assuming that its relationship with the immigration, emigration and net migration will be negative, positive and negative respectively. The model based on equations (3.6)-(3.8) has been also computed for the Czech Republic. The following table (table 3.4) summarizes our relevant findings for this country:

12 We report pooled regression model as county effects were found not to be significant in a panel data model.

87 Table 3.4: Determinants of regional migration: Czech Republic (1993-2003); Immigration flow (IM), Emigration flow (EM), Net migration flow (NM) Immigration Emigration Net migration Std. errors Std. errors Std. errors Disposable income .0407581*** .0403935*** .0003358 [.0041441] [.0028951] [.0020283] Unemployment rate -193.9096** -309.6038*** -45.91622 [77.5906] [-55.71] [37.97565] Constant 271.1135 149.1272 907.5348* [863.2292] [603.049] [422.4956] N 154 R-squared 0.39 0.59 0.009 Adjusted R-squared 0.38 0.58 -0.003 Source: own estimations. Note: Standard errors in brackets; * significant at 10%; ** significant at 5%; *** significant at 1%

The results show that for this specific model and data set the pattern that we can obtain for migration in the Czech Republic does not meet the criteria for migration being effective as a channel of regional adjustments. At the county level as the national level, the unemployment rate in the regions of the Czech Republic is not a significant driver of migration flows. Unemployment is negatively related to emigration indicating that the higher is unemployment, the lower is emigration. The coefficients of disposable income as an independent variable for immigration and emigration (between regions) are lower on average than those for the counties of Ireland. The comparison of the results obtained and reported in tables 3.3 (Republic of Ireland) and 3.4 (the Czech Republic) suggests two basic conclusions that can be made here: First, in comparison with the Irish population Czech population does not seem to react to the incentives or worsening the conditions throughout the country and internal migrations are likely to be effected by other factors than economic incentives in the neighboring region. Second, in some cases (e.g. worsening economic conditions at the place of immediate residence) people in the Czech Republic might rather decide to stay and explore some other alternatives than migration than to move to some other region in search of employment opportunities. The Irish people, under the same conditions as those prevailing in the Czech Republic, would have chosen emigration (at least a temporal one). This discussion could be objected by the fact that the Czech Republic is a small country and some people in hardships will simply travel to work to another region. We might object that Ireland is a same case of a small country (that is why both countries have been chosen for a joint analysis) and the processes are not the same. In short, there are two countries with wide differences in migration potential and both experience and analysis show that migration potentials differ greatly for each one. Simple as that, people in some

88 countries are just what we call “mobile” while people in the other countries are not. This mobility can be a good criterion in several cases, especially for politicians who often take the decisions of labor mobility basing their rhetoric on populism rather than some solid scientific grounds.

3.5. Conclusions

As we have seen from the elaboration and the model presented above, migration in different countries seem to be influenced by different factors. Apart from that there is a migration potential of every country’s population that plays crucial role in the international migrations. The specific focus is to be made on distinguishing the impact of economic differences between immigration and emigration country and economic processes in the country of emigration on the trends of international migration. Our analysis comparing the relationship of net migration, emigration and immigration with income and employment in the regions for Czech Republic and Ireland shows that migration potential of population of these two countries varies: regions with favorable economic conditions tend to experience high immigration as well as emigration, whereas depressed regions display generally low labor mobility. Our findings suggest that this variation will be in favour of countries with more “mobile” population such as the Republic of Ireland. More “mobile” population measured in regional emigration and immigration responsiveness to the changing economic conditions at home might mean that economic incentives abroad as well as opening of new foreign labour markets due to the political decisions (such as the EU Enlargement) will very likely lead to larger migration of labour searching for higher wages and employment opportunities. No doubt, there is a wide scale of factors, not just economic ones that can affect migrations: social and cultural factors (such as habits and language), housing stock and demographic characteristics and many more that can be considered in an analysis of this type. However, we wanted to show that the main motives are always economic. In this context it can be argued for the fact that EU Enlargement itself may not necessarily play a decisive role in triggering off the migrations from the EU-8 Member States. In order for this to happen, the emigrants should strongly react to the economic changes in the country and be highly mobile within their country. Without these two important factors the economic disparities going hand in hand with attractive employment opportunities, as those existing in the enlarged EU-25, may note result in such a large influx of migrants from the East. Accessing the migration potential in the CEECs we can see that they are quite low and thus it is unlikely that the recent trend of very modest migrations to the EU-15 is going to change.

89 3.6. References

• Bauer, Thomas, and Zimmermann, Klaus F., “Modelling International Migration: Economic and Econometric Issues.” In: Rob van der Erf and Liesbeth Heering, Eds., Causes of International Migration Proceedings of a workshop, Luxembourg, 14-16 December, 1994 , pp. 95-115. Luxembourg: Office for Official Publications of the European Communities, 1995. • Bauer, Thomas, and Klaus F. Zimmermann (1999): “Assessment of Possible Migration Pressure and its Labour Market Impact Following EU Enlargement to Central and Eastern Europe ”, IZA, Bonn. • Boeri, Tito, and Herbert Bruecker (2000), “The Impact of Eastern Enlargement on Employment and Labour Markets in the EU Member States ”, European Integration Consortium: DIW, CEPR, FIEF, IAS, IGIER, Berlin and Milano. • Borjas, George J. (1987), “Self-Selection and the Earnings of Immigrants,” American Economic Review 77 (4), 531-553. • Borjas, George J. (1994), “The Economics of Immigration,” Journal of Economic Literature, XXXII (December), 1667-1717. • Coulter, C. (1994), “Ireland: between the First and the Third Worlds”, Dublin: Attic Press. • De Jong, G., et al. (1983), “International and Internal Migration Decision-making”. International Migration Review, vol. 17, No. 3 • De Jong, G. et al. (1986), “Migration Intentions and Behavior. Decision-making in a Rural Philippines parish”, in J.T. Fawcet (Ed.), Population and Environment, Vol.8, No.1 • Dorigo, Tobler (1983), “ "Push-Pull Migration Laws", Annals, Assn. Am. Geographers, 73, 1 (1983); 1-17. • Keenan, J.G. (1981), “Irish Migration: All or Nothing Resolved?”’, Economic and Social Review, 12, 169- 186 • Farrell Niall (February 1991): “Emigration – the Galway Experience”, Galway Action on Poverty report (GAPR), Galway • Fidrmuc, J. (2002), „Migration and regional adjustment to asymmetric shocks in transition economies“. CPB Discussion Paper 007. • Geary, P. and O’Gráda, C. (1989), “Postwar Migration between Ireland and the UK” in “European Mobility: Trends and Consequences” (Eds. I. Gordon and A.P. Thirwall). • Harris, J.R. and M.P. Todaro (1970), “Migration, Unemployment and Development: A Two Sector Analysis,” American Economic Review 60 (1), 120-142. • Hazelkorn, E. (1990), “British labour and Irish capital”, Galway Labour History Publication, The emigrant experience, Galway • Mac Laughlin, J. (1994), “Ireland: the emigrant nursery and the world economy”, Cork University Press • Menken, J. (2005), “Advanced Data Analysis”, University of Colorado, Boulder. • NESC (National Economic and Social Council) (1991), “The Economic an Social Implications of Emigration”, prepared for the Council by J.J. Sexton, B.M. Walsh, D.F. Hannan and D. McMahon, Dublin: National Economic and Social Council. • Ravenstein (1885), “The Laws of Migration." Journal of the Statistical Society of London, Vol. 48, No. 2. (June, 1885), pp. 167-235. • Ravenstein (1876), "The Birthplace of the People and the Laws of Migration." The Geographical Magazine, Vol. 3, pp. 173-177,201-206,229-233. • Ravenstein (June, 1889), "The Laws of Migration." Journal of the Royal Statistical Society, Vol. 52, No. 2. pp. 241-305. • Todaro, Michael P. (1969), “A Model of Labor Migration and Urban Unemployment in Less Developed Countries.” American Economic Review 59 (1), 138-148. • Vavrejnova, M. (2004), “Mobilita pracovní síly p řed a po vstupu ČR do EU”, CERGE, UK v Praze. • Visek, J.A. (2006), “Econometrics: the lectures”, Charles University in Prague. • VÚPSV (leden 2001), Záv ěre čná zpráva: „D ůsledky vstup ů ČR do EU na vztahy s Rakouskem se zam ěř ením na zam ěstnanost, trh práce a migraci“, Praha. • Walsh, B.M. (1974), “Expectations, Information, and Human Migration: Specifying an Econometric Model of Irish Migration to Britain”, Journal of Regional Science, 14, 107-120.

90 • Walsh, B.M. (1987), “The Impact of Demographic Variables on Unemployment”, in “The Challenge of Unemployment”, Administration, vol. xxxv, No. 3, Dublin • Walsh, B.M. (1987), “Why is Unemployment so high in Ireland?” in “Perspectives if Economic policy”, Centre for Economic research, UCD, Dublin • Walsh, B.M. (1989), “Testing for the Existence of Macro Economic Feedback from Large-scale Migration”. The Economic and Social Review, Dublin. • Statistics and National Statistics: Austrian Statistical office: www.statistik.at , Irish Central Statistical Office (www.cso.ie ), Irish Ministry of Agriculture ( www.agriculture.gov.ie ), Czech Statistical Office ( www.czso.cz ), German Statistical Office ( www.destatis.de )

91 3.7. Appendix A

Table A : A number of studies on migration from the CEE countries to the EU. Authors/Year Predicted value Countries included Methods/Remarks Layard et al. 1992 130 thousand immigrants per year in Estimates for Poland, Method of potential estimation of South- Western countries Czechoslovakia and Hungary North migration flows (3 per cent of (projected on the other CEE population in between 1950-1970) countries) Brueker/Franzmeyer Depending on the scenarion: (1) Five countries: Poland, Gravitational model (embodying 1997 (1) 340-680 thousand per year in EU. Hungary, Czech republic, different economic parameters, the (2) 590 thousand – 1 million 800 Slovakia and Slovenia importance is given to wage thousand per year in EU (2) All candidate countries. differentials). Fassmann/Hintermann 721 thousand – actual migration Czech republic, Poland, Slovakia Representational research of Gallup 1997 potential, 320 thousand to Germany, 150 and Hungary. Institute in all four countries included in thousand to Austria. the prognoses. Aintila 1998 About 13 thousand immigrants to Baltic countries and Poland. Same as Layard 1992 Finland each year Birner/Huber/Winker (1) 24 100 Czech republic, Poland, Slovakia, (1) The year of accession is settled on 1998 (2) 21 700 Slovenia and Hungary. 2004. Regional flows of migrants in the year of (2) The year of accession is settled on accession (to Austria). 2010. Methods taken from Walterskirchen/Dietz for Austrian border regions. Hofer 1998 25–40 thousand per year Same as Brueker/Franzmeyer 1997 Estimations of Brueker/Franzmeyer results Lundborg et al. 1997 628 thousand – 1 million 885 thousand Baltic countries and Poland. Same as Layard 1992. Lundborg 1998 of workers (with the families) to EU countries in 15 years, 126 thousand. 20- 30 thousand of workers to Sweden alone. Sujanova/Sujan 1997 39 thousand in 2005-2010 to EU. Czech republic Econometric model (later Huber-Pichelmann 1998, Hofer 1998) Huber/Pichelmann 1998 140-200 thousand to EU. CEECs Based on Sujanova/Sujan method used for all countries. Sik 1998 (later Huber Migration potential in the border regions Hungary Panel research. 1999, Salt 1999) is irrelevant. Walterskirchen/Dietz (1) 42 thousand. Czech republic, Poland, Slovakia, Same as Brueker/Franzmeyer 1998 (2) 31.600 to Austria (workers and Slovenia, Hungary. (1) if free movement is to be commuters), 150-200 each year for implemented in 2005. five years (workers), 150 thousand (2) If it is to be implemented starting for commuters in the long run. from 2015. Wallace/IOM 1998 No exact numbers, but the explanation of Poland, Czech republic, Slovakia Representational research base on migration and the preferable countries and Slovenia, Hungary, Romania, questioning of about 1000 people in each (Germany). Bulgaria, Croatia, former country. Yugoslavia, Ukraine and Belarus. Bauer/Zimmerman 1999 About 3 million in the next 10-15 years, Bulgaria, Czech republic, Hungary, Same as Layard 1992. Two scenarios: 200 thousand each year to EU. Poland, Romania, Slovenia, transition periods and immediate free Slovakia. movement. Fertig 1999 (Huber (1) 38-31 thousand each year to (1) first candidates: Czech republic, Hatton model (1995) used, estimations 1999) Germany for the next 20 years, Estonia, Hungary and Poland. base on the statistics of the Bureau of (2) 39-33 thousand each year. migration in Germany and with implementation of the assumption of the economic growth of 2 per cent that the EU average in CEECs. Salt et al. 1999 41 thousand each year to EU. Czech republic, Estonia, Hungary, Projection of the basic immigration Poland and Slovenia. factors of chosen EU countries (1985- 1995). Source: see references

92 4. What living is worth leaving? Accessing the impact of economic incentives on migration flows: the case of Czech-Slovak labor migration after the split-up of Czechoslovakia

4.1. Introduction

Generally, primary determinants of migration decision are expressed through differences in wages and other sources of income between the host and the source countries (see for example Sjaastad, 1962; Todaro, 1969; Harris and Todaro, 1970; Walsh, 1974). However, this approach would not suffice to explain one interesting implication of modern migrations: if economic disparities between countries or regions prevail, why does the volume of migration remain considerably small? As Faini et al. (1999) point out international migration is the “great absentee” in the liberalization of global goods and factor markets. Net annual emigration rates in the developing countries represent less than 0.1 per cent of total population, although the GDP per capita in these countries is less than 1/10 of that in the developed world (World Bank, 2005). Even in case of poor countries and regions that are situated very close to substantially wealthier regions (and therefore the costs of emigration are very low) such as Eastern and South-Eastern Europe, Northern Africa and Central America the net annual emigration rates are about 0.15 per cent (Brücker and Schroder, 2006). Perhaps, one possible explanation to that moderate migration flows in the situation of high economic incentives is that there are high legal and administrative barriers to migration (Brücker and Siliverstovs, 2004). The existence of these barriers is heavily defended by policy-makers and the populations in the migration-receiving countries. Generally, there are fears that that opening the barriers to immigration will involve a massive influx of migrants. However, this was not the situation of European Union (EU) recent Enlargements. The introduction of free movement in the context of the EU Southern Enlargement did not increase immigration from the South, and opening the labor markets to citizens of the New Member States (NMS) in 1/3 of the Member States of the EU and the European Economic Area (EEA) after May 2004 has resulted in a net migration of less than 0.2 per cent of the population of the NMS, despite the GDP per capita in the NMS being approximately 25% of the average of the old EU Member States (Brücker and Schroder, 2006). This essay attempts to show that opening the borders to migration does not necessarily case increased influx of immigrants, even in case economic disparities prevail. It also advocates the idea that although it is very likely that some labor migrants would originate from the country with lower

93 economic performance to the country with higher economic performance, economic differences start to influence labor migrations only when they exceed some critical level.

4.2. Methodology

4.2.1. International migration and the role of economic incentives

It remains largely unclear why migrations happen and what causes them when the economic incentives are not involved. There is an on-going debate in migration literature that is trying to defend or refute linking international migration to economic factors. Wage differentials, differences in GDP per capita or differences in unemployment between the countries with lower level of economic performance (emigration countries) and countries with higher level of economic performance (immigration countries) are claimed to provide the basis for migration, or, be its decisive factors (see for example Ravenstain, 1889; Hicks, 1932; Sjaastad, 1962; Todaro, 1969; Harris and Todaro, 1970; Walsh, 1974; Geary and O’Grada, 1989). Although it seems unreasonable to underestimate the influence of those factors, however it becomes apparent that quite often economic incentives give the way to so-called “social factors” that most of the migration literature tries to omit (see for example Granovetter, 1982; Granoveter, 1995; Bauer et al., 2002). The so-called “social factors” include influence of the family, friends, and language barrier and positive or negative externalities provided by the immigrants’ networks. No doubt that international migration would have never happened without economic “pull-factors” 13 , however it can be assumed that there is a strong influence of these “social” or “invisible” factors that has somewhat stronger impact on inducing migration from poor regions or countries to wealthier ones.

