Eastern Migrations vs Western Welfare States-( Un ) Biased Fears

This inquiry considers some effects of migration on the labour markets and the welfare systems found in the EU-15, and from the perspectives of sustainability of the current welfare state regimes. Our inquiry aims to determine whether and to what extent different approaches in regulation of migration flows between the new and old member states are compatible with related economic and demographic findings. Within this context, our research considers regulations affecting migration flows. Our findings suggest that some effects of migration from the EU8+2 on the labour markets and social protection systems found in the EU-15, both with respect to level and structure, do indeed generate effects on migration, especially considering whether migration is based upon economic or welfare decisions. In addition, our inquiry considers perspectives upon restrictive versus liberal migration policies.

With the inclusion of the new members from Central, Eastern, and South-Eastern Europe, known as the EU-8+2 countries, indeed, the European Union (EU) has lost its former economic and social homogeneity. Nevertheless, the EU has more background in institutional and experiential terms. This suggests the EU can better face challenges associated with its enlargement that can be channelled towards solving internal problem, its demographic deficit and the lack of flexibility in the labor markets emerging as a pressing problem. In short, the removing of borders for workers from new member states is not only an economic issue but also has strong political, cultural and sociological dimensions that often blur the individual and collective rationality of preferences towards the types of immigration policies.
Data on migration flows from the New Member States (NMS) to the EU-15 suggest that the fears regarding the great influx of workers from the Eastern regions, after the EU 2004 and 2007 enlargements were indeed exaggerated. Although the share of workers from the NMS in the working age population of the EU-15 doubled, compared to the pre-enlargement period, the mobility of workers from the NMS has not yet reached the level of mobility of workers from the EU-15, and it is far behind the inflow of migrants from third (extra EU) countries into the EU-15. PANOECONOMICUS, 2013, 3, Special Issue, pp. 323-345 However, it remains important to consider that the NMS do not represent a homogenous group with respect to generating migration flows to the EU-15. Effects of enlargement processes on migration flows is most felt by the poorest member states: namely, Romania and Bulgaria, which, to some extent, indicates that fears of a social dumping phenomena between the poorest and richer member states, although these were exaggerated, were not completely unfounded.
In our analysis, the welfare state serves as both the starting and end points, especially when considering the desirability of restrictive or liberal migration policies. This inquiry attempts to answer a few questions. For one, do generous welfare policies of the highest income EU states serve as a magnet for workers from poor eastern countries? If so, what is the net impact on the earnings and the social spending of domestic workers? Then, is it possible to reconcile the free movement of labor and existing regimes of welfare states?

Review of the Selected Viewpoints in the Literature
Some dimensions of these questions have been considered in the research that examines links between migration and welfare states, and the structural nature of the current crisis of public finance most European economies face (see, for example, Kosta Josifidis, Alpar Lošonc, and Novica Supić 2010). However, links between migration and welfare states has become even more pressing in recent years, especially in light of the current crisis, what is also known as the "Great Recession" that got started in the Autumn of 2007. Along with this crisis, we also need to consider the levels of generosity and universality of the European welfare-state model. As a result, the literature in this field is becoming more and more diverse with the inclusion of various aspects of migration as well as when considering several aspects of the welfare state.
Migration represents a complex phenomenon that can be studied by relying upon selected theoretical frameworks. Within this context, Lucia Kureková (2010) provides a systematic and critical review of key migration theories applicable for the understanding of the movement of workers between the old and new EU member states. By using as examples the EU-8 countries, on the one hand, and Great Britain and Ireland, on the other, Kureková (2010) points to the inability of neoclassical migration theory to explain the considerable variation of the immigration rates of workers from countries with similar levels of living standards and wages. This challenge gives rise to the need for the development of new approaches for the study of migrations from the countries of Central, Eastern South-Eastern Europe into the higher per capita income countries found in the EU of Western Europe.
Migrations from the east to western welfare states are not uniform, and this is stressed in the contribution of Adrian Favell (2008). The emergence of the "new migration system in Europe" could be characterized as a temporary liberalization of movement that is related to the introduction of an informal labour market, giving rise to a new culture of immigration and transnational networks. Through studying the demographic and socio-economic background of the immigrant population in the EU, research of Luca Barbone, Bontch-Osmolovsky Misha, and Zaidi Salman (2009) suggest that migration flows tend toward an increased concentration in both low and high skilled workers. Based upon data derived from their European Union Survey of PANOECONOMICUS, 2013, 3, Special Issue, pp. 323-345 Income and Living Conditions (2006) the authors found no evidence supporting the assumption that immigrant workers pay fewer taxes and gain more benefits from welfare state. And, not all migrants generate the same impacts on the welfare state. Thus, Tito Boeri (2009), starting with the immigration towards the simple model of expected fiscal effects, and on the basis of EU-SILC data set, found no evidence that the legal migrants, especially educated migrants, are the net beneficiaries of government transfers. However, there is evidence of the so-called "residual dependency" of migrants from non-contributory transfer which is more evident in more generous welfare states.
As an addition to the points established in the two papers noted above, we could also add the research of Giacomo De Giorgi and Michele Pellizzari (2009) that considers roles played by welfare transfers on attracting migrants after the EU enlargement. According to the data of the European Community Household Panel (ECHP), these co-authors found that the generosity of the welfare state could indeed serve as act as a magnet for migrants from new member states. Such findings confirm the validity of concerns about the potential dangers that EU enlargement poses to the Western welfare state. At the same time, the same research reveals that the number of migrants attracted by the welfare state transfers is relatively small compared to the impact that migrants have on the labour market circumstances, such as unemployment and wage rates.
