Empirica

, Volume 39, Issue 4, pp 461–485

The effect of a culturally diverse labour supply on regional income in the EU

Authors

    • Institute for Employment Research
  • Hanna Brenzel
    • Institute for Employment Research
Original Paper

DOI: 10.1007/s10663-012-9201-z

Cite this article as:
Brunow, S. & Brenzel, H. Empirica (2012) 39: 461. doi:10.1007/s10663-012-9201-z

Abstract

Because of an inflow of people into the EU but also because of the freedom of workplace choice within the EU, European regions are becoming more diverse in cultural terms. Despite the redistribution of labour and changes in regional labour supply, the ultimate question raised is whether there are additional gains or losses as a result of immigration flows. This paper therefore focuses on the impact of migrants on regional Gross Domestic Product per capita for European regions. Does the proportion of foreigners in the labour force increase or lower regional income? Does the composition of non-natives with respect to their countries of origin matter? We provide evidence that immigration and a higher degree of cultural diversity raise regional income, while controlling for endogeneity. We show that cultural diversity promotes income gains for destination countries. Whereas the presence of dominant groups reduces the costs of interaction and integration, diversity among foreign-born people increases the supply of different skills, knowledge and tasks. Thus, in general immigration has a positive net effect on regional performance and the costs of immigration in destination regions are balanced out. The regions of origin within the EU face a rise or a decline in income, depending on the labour market status of movers.

Keywords

Regional incomeCultural diversityEffects of immigration

JEL Classification

F22J15R11

1 Introduction

Most of the European countries face the so-called demographic change, which is frequently discussed in the scientific and policy debate. Beside other issues, labour shortages are expected when the available stock of labour declines, the level of production remains on a certain level and adjustment processes last longer. Then immigration could compensate for labour shortages and reduce some undesired impacts of the demographic change on welfare. Nowadays every European citizen can freely choose in which country of the EU he or she wants to provide its working capacity. Although language barriers exist, the work permission is given by the European law. We observe migration flows between member states but also from other regions of the world into the EU. This redistribution of workers partly solves issues of the demographic change. In addition to that there is another consequence of immigration: the labour force gets more and more diverse in cultural terms. One may hypothesize that international migrants have different skills and different approaches to solving problems, which is advantageous when they work together with people from the host country, and may then increase productivity. These migrants, of course, have detailed knowledge of the cultures in their home countries. Host-country firms may want to enter foreign markets and therefore have an interest in employing migrants of the respective nationalities. As a result of country-specific knowledge the firm may gain an advantage and market entry may potentially be more successful. These two examples make it clear that employing foreign-born migrants may increase productivity. However, negative aspects may also occur. For example in the presence of language barriers or cultural misunderstandings, potential productivity gains may melt away and the net effect on productivity could be zero or even negative.

In economic literature it is argued that non-native migrants have a higher risk of unemployment. Additionally, they may suffer from overqualification when their skills and educational levels are not (fully) accepted and they therefore work in occupations that demand a lower qualification level. In this case, self-employment is a strategy for the migrants concerned to earn an income. They may provide ‘cultural’ consumption goods such as specialized food, work as specialist hair dressers or Bohemians. Then migrants increase the variety of (local) consumption goods in a region. The increase in heterogeneous products can be seen as consumption amenities such that household utility and welfare may increase. In contrast, the native population might be afraid of foreigners and possibly expect ethnic conflicts or higher crime rates and therefore face a disutility because of immigration. With regard to the higher unemployment risk, immigrants might accept lower wages compared to natives in order to achieve an income. If natives and immigrants now compete for jobs and employers therefore hire non-natives, average incomes will decline. As was the case for the production side, not only the total number of immigrants but also the combination of different nationalities or the cultural backgrounds of migrants may matter. For instance, language barriers or ethnic conflicts could lead to inefficiencies in production that reduce regional income on average.

The net effect of gains and losses resulting from a culturally diverse population is unclear from a theoretical point of view and empirical evidence should therefore be provided. In the following we focus on the impact of migrants on regional economic performance by analysing the impact on Gross Domestic Product (GDP) per capita. The structure is as follows. The next section reviews related literature. Section 3 provides a theoretical outline of how the cultural diversity broadened by immigration can explain differences in GDP per capita. We adopt an augmented Solow model and derive an empirically testable model. Section 4 introduces the data set and additional control variables and is followed by a descriptive analysis in Section 5. Section 6 shows the regression results and discusses the results of the estimates. Finally, the paper closes with a conclusion.