With regard to the above, it seems interesting to test how migration flows would behave in case almost all possible economic and administrative barriers to migration are removed and the impact of “social” or “invisible” factors is neglectable. This should involve a natural experiment: a situation when two countries apply no restriction to labor migration flows from another party. In addition, it would be covetable that both countries have a highly-developed level of language proximity and share same cultural and historical experience.

13 The notion of “push” and “pull” factors that are leading to migrations can be often found throughout migration literature. Among the most frequent push factors are for instance unbearable or threatening conditions in the home country, while pull factor cane be for instance represented by the incentives in the country of immigration (see for example Ravenstein 1876, 1885, 1889; or Dorigo, 1983).

94 Analyzing the recent development on the European continent one such experiment can be found: the split-up of Czechoslovakia and the creation of the Czechoslovak customs union (1993-2004). After the disintegration of the common state both countries gained independency and were co-existing side- by-side. While the Czech Republic was more advanced in pursuing economic reforms and proceeding with economic transformation, Slovak Republic was less developed and more dependent on mutual trade.

At the face of various economic problems many Slovaks left their country and resided in the Czech Republic, both temporarily and permanently, in a search of higher wages and better level of life. Transaction costs of this labor migration were marginal: citizens of the Slovak Republic did not need special residence permits or working visas to reside and work in the Czech Republic (a registration at the police department was required), language and cultural barriers were irrelevant due to the common cultural background (the existence of Czechoslovakia in the past) and social ties were much more stronger than with any other country. In other words, the Slovaks were not viewed as foreigners by the largest part of the Czech population. The case of the Slovak Republic is of special interest, as far as it was easier for the Slovak workers to move to the Czech Republic in order to improve their economic position than anywhere else in Europe. The lack of natural and physical barriers makes this observation to be a very significant one.

There is not much evidence about regional migrations in Federal Czechoslovakia prior to the Velvet Revolution of 1989. Most of the labor migrations were temporary and were not recorded. Quite often it happened that citizens of one Federal state were living and working in the other state for a number of years and then were returning back. Besides, given the formal status of Czech or Slovak citizenship in the Federation and overall similarities in the Czechoslovak socialist economy, these migrations cannot be well traced. It can only be assumed that some periphery-center migrations (from rural areas to urban centers and especially Prague) were taking place as they were happening in every country. In the beginning of 1994 there were 17.000 Slovak citizens in the Czech Republic (16% of all foreigners residing in the country). Three years later, in 1997 their number increased to 52.000 (Czech Statistical Office, 2007). Migration over the Czech-Slovak border became a considerable source of mitigation of the Slovak unemployment problem. For instance, 2.8% of the Slovak labor force was employed in the Czech Republic in the 1996 (Tang et al., 2000). The similarity of this case with the EU 2004 and 2007 Enlargements calls for deriving some comparisons. In spite of some methodological difference some critics point out (i.e. the fact that Czechoslovakia was divided instead of unified, or that the language and cultural proximity is closer)

95 the story of the Czech-Slovak migration is very similar to the situation of EU-8 countries after the EU accession (limited freedom of movement of labor in EU-15 states).

4.2.2. Countries’ profiles: the split-up of Czechoslovakia and the successor states

Czechoslovak Federation disappeared on the 1 st of January 1993, after more than seventy years of existence, pre-followed by the raise of nationalistic and separatist tendencies in both parts of the country.

The split-up of the Federation met with a whole range of pessimistic prognoses for both successor states. In case of the Slovak Republic the threats were the weakening caused by the division of common property and bearing relatively high costs of building a new state infrastructure. In the case of the Czech Republic it was the loss of broader internal market which was suddenly substituted by the international trade between two independent states (Dedek et. al., 1997). Although, none of the pessimistic scenarios ever materialized, both states had to minimize the impact of disintegration of a superior international body, the Federation that ceased to exist.

Nevertheless, the split-up of the Federation had an impact on both states, especially in its first phase. In the year that followed the split-up, both GDP of the Czech Republic and independent Slovakia decreased by 0.9 and 4.1 per cent respectively (Table 4.1). Although this was the most serious decline in GDP in the whole course of transformation process, the next year GDP in both countries increased substantially.

The major impact was, of course, the decline in mutual trade between both countries. The decrease of GDP in both states was mainly caused by its stagnating volume in the first year after the split-up. The exports of goods and services from the Czech Republic to Slovak Republic (in fixed prices) declined by 23 per cent, while from Slovak Republic to the Czech Republic the decline was 15 per cent (Czech Statistical Office, 1995).

96 Table 4.1: Development of GDP of the Czech Republic and Slovak Republic and its components in 1993-1994 (growth rate in %, expressed in 1984 fixed prices) Year Year

Czech Republic 1993 1994 Slovak Republic 1993 1994

GDP - 0.9 2.6 GDP - 4.1 4.8

Private consumption 2.9 5.3 Private consumption - 0.6 2.5

Public consumption - 0.1 - 2.3 Public consumption - 0.4 - 3.5

Gross fixed capital formation - 7.7 17.3 Gross fixed capital formation - 3.4 - 2.1

Change in inventories - 2.7 - 2 Change in inventories - 2.9

Exports of goods and services 7.5 0.2 Exports of goods and services - 1.9 12.4

Imports of goods and services 10.4 7.8 Imports of goods and services 1.6 7.8

Source: Czech Statistical Office, Slovak Statistics and Infostat Bratislava.

While the inflation remained at the pre-division level (around 20 per cent) the main difference was the rate of unemployment. Being quite high in Slovak Republic at the year of the split-up (10.4 per cent) it was growing in the next years by 40 per cent (Table 4.2).

Table 4.2: Unemployment rates in Czech Republic and Slovak Republic in 1992-1994 (in %) 1992 1993 1994

Czech Republic 2.6 3.5 3.2

Slovak Republic 10.4 14.4 14.8

Source: Czech Statistical Office, Slovak Statistical Office

According to the Czech Statistical Office, the total number of unemployed in Slovak Republic at the end of 1994 was 371.5 thousand people, in the same time there were 166.5 thousand unemployed in the Czech Republic. At the end of 1993 there were 48 unemployed per one vacant workplace in the Slovak Republic (this dropped to 28 in 1994) while the same ration in the Czech Republic was 4.2 and 2.2 people respectively (Czech Statistical Office, 1997).

97 It can also be noted that the split-up of Czechoslovakia caused more problems to Slovak economy in terms of labor productivity. The growth rates of both productivity of labor, nominal wage and real wage slowed down in the first year after the division to grow with a slow rate (Table 4.3).

Overall, it can be stated that the Slovak Republic was more affected by the split-up of the Federation. Always being less economically developed part of the Federation with the prevailing share of agriculture, it felt upon the loss of the large internal market and had to substitute many channels of intra-national trade for international ones. Economic disparities and economic problems in the Slovak Republic which were apparent especially in the first years after the disintegration of Czechoslovakia made it searching for job in the neighboring Czech Republic to look like quite a desirable opportunity (chart 4.1).

Table 4.3: The development of labor productivity and average wage in economy and industry of Czech Republic and Slovak Republic in 1993-1994 (growth rates in %) National Industry National Industry economy economy

Czech Republic 1993 1994 1993 1994 Slovak Republic 1993 1994 1993 1994

Labor productivity 1.2 2.6 0.0 5.1 Labor productivity - 0.9 5.3 - 7.3 7.5

Nominal wage 25 17.1 23.8 15.7 Nominal wage 18.5 17 21 17.4

Real wage 3.5 6.1 2.5 4.8 Real wage - 3.8 3.0 - 1.7 3.3

Source: Czech Statistical Office, Slovak Statistical Office

All these, accompanied by the language proximity and the lack of administrative barriers between the Czech Republic and Slovak Republic, created an interesting natural experiment in which: (i) there existed economic incentives for migration from a poor (Slovak Republic) to a wealthier country (Czech Republic), (ii) there were no administrative barriers to migration between the two countries in question, (iii) social factors were minimized by common cultural background of both countries (common Federation for over seven decades) and language proximity of Czech and Slovak language.

98 Chart 4.1: Structure of Slovak citizens’ employment in the Czech Republic (1995-2004)

90000

80000

70000

60000

50000

40000

30000

numberof persons 20000

10000

0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 year

Employees Traders

Source: RILSA (2006); Horaková (2005)

These conditions were of a unique nature: never before a situation like this could be observed and monitored for scientific purposes. The uniqueness of the Czech-Slovak migrations makes it a very interesting case that can be applied to explaining migration flows caused by economic incentives. For instance, this case shows that even in case of economic disparities and very small barriers to migration, mass influx of immigrants in the wealthier country should not necessarily be the only way the things might develop.

3.2.3. The implications of Czech-Slovak migrations for the European Union enlargement process.

The case of the Czech-Slovak migrations that will be used for building the empirical model further in this paper has been used mainly to simulate the behavior of migration flows when no administrative or social barriers are present. This is especially relevant in the case of the EU enlargement process and the introduction of the free movement of labor. The EU Eastern enlargement in 2004 was accompanied by many controversies mainly related to the fears of uncontrolled migration. As a result, until today only seven EU-15 Member States have opened their labor markets to newcomers. Although these seven countries never declared any massive influx, the others opted out for the

99 maximal possible transition period. With pending accession of Bulgaria and Romania the debates on migration are on the rise again. With regard to this many like to remind that GDP level in Bulgaria and Romania is well below of those of the EU which will inevitably lead to mass migrations once the barriers are removed (see for example Sinn, 2000; Dustmann et al., 2003; Delegation of the European Commission to Japan, 2006). Indeed, such measures of the country’s economic performance as difference in GDP growth or GDP per capita growth rate are of considerable importance and might influence the migration of workers following the EU enlargement. However these gaps are reduced very fast with for instance EU-8 NMS growing faster on average than EU-15. The economic growth of many of the NMS, for example, Poland is 2-3 times higher than for example in Germany. Nevertheless, the high rate of economic development most of the CEE countries experience does not necessarily mean that the levels of life would be leveled soon (it might as well take as long as 30 years or so). However, even the slow leveling of wages, incomes and the level of life is going to influence the decrease of the working migration in the long run. It should be also noted here that leveling the incomes is not always the relevant factor in explaining the decrease in the working migrations (see for example Faini and Venturini, 1994; Bauer and Zimmerman, 1999). The example of the EU Southern enlargement is exactly that case: although the level of incomes in Greece, Portugal and Spain was three times lower than this of the EU, the improvement of the economic situation, favorable prognoses concerning the future and the other gains from the EU accession have brought the working migration from these countries to the EU to the marginal level. Another serious issue is the demographic situation in the EU. EU-15 Most of the EU-15 Member States experience problems with demography: population is quickly aging. It can be shown that NMS have same problems: the population of Hungary is aging fast, Poland is characterized by the relatively young population with high demographic increase expected, while the Czech Republic is somewhere in between. It is labor migrations from the NMS that are to join the EU after 2007 especially Bulgaria, Romania and Turkey that might become a solution for this problem. The unemployment rate both in the country of emigration and immigration may also be of a relevant influence on the size and the structure of working migration (see for example Fields, 1979; Goss and Schoening, 1984; Pissarides and McMaster, 1990). Comparison of the current situation points out certain structural similarities. In both group of countries long-term unemployment as well as the unemployment among the young people prevail. The regions suffering from the high rate of unemployment in Poland and the Czech Republic are usually those situated close to the German or Austrian border. The problems are caused by the structural peculiarities, such as the high rate of agricultural production. The changes in the ownership as well as the transformational changes create the high pool of unemployed low-qualified agricultural workers. However, most of these individuals

100 are of low qualifications and skills, which together with strong ties to the countryside they are living in make them unable to search for a job abroad. Thus their readiness to emigrate is equal to zero. The accession of the candidate countries might also influence the rate of commuting or “pendling”. It is clear that commuting creates more problems than the predicted mass flows of workers from the countries of Eastern and Central Europe. Many people would prefer to work in a wealthier country and spend their incomes at home. The free movement of workers, if allowed, might lead to its increase. However, commuting is of a high rate now and the opening of the markets to the two sides of the Schengen border would rather eliminate regional differences and destroy the grounds for commuting. It should not be also forgotten that the working migrants from the CEE countries might be seen replaced by the workers from Bulgaria and Romania. The cheaper labor force from these countries occupying the free places on the EU labor market and demanding less money for the same jobs might lead to the fact that the workers from the Central and Eastern Europe would be thrown over the board, which will diminish working migration considerably. The last factor, which should be present in the process of evaluation of migration flows, is the one describing cultural differences and the language barriers. The experience shows that migrants from the Czech Republic, Poland and Hungary can adapt themselves to the new environment (Germany, Austria and Benelux) much more effectively than emigrants from Turkey or southern Europe. Nevertheless, poor language command creates many problems and few working migrants are able to learn the language to such extent that they would be able to find a decent job. There was an interesting example of offering job positions and work and residence permits in Germany for the limited number of specialists in information technologies from the CEE countries. It turned out that such specialists are in great need in their own countries and the German initiative ended up in a fiasco. The EU has to be ready to make all its Member States to open their labor markets to the new accession states. Without complete free movement of labor the grounds for the completion of the Single Market in all EU Member States are weak. Free migration of labor is the key factor for finalizing this process of economic consolidation and thus should be treated with care. Fears of massive and uncontrolled migration might be kept by politicians however they cannot be regarded for in relevant decision-making and used for justifying policy-sensitive measures. The case of the Czech- Slovak migrations might be used as a reference for considering somewhat different treatment of this problem.

101 4.3. Data

The data on the Slovak citizens employed in the Czech Republic include only officially registered Slovak citizens working on the territory of the Czech Republic on the legal basis in the years 1993- 2004. There is no need to go beyond that data, as far as EU Accession in 2004 that brought both the Czech Republic and Slovak Republic to the European Union, has also eliminated the previous provisions and opened the door to the free movement of labor in both states (neither the Czech Republic, nor Slovak Republic opted for the transition periods for the free movement of labor within the EU). Thus the data comes from the reliable source and can be accepted almost without any hesitation due to several reasons: first, at the time under consideration there were no residence permits or visas for Slovak citizens in order to enter the Czech Republic; second Slovak citizens did not have to have work permit in order to start a new job in the Czech Republic; third, the registration process was simplified to minimum (informing the neighboring Police department for foreigners). All this provided a very strong incentive for the citizens of Slovak Republic living and working in the Czech Republic not to conceal their actual activities and presence in the country. Indeed, simple registration and the lack of administrative barriers made illegal unemployment useless. Slovak citizens themselves were in favor of their actual numbers in the Czech Republic being counted and registered by the organs of public administration. In addition, it is clear that those Slovaks who, in spite of all incentives, choose to be employed illegally were putting themselves to higher risk than their law-obeying countrymen. All that made the number of illegal Slovak workers irrelevant. In addition to that and due to the reasons listed above Slovak labor migration to the Czech Republic can be regarded as one of the most accurately measured and described of its kind. The data also include: the annual difference in the GDP levels between the country of emigration and immigration calculated according to the respective exchange rates (measured by calculating the annual level of GDP in both the country of emigration and immigration in USD, then obtaining the differential of GDP levels in both countries calculated as a percentage change to the basic year of 1993); annual observations on size of economically active population (to the basic year of 1993) and measured as the people between 15 and 64 years of age (an approximation of the size of the labor market); GDP growth rate in Slovak Republic (to the basic year of 1993) and annual unemployment rates in both countries. The data pool crucial for testing the impact of economic incentives on labor migration from the Slovak Republic to the Czech Republic in 1993-2004 was obtained at the Czech and Slovak Statistical Offices and Research Institute for Labor and Social Affairs in Prague.