Fears of welfare-motivated migrations from the NMS into the EU-15 appear to be over-emphasized. This is supported by the research of Christian Dustmann, Tommaso Frattini, and Caroline Halls (2010). By focusing on Britain, the authors present evidence on the lack of welfare migration of workers from EU-8 to the UK after the EU's enlargement. Migration from the EU-8 had a positive effect on public finances, despite the fact that in the last few years, the United Kingdom was facing a budget deficit. The explanation for such a situation should primarily be sought in the high rate of participation of immigrants in the labour force, proportionally larger immigrant payments towards the state through indirect taxes, as well as in the significantly lower use of welfare benefits and public services. By relying upon a VEC model, research findings of Joan Muysken and Thomas Ziesemer (2011), have shown that the Netherlands (from 1973 to 2009) offers an example of the positive impacts of migration flows on the age structure of this country over time. Their findings suggest that immigrants were able to achieve at least the same rate of labour force participation as the domestic population. The contribution to the economic growth depends on the education level of immigrants, and especially their willingness to attain yet higher levels of education.

Institutions and Bounded Rationality
The theoretical framework developed in this paper represents an attempt of expansion (restriction) of the neoclassical explanation of migrations with the institutional rigidities of the welfare state that includes the idea of bounded rationality. PANOECONOMICUS, 2013, 3, Special Issue, pp. 323-345 Institutional rigidity of the welfare state. The labour market is not observed as an autonomous mechanism that determines the size and dynamics of migrations. The welfare state, in addition to the labour market, also serves as a magnet for migrants. The specificity of the welfare state in relation to the labour market is reflected in the specific form of institutionalization of social consensus based upon the inert compromise between economic efficiency and social equity. Corporate capital accepts the necessity of a social justice while labour unions formulate their demands, taking into account impacts on economic efficiency. The institutionalization of arrangements is performed by the state through various social programmes starting from the active labour market measures to the design of the pension systems.
The institutionalization of relationships in the welfare state has a more permanent character in comparison to the institutionalization that exists in the labour market. Consequently, the changes in economy are more quickly reflected on the labour market than in the welfare state which has a significant impact on migration flows. The issue is the time-unsynchronized signals that the labour market and welfare state are sending to migrants. Generous social spending can attract migrants in the circumstances of the fall of the demand for labour. Similarly, the increase in the demand for labour increases the number of migrant workers who may be faced with a system of social protection that is not sufficiently comprehensive/generous. Bounded rationality. In general, bounded rationality is interpreted as decisionmaking under uncertainty. The shaping of migration flows, whether it concerns immigrant workers who decide to migrate or domestic workers in the role of the median voter making a choice of immigration policy, is characterized by a high degree of uncertainty arising from the socio-economic complexity of the migration process. The ability of rational judgment is not the same for domestic workers and migrant workers. When compared to domestic workers, migrant workers on average make more rational decisions. How do we explain this?
The choice of migration policy is a collective decision while the choice of the country of immigration is an individual decision. In the absence of knowledge about the future, individuals rely on their imagination and expectations (see, for example, Philip Arestis 1996). In the process of articulating the collective attitude the individuals harmonize their expectations with beliefs of others within the contexts of an unforeseeable future. The key group landmark is the previous experience with migration.
With domestic workers, positive experiences form preferences toward liberal policies on migration and vice versa, negative experiences -restrictive rules are chosen. As a result, the standpoint on migration is based on experiential rather than the anticipative rationality which would include the evaluation of migration potentials regarding the current and anticipated social problems.

(Ir)rational Immigrants
From an economic standpoint, the migrations of workers are primarily motivated by the differences in the level of prosperity in their origin and the destination country. The improvement of the economic status of immigrant workers can be achieved in PANOECONOMICUS, 2013, 3, Special Issue, pp. 323-345 two ways: (i) by employment and (ii) by using the welfare state programmes, independent of their employment status -the non-contributory scheme.
Among countries, there are significant differences in the level and structure of the non contributory schemes. In general, these differences are greater than the differences in earnings and employment. While earnings and employment are primarily the consequence of economic factors, the non contributory schemes, apart from economic factors, are affected by a number of non-economic factors, above all, the nature of the socio-political system. The attraction of non contributory schemes for immigrants depends on the generosity of these programmes and conditions that immigrants must meet in order to use them. The countries with more generous social spending, as a rule, have more generous and accessible non contributory schemes and, conversely, the countries with lower level of social spending are characterized by low generosity and accessibility of non contributory schemes. It follows that the difference in the generosity of welfare states in the destination and the source countries can affect level, dynamics, as well as, structure of migration flows.
Contributory social programs are excluded as motives for migration. Redistribution, which is achieved through contributory social programs does not change the economic position of workers gained through employment status. Moreover, migrant employees included in the welfare system contribute to the welfare state to a greater extent through taxes and contributions, even more than their feedback from the welfare program.
Starting from the given typology of the economic motives of immigrants, there are two main types of migrations: (i) economic migration and (ii) welfare migration. The first type of migration is determined by the differences in earnings, employment rates and working conditions between the source and destination countries. The second type -welfare migrations depend on the differences in the generosity of social programs, levels and trends of economic inequalities and poverty in the start to the destination country.