2 Review of existing literature on cultural diversity

There are various reasons for the existence of cultural diversity. One reason is historical processes in which national borders do not necessarily represent boundaries separating ethnic groups with distinct cultural backgrounds. Another reason for cultural diversity in a region or country is an inflow of people from another country (immigration). In this case, people with different cultural backgrounds may interact with each other in a specific area. Irrespective of the reasons for the occurrence of cultural diversity, our view of cultural diversity relates to the interaction between groups of people who differ with regard to their regions of origin and their distinct cultural backgrounds.

There is a growing stock of literature analysing the influence of cultural diversity on economic performance, mainly using cross-country approaches. An early study in this line is the paper by Easterly and Levine (1997). They pay explicit attention to the remarkable effects of ethnic diversity across countries on economic growth. They argue that Africa’s growth failure is deeply rooted in the existence of ethnic conflicts and that per capita GDP growth is inversely related to ethno-linguistic fractionalization. For their measurement of ethnic fragmentation they use indices based on an ethno-linguistic classification derived from data from the former Soviet Union. Subsequent work confirms their results. Alesina et al. (2003) broaden the empirical approach of Easterly and Levine (1997) by introducing new measures of cultural diversity that permit a differentiation between ethnic, linguistic and religious fractionalization. They provide substantially different evidence depending on the classification they apply. By analysing the influence on economic growth they broadly confirmed the results obtained by Easterly and Levine (1997) when ethnic and linguistic fractionalization are considered. Both types of fractionalization are associated with negative growth of GDP per capita. However, religious fractionalization does not affect growth rates significantly. Collier (2001) argues that cultural fractionalization has a negative effect on productivity and growth in non-democratic regimes whereas this is not the case for democracies. However, Collier cannot find any significant effects of religious diversity. Inspired by the evidence provided by Collier (2001), Alesina and La Ferrara (2004) revisit the effect of diversity on economic performance and, by employing a fractionalization index, they confirm Collier’s (2001) finding that religious diversity has no effect on economic growth. Furthermore they show that the negative effect of diversity is stronger for countries that exhibit lower income levels. Montalvo and Reynal-Querol (2005) argue that both ethno-linguistic and religious diversity may be a potential measure for a strong conflict dimension. They therefore suggest a new measure which aims to capture the potential for conflict in heterogeneous societies based on a polarization index instead of the fractionalization index. Their results indicate that a higher degree of ethnic and religious polarization has a large and negative impact on economic development through indirect channels such as civil war.

Besides the evidence on losses resulting from cultural diversity, there is also a strand of literature which substantiates the existence of benefits from heterogeneous societies. Ottaviano and Peri (2005) investigate the impact of cultural diversity on the economic life of US cities through the wages of the native population. Allowing for imperfect substitutability between natives and foreigners, the authors find a significant and robust positive correlation between cultural diversity and the wages of white US-born workers. They additionally point out that the benefits originating from migrants who have integrated are larger than those from new immigrants that have not yet integrated in the host country. Similarly, Bellini et al. (2008) follow the same idea that cultural diversity may affect both production and consumption through positive or negative externalities. To identify the dominant effect they analysed the joint estimation of a price and income equation. Their results are consistent with those obtained by Ottaviano and Peri (2005) for US cities. They focus on NUTS-3 regions of 12 European countries and provide evidence that diversity is positively correlated with productivity and that the causality runs from the former to the latter. Bellini et al. (2008) and Ottaviano and Peri (2005) measure cultural diversity on the basis of the characteristics of language -spoken -at -home, country of birth, citizenship and race. They employ Herfindahl-like indices.

D’Amuri et al. (2010) investigate the labour market impact of immigration on wages and employment in western Germany. The group of new immigrants mainly affects the employment levels of those in the previous immigration waves. The effect is statistically and economically significant. Interestingly, the impact of (substantial) immigration inflows on the wages and employment levels of natives is relatively small. These asymmetrical results are mainly driven by a higher degree of substitution between ‘old’ and ‘new’ migrants in the labour market, for instance due to rigid wages. Suedekum et al. (2009) study the impact of increasing diversity on native employees in western Germany. The analysis is conducted at local level and concludes that diversity raises productivity at this level. Additionally, the study reveals the importance of distinguishing between high- and low-skilled foreign workers. For high-skilled foreign workers, they found that both the size of the group and the diversification into different nationalities increase the local wage and employment for native workers. However, the effect is negative for low-skilled foreign workers. They argue that the presence of high-skilled foreign workers can be regarded as a positive production amenity from a regional perspective. Nathan (2011) reaches a similar conclusion for the UK based on a panel period lasting 16 years. Average productivity and wages rise for UK-born people on average due to immigration. However Nathan (2011) also provides evidence that natives are crowded out when they compete for similar jobs. Based on the findings obtained by D’Amuri et al. (2010), as well as those of Suedekum et al. (2009) and Nathan (2011) one might hypothesize that immigration does not necessarily lead to lower average income.