102 4.4. Empirical model

In this section of the paper an econometric model which is trying to capture the impact of economic incentives on Czech-Slovak migrations will be presented. The usual form of most of the econometric models that aim at testing the same causality is to relate the rate of migration to such economic factors as wages, employment and GDP per capita. For example Walsh (1974) built one of the first models of fluctuations in year-to-year migrations for Ireland. The general approach to the model adopted by Walsh in his paper and followed by others was in the following form:

= α +β +β + Mijt 1Yit 2 Yjt u )1.4( where M ijt is the annual migration flow rate from country i to country j in period t and Yit and

Yjt are the measures of the expected lifetime income in each country (Walsh, 1974). Further, Walsh extended the model described by (1) to the equation with annual migration flows on the right-hand side and average earnings (weekly or annually) and average annual percentage of workers registered as unemployed in both the emigration and immigration countries (Walsh, 1994) Whilst most of the literature already assumes that there is a certain impact of economic factors on migration and merely tries to test the strength of this impact, this paper examines weather and to what extent economic factors have influenced migration rate (flows of migrants expressed as net migration flows) between two countries, represented by the Czech Republic and Slovak Republic. Following Walsh (1974), Geary and O’Gráda (1989) and Strielkowski and O’Donoghue (2006) the model used in this paper can be presented in a form of a multivariate statistical model in the general linear form 14 :

= α+β +β +β +β +β + = Mijt 1Ydiff ijt 2DEM jt 3Yjt 4 U jt 5 Uit u k , k 1,..., n )2.4( where Ydiff is the annual difference (in %) in the GDP levels of the Czech Republic and Slovak Republic calculated using the respective exchange rates and the year of 1993 as the baseline; DEM is the labor market absorption capacity in immigration country (Czech Republic) expressed as the change

14 The model has been tested for linearity: both by checking the relationships of dependent and independent variables (plotting them into a graph) and by plotting a graph of residuals of independent variables. Both methods provide an efficient proof that for this particular model and this particular data set linearity of the model can be proved and accepted.

103 in the number of economically active population (to the basic year of 1993) at the age between 15 and 64 (in %); Y is the rise of GDP growth rate (in % to the baseline year) in emigration country (Slovak Republic) introduced in order to capture the improvement of economic conditions at home for Slovak immigrants that might become a considerable inducement for the reverse migration; U jt is the unemployment rate (in %) in the country of emigration (Czech Republic); U it is the unemployment rate (in %) in the country of immigration (Slovak Republic) and u k represents disturbance or noise.

The hypothesis that it to be tested here can be articulated in the following way: the particular model and data would lead to conclusions that economic differences between the two countries (Czech Republic and Slovakia) might influence migration. The hypothesis can be either supported or rejected on the basis of estimation, the latter showing that the economic differences cannot influence migration of workers of any kind, al least not to the considerable level. Rejecting the hypothesis would then mean that in for the model explained above different levels of economic performance in countries that produce net outflows of labor migration and counties that are net recipients of labor migration is unlikely to be the main explanatory factor for migrations (especially labor migrations).

Although the independent variables might seem self-explanatory, it is necessary to discuss their expected signs and impact on the dependent variable:

a) GDP difference - the more is the difference in GDP between the two countries, the more people in the poorer (emigration) country would tend to seek jobs in the wealthier (immigration) country. Hence, the expected relationship between the GDP difference and the migration flows is expected to be positive, b) Labor market capacity of immigration country to “absorb” workers from the emigration country – the more people of working age there are in the country of immigration, the larger is the labor market and therefore the more working migrants it can absorb. The expected relationship should be positive again, c) GDP growth rate in emigration country is expected to have negative relationship on migration – the higher the growth rate of GDP, the less workers would be eager to migrate abroad in a search of better living conditions as far as they might expect the improvement of the economic situation at home, d) Unemployment rate in the immigration country is expected to have negative relationship on labor migration flow. When there is a high rate of unemployment in the target country, few employers would be willing to employ foreign workers and the competition for a working place becomes very high,

104 e) Unemployment rate in the emigration country is expected to have positive relationship on migration. The higher the unemployment (and therefore job seeking becomes very hard) the eager would the workers from the emigration country to move on and to start looking for the new employment opportunities In addition to that, independent variables can be divided into three main groups: (i) social and political, (ii) demographic and (iii) economic and financial. This is summarized in the following table:

Table 4.4: Variables influencing the emigration rate by categories of variable and expected impact

Categories Variables Nature of impact Social and political GDP difference +

Demographic DEM IMM +

Economic and financial GDP EM -

UNEMPL IMM -

UNEMPL EM + Source: own estimations. Note: “+” = enhancing; “ -“ = weakening

Although the model presented in (4.2) can be estimated using the multiple regression model, it might be that more reliable arguments can be provided from individual equations.

First and foremost, it should be realized that there is a certain issue with the model which refers not to the accuracy of the model or the data but to the methodological grounds. There are five independent variables in the model but only twelve observations which considerably reduces the degrees of freedom and thus the predictive power of the model. This is due entirely to the nature of the data: Czechoslovakia ceased to exist in 1993, so today there is data available for just 14 years (and in this paper migration flows until 2004, thus until the EU accession of both Czech Republic and Slovakia, are considered). The small data pool might have no use in the regression because the number of observations is too low for the interpretation of the regression results.

Second, a backwards stepwise regression was run for the “full” model as specified in (4.2). The results of the stepwise regression (table 4.5) show that just two variables that belong to the categories of economic and financial and demographic variables (i.e. economically active population in the country of immigration and the unemployment rate in the country of immigration) appear to come through as significant (significant at 1% and 10% respectively) in the estimation. This might suggest that just these two variables might be of some value in determining the influence of various variables on the valume of labor migration

105 from Slovakia to the Czech Republic. However favorable this conclusion can be for our results derived in this essay, we still might need more clarification of the relationships between labor migration from Slovakia and various social and political, demographic and economic and financial variables that stem from the literature analysis in preceding sub- chapters.

Table 4.5: Determinants of labor migration from Slovakia to the Czech Republic; Number of officially registered labor migrants from Slovak Republic in the Czech Republic (1993-2004), (M ijt ); backwards stepwise regression Labor migration from Slovakia to the Czech Republic in 1993-2004 (M ijt ) Std. Errors Economically active population in the -103681.3*** country of immigration (Czech Republic) [16694.61] Unemployment rate in the country of 5992.401* immigration (Czech Republic) [2681.96] 97112.85 Constant [12617.47} N 12 R-squared 0.84 Adjusted R-squared 0.79 Source: own estimations. Note: Standard errors in brackets; * significant at 10%; ** significant at 5%; *** significant at 1%

Therefore, taking into account all the problems raised above it might be more precise to try using alternative measures of the relationships between the dependent and the independent variables in turns, taking the independent variables one by one and analyzing single cases (in fact, analyzing simple regression models with one dependent and one independent variable).

In order to do so, equation (4.2) has been modified in order to yield the following 5 equations:

= α +β + = Mijt 1Ydiff ijt uk , k 1,..., n (4.3)

= α +β + = Mijt 1DEM jt uk , k 1,..., n (4.5)

= α +β + = Mijt 1Yjt uk , k 1,..., n (4.6)

106 = α +β + = Mijt 1Ujt uk, k 1,..., n (4.7)

= α +β + = Mijt 1Uit uk , k 1,..., n (4.8)

Thence, it seems reasonable to proceed in this manner, computing the regressions with the help of the ordinary least squares method using one dependent (logarithm of labor migration flows from Slovak Republic to the Czech Republic) and one independent variable.

The results of estimations of equations (4.3)-(4.8) are presented in the table 4.6. The separate impact of each independent variable on the labor migration flows is shown using the sign and the value of the coefficient, the significance levels (if applicable, i.e. in case of significant variables) and the standard errors.

Table 4.6: Results for regression analysis: Labor migration from Slovakia to the Czech Republic in 1993-2004 (M ijt ).

Mijt (number of officially registered labor migrants from Slovak Republic in the Czech Republic, 1993-2004) Std. Errors -526.030* GDP difference [237.418] Economically active population in the country -75687.75*** of immigration (Czech Republic) [14486.303] GDP growth rate in the country of emigration -2352.093** (Slovak Republic) [921.027] Unemployment rate in the country of 1627.522 immigration (Czech Republic) [2006.485] Unemployment rate in the country of 546.997 emigration (Slovak Republic) [1393.338] Constant 58699.110*** 87163.458*** 86435.267*** 48581.636* 46408.482*** [3422.764] [6144.026] [11948.731] [21857.944] [13658.356] N 12 R-squared 0.32 0.73 0.39 0.01 0.06 Adjusted R-squared 0.26 0.7 0.33 -0.08 -0.03 Source: own estimations. Note: Standard errors in brackets; * significant at 10%; ** significant at 5%; *** significant at 1%

Although the coefficients of determination vary considerably from case to the case (from 0.04 being the lowest to 0.72 being the highest) three variables have appeared to have impact on the flow of labor migrations: GDP difference, size of the labor market in the Czech Republic (economically active population) and GDP rise in Slovak Republic (incentive to stay at home). Unemployment rates in both

107 the Slovak Republic (country of emigration) and the Czech Republic (country of immigration) seem to have no significant impact on determining the flows of labor migration of Slovak citizens.

The difference in GDP between the two countries has a negative impact on labor migration, the estimated coefficient is small but the variable is of the high significance (significant at 10% interval, adjusted R-squared is 0.32). This would mean that increasing or decreasing gaps in the GDP levels between the Slovak Republic and the Czech Republic do not influence the numbers of the Slovak citizens legally employed in the Czech Republic.

Contrary to the expectations summarized in table 4.4 there appears to be a negative relationship between the approximate size of the labor market in the Czech Republic and the Slovak labor migration (significant on 1% interval, adjusted R-squared is 0.7). So, the larger the labor market, the less are labor migrations. This startling result might be explained by two things: (i) it might be that the size of the labor market cannot be approximated by the number of people of the working age, when it is not adjusted for the unemployment rate it might easily be that not all these people are in the job or want to take a job; (ii) it might be that the number of people of the working age in the immigration country is actually playing a reverse role – the more are the people of the working age, the higher is the competition and the harder it is for the immigrants to find a job, so many prefer to stay at home to trying their chances elsewhere. In fact, many Slovak immigrants in the Czech Republic applied for low-qualified and low-paid jobs, often these ones, the Czechs rejected to take. It might be that due to the increase of population in the working age these jobs might be already taken. In addition to the previous, it might as well be that a large share of immigrants engages in illegal jobs that have no record in the model and therefore cannot be controlled for.

The assumption concerning the negative relationship between the dependent variable and the GDP growth rate in the country of emigration has proved to be correct (the variable is significant at 5% level, adjusted R-squared is 0.33). This means that the rise of the GDP in the Slovak Republic (the economy will thus be better off) might reassure many Slovaks to look for a job in the Czech Republic and make them stay at home instead.

Unemployment rates in the country of immigration (the Czech Republic) and the country of emigration (the Slovak Republic) have proved to be insignificant in their relation to pre-determining the flow of labor migration and both having positive signs. This would lead to the implication that the higher is the unemployment in the Slovak Republic, the higher is the outbound labor migration, but higher unemployment rate in the Czech Republic would still mean high migration. The explanation to that might be that Slovak migrants are highly sensitive to worsening the economic conditions at home, while they are indifferent to worsening of the economic situation in the target country (e.g. the labor

108 market is so large that they expect to find employment anyway. However, very low R-squared and adjusted R-squared and no significance in the predictive power of the variables makes these speculations irrelevant. As far as the case of Czech-Slovak migrations is considered unemployment simply cannot be regarded as the relevant determinant of labor migrations.

Overall, it seems that in the model specified above and using the data outlined in sub-sections 4.3 and 4.4 most of the variables have no or limited impact on the rate of labor migration flows of Slovak citizens to the Czech Republic. Just one variable (rise of GDP growth rate in Slovak Republic) behaved as expected according to economic rationale, while the other four appeared to have different signs (GDP difference between two countries, approximation of the labor market size in the Czech Republic expressed as the number of people of the working age) or to have no significance at all (unemployment rates in the Czech Republic and Slovak Republic).

To sum it up, for the model specified above migration flows are not sensitive to economic differences between the countries, even in case extremely favorable conditions for migration prevail (absence of language barriers, social and economic similarities of the countries, cultural proximity, etc.) as it is in the case of Czech Republic and Slovakia. There is an implication that stems from that conclusion: there might exist a possibility that economic differences within some magnitude have no impact on migration between two countries – a country with a lower economic performance and a country with higher economic performance. However, these economic differences might start to influence migration after they exceed some critical level. There is a legitimate hypothesis that this “critical level” can be different for different countries (having different country-specific propensity to migration).

In order to test this hypothesis, an analysis similar to the one described in Walsh (1974) can be implied. The model can be specified as the following:

U W = α +β it +β it + = Mijt 1 2 u k k 1,.., n )8.4( U jt Wjt

where M ijt is the annual rate of migration flow from country i to country j in period t (number of U officially recorded migrants – people with residence visa), it is the unemployment rate ratio of U jt W emigration (i) and immigration (j) country in period t and it is the wage ratio. W jt This model can be tested, using the data on migrations to the Czech Republic from Slovakia, Ukraine and Russia (three top immigration countries) (Czech Statistical Office, 2006). The variables

109 used in the model are self-explanatory; however some clarifications can be helpful. The annual migration flows are represented by the number of registered migrants (people who applied for residence permit in the Czech Republic) in 1995-2006. It has been already explained that in case of Slovak citizens avoiding registration does not represent a reasonable option. In case of Ukrainian and Russian migrants, the records kept by the Czech Ministry of Foreign Affairs can be underestimated, however due to the distance of both countries and visa regulations the official estimate can be accepted as the best existing data source. Further, the unemployment rate ratio in the country of emigration and immigration is represented by a ratio of official unemployment rates (annual average percentage of all people in working age registered as unemployed) in emigration and immigration countries in 1995-2006. The wage ratio has been calculated using the average annual earnings of workers in the country of emigration and immigration in 1995-2006 estimated as their corresponding values in USD (according to the respective exchange rates). It is, of course, postulated that the signs of coefficients of unemployment ratio will be opposite in sign to those of wage ratio. The results of the estimation are presented in table 4.7 that follows.

Table 4.7: Regressions’ results using immigration to the Czech Republic from Slovakia, Ukraine and Russia in 1995-2006; M is number of registered migrants.