Earnings, employment and working conditions, on one hand, and social spending, inequality and poverty on the other hand, are sufficient to describe the average economic in comparison to the welfare immigrant. Economic migrants are mostly young, educated, unmarried workers with the prospects growth in earnings in the future. Welfare migrants are characterized by low levels of human capital and productivity, numerous household with very low income growth perspective. For the source country, economic migrations have a negative sign whereas welfare migrations have a positive sign. The situation is opposite in the destination countries, economic migrations contribute to the economy while the welfare migrations have harmful effects. Economic migrants are better integrated into the labour market and can have a positive impact on the social security system as they are less likely to be a burden on the national social security system and receive social transfers (Peter Huber 2011).

(Ir)rational Immigration Regimes
The structure of immigrants depends on the degree of labour market regulation in destination country and difference in the generosity of welfare states in destination PANOECONOMICUS, 2013, 3, Special Issue, pp. 323-345 and source country and it is formed by the implementation of immigration policies. There are two basic ways of implementing immigration policy: (i) liberal; (ii) restrictive.
From a theoretical perspective, liberal migration regimes are based on the open doors principle, applicable to all immigrants. The labour market and welfare state are influenced by internal and external factors associated with migrations. In the long run, liberal migration policy should lead to the convergence of various economic and social models following the logic of common market. The downside of liberal migration regime is the potential for spillover of economic and social issues among countries i.e. the emergence of social dumping. Consequently, the liberal regime is recognizable in relation to the countries with similar levels of economic and social development. The key objective of liberal migration policy is to increase the flexibility of labor market and, on that basis, to increase the competitiveness of domestic economy. As a result, migration is viewed in a positive sense and implies the positive effects of migration on destination country.
On the other hand, the restrictive migration policy closes labour market and welfare state to foreign workers. The functioning of labour markets and social protection system is primarily determined by the actions of internal factors. The advantage of this approach is reflected in the prevention of spillover effects of economic and social problems among countries. The implementation of restrictive policies in the long run leads to the creation of rigid structures which can threaten the global competitiveness of domestic economy and the sustainability of welfare state model. Restrictive regimes are applied in the case of significant economic and social disparities among countries. The key objective of restrictive migration policy is the protection of the labor market and the social security system from the disorder which can occur due to unregulated influx of immigrants. Consequently, migration is viewed in a negative context because it is assumed the negative impacts of migration.
In practice, immigration policies do not occur in a purely restrictive or liberal form so we are speaking about the greater or lesser degree of restrictiveness/openness of migration policies. The objective of real immigration policy is the selection of migrants in order to avoid the negative consequences of uncontrolled migration (the theoretical framework of restrictive policies) and achieved the positive effects of greater labor market flexibility (the theoretical framework of liberal policies).
Since the immigration policy is part of the public rather than the economic policy, the choice of the immigration regime depends on the perception of the effects of migration flows from the viewpoint of the interests of labour and capital. Domestic workers, as the medial voters, form preferences on the basis of beliefs or expectations (though irrational) about the impact of immigration on the destination economy, especially earnings, employment and financing of welfare state. The corporate capital is lobbying towards the immigration policy that provides greater flexibility of the labour market, i.e. the one leading to the reduction in earnings and productivity growth. Consequently, the immigration policy is the reflection of the social consensus of labour and the corporate capital in given institutional frameworks.   Figure 1 shows a hypothetical link between immigration regimes and welfare state sustainability. Liberal migration policies can hinder as well as support the sustainability of the welfare state. If the free movement of workers is accompanied by a larger influx of welfare immigrants rather than economic immigrants the sustainability of social programs is under threat. The opposite is true when the motives of immigrants are to a greater extent of economic nature rather than the welfare nature. Restrictive migration policies close the welfare state in the sense of the potential opportunities as well as hazards of the free fluctuation of workers between countries. The demographic pressure increases due to lower immigrant influx, which threatens the sustainability of social spending but simultaneously also reduces the risk of social dumping i.e. spillover from the social into economic problems from the environment.
Based on these assumptions, the two hypotheses are derived: Hypothesis I. The key issue of the welfare state in the EU-15 regarding migrations is the issue of selectivity of the immigration policy towards attracting more economic migrants and not the restrictiveness of the migration policy.
Hypothesis II. The convergence of earnings and social spending of the starting -poorer countries with earnings and social spending with the destination -richer countries, leads to changes in the structure of immigration flows towards the increase participation economic and reducing the share of immigrant welfare.

Convergence -A Landmark of (Ir)rationality
How do domestic workers, corporate capital and the state form preferences according to the types of immigration policy and to what extent are these preferences a reflection of economic rationality? The thesis that is promoted in this paper is that domestic workers as opposed to corporate capital and the state are not anticipatory rational but their preferences are primarily derived from past experiences. Positive experiences speak in favour of liberal immigration regimes while negative experiences are directed towards more restrictive rules. If there is a convergence of earnings and social spending in the country of emigration in relation to earnings and social spending in the last country in which the barriers of free movement of workers were lifted, PANOECONOMICUS, 2013, 3, Special Issue, pp. 323-345 then the beliefs of domestic workers in similar effects of liberalization of migration flows are reinforced. As a result, the liberal migration policy is chosen. In case of divergence, it is the other way around -the restrictive rules are preferred.