Ratna et al. (2009) and Sparber (2010) analyse the macroeconomic effects of social diversity in the US based on the state level and using cross-sectional data. The empirical investigations yield mixed results. Whereas Sparber (2010) was unable to find any causal relationship between diversity and gross state output per worker, Ratna et al. (2009) find evidence that racial diversity reduces GDP growth while linguistic diversity raises GDP growth. They justify their results with the fact that English is frequently used by non-native speakers and so the barriers to communication based on race are more pronounced and enduring than those based on linguistic differences.

Cheng and Li (2011) consider regional and sectoral firm formation and the role of the composition of foreigners in terms of racial and cultural diversity. They especially identify specific sectors where the effect of fragmentation on firm formation is greater. Cheng and Li (2011) highlight service sectors with special cultural needs in production to supply culturally diverse products. This evidence supports the arguments of Ottaviano and Peri (2005) as to why cultural diversity might matter in a positive manner and why foreign-born workers offer different skills.

The empirical contributions cited above focus on a country or regional level. There is a branch of literature focusing on firm level in general or on sub-groups of the labour market, for instance the impact of high-skilled workers on innovation. Niebuhr (2010) investigates the impact of cultural diversity in the workforce involved in R&D on regional innovation output. The regression results support the hypothesis that differences in the knowledge and capabilities of workers from diverse cultural backgrounds may enhance the performance of regional R&D sectors. Ozgen et al. (2011) discuss various effects of immigration on the innovativeness of European regions. They base their measures of innovation on the means of the number and types of patent applications and argue that regions with many immigrants might also have a larger number of patent applications. However, they suggest that there might be an optimum level for cultural diversity, because the benefits gained from diversity appear to decrease when a value of the fractionalization index exceeds a critical point. The work of Parrotta et al. (2011) also confirms the positive impact of cultural diversity on innovativeness within firms, explaining the incentives for patenting, the number (mass) of patents and the ability to patent in various, distinct fields.

Besides the impact of innovation on firm performance, Brunow and Blien (2011) and Parrotta et al. (2010) focus on the impact of cultural diversity on establishment productivity. Brunow and Blien (2011) find evidence of productivity gains when the employed labour force is more diverse. Diversity is measured on the basis of the information about the employees’ nationalities. They also find negative effects, however, which they relate to the “Babel” effect. The more foreign nationalities are employed the lower productivity is. The study by Parrotta et al. (2010) partially supports these findings. In this work, positive effects are due to human capital diversity, especially in skills and education. Ethnic diversity has no or only an insignificant impact on firms’ total factor productivity.

So far, the focus has been on the regional or firm level. Longhi (2011) analyses the impact of cultural diversity on individual wages and on various aspects of job satisfaction and finds positive but also insignificant results.

Based on the evidence in the literature we conclude that the effect of cultural diversity on productivity or growth is unclear and depends on the measure applied, the level of aggregation and the underlying background (racial, ethnic, linguistic etc.). Most studies identify gains as long as conflicts are not considered, but the literature also shows that negative effects occur as well. Thus, based on the review we expect a positive, a negative or an insignificant impact of cultural diversity on regional income. Most studies in this field use cross-sectional data to identify the effect. However, Islam (1995) discusses a serious parameter bias when country- or region-specific effects are not taken into account.