MSLCZ MUACZ MRUSCZ -6590.141 28756.99** -2877.883*** Unemployment ratio (U EMIG /U IMM ) [9521.405] [9753.16] [863.512] -64359.39 156989.2*** -9846.538

Wage ratio (W EMIG /W IMM ) [62730.97] [85034.76] [5723.557] 120495.6** 63759.61*** 15463.6*** Constant [41124.95] [18681.82] [2187.753] N 12 R-squared 0.23 0.55 0.62 Adjusted R-squared 0.06 0.45 0.53 Source: own estimations. Note: Standard errors in brackets; * significant at 10%; ** significant at 5%; *** significant at 1%

It is interesting to observe that although the signs of coefficients of unemployment ratio variables are negative for migration from Slovakia and Russia, the sign of unemployment ratio in the case of Ukraine is positive (and significant at 5% level). Also the wage ratio variable coefficient for migration from Ukraine is positive (and the only one significant amongst the three countries in question). This indicates that increasing the rate of unemployment and the wage in the country of emigration (Ukraine

110 in this case) would lead to the increase in migration from Ukraine (migration is sensitive to economic push and pull factors – worsening of economic conditions at home and incentives in the Czech Republic). In the case of migrations from Russia the results does not lead to conclusion that economic differences between this country and the Czech Republic has reached the “critical level” discussed above. However limited the model presented here might be (due to the shortage of data and the limitation with the time period for which the model is tested), it might be used to suggest that there is some critical level of economic differences between two countries crossing with would lead to these differences influencing migrations including labor migrations. Therefore, for the model and the date specified above, it appears that for countries with a very low economic performance relative to the economic performance of the Czech Republic (such as Ukraine) economic differences might have some impact on migration. On the other hand, in case when economic performance is lower but has not reached the “critical level” (as in the case of Slovakia), economic differences might be of no relevance. Even though the estimates of the model outlined in (4.8) should be taken with some reservations, they provide some interesting results and put the analysis of labor migrations on a rational basis.

The outcomes in this study can be strengthened by some extended empirical analysis of “critical level” values in different countries. This might become the core point of further research on the issue of labor migrations and economic differences between countries. These topics are however beyond the intended scope of this essay.

4.5. Conclusions

It becomes very clear that in a quite number of cases migration flows are not sensitive to economic differences between countries even if extremely favorable conditions for migration prevail. An example of labor migrations of Slovak citizens to the Czech Republic in 1993-2004 was used to testify this case. The results of particular model and data set used in the paper does not lead to conclusion that migrations from Slovak Republic (a country with lower economic performance) migrations to the Czech Republic (a country with higher economic performance) under very favorable administrative conditions for migration (no visas, working permits or other administrational obstacles) as well as language and cultural proximity between the two countries was affected by the economic incentives and economic differences. Truly, economic differences might start effecting migration after they reach

111 some critical level (which has not been reached in the case of Czechoslovak migrations). However, it seems that in the EU context this level should be very high. The analysis of the available data showed that this critical level can be reached, for example, in the case of Ukrainian labor migrations to the Czech Republic.

The analysis of Czech-Slovak labor migrations calls for several important implications. For instance, knowing that economic differences in Czech Republic and Slovak Republic would be creating incentives for workers from the Slovak Republic for emigration to the Czech Republic in a search of better jobs and knowing that barriers to this migration almost do not exist, it would be logical to assume high rate of migrations. However, knowing that under the conditions described above there were no massive labor migrations from the Slovak Republic to the Czech Republic, it seems puzzling why migration expectations from EU-8 and EU-2 (Bulgaria and Romania) are so widely discussed by the general public and justified in television and press. Instead, another postulate seems to be true: the labor market in the conditions of the European Union is sticky and expectations of inflow of labor from poorer Member States to richer Member States are not sufficiently justified.

The case of the Czech-Slovak labor migration was just an illustration of what the trends might be like after the EU accession; this specific case clearly shows that the situation does not incline to massive migration moves of workers. The thing is that even in the case of the Czech-Slovak migration the large gaps in income and problems with demography have not brought more than 2% of the emigrant-country workforce on the move. When the case of labor migration most similar to the situation in the enlarged EU provides such small impact of economic difference on the labor migration, why would it be different after the EU Eastern enlargement?

It should not be forgotten that apart from the economic factors there are other factors, such as cultural, linguistic and social ties which prevent massive flows of workers from rising. It has to be remembered that these factors are often of more sense and should be taken into consideration first when talking about the labor movements. These factors and the proximity to migration as such can differ in various EU Member States.

It is more than likely that the fears of massive labor migration to the EU-25 from the new accession states will not come true as they did not come true in the case of the EU Eastern enlargement and the transitional periods, which are likely to be implemented, will be merely caused by the political tensions and demands, not any considerable economic and demographic grounds.

112 4.6. References

• Bauer T., G. S. Epstein and I. Gang (2002), “Herd Effects or Migration Networks? The location Choice of Mexican Immigrants in the US”, CEPR Discussion Paper Series , n. 3505, August 2002. • Bauer, T., Zimmermann, K. F. (1999), “Assessment of Possible Migration Pressure and Its Labor Market Impact Following EU Enlargement to Central and Eastern Europe“, IZA, Research Report No. 3, Bonn, July. • Brücker, H. and Boeri, T. (2000), “The impact of Eastern Enlargement on Employment and the Labor markets in the EU Member States”, commissioned by the Employment and Social Affairs DG of the European Commission, European Integration Consortium, Berlin and Milan, 2000. • Brücker, H. and Shroder,P. (2006), “International Migration with Heterogeneous agents: Theory and Evidence”, a paper presented on the Annual Labor Economics Association conference in Prague, September 2006 • Brücker, H., Siliverstovs, B. (2004), “The Macro-Determinants of International Migration in Europe: Evidence from Germany”, Paper presented at the conference “Immigration Issues in EU-Turkish Relations”, Bogazici University, October 8-9, 2004. • Dedek, O. et al. (1997), “Ekonomické aspekty zániku Československa“, Praha, Fortuna. • Delegation of the European Commission to Japan (2006), “European Union Enlargement: the implications for Japan”, http://jpn.cec.eu.int/data/current/IMPLICATIONS_0608.pdf • Dorigo, T. (1983), “Push-Pull Migration Laws", Annals Assn. Am. Geographers, 73 (1), pp. 1-17. • Dustmann, C. et al. (2003), “The impact of EU Enlargement on migration flows”, Home Office Online report, 25/03. • European Commission: “Towards a Europe for all ages”, Communication from the Commission to the Council (COM(1999)221 final), Brussels, 21 st May 1999. • Faini et al. (1999), “Migration. The Controversies and the Evidence”, Cambridge: Cambridge University Press. • Faini, R., Venturini, A. (1994), “Migration and Growth: The Experience of Southern Europe“, CEPR Discussion Paper, No. 964, London, May, www.cepr.org/pubs/dps/DP964.asp • Fields, G. S. (1979), “Place-To-Place Migration: Some New Evidence“, The Review of Economics and Statistics, Vol. 61, No. 1, pp. 21 – 32. • Geary, P. and O’Grada, C. (1989), “Postwar Migration between Ireland and the UK” in “European Mobility: Trends and Consequences” (Eds. I. Gordon and A.P. Thirwall). • Goss, E. P., Schoening, N. C. (1984), “Search Time, Unemployment, and the Migration Decision“, The Journal of Human Resources, Vol. 19, No. 4, pp. 570 – 579. • Granovetter, Mark S., (1982), “The Strength of Weak Ties: A Network Theory Revisited.” In Social Structure and Network Analysis, edited by P. V. Marsden and N. Lin. Beverly Hills, CA: Sage Publications, pp.105-30. • Granovetter, Mark S., (1995), “Afterword 1994: Reconsiderations and a New Agenda” in Getting a Job , 2d ed., Chicago, IL: University of Chicago Press, pp.139-82. • Harris, J., Todaro, M. (1970), “Migration, Unemployment and Development: A Two-Sector Analysis”, The American Economic Review, Vol. 60, No.1, pp. 126-142. • Horaková, M. (2005), “Labor migration trends in the Czech Republic during the last decade”, RILSA, Praha, www.vupsv.cz /Horakova-TREND.pdf • O’Gráda, C. (1986), “Determinants of Irish Emigration: A note”. International Migration Review, vol. 20, 650-656. • Pissarides, Ch. A., McMaster I. (1990), “Regional Migration, Wages, and Unemployment: Empirical Evidence and Implications for Policy“, Oxford Economic Papers, Vol. 42, pp. 812 – 831. • Ravenstein (June, 1889): "The Laws of Migration." Journal of the Royal Statistical Society, Vol. 52, No. 2. pp. 241-305.

113 • RILSA (Research Institute for Labor and Social Affairs, VÚPSV (cz)) (2006), surveys and statistical data, www.vupsv.cz • Sinn, H.-W. (1999), “EU Enlargement, migration and lessons from German unification“, Center for Economic Studies, CESifo Working Paper Series, Working Paper No. 182. • Sinn, H.-W. (2000), “EU Enlargement and the future of the welfare state”, CESifo working paper, 207. • Sjaastad, L.A. (1962), “The costs and returns of human migration”, Journal of Political Economy, Vol. 70(5), pp. 80-93. • Tang et al. (2000), “Winners and Losers of EU integration: Policy Issues for Central and Eastern Europe”, The World Bank, Washington D. C. • Todaro, M. P. (1969): “A Model of Labor Migration and Urban Unemployment in Less Developed Countries.” American Economic Review 59 (1), pp.138-148. • Walsh, B.M. (1974), “Expectations, Information, and Human Migration: Specifying an Econometric Model of Irish Migration to Britain”, Journal of Regional Science, 14, 107-120. • Walsh, B.M. (1989), “Tests for Macroeconomic Feedback from Large-scale Migration. Based on the Irish Experience: A Note”, Economic and Social Review, 20, pp. 257-266. • World Bank (2005), “World Development Indicators”, Word Bank, Washington, D.C.

114 5. People of the road: the role of ethnic origin in migration decisions. A study of Slovak Roma asylum-seekers in the Czech Republic in 1998-2006

5.1. Introduction

Migrations, either between regions or countries (i.e. on the regional, national and international level) are accounted for a wide spectrum of factors. Generally, we can distinguish economic, cultural and socio-demographic factors. Ethnicity might also be one of migration determinants. Trovato and Halli (1983) analyze three hypotheses of the existence of substantial ties between ethnicity and migration: i) the social characteristics hypothesis, ii) the ethnic effect hypothesis, and iii) the structural-insecurities hypothesis (Trovato and Halli, 1983). The first hypothesis represents the main antidote to the hypothesis of migration influenced by the economic factors. In our analysis of factors pre-determining migration decision we are often left with some “unexplained” issues (those issues are called “influence of family or friends”, “social ties” – in short those can be called “social factors”). Having its roots in the demographic literature, the social characteristics hypothesis states that migration, a phenomenon being demographic in its nature, is affected by factors like gender, age, income, education, religion, etc. (Shaw, 1975; Bogue, 1959). For instance, education, age and income are showing significant determination of the migration volume. Some authors show that people in advanced social positions are more aware of the opportunities migration can bring (Lee, 1966). The second hypothesis highlights ethnicity as a supplement to other social and cultural characteristics that may have an important effect on migration patterns. The literature describes in which ways ethnicity can affect migration: through reflecting sub-cultural norms which encourage or, on the contrary, discourage, members of specific ethnic groups to migrate; through familistic orientation, social ties and feeling of belonging to community and friends which can also be characteristic for some ethnics (Ritchey, 1976; Mueller, 1973). Ethnicity and migration also have one more important issue: ability or inability to assimilate related to the command of languages. Historical effects (traditions of certain ethnic groups to settle in specific places – i.e. Vietnamese immigrants in the Czech Republic) should not also be forgotten (Korbrin and Goldscheider, 1978). The third hypothesis considers the “pressure” of the larger society on minor ethnic groups: as an effect it leads both to the establishment of ghettos (either induced by the pressure of ethnic “majority” on ethnic “minority” or, alternatively, by the initiative decision of minorities’ leading elites) and to

115 inducing migration (Lee, 1966). There is an insecurity an ethnic minority is experiencing within a larger and dominant nation. Migration is one of many channels to break this insecurity barrier. While the majority of the research literature focuses on economic determinants of migration regardless on the ethnical origin of migrants (usually, a migration from a poorer to a wealthier country is explored), this paper is concerned with the ethnical aspects of migration. Specifically, it tries to explain the migration decisions of ethnical minority. As an example for empirical analysis a story of Slovak Roma asylum migrations to the Czech Republic in 1998-2006 was used. After the collapse of Communism in Central and Eastern Europe large numbers of local Roma started to migrate to a number of wealthier countries: Austria, Canada, Finland, Germany and United Kingdom, just to name a few. Some cases (i.e. massive migration of Romanian Roma to Germany in the 1990 or Czech Roma migration to Canada in the late 1990s and early 2000s) were widely described in press and research literature and contributed to the creation of a negative image of Central and Eastern European countries in the Western media (in the case of the Czech Republic a humiliating introduction of entry tourist visas for Czech citizens wishing to travel to Canada and New Zealand followed soon after). The case of Slovak Roma asylum migrations has been chosen due to the two main reasons: (i) Although almost all EU Member States have some Roma population, substantially large Roma communities can be found in the new Member States of Central and Eastern Europe (notably the Czech Republic, Hungary and Slovakia) (EUMC, 2005)) which makes it a more interesting and rich topic for the analysis; and (ii) Roma minority issues have become a well-known problem of the former Communist countries of Central and Eastern Europe (CEEC), and of all countries migrations of Roma from Slovakia to the Czech Republic have been popularized in mass media and used by politicians and the general public as a classical example of Roma nomad behaviour. It seems interesting to use the publicity and view this problem from slightly different angle, namely by seeing what factors were behind the decision of Slovak Roma to migrate to the Czech Republic using the asylum-seeking procedures. The main value-added of this paper is in using economic and statistical analysis for studying the Roma asylum migrations in CEECs. The paper is trying to investigate whether in case of ethnical asylum migrations migration decisions can be attributed to “push” or rather “pull” factors 15 of migration. The essay is organized in six main parts. The first part provides an overview of economic and sociological literature related to ethnicity and migration. The second part is dealing with the historical

15 “Push” and “pull” factors are the terms used in sociological and demographic literature. Amongst “push” factors are worsening of living conditions in the home country, “pull” factors usually include emigrants’ expectations to obtain higher earnings as well as to improve their economic (and social) well-being in the country of immigration (see for example Ravenstein (1876), (1885), (1889) or Dorigo (1983)).

116 excurse to the history of Roma in Central and Eastern Europe with a special focus on Roma in Czechoslovakia. The third part explains the background of Roma migrations from Slovakia to the Czech Republic after the split-up of Czechoslovak Federation. Part four provides an overview of the data for the empirical model. Part five outlines the empirical model and summarized its main results. Finally, part six brings overall conclusions and implications.