Economic theory suggests a number of methods for determining the convergence of economic and social indicators between countries. In this paper we have chosen the Robert J. Barro and Xavier X. Sala-i-Martin (1995) and Sala-i-Martin (1996) approach which distinguish between two types of convergence: (i) sigma convergence and (ii) beta convergence. Sigma convergence exists when the dispersion of earnings/social security spending in the observed group of countries decreases over time. This happens if the following condition is satisfied: where the T t   -is the standard deviation of earnings/social spending in the period t+T; t  is the standard deviation of earnings/social spending in the period t.
Beta convergence is present if there is a negative relationship between the growth rates of earnings/social spending relative to the initial level of earnings/social spending between the countries. In other words, beta convergence is present if the poorer countries, in terms of net earnings/social spending, progress faster than richer countries.
Depending on the choice of statistical measures of sigma convergence, Peter Cornelisse and Kees Goudswaard (2002) distinguish between: (i) absolute convergence/divergence, in which the standard deviation i is used as a criterion and (ii) relative convergence/divergence, which is based on the coefficient of variation. The easiest way to calculate the sigma convergence is the use of standard deviation as a measure of convergence. By using this method it is possible to determine how the dispersion between the level of earnings and social spending alters, i.e. how the differences in earnings and social spending within the group of countries change in relation to average values. The standard deviation has a characteristic i.e. its value increases with the increase in the average value of the data set. In order to take this into account, as a measure of sigma convergence the coefficient of variation is generally used: standard deviation divided by the arithmetic mean of the observed data set (Koen Caminada, Goudswaard, and Vliet van Olaf 2010).
For research needs, we chose to use the relative sigma convergence, which is determined by applying the following formulas: -is the deviation of net earnings/non-contributory in the source country i in relation to the average of the destination countries j; Similarly to the sigma convergence, the beta convergence may also occur in two forms: (i) absolute and (ii) conditional convergence. Usually, the hypothesis about the absolute beta convergence is tested when observing a group of countries with a high degree of homogeneity in terms of institutional characteristics and preferences. Conditional convergence is derived from the assumption that the process of convergence can be observed by excluding some socio-economic characteristics of countries from the analysis. In other words, in models of conditional convergence certain socio-economic variables are treated as constants in order to detect general tendencies of harmonization.
The usefulness of the concept of absolute beta convergence is particularly evident in the circumstances of the economic integration of countries. For research purposes, two models of absolute beta convergence were evaluated: where the symbols signify: α -a constant, -annual growth rate of net earnings/non-contributory social spending in the country i in the period between t and t+T according to the formula The hypothesis that the coefficient β is negative is tested by using the method of ordinary least squares. If the coefficient β is negative (positive) we conclude that there is absolute convergence (divergence) in net earnings/non-contributory social spending across countries. The higher the value of the β coefficient, the faster the gap between richer and poorer countries is reduced, through convergence toward higher levels of earnings/social spending.

Empirical Evidence
Do migration trends in the EU reflect the assumptions set out in a previously given theoretical framework? To answer this question, in the first step, it is necessary to show and explain the situation and the dynamics of earnings, social spending and the rate of migration in the EU-8+2 countries in relation to the EU-15 countries, before and after their joining to the EU. In a second step, it should determine whether and to what extent there is the convergence or divergence in the labor market indicators (shown by differences in net earnings) and the welfare state (expressed by differences in non-contributory social spending), in the EU-8+2 countries compared to the average of the EU-15, essential for decision making about immigration. In this way, it is possible to test the degree of rationality in the formation of preferences towards immigration regimes in the EU-15 countries, particularly, the labor market liberalization during the period of transitional arrangements. Finally, in the third step, exploring the inter-relationship among earnings, social spending and immigration rate, it is possible to recognize the dominant character of the process of migration in the context of economic vs welfare state immigrations.