3 Theoretical background

As highlighted in the introduction several mechanisms suggest that a culturally diverse population may yield gains or losses. They might occur on the production or the consumption side. Some studies focus on firm or establishment data to reveal these effects from a production-side perspective. However, these studies cannot focus on the consumption side directly. We are interested in the general effect on income at regional level. The theory of economic growth describes relevant variables that explain income differences between countries and regions. We adopt the model suggested by Mankiw, Romer and Weil (1992). The production function produces output under constant returns to scale and employs physical and human capital and labour. Ottaviano and Peri (2005) suggest that a culturally diverse labour force yields higher levels of productivity, when gains of diversity override negative effects. They model both effects and assume that labour is diverse with respect to the cultural background. The impact of cultural diversity on the estimation equation is discussed below. We additionally adopt the approach of Brunow and Hirte (2006) who introduce labour market variables into the model of Mankiw, Romer and Weil (1992), namely participation p and the unemployment rate u. It is based on a decomposition of total population to participants of the labour force, employed and unemployed. Following Mankiw, Romer and Weil (1992), the steady state on per capita income depends on the investment rate of both types of capital sk and sh, the growth rates of the population and technological progress, n and g respectively, but also on the depreciation of both types of capital δ. Brunow and Brenzel (2011) derive the steady-state per capita income for this augmented Mankiw-Romer-Weil approach, that leads to our regression model. The advantage of the steady-state is that it includes regional investment rather than the (unknown) capital stock as explanatory variable. In our empirical analysis the investment rate includes gross investment flows from the region of origin but also investment inflows from other regions. With this respect we differ from the model of Brunow and Brenzel (2011) who assume that only regional saving equals regional investment.
$$ \begin{aligned} \ln y_{r} &=\alpha _{0}+\alpha _{1}\ln \left( n_{r}+g+\delta \right) +\alpha _{2}\ln s_{kr}+\alpha _{3}\ln h_{r}\\ &\quad+\alpha_{4}\ln \left[p_{r}\left( 1-u_{r}\right) \right] +\alpha _{5}\ln DIV_{r}+\mu _{r}+\mu _{t}+\varepsilon _{r}, \end{aligned} $$
(1)

This equation describes the influence of theoretically relevant variables on differences in steady-state values. For instance, a larger share of human capital employed raises income, whereas a relatively lower participation rate reduces regional GDP per capita. Equation (1) considers the level of human capital employed, h, rather than investment in new human capital, 1sh, for the following reason: as our time period is rather short, investment in new human capital is not expected to influence current per capita income.

Note that the net effect of cultural diversity DIV simply adds as an additional variable to the steady state level. The net effect of cultural diversity could be positive and negative. In the case that the two opposed effects cancel each other out, the resulting net effect is 0. Using the labour market decomposition has another advantage when studying the effect of immigration: suppose there is only a labour supply effect of immigration that solely affects the participation and unemployment rates. Then the effect of cultural diversity will be irrelevant in production. To put it another way, since labour market effects of immigration are captured in \(p\left( 1-u\right),\) the effect of cultural diversity therefore relates to productivity gains and losses. Thus, the decomposition as suggested by Brunow and Hirte (2006) also leads to the productivity effects of cultural diversity being separated from the labour supply effects of immigration.

Mankiw, Romer and Weil (1992) additionally derive a growth regression where regional income growth is explained by the income level at the beginning of the time period. As is shown below, our data set covers a relatively short time period. This means that we cannot take sufficiently into account the endogeneity problem emerging in growth regressions as described by Caselli et al. (1996). Therefore we keep to our approach and answer the question of whether differences in regional income are additionally explained by distinct levels of cultural diversity. Islam (1995) points out that parameter estimates are potentially biased when region-specific effects are not taken into account. Since our research fields are regions which we observe over time, we control for region-specific effects μr and time fixed effects μt.

With the error term \(\varepsilon_{r}\) we examine Eq. (1) empirically in the next sections. The method used to measure and capture the cultural diversity DIVr term is examined in the following section.

4 Variables and data

We combine the Eurostat regional database with the European Labour Force Survey (ELFS), both provided by Eurostat, the Statistical Office of the EU. The regional classification is based on the NUTS 1 and NUTS 2 level of aggregation. The advantage of the NUTS 2 over the NUTS 3 level is that it overcomes strong spatial interdependencies emerging at the NUTS 3 level due to a common labour market and commuter flows between regions or vertical linkages of upstream industries located nearby. Unfortunately, some countries only report in the ELFS data the NUTS 1 regional level for data security reasons. We have to take this limitation as it is.