5.2. Literature review: ethnicity and migration

Economic literature unanimously agrees that ethnical (or national) origin has an impact on both the economic performance and composition of emigrants in a given country (see for example Ritchey, 1976; Trovato and Halli, 1983; Borjas, 1994; Borjas, 1999). Running an analysis of different ethnic groups of emigrants Borjas came to a conclusion that: “the decline in immigrant economic performance (in the U.S.) can be attributed to a single factor, the changing national mix of the immigrant population” (Borjas, 1999). Clearly, different ethnic groups tend to work, migrate and use (or abuse) the incentives a state provides to its citizens in different ways. Canadian government took that in mind when introducing the point-based system of immigration policy in 1961. As a result more skilled immigrants flow has been attracted into the country than in the case of U.S. (Baker and Benjamin, 1994). Generally, in the absence of targeted immigration policy, the stream of immigrants will always follow the same rule: immigrant skilled workers will be from the countries where the returns to skills are low and immigrant unskilled persons will be from the countries where the returns to skills are high. An example from Borjas (1999) can be used as an explanation: imagine people who live in a country where the payoff to human capital is low, so that high-skilled workers do not earn much more than less-skilled workers. An example of such a country can be any EU Member State. Indeed, social traditions in most European countries tend to equalize the incomes by taxing the high-skilled persons and subsidizing the surplus to the less-skilled. In this case the workers who profit from migrating to the countries with better economic opportunities are those with their skills above average. On the other side of the spectrum there are countries where the payoff to human capital is high. Most developing and third-world countries can be used as an example. It is quite cumbersome to obtain a decent education and skills in these countries, that is why those qualities are highly-praised and their bearers enjoy wide social recognition and public respect, which is also reflected in their earnings. Highly-skilled people in developing countries earn much more that unskilled which leads to

117 uneven income distribution. The workers who profit from migration from the countries with better economic opportunities are those with their skills far below average. Attributing this analysis to Roma migrations in the Central and Eastern European countries (CEEC) arises several major issues: i) All Roma in the CEEC migrate between countries with more or less similar level of economic development and high payoff to human capital. Although some countries offer more economic incentives (Czech Republic, Hungary) and some less (Slovakia, Romania) all of them have been transition economics burdened by the communist heritage. ii) Roma in CEEC usually accept the low-paid jobs and are generally unskilled, therefore their level of human capital is low. Unlike another persecuted social group – the Jews – Roma have always neglected the possibility to face the atrocities and misfortunes that can happen to their ethnic (pogroms, discrimination by majority, Holocaust) by investing in the education of their children. iii) With regard to the two points above it can be concluded that Roma are likely to constitute the least skilled immigrant workers any country in CEECs can attract. As a result, Roma economic performance is likely to be very low. iv) As follows from the above, Roma migrants will be earning less than the “majority” population. This fact is often attributed to the existence of economic discrimination. However, it has to be considered that Roma earn less not because of discrimination but because they are less skilled. Moreover, it has to be noted that the majority of Roma immigrants in the CEECs are economic migrants: absence of wars or major natural catastrophes, stable political climate and moderate racial tensions seem to support this thesis. In addition, what has to be considered in the debate on Roma migrations are the costs of their emigration: such as transportation, settling down in a new place, searching for a job as well as bearing various social burdens of leaving familiar environment, family and friends. If we put aside people driven out by war conflicts, famines, diseases and natural catastrophes, emigration is a self-defining act equal to the process of self-selection. As a rule, people choose to emigrate by their own will. It might as well be that some people tend to choose emigration more often than the others. Taking into consideration all of the above, it will be very unlikely that anyone in that conditions would prefer emigration to staying at home. Emigration to the country with similar economic conditions (low payoff to human capital) and facing the same economic discriminations seems to be illogical. However, in spite of all these Roma do migrate. Perhaps, the main reason for Roma migrations is an attempt to abuse the welfare state. However, would this be enough to undertake a

118 long, exhausting trip, to face uncertainty and hardships in the country of immigration? Economic reasoning seems to speak against this assumption.

5.3. A short excurse to the history of Roma in Central and Eastern Europe

5.3.1. Destinies of Roma population in Central and Eastern Europe

Roma trace their origin to India where they first appeared about 3500 years ago as an ethnic called the Dom. After the conquest of India by the Indo-European tribes, the Dom became one of the lowest casts in the Indian hierarchy: their typical occupational domains were professions like sewage cleaners, fur and skin sellers, bear handlers and the like. Around the 11 th century AD some small groups of Roma had started to move through Persia and Syria to the Balkans, from were they were spreading all around modern Europe. Although one of the first well-documented presence of Roma in Europe comes from the 14 th - century (1362) Republic of Ragusa (now Dubrovnik in Croatia), it is apparent that they had been living there permanently for more than a century before that date (Petrovic, 1976). Traces of what is now identified as Roma population can be also found in Hungarian, Slovak and Moldavian- Wallachian records from the same period. Some of the Roma were captured and sold as slaves, although most of the Roma in Eastern and Central Europe were free and worked as musicians, soldiers and craftsmen (predominantly as blacksmiths). Many Roma were joining the military forces of Hungarian kings and Lithuanian princes, although there had been always seen as strangers with and treated with a mixture of awry and disbelief. However, it was not until the growing threat of the Ottoman Empire and the upheavals of the Protestant Reformations in the 16 th century that the attitude towards Roma in Europe began to change dramatically. Growing hatred and suspicion (many regarded Roma as the Turk’s “fifth column”) caused restrictions to Roma lifestyle and trade. These had played a major role in Roma ethnic adopting a nomadic way of life as a means of survival which in its turn had caused a prejudiced attitude towards Roma minority in most of Europe. Roma started to be placed at the lowest social ranking in social hierarchy not only by the Christian population of Europe but also by the Turkish invaders: many Roma in were bought and sold as the slaves (called “robi”) and a serious of restrictive laws regulations was issued to govern the Roma minority. Although the destinies of Roma were following a somewhat slightly different path in various European countries – from the plight and mistreatment of Roma in Wallachia and Moldavia through

119 harsh policies of the Habsburgs in the Czech lands to the relatively moderate treatment in Hungary and Slovakia – one trait always remained the same: a prejudiced attitude of the national “majority” to all Roma linked to the persecutions and disdain. Throughout the centuries Roma minority in Europe had been vexed and restricted in their free lifestyle. Various bans pillorying Roma as “vagabonds” and “beggars” had existed with different degree of intensity until the WW II. Centuries of European history of wars, conflicts and turmoil had always been marked by general mistreatment of the Roma minority: in the times of war or piece they were the ones that had been more often marked as a scapegoat. There had never been serious attempts to integrate this minority: partly due to the lack of interest of Roma themselves, partly due to the lack of interest of the ethnic “majority”. Roma remained a strange and inadaptable ethnic that was let to live its own life as soon as it did not contradict with the interests of the population “majority”. There had been attempts to include Roma into the social and political life in different periods of history in the 19 th and early 20 th century, however the major efforts can accredited to the Communist regimes in the CEECs and Russia: many programs and initiatives including forcible parting Roma children from their parents and obligatory placement of Roma into newly built housing blocks are among the few to mention. Nevertheless, none of them bore remarkable fruit and most of them did more harm than good. Although notable Roma minorities are spread all over present Europe, nowhere is their concentration so high as in the Central and Eastern European countries: according to the 2001 census there were 535.000 Roma in Romania (INSSE, 2002), 190.000 in Hungary (Hungarian CSO, 2001), about 200.000 in the Czech Republic and 450.000-500.000 Roma in Slovak Republic (Czech Statistical Office and Slovak Statistics, 2006). Even though the official statistics on Roma population of most of the Central and European countries already seem quite high, they fail to reflect all actual Roma: many people do not identify themselves as Roma in the census, or, in many cases, a mother tongue is used as a criterion of nationality (many Roma do not use their language as the first language of communication). Taking into account these factors, the predictions of actual Roma population in CEECs might be much more higher: for example Romanian sources mention that it might be up to 2.3 million Roma in Romania instead of officially proclaimed figures that were much more moderate (Ionescu, 1992; Liebich, 1992). With Central and Eastern European countries shifting towards democracy and market economy in the early 1990s Roma population found itself on the margin of the society. The problems and bad economic situation in the first years of transformation reflected on growing ethnical tensions and worsening situations of Roma minority. Once again, like many times before, Roma were blamed for growing crime rates, abusing welfare systems and creating disorder. This resulted in many Roma

120 immigrating far and away, often to the wealthier countries of Western Europe and beyond (Finland, UK and Canada just to name a few).

5.3.2. Roots and destinies of the first Roma settlers in Czechoslovakia

The first record of Roma appearing in what is now Czech Republic and Slovakia dates back to the late Middle Ages. Emilia Horvathova, a Slovak gypsiologist, states that Roma first entered Bohemia via Hungary with the army of the king Andrew II (1205-1235) after he returned from the Crusade in the Holy Lands in 1217-1218 (Horvathova, 1962) .When the Tatars invasion to Hungary came in 1241 and especially after the defeat of King Bela IV at Muhi, Roma fled to Bohemia to escape the butchery (Crowe, 1995). According to a prominent Czech gypsiologist, Eva Davidova, the first undisputed reference to Roma comes from 1399 at the Book of Executions of the Lord of Ruzomberok where a certain Gypsy is mentioned as a member of a robber band (Davidova, 1970). First Roma settlers in what is now Czech Republic and Slovakia were musicians, metal workers and some even became warriors in the armies of Hungarian kings. They co-existed with the original population in quite and peaceful atmosphere, only sporadically facing some racial tensions caused by religious and cultural differences. It was not, however, until the Hungarian army defeat at Mohacs on the 28 th of August 1526 that strong anti-Gypsy policies started to emerge. The general public was seeing Roma as “the spies of Tatars”, a sort of “fifth column preparing the grounds for invasion” and they were banned from traditional professions and limited in their travels. The anti-gypsy legislation was issued in Moravia in 1538 by king Ferdinand I Habsburg as a result of fear of Turkey military power and the destabilizing atmosphere brought by Protestant wars. This was the period when all Roma troubles had began: regarded as outlaws both by Christians and Muslims (Muslim Turks respected Christianity considering Christians and Jews to be “people of the book” while Roma were exterminated severely), driven out of big cities, suspected of espionage, crimes and theft they began to wear a label that is to remain on this ethnic for centuries.

5.3.3. Roma in the Habsburg monarchy and Austrian-Hungarian Empire

Austrian emperors Leopold I and Joseph I issued decrees proclaiming Roma to be outlaws (“vogelfrei”) and ordering mass hunts and killing of the ethnic. Maria Theresa first ordered all Roma to be driven out from the land (1749) and then outlawed the use of word “Rom” or “Cigan” and

121 decreed them to be called “new citizens” or “new Hungarians” (Crowe, 1991). Her enlightening policies on Gypsies included the creation of Rom settlements, census taking and placing Rom children in foster homes, schools or jobs. In 1780 there were 43609 Roma in Slovakia (without female) and Hungary; this figure dropped to 38.312 in 1781 but rose to 43.778 in 1782 as a result of Joseph’s II policies of emancipation and Serfdom Patent (Horvathova, 1962). After the revolutions in Europe in 1848 and the creation of Austrian-Hungarian Empire most of the Roma found themselves in Hungary and were exposed to magyarization and oppression. The data on those remaining in Bohemia remain scarce, however it is easy to assume that they had also been naturalized and attempted to turn to the “good citizens”.

5.3.4. Roma in the independent Czechoslovakia in the after-WWI period

After the WWI and the declaration of Czechoslovak independence in 1918 all Roma were given Czechoslovak nationality ipso facto . According the 1921 Czechoslovak census there were 61 Roma in Bohemia and 7.967 in Slovakia (Davidova, 1970). According to specialists this was based on people speaking “mother tongue”, because about the same time in 1924 another statistics show that there were 18.257 Roma in Kosice region and 11.066 in Bratislava (Horvathova, 1962). The mid-war period was marked by various reforms aimed to improve the life conditions of peasants in Slovakia, those including the Roma. Various nomad passes had been issued for them and notorious law nr. 117 took described special measures how to treat “nomadic Olašti Rom” and “people living by nomadic style” (for instance requiring them to ask permission to stay on the lands from the local majors or prohibiting them from entering some territories such as spas) (Davidova, 1970). There were several attempts to organize schools for Roma children or commit them to sports (for instance the creation of a football club SK ROMA which gained some fame in those times), however the new Republic had many other problems to face.

5.3.5. Roma minority in Czechoslovakia during the WWII

After the Munich Treaty, occupation of Czechoslovakia by Nazi Germany and the outbreak of the WWII those Roma remaining in Bohemia (now called Protectorate of Bohemia and Moravia) suffered the most (their fate being similar to that of the Jews), while those living in Slovakia (independent Slovak state) found themselves in a more tolerant environment.

122 Labor camps Léty and Hodonín were created in the Protectorate in the 1941 especially for the Roma and many of them found their death there or in Auschwitz or Buchenwald. The Roma Holocaust (called Po řajmos) took away from 6.000 to 8.000 Rom from the Czech lands and few hundreds of their more fortunate Slovak compatriots. According to the Census in 1947 there were 84.438 Roma in Slovakia and 16.752 Roma in Bohemia (Ne čas and Miklušáková, 2002).

5.3.6. Attempts of Roma assimilation in the Communist Czechoslovakia

The post-war period was marked by the Communist takeover in the 1948 and the interest to Roma minority was bleak. In 1950 president Gottwald called up a commission to find a solution for “Gypsy question”. A report on “Position of Individuals of Gypsy origin in the Work Process” was published, however not a single major effort was taken until the 1958 when the Communist party published a declaration demanding “unconditional solution to the Gypsy question” and dividing all Roma into three categories: nomad, semi-nomads and completely sedentary. Those of nomadic and semi-nomadic nature were then subjected to the Law No.74, (October 1958) “On Permanent Settlement of Nomadic People”. In addition to that, perhaps to avoid the charges of racism, Roma were given a special status of socio-ethnic group which only widened the gap between them and the Czechs and Slovaks. In 1965 the Communist party prepared a long-range assimilation plan crowned by the Ordinance No. 502 (June 1965) which prescribed “full-employment of all Gypsies, liquidation of Gypsy illegal settlements and dispersion of the Roma all around Czechoslovakia” (Crowe, 1995). However the plan failed due to the growing resentment towards the Roma presence in the Czech lands, inadequate funding and Roma unwillingness to live in Bohemia and Moravia. As Czechoslovak demographic literature suggests, in the 1960s Czechoslovakia had one of the largest Roma population in the Socialist Bloc. The data from the 1967 estimated 223.993 Roma in the country (with 164.526 in Slovakia and 59.467 in the Czech lands) (Ul č, 1990). Interrupted by the occupation of Czechoslovakia, the campaign of Roma assimilation continued in the 1970s under the rule of the new party leader Gustav Husák. It included more severe policies, including sterilization of Roma women for financial benefits offered by the local health workers and intended to reduce Roma birth rate (the average number of live births for a Czechoslovak woman was 2.4 and for Roma woman it was 6.4) (Srb, 1985). Roma population grew from 219.554 in 1970 to 288.440 in 1980, a 31.4 per cent increase, while the general population grew by only 6 per cent during the same period. The Roma population in the Czech Republic grew by 47 per cent during 1970 and 1980, while the same for Slovakia was 25.5 per

123 cent (Ne čas and Miklušáková, 2002). Quite understandingly, some fears of over-population of Gypsies were quite common within the general public which resulted in day-to-day discrimination and despise. In the same time there began an increase in the Roma population Bohemia caused by the destruction of Roma housing in Slovakia and forced movement of Gypsies to Bohemia and by voluntarily movements of Gypsies searching for better economic opportunities and higher standard of leaving. During the decade of 1970-1980, over 4.000 Roma dwellings had been destroyed in Slovakia (Ne čas and Miklušáková, 2002). Another part of the assimilation campaign led by the government was forced education of Roma youngsters. In 1971 only 10 per cent of eligible Roma children in Czechoslovakia attended kindergartens. Throughout the 1970s the number of Roma children graduating from schools rose from 16 per cent to 25.6 per cent (Crowe, 1995). Generally, school drop-out rate was increasing with the grade and only a third of Gypsy children finished eights grade in 1983-1984, while merely 4 per cent went on to University (Ul č, 1990). Gypsy employment during the period of assimilation policies was a mixture of failure and success: in 1981 in Czechoslovakia 87.7 per cent of male Roma (91.9 per cent average for the Czechoslovak males) and 54.9 of female Roma (87.7 for Czechoslovak females) were employed. In addition, 80 per cent of all Roma were in low-paid and low-skilled jobs (in agriculture, construction and industry) (Crowe, 1995).

5.3.7. Roma in Czechoslovakia after the collapse of the Communist regime in the 1989

The Velvet Revolution on the 17 th of November 1989 and the liberalization of economic, social and political life brought two different effects for the Roma minority: alongside with the opportunity for the Roma to form their own political and cultural organizations, a new atmosphere for open expression of the prejudice towards the Roma was created. Right after the collapse of the Communist regime the government estimated that there were some 400.000 Roma in Czechoslovakia, a 36 per cent increase for over a decade. A sharp rise in crime level (by 52 per cent in the Czech Republic and 17 per cent in Slovakia) that engulfed Prague (by 181 per cent in Prague alone in 1990) and Bratislava, a thing unknown in the times of the police state maintained by the “iron hand” of the Communists, was attributed to Gypsies, especially those 18 years old and younger (Orbman, 1991). This resulted in skinhead attacks that were the main reason the Rom emigrating to Canada and the UK were declaring in the asylum forms.