The Eastern EU enlargements in 2002 and 2007 generated fear of excessive migration of workers from the NMS, who could jeopardize the labor market and the system of social protection in EU-15. Reasons for concerns were based on the existence of significant differences in the amount of net earnings and generosity of social programs. On the eve of the enlargement in 2004, the average net earnings of workers from the NMS, expressed in purchasing power parity, amounted to one third of average net earnings of workers from EU-15. The situation was nearly identical with the non-contributory social programs. In order to protect themselves from potentially negative effects of the inflow of migrant workers from the East, EU-15 countries have adhered to the sequential liberalization of the labour market by applying the socalled transitional arrangements. Table 1 shows the ratio of net earnings in the EU-8+2 to the EU-15 average. The EU-8+2 countries are characterized by lower net earnings than the average in the EU-15. Thus, in 2010, the workers from the NMS on average had 60% lower net earnings compared to the workers in the OMS. However, in the period from 1998-2010, in eight of the ten NMS a trend of narrowing the range of net earnings was reported, most prominently in the Baltic countries (over 10 percentage points) while a slight worsening trend was recorded in Poland and Bulgaria (under 1 percentage points). Regarding the impact of the enlargement on the fluctuation of net earnings, it was positive in all countries, including Poland and Bulgaria. The effect of enlargement on the net earnings was highest in the Baltic countries and Slovakia (over 8 percentage points) and lowest in Hungary (under 2 percentage points). Table 1 clearly shows the existence of significant sigma convergence between NMS. In the period from 1998 to 2010, the standard deviation of the differences in net earnings of the NMS declined by 15% and the coefficient of variation for 33%. The average net earnings in EU-8+2 for the same period increased compared to the average EU-15 by 9%, indicating that there was no downward trend, but quite opposite, the convergence had an upward trend towards higher levels of net earnings.    Table 2 shows the results of β convergence of net earnings in the EU-8+2. Convergence is determined by the regression of the logarithm of the annual growth rate of net earnings in the period from 1998-2010 in relation to the logarithm of net earnings in 1998. The results indicate the existence of β convergence of 3.1% per annum. This means that the earnings gap between the individual countries was reduced when compared to the group earnings of the NMS average for 3.1% annually. Table 3 shows the fluctuation of non-contributory social spending of the NMS towards the average of EU-15. Similarly to the net earnings, the NMS are significantly lagging behind the EU-15 in terms of the non-contributory social spending. Thus, in 2010, socially vulnerable groups in the NMS received on average 60% less welfare when compared to the average allocations for the same purposes in the EU-15. However, the gap between the EU-15 and the NMS narrows over time. Between 1998 and 2010, a drop in the trend compared to the average EU-15 was recorded only in three countries; however it did not exceed 2 percentage points. At the same time, five countries recorded an average growth of non-contributory social programs of over 15%. When the impact of enlargement on the dynamics of non-contributory social programs is concerned, a visible decrease was recorded only in Slovenia -6.2 percentage points.  Notes: NA -not available.The data from the first and second column show the absolute value of non-contributory social spending. According to Eurostat methodology, non-contributory schemes are social protection schemes in which eligibility to benefits is not conditional on the payment of contributions by the protected persons or by other parties on their behalf.

Source: Authors' calculation based on Eurostat database (2012).
The EU-8+2 countries recorded a faster growth in social spending compared with the EU-15 countries. Thus, in the period from 1998 to 2010, the average growth of social spending in the EU-8+2 countries was two times higher than the average growth of social spending in the EU-15 (96.91% : 48.3%). The higher growth of social spending in the EU-8+2 should be interpreted primarily as a result of their low starting point. The nature of this growth could be recognized, to some extent, by the analysis of sigma and beta convergence of social expenditure in the EU-8+2 countries. The data on sigma convergence fluctuation point to the reduction of the gap between the NMS in the area of non-contributory social spending. During the period from 1998-2010 the standard deviation was reduced by 8% while the coefficient of variation decreased by 36%. The 40% growth of the average non-contributory social programs in the same period from suggests that the convergence was directed to-PANOECONOMICUS, 2013, 3, Special Issue, pp. 323-345 wards higher levels of social spending. Regarding the β convergence, the coefficient has the expected sign, which indicates the approaching to NMS, but the value is not statistically significant. The explanation for such a situation should primarily be sought in a short time interval and the unavailability of data for Bulgaria. Table 5 provides information on the rate of immigration of workers from EU-8+2 into EU-15. In 2010 workers from the NMS accounted for 3.3% of the EU-8+2 in the EU-15. This is four times more when compared to the situation in 1998. The impact of enlargement on the rate of immigration is most evident in the poorest countries in the EU: Romania and Bulgaria, where the rate of immigration compared to the situation prior to accession to the EU increased by 4.29 percentage points i.e. 1.64 percentage points respectively. The example of workers from Romania is particularly illustrative, whose number in the EU-15 grew by over 1.2 million after the accession. Out of EU-8 countries, the highest rate of immigration after the enlargement was in the Baltic countries, Poland and Slovakia while Slovenia recorded the lowest emigration rate. Are there any similarities between the migration of workers from the NMS into the EU-15 and workers from third countries and internal migrations in EU-15? According to Eurostat data (2012), in the period from 2005-2010, the participation of workers from EU-8+2 in the working age population of the EU-15 was doubled from 0.7% to 1.5%. In 2010, the mobility of workers from the NMS almost reached the level of mobility of workers from EU-15 -1.9% : 1.5%. However, compared with PANOECONOMICUS, 2013, 3, Special Issue, pp. 323-345 workers from third countries, the mobility of workers from the NMS and the OMS is very low. In 2005, the share of workers from third countries in the working age population of the EU-15 was four times higher (7.1% : 1.7%) compared to migrant workers from EU-15 i.e. ten times more (7.1% : 0.7%) when compared to workers from the NMS. Over time, the difference has increased when compared to workers from EU-15 (9% : 1.9% in 2010) however it was reduced when compared to workers from the NMS (9% : 1.5% in 2010). In all EU-15 countries, except Ireland, the share of workers from third countries is higher than the share of immigrant workers from EU-15 and EU-8+2.