The ELFS data come from a household survey which basically gathers labour market characteristics and individual information about household members. Our panel spans the time period from 2003 to 2008. We follow Ottaviano and Peri (2005), and use the country of birth and the nationality to capture the cultural background of individuals. Unfortunately, both of these variables are only available from 2004 until 2008. Therefore the data from 2003 are only needed for the construction of lagged values that do not relate to our measures of the cultural background. In individual years there is a lack of data on the country of birth and nationality for some countries, which means that we are considering an unbalanced panel. There are no data on the cultural background for Polish regions, so we have to exclude Poland from our sample entirely. Furthermore, we cannot consider Greece, Norway and Iceland due to missing values on investment data. Because of unreasonable values for the share of high-skilled workers and the share of foreigners for 2006 and 2007 we have to exclude the French region FR83 for both years. We also exclude some Spanish regions (the exclaves ES63, ES64, and the Canary Islands ES70) and the Portugese islands (Azores PT20 and Madeira PT30). Given the limitations of some years, the countries included in our sample are:2 Austria, Belgium, Czech Republic, Denmark, Estonia, Finland, France, Germany, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Netherlands, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, and the UK.

Because the ELFS is a household survey it does not necessarily represent the regional population. Each respondent is therefore assigned an individual weighting factor in order to ensure representativeness. The factors are provided along with the ELFS data. We take these weighting factors into account when we construct and aggregate variables at regional level.

From the regional data basis of Eurostat we use data for the GDP divided by the total regional population to construct the GDP per capita measure as a proxy for \(\ln y_{r}.\) The population growth rate nr is constructed using the difference between births and deaths relative to the population. These data are not available for the UK, so we calculate nr as the change in the population instead. To define the variable \(\left( n_{r}+g+\delta \right)\) we assume that g = 0.03 and δ = 0.05.

On the basis of the regional gross investments we compute the investment share to cover the capital investments skr. Unfortunately, capital investment data are not available for all regions. We therefore replace missing values with data provided by Cambridge Econometrics.3 To avoid problems resulting from different data sources, we use exclusively one data source within each region. Thus, measurement errors between both data sources is then relaxed by the fixed effects transformation. We carefully checked data quality and comparability. It emerged that the absolute values in the two data sets of investment data differ but the correlation is 0.99. The correlation pattern is still high for the within-transformed data, namely 0.81. It is reasonable to assume that the revenues from investment are achieved some time after the investment was made. We therefore use the lagged value of investment to explain income differences.

The ELFS collects information about the educational level of respondents. We use this information to construct the proxy for human capital hr, measured as the proportion of people with a university degree. As was the case for capital investment we use the lagged values of hr. The term \(p_{r}\left( 1-u_{r}\right)\) which describes the labour market is also calculated on the basis of the ELFS data, and is covered by the questions about participation and unemployment. The advantage here is that it explicitly considers the effect of foreign-born people and their labor status. Variables capturing the cultural background are also taken from the ELFS, which provides two types of information on this issue. First, the respondents are asked to report their country of birth, and second their nationality. The questions are designed to distinguish between natives and non-natives. If they are not natives, respondents can report one of 8 macro-regions of origin or nationality, respectively. Distinguishing between natives and non-natives makes it possible to study not just the effect of international migration but also the effect of migrants within the European Union. Intra-country migration, however, cannot be taken into account. The group of migrants therefore only consists of international migrants who are foreign-born. The roughly classified macro-regions are: EU 15; New Member States 12; Europe outside EU 27; Other Africa; North Africa and Near/Middle East; East and South Asia; Latin America; and North America and Australia. Since we are more interested in cultural differences and not in the legal status, we use the country of birth to compute diversity measures. Because in some countries the country of birth was not surveyed or respondents did not answer, we use nationality as a weaker proxy instead.4

As our model suggests, the cultural background of employees or self-employed individuals is mainly of importance here, since we focus on the effect of regional labour supply. We therefore do not consider children up to the age of 15 or pensioners.5 Another reason to exclude children is that some countries do not report individuals under the age of 16.