124 Before the formal separation of Czechoslovakia in 1993, Slovakia had an estimated Gypsy population of 400.000, and the Czech lands had 150.000 (Crowe, 1995). In anticipation of a breakup of a federation a number of Slovak Roma began to move to the Czech lands because they felt the political and social atmosphere would be more tolerant there, and economic conditions better. This sudden move was a surprise for many Czechs who attributed a sharp increase in crime by 90 per cent as caused by these migrations. After the separation of the Czechoslovak federation, many Roma were having troubles with obtaining the Czech citizenship. Due to their inability to fulfill the demanded criteria the majority of Roma living in the Czech Republic by 1993 became Slovak citizens. In general, Roma living standard in the Czech Republic can be regarded as those superior to the ones in Slovakia, especially in the eastern regions of the country. This, in combination with lower wages, social and racial segregation and chronic debts many Gypsy families have, led to the natural trend of inducing more or less substantial migrations of Slovak Roma to the Czech Republic. As the report of International Organization for Migration (IOM) dealing with the emigration of Slovak Roma to the Czech Republic proclaims, those migration waves could have reached some 10.000 Roma in the 2000-2003 (IOM, 2003). However, openness of the borders and labor markets between Czech Republic and Slovakia for the whole period after the split-up of the Federation reinforced by the traditional Roma secrecy and lack of belief in authorities made it very difficult to obtain the real data on migrations.

5.3.8. The European dimension of the place of Roma minority in Central and Eastern Europe after the collapse of the Communist regimes.

The situation of Roma minority in other Central and Eastern European countries after the fall of Communist regimes was quite similar to the one in Czechoslovakia as presented above: Roma are viewed as the symbol of all gone awry in the new democracies of the Eastern Europe and, to a lesser degree in Russia (Crowe, 1995). These attitudes are often fueled by the fears of Roma population immense growth rates that usually outnumber those of the national “majority” as well as blaming the Roma for the threatening crime rates. Roma minorities in different European countries were meeting with varying attitudes and treatment. For instance, the situation of Roma in Hungary can be used as an example of moderate tolerance and even attempts to let Roma participate on the political decision-making. The political climate in Hungary was moderate even in socialist times (especially in the 1970s and 1980s) and some form of Roma council (Ciganyszovetseg) had existed since 1974. Hungarian Roma were granted an

125 ethnic group status in 1979 by the Council of Ministers’ resolution and there were attempts to create a purely Roma political organization in 1985 (Puxon, 1975). After the political reforms on 1990 several political organizations consisting of local Roma were created: Roma Brotherhood, Democratic Alliance of Hungarian Gypsies, Hungarian Gypsy Party (HGP) and Hungarian Gypsy’s Social Democratic party (HGSDP) with an astonishing membership of 15.000 in 1990 (Barany, 1992). Although these parties never gained enough votes to enter the parliament, they paved the way for other Roma political initiatives: in March 1990 Association of Free Democrats (AFD), the second most- important party in the country, allied with several Roma groups and promised Roma nationality status and government aid supporting Roma candidates for parliament (Bruszt and Stark, 1992). Roma attempts to get political voice and claim their social presence in Hungary caused some anti-Roma feelings and extremists’ attacks (which were quite few), but politicians were working to strengthen the minority rights for all groups. In 1993 the Law on the Rights of National and Ethnic Minorities (which specifically mentioned Roma) was passed by Hungarian parliament stating that outlawed any form of discrimination against minorities and guaranteed equal opportunity for all minorities (this was particularly significant in Hungary where different ethnic groups constitute over 30 per cent of the population) (Crowe, 1995). On the contrary, the situation of Roma in Romania can be used as an example of constant oppression and devastation. Bearing a considerable Roma community (in fact, the largest in the country), Romania suffered greatly after the collapse of Ceausescu dictatorship in 1989: poor economic performance and constant protests (miners’ marches on the capital city Bucharest) burdened the country’s way into capitalism. Roma became the national scapegoat for Romania’s immense problems. There were several riots in Bucharest (for instance Jiu Valley miners’ riot in June 1990) that resulted in pogroms of Roma quarters and beating and abusing Roma on the streets. In addition to that many Roma were arrested by the police and blamed as the starters of the unrests (King, 1991). In 1991 the International Helsinki Federation for Human Rights and International Labor Organization (ILO) raised the topic of ethnic violence against Roma in Romania. These outbursts of violence of equally deprived Romanian “majority” population combined with Roma extreme poverty and discrimination lead to massive Roma emigration from Romania. The majority of Romanian Roma chose Germany as their target country of immigration (about 35.000 Roma that entered Germany among 130.000 Romanians in 1990 alone) (Ionescu, 1991). The problem of Roma orphans in Romania was also a very grim legacy of the Communist regime that no one until today knows how to solve. Following Ceausescu’s decree of 1966 banning abortions for women under 45 many Roma children had been put away by their careless parents. In 1990 it was discovered that Romanian orphanages were filled with

126 100.000 children (Crowe, 1995). This situation has been criticized by the European Union (EU) several times and many improvements have been made. Overall, if the problems of Roma community in several Central and Eastern European countries are compared, there are some similarities that can be pointed out. First of all, due to the low level of education Roma were the first to loose jobs when large socialist companies were downsizing at the beginning of economic transformation that followed the collapse of Communist regimes in 1989-1991. This contributed to high unemployment and growing social deprivation amidst Roma minority. Second, as a result of economic downfall and growing social unrest in the former CEE Communist countries Roma were often becoming a target for racially-motivated “ventilating” attacks. Third, in spite of efforts to establish in politics, Roma representatives have never been capable of putting together a considerable political force. Fourth, as a result of institutional changes the majority of Roma found themselves at the margins of the society which led to further social deprivation, growth of usurers’ networks and finally to choosing migration (usual migration or asylum migration) as the only possible solution to all problems.

5.4. Types and reasons of Roma migration from Slovakia to the Czech Republic after the split-up of Czechoslovakia

In the report of the Slovak Institute of Informatics and Statistics (Infostat) entitled “Projection of Roma population in Slovakia until 2025” Roma migration across the Slovak borders is considered “minimal” with “no significant changes expected in the future” (Infostat, 2002). Drawing possible scenarios of the position of Roma population in Slovakia external Roma migration is considered to be zero: authors support that assumption by noting that in the enlarged Europe the main type of migration is labour migration and Roma people do not have the tendency to migrate for labour (Infostat, 2002). Apart from that members of the Roma ethnic group are distinguished by the low level of education which makes their adaptability and their chances for perspective employment in Europe quite low. Prior to 2000 there were no major Roma migrations from Slovakia to the Czech Republic, or even if they were, no attention from mass media and the general public was expressed. In fact, Roma from former Czechoslovakia preferred asylum migrations to more “wealthier” countries, such as Canada or United Kingdom. These migrations from former Czechoslovakia started much earlier and were allegedly caused by the Czech Citizenship Law (40/1993 of the Czech National Council) adapted in 1993 according to which people who previously had federal citizenship (and not the one of two federal states) and who were Slovak citizens in 1969 or were born in a family of such people were denied

127 Czech citizenship – notwithstanding their permanent residence on Czech soil (O’Nions, 1999). This definition fitted many Roma whose parents came to the Czech Republic from Slovakia and as a result from four hundred Roma to 77.000 Roma were affected (Tolerance Foundation, 1994b). This put together with worsening employment opportunities for poorly-educated Roma and (to some extent) a raise of racially motivated attacks on Roma lead to predominantly economically-motivated migrations of Roma from former Czechoslovakia to the United Kingdom. In October 1997 the information appeared in British press that 3.000 Roma from former Czechoslovakia came to the UK.

Table 5.1: Top 10 asylum-seeking nations in the Czech Republic in the 2002 Country Nr. of applications % Ukraine 1674 20 Vietnam 891 11 Slovakia 843 10 Moldavia 724 9 Georgia 678 8 Russia 628 7 China 511 6 Armenia 452 5 India 364 4 Belarus 311 4 Source: Ministry of Internal affairs of the Czech Republic, 2005

The UK Home Office minister, Mike O’Brien had to reduce the period of application for asylum- seekers from 28 days to 5 days as a measure of dealing with these migrants (CTK, 1997). Throughout the 1995-2002 thousands of Czech and Slovak Roma applied for asylum in the UK – for instance in 2000 alone it was 1200 cases (Migration Watch, 2005). Thousands were deported after their application of asylum had been rejected. Nevertheless, welfare benefits and free housing during the waiting period made the UK asylum migration a very attractive opportunity for many Roma. Czech and Slovak accession to the EU made the asylum-seeking legally impossible and brought a halt to these issues of controversial asylum migrations. Many Czech and Slovak Roma did not hesitate to travel further for the asylum and resulting economic and welfare benefits – their target countries were Canada and New Zealand. The Helsinki Commission noted 1100 arrivals in September 1997 alone and by the end of August of 1997 the homeless shelters in Toronto were filled with Czech Roma asylum-seekers (Helsinki Commission, 1997). The Czech Republic was merely watching the process and had to agree with placing British

128 immigration officers in Prague Ruzyne airport in order to filter out potential asylum-seekers. The media took the subject as an interesting issue and the reports on Roma emigrants were often broadcasted in the media. All that changed for the Czech Republic when it experienced its “own” Roma asylum immigration for the first time in 2000. The wave of Roma asylum-seekers did not fade in the successive years. For instance, in 2002 alone Slovak Rom submitted 843 applications for asylum in the Czech Republic (Table 1), allotting Slovakia on the third place in the ranking according to the number of asylum-seekers (right after Ukraine and Vietnam). The statistics on the actual numbers of the Rom migrants in that times have never been collected (Ministry of Interior of the Czech Republic, 2002), although many Roma activists and NGOs at that time proclaimed that “thousands of Slovak Roma were on the move”, a number too far exaggerated.

Chart 5.1: Slovak (Roma) asylum-seekers in the Czech Republic in 1999-2006

350

300

250

200

150

100

50 Number of asylum-seekersNumber of (persons) 0

3 6 99 0 00 0 2 02 3 0 3 04 5 05 6 0 - -0 0 - -0 - 0 -0 n-99 p-99 n y- -0 n- p-02 n -0 n- y- p-05 n a ay a ep ay ep ay a J M Se Ja M S Jan-01 May-01 Sep-01 Ja M Se Ja May- S Jan-04 M Sep-04 Ja M Se Ja May- time period

Source: Czech Statistical Office, 2005 and Ministry of Internal affairs of the Czech Republic, 2005.

According to the International Organization for Migration (IOM, 2003) factors that influence Roma migrations from Slovakia into the Czech Republic can be split into push and pull factors. The IOM report on the emigration of Slovak Rom to the Czech Republic names a number of push factors in the domicile (Slovakia) and pull factors in the target country (Czech Republic). Table 4.2 below provides an outline of major quantifiable push and pull factors.

129 Table 5.2: Major quantifiable „push“ a „pull“ factors influencing emigration of Slovak Roma to the Czech Republic „Push“ factors „Pull“ factors Increase of racially-motivated attacks in Difference between the unemployment rate Slovakia in the Czech Republic and Slovakia Dynamics of minimal hourly minimal Increase of unemployment in Slovakia wage in the Czech Republic Dynamics of social welfare in Slovakia (for Dynamics of social welfare benefits in the instance, the decrease of the maxium amount Czech Republic of social welfare benefits per one family to 10.500 SKK) Source: IOM, 2003, own estimations

Among major quantifiable push factors are such worsening conditions at home as the increase in the number of racially-motivated attacks in Slovakia, hard economic conditions at home represented by the high rate of unemployment, decrease in social welfare benefits for families (the maximum amount was decreased to 10.500 SKK). Other push factors, also mentioned by the IOM report but rather more complicated to be quantify and thus to verity are: congestion of Roma in their settlements, high fertility of Roma women, increase in the number of usurers and the tolerance to usurers from the state. On the other hand, major quantifiable pull factors for Roma migrants are the incentives in the country of immigration represented by the employment opportunities (difference between the unemployment rate in the Czech Republic and Slovakia), dynamics of the hourly minimal wage in the Czech Republic (this is especially relevant for asylum migrants who cannot engage in full-time employment but who, until recently, could work part-time while waiting for the decision about their asylum applications) and dynamics of social welfare benefits in the country of immigration (Czech Republic). Other pull factors include networks of families and family ties, opportunities to increase personal income (those including illegal activities), food and shelter offered by NGOs and asylum facilities (IOM, 2003). All factors influencing migrations of Slovak Roma to the Czech Republic can further divided into three main groups according to their nature of impact: “economic”, “social” and “cultural” factors. This division can be better described using the following points:

130 A. “Social” factors : a) Decrease in welfare benefits for Roma : starting from January 2003 the maximum amount any given family could obtain from social welfare was lowered to 10.500 SKK. This is, however, doubtful, as far as many Roma families could not obtain all possible welfare benefits (such as housing support) because just a small fraction of them actually own the housing they are living in. Besides, the majority of social welfare goes for repaying the debts to the usurers and thus do not affect the families’ income in any substantial way. b) Employment opportunities : emigration and temporary migration in search for jobs from the Eastern Slovakia is not just the typical feature of Roma community. Many native Slovaks do that too, this factor cannot be attributed only to the Roma, but rather to all Eastern Slovakia population. Besides, the majority of Slovak Roma engages in illegal unemployment which brings the risks of not getting paid by the employer. This, of course, does not concern asylum-seekers who, on most of the occasions, cannot take a paid employment. c) Spatial segregation : many Roma settlements in Slovakia are segregated from the Slovak population – there are schools with Roma children, ghettos where only representatives of Roma ethnic minority live, etc. In such environment crime, drugs and alcohol are vivid on the daily basis. This leads many more advanced Roma to attempting to break the “vicious circle” by emigrating. d) Discrimination : typical discrimination can be divided into racial and employment. Racial discrimination, as the IOM report states, is now on the decline but employment discrimination is quite common (a very common story told by many Roma that employment was offered to them in the phone conversation but the offer was withdrawn when the employer saw who he is about to employ had been verified by mass media many times in the past). i) Conflict : many Roma choose to escape the conflict situation, such as attacks from usurers by migrating. This fact is supported by the inability of the state to intervene in the illegal usurer’s practices. e) Deprivation : deprivation and feeling useless are quite common features for many Slovak Roma. Locked in the claustrophobic environment of ghettos and settlements, many Roma feel deprived and quite often try to break the circle by leaving anywhere to the West.

B. “Cultural” factors : a) Family networks : Roma ethnic minority has always been known for supporting extended family ties and helping each other out. Although many gypsiologists say that the traditional family ties are in

131 decline right now, many Roma from Slovakia can always rely upon their relatives living in the Czech Republic. The help provided by the Czech branch of the Roma families might include provisional and long-term accommodation, food, help in finding employment, assistance in engaging in a marriage with a Czech Roma in order to obtain Czech citizenship, etc. b) Usurer networks : usurers (borrowing money with a high per cent) can be both a catalysis and a slowing factor of migrations: on the one hand usurers often induce Roma debtors to engage in employment abroad or apply for asylum in order to repay the money they have previously borrowed; on the other hand, even if some debtors wanted to leave or even escape the usurers through emigration, they are forced to stay in ghettos and settlements with practically all the social welfare benefits going to the usurers as a part of the debt repay. c) Roma status and social structure : as it has been already said in this paper, Roma migration can be characterised as the migration of the lower middle strata. While the upper-class Roma do not need to improve their social status (they benefit from living in a society with a high return to human capital), lower-class Roma have nothing to lose and thus engage in migration of any kind.