Migration of workers from EU-8+2 to EU-15 had both the welfare and economic background. This is indicated by the results of the analysis of the Spearman's rank correlation coefficient. The ranking is conducted by comparing the net earnings, non-contributory social spending and the rate of immigration of workers from EU-8+2. Differences in earnings/social spending are expressed through the rate of depreciation relative to the EU-15 average. Table 6 reveals a statistically significant correlation between earnings/social spending and the immigration rate. The correlation coefficient between the two indicators is in the range of 0.5<|r|<0.8 indicating a medium-strong, positive correlation. In the period of 2004-2010, the immigration rate of workers from EU-8+2 was higher in countries where the earnings and non-contributory social spending are significantly below the EU-15 average. By comparing the values of correlation coefficients, we can conclude that in the observed period migrations were greatly determined by differences in earnings rather than social spending, which gives the migrations a predominantly economic feature.  The primarily economic character of migrations is further indicated by the fluctuation of the value of the correlation coefficient over time that moves in the direction of the increase of the degree of correlation of earnings and immigration rates and the weakening of links between social spending and immigration rates. Thus, in 2010, earnings rank correlation coefficient and the immigration rate is in the range of 0.8<|r|<1, therefore, a strong positive correlation, while for social spending the coefficient value is in the range of 0.5<|r|<0.8, which indicates a medium strong positive correlation. Thus, motives for migrations of workers from EU-8+2 should primarily be sought in economic reasons rather than welfare.
The data on the qualification and age structure of migrant workers from EU-8+2 speak in favour of the domination of economic migrants in the total migration flows. In the EU-15 the tendency of positive selection of immigrants is evident in the sense that the immigrants are better educated, more productive, have higher entrepreneurial potential and that the younger than the average of the country of origin. According to the data of the European Labour Force Survey (2012), most immigrant workers from EU-8+2 have upper secondary education. The share of low skilled workers in the total number of immigrants is less than their share in the population of the country of origin. Also, in some countries (like Austria) the share of highly skilled migrants in the total number of immigrants is higher than the total share of highly qualified population in their own countries. Immigrant workers from EU-8+2, whether they are staying legally or illegally in the EU-15, on average, perform lower qualifications jobs compared to their level of formal education. The main source of jobs for legal immigrant workers is the industry, followed by the service sector and construction. With illegal immigrants and season workers the service sector and construction dominate. The average age of immigrant workers from EU-8+2 is less than the average of the countries they come from.
In order to gain a more complete picture of the profile of Eastern migrations, we should briefly address the data on the implications of migration flows on key indicators of the economy of the EU-15. For the EU-15 countries, the macroeconomic impact of population flows from EU-8+2 is relatively small. According to the research Huber et al. (2011), the growth of potential GDP of 0.5% is expected in the long run. The first countries that started to implement an open door policy to workers PANOECONOMICUS, 2013, 3, Special Issue, pp. 323-345 from EU-8 could have the greatest benefits: Ireland and the UK. For Ireland, the projected growth rate of potential GDP as a result of migrations is 3% and for the UK just over 1%. Regarding the effects of migration of workers from EU-8+2 the GDP/pc EU-15, it is realistic to expect a positive but not a significant effect. The projected effect of the outflow of capital based on the information from abroad on the EU-15 economy is marginal. The impact of migrations from EU-8+2 on the rate of unemployment in the EU-15 as a whole is marginal. A significant rise in unemployment, which can be attributed to migrations and in specific sectors, is evident in the UK, Ireland and Spain. When we speak of real wages, it is realistic to expect a negative impact as a result of relatively unchanged unemployment rate however the fluctuations are within the minimum boundaries.

Discussion
To what extent are immigration policies of the EU-15 countries towards workers from the NMS a reflection of the actual impact of migration flows in the labour market and welfare state in the EU-15 or a result of social consensus of labour, the corporate capital within the EU institutional framework? To answer this question, it is necessary to point out the specifics of immigration policy in the EU and to comment on the effects of migration on the economy, labour market and welfare state in EU-15 in terms of the relationship between the established theoretical framework and the presented empirical findings.
The institutional framework. The liberalization of labour markets towards new member states was not automatic. By following the practice of previous enlargements (with the exception of the enlargement in 1995), with the Enlargement Agreements in 2004 and 2007 a selective liberalization of labour markets towards workers from the NMS was planned. The key mechanisms in this area are the transitional arrangements. It is the so-called 2+3+2 formula according to which the full application of the principle of free movement of workers may be deferred for a maximum period of 7 years. Initially, only three countries (Sweden, Great Britain and Ireland) opened its borders to workers from EU-8. By 2007, another 5 countries did the same (Netherlands, Finland, Greece, Italy, Portugal and Spain). France, Belgium, Denmark and Luxembourg removed the barriers before the deadline. Germany and Austria, where population of workers from EU-8 is the largest, decided to keep the restrictions until 2011.
In terms of enlargement in 2007, most of the countries decided to retain restrictions on workers from Romania and Bulgaria. Currently, 9 countries of the EU-15 require a work permit for workers from EU-2. In this context, examples of Spain, Great Britain and Ireland are particularly illustrative. In 2011, Spain requested the European Commission to close the formerly liberalized labour market (since 2009) for workers from Romania due to the large influx of immigrants from that country (in 2011 that number exceeded 800,000). During the second wave of enlargement in 2007, the UK and Ireland did not open their labour markets to workers from EU-2 as they did in 2004 for workers from the EU-8.
Restrictive rules on workers from EU-2 should be interpreted by a relatively high immigration potential: (i) in the case of Bulgaria based on the experience of a PANOECONOMICUS, 2013, 3, Special Issue, pp. 323-345 great wave of immigration even before the EU accession; (ii) or in the case of Romania -a low level of GDP (even compared to the NMS in 2004) and population (Klára Fóti 2011). According to Eurostat data (2012), in 2010 over 2 million Romanians and nearly half a million of Bulgarian nationals were living in EU-15. The average share of Romanians and Bulgarians who live in the EU-15 increased from 0.2% in 2004 to 0.6% in 2010.
Transitional arrangements have had a significant impact on the regional distribution of migration flows of workers from the NMS. The application of restrictive regimes by the traditionally most attractive countries for immigration -Germany and Austria led to the diverting migrants to new destinations -Great Britain and Ireland which in the past had no significant populations of workers from EU-8. Different dynamics in the liberalization of the labour market will have a lasting impact on labour mobility in the EU-27. It is the so-called "migration network effect": the probability of migration of workers to a specific country is higher if the Diaspora population is more numerous in the country. Relations that facilitate the migration process may be kinship, friendship or belonging to the same general culture pattern. Moreover, immigrant inflows in countries that have used restrictive rules during the transition period were based precisely on the networking phenomenon.
Starting from the findings that indicate a net positive impact of migrations on the macroeconomic indicators of the EU-15 countries, it is realistic to expect that the selective liberalization plus the networking effect will result in a more equal distribution of benefits of migrations between the EU-15 during the following period, which should reduce the demographic pressure on the welfare state.
Ir(rational) migrants. The key motives of migration for most immigrant workers from the NMS are the differences in net earnings i.e. employment. The analysis of Spearman's rank correlation coefficients (Tables 6 and 7) indicates that correlation of differences in net earnings and the rate of immigration are stronger than the relationship between the difference in social spending and the immigration rate. The convergence of net earnings and social spending towards higher levels (Tables 1 and 3) greatly weakens the incentives of welfare rather than economic immigrants. Welfare immigrants have, on average, higher costs of immigration because in the same conditions of earnings and social spending the welfare immigrants mostly decide to stay in the country unlike the economic immigrants. The age and education structure of immigrants which is more favourable than the average values for the population in the country of origin refers to the same conclusion. Young and educated workers have the dominant share in the immigrant population (European Labour Force Survey 2012).
The decision of the migration of workers is based on the comparison of net earnings between the source and destination country. The choice of immigration countries has been narrowed by the implementation of the transition agreements; however, this does not question the assumption of economic rationality. Opening the borders for workers from the NMS were followed by the convergence of earnings in EU-8+2 towards higher levels (Tables 1 and 3). The primary reason is certainly the growth of the NMS economies based on the privileges of a common market, but we should not lose sight of the impact of migration on the labour market. The reduction PANOECONOMICUS, 2013, 3, Special Issue, pp. 323-345 of labour supply in the countries of emigration resulted in the growth of real net earnings. The fact that the countries with high rates of immigration experienced an increase in net earnings above the group average is very indicative (Tables 1 and Table  5).
We cannot fully rule out differences in the non-contributory social programs as a motive for migrations from EU-8+2 in the EU-15. However this reason is not crucial for the majority of immigrants. The correlation between the migration rate is greater than the difference in net earnings, when observing the differences in social spending (Tables 6 and 7). Also, age and education structure of immigrants from EU-8+2 deviates from the usual description of welfare migrants. Young and skilled workers are more present then the older and low-skilled workers (according to the data of European Labour Force Survey 2012).
Similarly to net earnings, social spending in NMS records a continuing growth throughout the entire period starting from 1998 to 2010. Parallel increase in social spending and the immigration rate refers to the less prominent welfare feature of migrations (Tables 3 and 5). If immigration flows were predominantly made up of welfare motivated immigrants that would lead to a decline in the number of socially dependent population in the country of origin and the dynamics of social spending would slow down. Moreover, the reduction of the gap in non-contributory social programs is more intense when compared to the differences in net earnings, which indicates that the differences in earnings are a more influential determinant of migrations than the differences in the generosity of the welfare state (Tables 1 and 3).
Ir(rational) migration regimes. Were the positive trends of immigration followed the economic -welfare rational immigration policies? The dynamics of the opening of labour markets to workers from the NMS shows that the United Kingdom, Ireland and Sweden were the first to take this step. There is a significant difference between them in terms of the labor market and welfare state models. The UK and Ireland are the countries with flexible labor markets, weak unions and with a not so generous welfare state. The constellation of these variables opens the doors to economic and closes the door to welfare immigrants. On the other hand, Sweden is characterized by a rigid labor market, strong unions and a generous welfare state. Unlike Great Britain and Ireland, Sweden hasn't introduced restrictions on workers from EU-8 regarding the use of the welfare state programmes. It is an environment favourable to economic but also very attractive for the welfare immigrants.
What does the statistics data say? In the period of greatest influx of immigrants from the NMS into the UK and Ireland, from 2005-2008, immigrant workers from EU-8 had a higher rate of employment than the domestic workers, migrant workers from EU-15 and third countries. At the same time, the earnings of workers from the NMS were below the average of domestic workers (European Labour Force Survey 2012). Lower earnings suggest a higher concentration of immigrants in jobs with lower qualifications. However, this does not mean that immigrant workers are less educated than the domestic workers. Moreover, immigrant workers, on average, have a higher level of education than the local workers in the same positions. This situation is a result of the employment of immigrant workers in occupations below their qualifications and the fact that immigrant workers are mostly young people entering the labor market with a higher level of human capital. did not reveal any significant indication that the migration caused further decline in employment and earnings of domestic workers. Moreover, if there is a connection between the amount of immigration flows from EU-8 and the earnings and employment of domestic workers it would be positive rather than negative. We have similar observations in the case of Ireland (Alan Barrett, Adele Bergin, and David Duffy 2006;Nicola Doyle, Gerard Hughes, and Eskil Wadensjö 2006). The absence of a negative impact on wages and employment is the result of high rates of economic growth in countries of immigrationabove the euro zone average, the effect of the over-qualification of immigrant workers in relation to the tasks they perform as well as employment in the sectors for which the domestic workers are not interested. The favourable structures of migrants, including the limitations concerning the prospects of possibly receiving social assistance not only prevented a negative impact on the welfare state programs but have also had a positive impact on public finances.