C. Economic factors: a) Social welfare benefits : the amount of social welfare has not proved to be an important factor for Roma migrations. As mentioned before, most of the social welfare benefits end up in the hand of usurers, thus Slovak state cannot do anything to halt of reverse the emigration. b) Paid medical care : paid medical care in Slovakia, when a patient has to pay some symbolic amount for visiting the practitioner, is often mentioned by Slovak Roma as the major push factor. However, it is quite difficult to attribute emigration to that factor alone, especially when estimating such medical expenses for one family to be about 500-600 SKK. c) Unemployment in Slovakia and in the Czech Republic : the difference in the employment rate plus the language and geographical proximity make the Czech Republic to be a very perspective destination for many Roma (only those who migrate for labour, asylum migrants cannot be considered). Even though many Roma emigrate in order to make additional money, the labour migration from Slovakia is not their domain alone. Many Slovaks do the same with the only difference being in obtaining all papers for registering for legal employment. Roma low trust in the power of state and police often lead them to look for illegal unemployment with doubtful results at the payday. Both social and cultural factors constitute and interesting framework for the analysis of determinants of Slovak Roma migrations. Alongside with economic factors that are used most often in these types of the analyses, they might provide a starting point for further research. Issues like racially

132 motivated attacks or welfare policy in the country of emigration and immigration might be analyzed in order to determine their impact on the volume of Roma migration, however their “immeasurableness” and purely qualitative nature make them difficult to be used for the approach envisaged in this paper. This provides a rationale to leave them behind and focus on the quantifiable determinants (represented by the economic factors only). Perhaps, to give a hint of the direction of future research would be utilizing the framework described for example in Gramlich et al. (1984); Friedberg and Hunt (1995); Borjas (1999) and Meyer (2000) which all analyze the impact of welfare on migration. Quite true, an issue whether welfare attracts potential immigrants and to what extent it is attractive is an interesting issue to consider. However, this analysis will therein limit itself to another, no less interesting aspect that includes ruling out economic determinants of Roma migrations. When it comes to the types of migration of Slovak Roma, IOM report (IOM, 2003) distinguishes four basic types of immigration:

1. Unregistered immigrants : those immigrants from Slovakia usually contain singles or deprived individuals (quite often teenagers and orphanage-leavers) who are not interested in supporting their families in Slovakia and thus do not try to look for employment. These individuals are looking for relatives and often use them as base and source of pocket money. These types of migrants often engage in petit crime: stealing, drug dealing and the like. 2. Temporary labour immigrants : this type of immigration of Slovak Roma into the Czech Republic is the oldest. Open labour market for Slovak citizens, inexistence of the language barrier, similar environment and culture make employment in the Czech Republic to be very attractive opportunity to make. The vision of employment in the Czech Republic is so tempting that many Roma accept illegal employment. This type of migration is temporary: Roma workers come back home in short intervals and also send the majority of their income back to their families. In many cases, temporary labour immigration has seasonal nature. 3. Asylum seekers : this type of immigration is becoming very popular in the recent years. Although almost every asylum seeker is sure his or her application will be rejected by the Czech authorities, long processing of the application (considered to be a disadvantage for many) which can take up to two years, combined with free housing and accommodation (sometimes referred by the Slovak Rom as “holiday”) and, until recently, employment opportunities for asylum-seekers make the Czech Republic to be one of the favourite places for Slovak Rom. Even though the applications for asylum are rejected the individuals can return back and apply again every two years. I will thereafter concentrate on that type of migration in analysing the interdependence between ethnicity and migration in the Czech

133 Republic: the data are readily available and, contrary to the other types of migration, are precise and thus can be used in the econometric analysis. 4. Multiply immigrants : this type of migration includes Roma migrating from Slovakia to the Czech Republic back and forth. Sometimes they are taking care of the elders or children, in many cases there are Roma born and raised in the Czech Republic but given Slovak citizenship after the split-up of the Czechoslovak federation. Quite numerous group is unsuccessful asylum-seekers who have just returned from the Western Europe or North America: with the shutdown of the social welfare benefits (usually for half a year for those who leave the country) and the inexistence any means for living, those Slovak Roma choose to emigrate to the Czech Republic and use the asylum procedure to obtain free housing and accommodation in order to gain time and get new ideas (there are lots of “professional” asylum-seekers among them).

5.5. Data

The data have been obtained from the Czech and Slovak Statistical Office and Statistics of the Ministry of Internal affairs of the Czech Republic and Slovak Republic. The data is time-series nature. It represents quarterly observations of the number of Slovak asylum-seekers in the Czech Republic in 1998-2006 as well as quarterly observations of the following variables: number of racially-motivated attacks in Slovakia regardless of the victims’ origin or ethnical background (this indicator is kept by the Ministry of Internal affairs of Slovakia), number of unemployed in Slovakia in 1998-2006 (in thousands of persons), dynamics of social welfare benefits in Slovakia in 1998-2006 (in thousands of SKK), difference in unemployment rate in Slovakia and the Czech Republic in 1998-2006 (in %), the dynamics of hourly minimum wage in the Czech Republic in 1998-2006 (in CZK) and the dynamics of the social welfare benefits in the Czech Republic in 1998-2006 (in thousands of CZK). Illegal migrations of Slovak Roma have not been considered in the analysis, as far as the precise data on Slovak Roma migrations do not exist. Keeping such records has been considered unethical (or discriminatory). Two main approximations needed to be done in order to proceed with the analysis: i) All asylum-seekers from Slovakia in the Czech Republic were considered as those of Roma origin ii) Asylum migrations were considered to be the best obtainable approximation of the real migration flows (otherwise unquantifiable). A number of supporting arguments can be used in order to defend both approximations:

134 Ad i) The main reason for this approximation is the nature of the asylum-seeking procedure itself. According to it, asylum-seekers are to stay inside special refugee camps where they are provided shelter, food and small money allowances. This special regime does not allow them to travel freely, neither engage in any forms of long-term paid employment (until the year 2005) and no employment at all (after the year 2005, due to the changes in the Law of Foreigners of the Czech Republic). Such a regime might be advantageous mainly for individuals who are unlikely to get a job in the target country of migration (i.e. the Roma). Another reason is that asylum migration as such (applying for political asylum) does not present a reasonable economic option for Slovak citizens of non-Roma origin. Since the split-up of the Czechoslovakia the customs union had existed, which enabled Slovaks to enter the Czech Republic without any obstacles and engage in any kind of employment without restrictions. The majority of Slovaks are accepting low-paid jobs (in gastronomy or services) no Czechs are willing to take. With regard to this, it can be noted that Roma have difficulties with getting even low-paid jobs (the issue of employment discrimination elaborated on earlier). That is why asylum has more sense for them in economic terms. Besides, it has been noted above (see IOM, 2003) that Slovak Roma are likely to engage in this kind of migration in order to make some cash and take a break from difficult social conditions at home (IOM, 2003). Ad ii) Asylum migrations resemble regular migrations (people leave the source country for the target country, immigrant has to bear different costs and externalities of migration and benefits from immigrants’ networks and social ties). In fact, it can be described as the sub-section of “normal” migration, the flow is the same, just the means that are used are different. As it has been mentioned previously, asylum was one of the channels of temporary migrations of Slovak Roma. The absence of legal (open borders, labor markets) and cultural (similarity between Slovak and Czech languages, heritage of former Czechoslovak federation) combined with social benefits made it very profitable for Slovak Roma citizens to apply for asylum in the Czech Republic. This solution was of especial relevance for those who did not have any family or relatives already living in the Czech Republic or could not count of them providing food and shelter: many Slovak Roma simply abused the asylum system for free food and accommodation reinforced by illegal unemployment and sometimes a chance to escape the “vicious circle” of debts, alcoholism and unemployment in Slovakia. In addition to that, another interesting thing occurs when looking at the patterns of Slovak Roma asylum migrations (Chart 4.1): it seems that Slovak Roma asylum migrations follow “seasonal” nature: the main peaks of asylum migration are from May to September. It might be that this is due to the fact that many asylum-seekers perceive the whole process as some “paid holiday” and want to enjoy it full scale which can be better achieved during the summer months. Besides, there are more

135 opportunities to find some form of illegal job in summer as a contribution to asylum benefits paid by the Czech state (around 15 CZK (0.5 EUR) per person per day): from constructions to farming, gardening and petty crimes. However, the impact of seasoning on migration deserves to be studied further, which, unfortunately, overrides the scope of this paper.

5.5.1. Summary statistics for the main group of asylum-seekers in the Czech Republic.

The Czech Republic appears to be quite an attractive destination for the asylum-seekers from both the neighboring countries and countries being more distant. Relatively high standard of living, health and social security together with the ubiquitous heritage of socialism and cheap goods, housing and labor (in comparison to the West) attract more and more immigrants every year. Chart 5.2 provides a graphical overview of the main seven target countries and their citizens applying for asylum in the Czech Republic in the period of 1999-2006.

Chart 5.2: Asylum-seekers (by selected nations) in the Czech Republic (1999-2006)

500 450 400 350 Slovakia 300 Ukraine 250 Moldavia 200 Bulgaria 150 Vietnam numberof persons 100 50 0

9 9 9 0 0 0 1 1 1 2 2 2 3 3 3 4 4 4 5 5 5 6 6 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 /1 /1 /1 /2 /2 /2 /2 /2 /2 /2 /2 /2 /2 /2 /2 /2 /2 /2 /2 /2 /2 /2 /2 /1 /1 /1 /1 /1 /1 /1 /1 /1 /1 /1 /1 /1 /1 /1 /1 /1 /1 /1 /1 /1 /1 /1 1 5 9 1 5 9 1 5 9 1 5 9 1 5 9 1 5 9 1 5 9 1 5

time periods

Source: Czech Statistical Office, 2005 and Ministry of Internal affairs of the Czech Republic, 2006

In order to summarize the main findings on the asylum-seekers a basic summary statistics was generated on the most represented groups of immigrants. Table 5.3 below describes migration flows to the Czech Republic in the period of 1998 to 2006.

136 Table 5.3: Summary statistics for the asylum-seeking immigrants in the Czech Republic (1999-2006), no. of observations = 97 Asylum-seekers Mean Std. Dev. Min. Max. by country Slovakia (Roma) 47.75309 57.28864 0 307 Ukraine 144.9877 115.7125 0 464 Moldavia 54.12346 74.09173 1 269 Yugoslavia 11.97531 24.99949 0 140 Bulgaria 10.58025 9.303043 0 39 Belarus 20.79012 12.69421 0 52 Vietnam 51.16049 44.5616 0 193 Source: Own estimations

As far as we can see, in spite of Slovak asylum migrations (i.e. those by Slovak Roma) cannot be distinguished neither by the highest mean or standard deviation. Ukrainian, Moldavian and Vietnamese immigrants seem to have much more higher potentials for migration. However, Roma asylum-seekers rate on the forth place – a very alarming issue with respect to Czech-Slovak relations. Unlike Ukrainians or Moldavians, originating from countries devastated by economic transition, Czech Republic and Slovakia can boast EU and NATO membership and relatively high standards of living. High number of asylum-seekers shows that there is likely something wrong in the process of leveling of economic and social conditions in both countries.

5.6. Empirical model

In this section of the paper an econometric model estimating the dependence of Slovak Roma asylum migration on economic factors will be drawn. Generally, when migration studies attempt to determine the main factors of migration, they equate the rate of migration (annually, quarterly or monthly) with various economic variables (such as GDP per capita or per capita real personal income) that are used for capturing the economic incentives for migration, and socio-demographic variables (such as the size or the rate of change of population in the source or target country of migration) that are used to capture the size and the dynamics of the labour markets (see for example Walsh, 1974; Helliwell, 1997; Strielkowski and O’Donoghue, 2006). In this paper, selected “push” and “pull” factors of quantifiable nature that are defined by the research report of the International Organization for Migration (IOM, 2003) are used to investigate the impact of

137 economic and social-economic variables on the dynamics of asylum migrations. Those factors can be supported by empirical data that is available and easy to retrieve (see table 5.2) in section 5.4 of this essay. Therefore, the model investigates the dependence of asylum migrations of Slovak Roma to the Czech Republics in 1998-2006 on “push” and “pull” factors. The country of emigration is Slovak Republic and the country of immigration in the Czech Republic. The model is presented in the linear form (it has been checked for linearity 16 , heteroscedasticity and multicollinearity). The equation representing the “full” model can be written as the following:

= β + β + β + = AM i 0 1Y(ATT SK ,USK , BSK )PUSH 2Y(Udiff , MW CZ , BCZ )PULL ui , i 2,1 ,.., n (5.1) where AM is the number of asylum-seekers from Slovakia in the Czech Republic in 1998-2006 (approximation of Roma asylum migratio due to the reasons described in sub-section 5.5),

Y(ATT SK ,USK , BSK )PUSH is the cumulative of “push” factors which consist of a number of racially motivated attacks and Slovakia in 1998-2006 (ATT SK ), number of unemployed in Slovakia (U SK ) and dynamics of social welfare benefits in Slovakia (B SK ). Y(Udiff , MW CZ , BCZ )PULL is a cumulative of “pull” factors of migration which consist of the difference in unemployment rates in Czech Republic and

Slovak Republic (U Diff ), dynamics of hourly minimal wage in the Czech Republic (MW CZ ) and dynamics of social welfare benefits in the Czech Republic (B CZ ). The model has been estimated using STATA® software. The method of estimation is ordinary least squares (OLS) with robust standard errors (in order to control for heteroscedasticity which was detected during the data analysis). Multicollinearity testing was also performed. Using different samples (randomly dividing the sample of observations in two) leaves the coefficients stable. Apart from that, different specifications of a model using the same data were run without recording big shifts. Finally, correlations between pairs of estimated coefficients do not indicated possible collinearity problem. The results of the estimation are presented in the table 5.4.

16 Testing of the model for linearity has been done similar to the testing in previous two chapters: by checking the relationships of dependent and independent variables (plotting them into a graph) and by plotting a graph of residuals of independent variables.

138 Table 5.4: Determinans (push and pull factors) influencing Slovak Roma asylum migration to the Czech Republic (1998-2006); Numbers of Slovak asylum-seekers (AM), „full“ model „Push“ factors AM „Pull“ factors AM Std. errors Std. errors Number of racially- 5.171019*** Difference between the -.8974955 motivated attacks in (1.652574) unemployment rate in the Czech (1.434491)

Slovakia, ATT SK Republic and Slovakia, UDiff 1.323265*** Dynamics of hourly minimal -14.15545*** Unemployment in Slovakia (.5108181) wage in the Czech Republic (in (4.0775) (thousands of people), USK CZK), MW CZ Dynamics of social welfare .0000147 Dynamics of social welfare -.0000367*** in Slovakia (in thousands (.0000388) benefits in the Czech Republic (.0000101)

of SKK), BSK (in thousands of CZK), BCZ .

Constant 547.6878 (637.9694) N 34 R-squared 0.88 Source: own estimations. Note: Robust standard errors in brackets; * significant at 10%; ** significant at 5%; *** significant at 1%

There are four significant factors out of six „push“ and „pull“ independent variables outlined in table 5.2 in section 5.4. These are: number of racially motivated attacks and Slovakia in 1998-2006, number of unemployed in Slovakia in 1998-2006, dynamics of hourly minimal wage in the Czech Republic and dynamics of social benefits in the Czech Republic (all of them on 1% significance level). The remaining two variables (dynamics of social welfare in Slovakia and the difference between the unemployment rate in the Czech Republic and Slovakia) have no influence and are insignificant. Due to the fact that two variables of the model outlined in (5.1) seem to be insignificant and therefore having no predictory power in our model, it has been decided to check their significance using the backwards stepwise regression with robust standard errors. This approach provides grounds for a decision whether to leave these variables in the model or to compute a modified version of the model from (5.1). The results of stepwise regression are reported in table 5.4.1.

139 Table 5.4.1: Determinans of Slovak Roma asylum migration to the Czech Republic (1998-2006); Numbers of Slovak asylum-seekers (AM), backwards stepwise regression AM Std. errors Number of racially-motivated attacks 5.056864*** in Slovakia, ATT SK [.8779492] Unemployment in Slovakia 1.13512*** (thousands of people), U SK [.3205591] Dynamics of social welfare benefits -.0000327** in the Czech Republic (in thousands [8.34e-06] of CZK), B CZ . Dynamics of hourly minimal wage in -13.4668*** the Czech Republic (in CZK), MW CZ [4.171691] 732.0076 ** Constant [265.0833] N 34 R-squared 88 Source: own estimations. Note: Robust standard errors in brackets; * significant at 10%; ** significant at 5%; *** significant at 1%

It appears that all four significant independent variables from the “full” model (number of racially motivated attacks and Slovakia in 1998-2006, number of unemployed in Slovakia in 1998-2006, dynamics of hourly minimal wage in the Czech Republic and dynamics of social benefits in the Czech Republic) remain significant. In addition, the R-squared remains high and, in addition, the constant term becomes significant too (at the 5% interval). The results of stepwise regression give the grounds to re-run the “full” model specified in (5.1) in the following “adjusted” form:

= β +β +β + = AM i 0 1Y(ATT SK , USK ) PUSH 2 Y(MW CZ , BCZ ) PULL u i , i 2,1 ,.., n (5.2) where AM is the number of asylum-seekers from Slovakia in the Czech Republic in 1998-2006,

Y(ATT SK ,USK )PUSH is the cumulative of “push” factors which consist of a number of racially motivated attacks and Slovakia in 1998-2006 (ATT SK ) and the number of unemployed in Slovakia ( USK ).

Y(MW CZ ,BCZ )PULL is a cumulative of “pull” factors of migration which consist of the dynamics of hourly minimal wage in the Czech Republic (MW CZ ) and dynamics of social welfare benefits in the

Czech Republic (B CZ ).

140 This model can be computed using the OLS with robust standard errors. The results of the estimation are presented in table 5.4.2. What is to be noted are the highly-significant coefficients of the first two variables (both belong to „push“ factors of migration). Especially the number of racially motivated takes in Slovakia has a considerable influence on the increase of Slovakia Roma asylum migrations. The influence of the variable representing the unemployment in Slovakia on Roma asylum migrations has been expected, therefore the sign of its coefficient and it significance have been anticipated.

Table 5.4.2: Determinans (push and pull factors) influencing Slovak Roma asylum migration to the Czech Republic (1998-2006); Number of Slovak asylum-seekers (AM), „adjusted“ model „Push“ factors AM „Pull“ factors AM Std. errors Std. errors Number of racially- 5.056864*** Dynamics of hourly minimal -13.4668*** motivated attacks in [.8779492] wage in the Czech Republic (in [4.171691]

Slovakia, ATT SK CZK), MW CZ 1.13512*** Dynamics of social welfare -.0000327*** Unemployment in Slovakia [.3205591] benefits in the Czech Republic [8.34e-06] (thousands of people), U SK (in thousands of CZK), B CZ .

Constant 732.0076** [265.0833] N 34 R-squared 0.89 Source: own estimations. Note: Robust standard errors in brackets; * significant at 10%; ** significant at 5%; *** significant at 1%

The remaining two „pull“ factors - dynamics of hourly minimal wage in the Czech Republic and dynamics of social welfare benefits in the Czech Republic - are also highly significant, however they have a negative influence on Slovak Roma asylum migrations. This is quite surprising as far as it means that higher incentives in the country of immigration (Czech Republic) actually reverse the asylum migration. This can be explained using two main arguments: First, due to impossibility to undertake a full-time jobs for asylum migrants, they can only engage in part-time employment. The improvement of labor markets situation (higher hourly wages) in the Czech Republic might lead to the increasing competition amongst different social groups (workers from Ukraine, North Korea, Slovakia

141 or Poland) which does not motivate Slovak Roma to undergo the asylum-seeking procedure in the Czech Republic (they know that it would be hard to find a part-time jobs and make additional money). Second, the improvement of economic and social situation in the Czech Republic might make some of the Slovak Roma to consider other forms of migration (temporary labor migration or moving to re-unite with distant relatives using Roma social networks). Asylum migration might not longer seem interesting for them if there is a possibility to make themselves better off using other channels of migration. The coefficient of determination (R-squared) is quite high (0.89) which means that the great deal of variances in the model is explained. It means that the results of the model can be used can be used for estimating the relationship between the asylum migration of Slovak Roma and “push” and “pull” factors of migrations (the signs of the regression coefficients are important here). The overall result can be written in the form of the following equation:

__ = + + − − + AM i 732 .007 .5 056 *ATT SK .1 135 * U SK 13 .466 *MW CZ .0000327 *BCZ u i (5.3)

The important conclusion that comes from the empirical model presented above is that the only satisfying explanation of Roma asylum migrations from Slovakia to the Czech Republic is so-called “push” factors of social and economic nature (increase in unemployment and worsening social situation at home). This can be clearly seen from the results of the model estimation: the most important determinant of Roma asylum migrations appears to be an increase in the number of racially- motivated attacks in the source country. Although there are clearly other incentives that might influnce the motivation of Slovak Roma asylum seekers. Amongst most relevant ones that come to mind are for instance looking at the whole process of asylum in the Czech Republic as at some form of paid holiday or the wish of some Roma to turn asylum migrations into profitable business (if rejected in one country, they briskly move to another one and apply for asylum there; the Czech Republic is thus a „transfer country“). However, the most reliable explanation of Roma migrations is that thanks to their ethnical specifics Roma are very inclined to migrating if social and economic situation in Slovakia is worsening. For instance, it seems that in case the number of racially-motivated attacks in Slovakia is up, local Roma are prepared to escape to the countries with similar level of life and economic and social conditions. Thus, the situation similar to the one highlighted by the media and widely discussed by the general public in 2000, when a large number of Slovak Roma came pouring through the border to the Czech Republic can, in theory, be easily repeated.

142 5.7. Conclusions

It is quite clear that different ethnic groups have a different approaches to migration. An example of Slovak Roma asylum-seekers in the Czech Republic makes it clear that economic incentives (and „pull“ factors) account for just a small proportion of reasons for migration of various ethnic groups (including asylum migrations). „Push“ factors (worsening of economic and social conditions in the source country) seem to explain Roma migrations better. Notwithstanding, Roma propensity to migration is pre-determined by quite a large number of factors. First of all, Roma willingness to change the place of residence stems not only from their ethnic background or cultural specifics but rather from the historic necessity: in former Austrian-Hungarian emprise and later Czechoslovakia Roma minority has always been persecuted and outlawed; changing the place of residence became an integral part of survival. Second, it might be that the Roma are tempted to solve all the problems merely by moving to another destination – this can provide an explanation of their “nomadic” behaviour. Roma migrations might have very different grounds: alongside with economic they can be social, cultural, etc. We also can distinguish long-term or short-term migrations. What is the common point in the Czechoslovak context is that in many cases no one can be sure that we are dealing with migrations: quite often changing the place of residence is merely a “visit” to the relatives or members of the allied clan which can vary from several months to several years. Roma networks and families that preserve contacts and help each other make any estimates quite difficult. However, what is important here is the fact that Roma seem to be very sensitive to the development of social situation in their “home country” (however strange this term would seem in the case of Roma). In case the social and economic situation is worsening and Roma feel threatened by this development, their immediate response would be emigration, even if this emigration involves undertaking an unpleasant procedure of asylum migrations (staying in asylum camp, going through humiliating interviews and screening process).

5.8. References

Baker, M. and Benjamin, D. (1994), “The performance of Immigrants in the Canadian Labour Market”, Journal of Labour Economics 12, 369-405.
Barany, Z. (1992), “Democratic changes bring mixed blessings for Gypsies”, RFE/RL Research Report, Vol.1, No. 20, pp. 40-47.
Bogue, D. (1959), “Internal Migration” in “Study of population”, University of Chicago Press
Borjas, G. (1994), “The Economics of Immigration”, Journal of Economic Literature 32, 1667-717

143
Borjas, G. (1999), “Knocking on heaven’s door”, Princeton University Press.
Borjas, G. (1999), ”Immigration and Welfare Magnets”, Journal of Labour Economics 17(4): 607-637
Bruszt, L., Stark, D. (1992), “Remarking the Political field in Hungary: from the Politics of Confrontation to the Politics of Competition”, in “Eastern Europe in Revolution”, Banac, I. (ed.), Ithaca, NY: Cornell University Press.
Český statistický ú řad. Statistické ro čenky České republiky (1993-2005). Praha (www.czso.cz).
Crowe, D. (1995), “A history of Gypsies of Eastern Europe and Russia”, I.B. Taurus Publishers, London.
CTK (Czech News Agency), 27 th of October 1997 report (www.ctk.cz).
Davidova, E. (1970), “The Gypsies in Czechoslovakia. Part I: Main Characteristics and Brief Historical Development”, translated by D.E. Guy. Journal of the Gypsy Lore Society, Third Series. Vol. 69, no. 3- 4.
Davidova, E. (1995), “Cesty Rom ůu”, Vydavatelstvíi Univerzity Palackéeho, Olomouc
European Monitoring Centre on Racism and Xenophobia (EUMC), EUMC annual report 2005 (http://eumc.eu.int ).
Fidrmuc, J. (2001), “Disintegration and Trade”, LICOS discussion paper.
Gramlich, E. M. et al. (1984), "Migration and Income Redistribution Responsibilities". Journal of Human Resources 19(4): 489-511
Head, Keith (2003), “Gravity for Beginners”, University of British Columbia, Canada.
Helliwell, J.F. (1997), “National borders, trade and migration” NBER Working paper No. 6027.
Horakova, M. (2005), “Labor migration trends in the Czech Republic during the last decade”, RILSA, Praha.
Horvathova, E. (1962), “Cigani Cigáni na Slovensku”, Bratislava, Vydavetel’stvo Slovenskej Akademie veed.
Hungarian Central Statistical Office, www.nepszamlalas.hu , Census 2001.
Infostat – Institute of Informatics and Statistics (2002): “Projection of Roma population in Slovakia until 2025”, Bratislava, Akty
INSSE (Institutul National de Statistica), www.insse.ro , Census of population and dwellings, 2002
International Organization for Migration (2003), “Zprava o analyze soudobe migrace a usazovani prislusniku romskych komunit ze Slovenske republiky na uzemi Ceske republiky”.
Ionescu, D. (1991), “Violence against Gypsies escalates”, Report on Eastern Europe, June.
Ionescu, D. (1992), “Romania’s First Post Communist Census”, RFE/RL Research Report, March, pp. 55-59.
King, D. (1991) “Romania”, in Starr and Gregory (eds.) Yearbook 1991, Helsinki Watch report, pp. 336-339.
Korbrin, F. and Goldscheider, C. (1978), “The Ethnic factor in family structure and mobility“, Cambridge MA: Ballinger Publishing Co.
Lee, E. (1966), “A theory of migration”, Demography 3(1): 47-57
Liebich, A. (1992), “Minorities in Eastern Europe: Obstacles to a reliable count”, RFE/RL Research Report, Vol. 1, No. 20, p.38.
Markos, E. (1985), “Dim prosperity for improving the Plight of the Gypsies”, Radio Free Europe research, Hungary, no. 5, pp. 13-16.
McCallum, John (1995): “National Borders Matter: Canada-U.S. Trade Patterns” American Economic Review, Vol.85, No.3.
Meyer, B. D. (2000): "Do the Poor Move to Receive Higher Welfare Benefits?" WP no. 58. Northwestern University, Joint Center for Poverty Research Working Paper.
Migration Watch UK (2005), “The Roma case in the House of Lords”, Briefing paper 8.6 legal (www.migrationwatchuk.org ).
Ministry of Interior of the Czech Republic (2004), “Zprava o situaci v oblasti migrace na uzemi CR za rok 2003”, Praha
Ministry of Interior of the Czech Republic, www.mvcr.cz
Ne čas, C. and Miklušáková, M. (2002): “Historie Rom ů na území České republiky”, Cesky Rozhlas, retrieved from http://romove.radio.cz/cz/clanek/18785
Orbman, J. (1991), “Crime Rate among the young”, Radio Free Europe Research, Vol. 12. no. 32

144
OSCE/ODIHR (2000), “The report on the Reasons of the Migration of the Roma in the Slovak Republic”, Vienna.
Petrovic, D. (1976), “Tsigani u srednjovekom Dubrovniku”, in Zbornik filosovskog fakulteta, ed. Garashanin, M., Belgrade, Naucno Delo.
Puxon, G. (1975), “Rom: Europe’s Gypsies”, London: Minority group
Ritchey, P. (1976), “Explanations of migration”, Annual review of sociology, no.2, 363-404
Shaw, R. (1975), “Migration theory and Fact”, Philadelphia, Regional Science Institute.
Srb, V. (1985), “Zmeny v reprodukci ceskoslovenckych Romu”, Demografie, 30, no. 4, 305-309.
Štatistický úrad Slovenskej republiky (2000), „Štatistická sprava o základních vývojových tendenciách v hospodarstve SR“, č.1, Bratislava.
Štatistický úrad Slovenskej republiky (200 ř), „Vyvoj obyvatelstva v SR“, Bratislava.
Strielkowski, W. and O’Donoghue, C. (2006), “Ready to go? EU Enlargement and Migration potential: lessons for the Czech Republic in the context of Irish migration experience”, Prague Economic Papers, vol.1, 14-28.
Tolerance Foundation (1994a), “Report on Czech Citizenship Law” (Prague: Tolerance Foundation), www.ecn.cz .
Tolerance Foundation (1994b), “Analysis of 99 cases” (Prague: Tolerance Foundation), www.ecn.cz.
Trovato, F. and Halli, S.S. (1983), “Ethnicity and Migration in Canada”, International migration review, Vol. 17. No. 2
UN Common database/World Bank, 2005.
UUl č, O. (1990), “Communist National Minority Policy: The case of the Gypsies in Czechoslovakia”, Soviet Studies, Vol. 2, no. 4.
Vano, B. (2002), “The Demographic Characteristics of Roma Population in Slovakia”, Infostat, Bratislava
Vano, B., Haviarova, E. (2002), “Demografickée trendy romskej polulacie” in “Uhrnáa sprava o Romoch na Slovensku”, Institut pro verejnée otáazky (Infostat), Bratislava.
Walsh, B.M (1974), “Expectations, Information and Human Migration: Specifying an Econometric Model of Irish migration to Britain”, Journal of Regional Science, vol. 14, 107-120.

145

APPENDICES

146

1) READY TO GO? EU ENLARGEMENT AND MIGRATION POTENTIAL: LESSONS FOR THE CZECH REPUBLIC IN THE CONTEXT OF IRISH MIGRATION EXPERIENCE, Prague Economic Papers 1/2006 ( http://www.vse.cz/pep/abstrakt.php3?IDcl=273 )

2) A LIVING WORTH LEAVING? ECONOMIC INCENTIVES AND MIGRATION FLOWS: THE CASE OF CZECHOSLOVAK LABOUR MIGRATION, Prague Economic Papers 3/2007 (http://www.vse.cz/pep/abstrakt.php3?IDcl=308 )

147