Similar effects of migrations on the labor market and welfare state can be observed in the case of Sweden (Wadensjö 2007). Despite the fact that Sweden was the first of the EU-15 countries to fully liberalize the labor market and social policy towards workers from EU-8, the influx of immigrants was lower than in the UK and Ireland. The number of immigrants from the NMS was doubled when compared to the situation before enlargement in 2004, but is still low compared to the domestic population (0.5%). The greater influx of immigrants to Britain and Ireland is interpreted primarily by easier entry into the labor market, then the lower cost of living and the fact that English is the official language. Labour migrations from the NMS did not lead to disturbances in the labor market and the increase of the dependence on the welfare state. The immigrants are mostly young and educated workers in the labour market who may achieve more than just rely on social assistance. Also, the lack of Diaspora provides difficulties for welfare migrants who need somewhere to stay before they can exercise their rights to welfare-state transfers.
In relation to the EU-2, United Kingdom and Ireland have changed the course of immigration policy for workers from Romania and Bulgaria while Sweden remained in line with the liberal regime. Restrictions on the EU-2 are explained by the strengthening of anti-immigration lobby due to the growing number of immigrants in the past. During the liberal regime a Diaspora of economic migrants from EU-8 was created, which seems to be sufficient to meet demand for new workers from EU through a positive migrant network effect. There is a fear that the same treatment of workers from Romania and Bulgaria could create additional pressure on the labor market due to generally unfavourable immigration potential, significant differences PANOECONOMICUS, 2013, 3, Special Issue, pp. 323-345 in net earnings and social spending and a relatively high percentage of illegal immigrants from Romania. The presence of illegal immigrants increases the risk of negative migrant network effects.
By comparing the examples of Great Britain and Ireland on one side and Sweden on the other, and on the same basis the immigration policies of the remaining EU-15 countries, we can identify the elements of social consensus of labour and corporate capital within a given institutional framework of the EU. Unions oppose the process of the policy of free immigration while the interests of the capital favour greater flexibility of the labour market. The opinions of unions are economically irrational because there is no apparent linear negative effect on employment and earnings. Corporate capital achieves benefits from the over-qualification of the immigrant workers which increases the productivity of labour while the sector distribution of immigrants fills the empty niches in which domestic workers are not interested. The structure of immigrants, present trade unions, and the regulation of the labor market do not leave room for significant reduction of earnings. The implementation of the transitional arrangements allows an even distribution of migration flows between countries, creating a critical immigration mass, both in terms of benefits and potential pressures for welfare state. Fears regarding the disturbances in the labour market and pressure on the welfare state have proved to be unjustified. Bearing all this in mind, the limited rationality through the institutionalized social consensus is transformed into a form of channelled rationality.

Conclusion
Experiences with migration flows from EU-8+2 in the EU-15 suggest that the convergence of net earnings and social spending towards higher levels in the countries of emigration has a strong influence on the immigration regimes in countries of immigration. A greater degree of restrictiveness of the immigration policy towards the EU-2 is obvious, where the differences in net earnings and social spending are more prominent when compared to those members of the EU-8.
The countries exhibiting lower levels of social spending, more flexible labour markets, weaker unions, and a smaller Diaspora of workers from Eastern Europe: register as more quickly opening their borders to workers from EU-8. The change of orientation of the immigration policy, with Great Britain and Ireland serving as examples, may be associated with reaching critical levels of economic migrant Diasporas through the migrant network effect which could indeed attract a sufficient number of qualified -economic migrants in the future. The pressure of domestic workers to take a sharper course towards the immigrants is closely related to the large numbers of immigrant populations in their country, and, so, there is no economically rational foundation with respect to impacts of migration flows on the labour market and social spending.
The profile of the typical migrant worker from EU-8+2, who is younger and better educated than the average in the country of origin, favourably supports the acceptability of liberalization for the EU-15 labour market from the perspective of the welfare state. The impact on earnings and employment of domestic workers is insignificant, mostly positive, and relatively independent of the type of labour market PANOECONOMICUS, 2013, 3, Special Issue, pp. 323-345 due to the effects of over-qualification of the immigrant workers. The relatively high cost of living and access to labor markets in countries of immigration, on one hand, and increases in social spending and earnings in the countries of emigration, on the other, reduce the levels of benefits of welfare migration, while also increasing migration flows based upon economic considerations. Consequently, the risk of a demographic deficit is reduced, and without any additional pressure on the system offering social protection.
In sum, immigration policies of the EU-15 countries reflect the social consensus of labour as well as the interests of corporate capital invested within the EU institutional framework. The presented findings along with the discussion integral to this inquiry clearly support accepting the hypothesis posed. Accordingly, key dimensions of welfare states forming the EU-15 and also the in-migrations from citizens stemming from the EU-8+2 depends upon the convergence of earnings and social spending between the starting poorer and destination richer countries.