To approximate cultural diversity, DIVr, we employ two variables. D’Amuri et al. (2010) and Brücker and Jahn (2011) provide evidence that the elasticity of substitution between natives and non-natives is distinct from the substitution elasticity among non-natives in their respective destination regions. To distinguish between the two effects we compute the proportion of foreigners smigrantsr to control for the substitution pattern between natives and non-natives and additionally, we also compute measures that capture the degree of diversity among foreigners. Since we cannot distinguish between positive and negative effects, we may obtain positive or negative estimates. The distinction between smigrantsr and the diversity among foreign-born people also makes it possible to separate a general effect capturing the presence of foreign-born people from a diversity aspect of immigration. As outlined in the literature review, various measures are suggested for capturing diversity. It is worth noting that there is no best proxy and we therefore compute three common measures, a Herfindahl-like index, the fractionalization index and finally a polarization index. Let sm be the proportion of the m-th of M foreign-born cultural groups, then the different measures are calculated as follows,
$$ Herfindahl =\sum_{m}^{M}s_{m}^{0.66}, $$
(2)
$$ Fractionalization =1-\sum_{m}^{M}s_{m}^{2}, $$
(3)
$$ Polarization =1-\sum_{m}^{M}\left[ \left( \frac{0.5-s_{m}}{0.5}\right)^{2}s_{m}\right]. $$
(4)

There is a crucial difference between the first two measures and the third one. The fractionalization and the Herfindahl-like measures increase with the degree of cultural diversity and especially the more equally distributed the shares are. The polarization index, on the other hand, increases in the presence of two dominant groups. As can easily be seen, when there are two groups, each with a share of 0.5, the index reaches its maximum at 1. Thus the polarization index identifies the presence of two dominant groups out of M distinct groups.

5 Descriptive analysis

The upper part of Fig. 1 shows the income distribution on the left and the proportion of human capital on the right. The second row displays the distribution of non-natives as a proportion of the total population and the diversity among non-natives. The band width is chosen in such a way that each class contains approximately the same number of observations, so that each colour can be interpreted in terms of percentiles. The regions coloured yellow are those not included in the data set. The data relate to the regional average during the sample period.
https://static-content.springer.com/image/art%3A10.1007%2Fs10663-012-9201-z/MediaObjects/10663_2012_9201_Fig1_HTML.gif
Fig. 1

Regional distribution of main variables

As can be seen, the proportion of foreign-born migrants is not necessarily larger in wealthier regions, although there is still a clear pattern in which regions with higher incomes are in favour of immigration from abroad and therefore the proportion of non-natives increases. Interestingly, even more than 15 years after the political change in the former centrally planned economies, the proportion of foreign-born migrants is still small in these regions.6 The cultural mix among foreign-born people is shown in the lower right panel where we employ the fractionalization index. It reveals that regions with a relatively low level of non-natives could nevertheless be highly diverse in cultural terms. There are also regions with a large proportion of foreign-born migrants and a high degree of diversity. In contrast, a large proportion of non-natives and low diversity means that there has to be a dominant group of foreign-born people.

Figure 2 plots the proportion of non-natives against the log of GDP per capita within regions to obtain deeper insights into a potential correlation. The larger the proportion of immigrants the higher GDP per capita is, providing a first indication that international migrants may improve regional productivity and income. Interestingly, this pattern holds for different European regions in which the average income and the immigration history are quite distinct. However, endogeneity issues also arise because of the selection problem: a well-performing region, whether wealthy or not, may offer higher wages, making this region more attractive for immigration relative to other regions. We should therefore focus on the immigration structure and the distribution of foreign-born people.
https://static-content.springer.com/image/art%3A10.1007%2Fs10663-012-9201-z/MediaObjects/10663_2012_9201_Fig2_HTML.gif
Fig. 2

Correlation between GDP per capita and the proportion of non-natives

Table 1 provides a descriptive overview of the proportion of foreign-born migrants as a percentage of the population in EU regions for the years 2004 and 2008. Besides the average proportion of the total population, the relative average proportion of the foreign population is also reported. As can be seen, the data only allow the observation of 8 distinct groups of migrants. The diversity measures are calculated from these groups. For instance, in western European regions the average proportion of EU 15 foreigners in the total population is 3.6 % and these 3.6 % are 35.5 % of all foreigners in 2004. As shown, the cultural mix increased in all three macro-regions during the sample period. Interestingly, people from formerly centrally planned economies seem to settle more frequently in the southern parts of Europe. On the other hand, eastern European regions mainly attract people from the EU itself but not from the rest of the world. Focusing on the relative proportion of all non-natives reveals that the regions in formerly centrally planned economies mainly attract foreigners from regions in other former centrally planned economies. This could be because of language similarities (Slavic languages) which possibly reduce the costs of integration. The descriptive table does not immediately confirm the fact that migrants prefer only regions with higher income levels for immigration, as smigrantsr rose in all sub-groups of European regions.
Table 1

Average and relative share of migrants in the population within EU regions

 

European regions

Westa

Southb

Former centr.c

Shared

Relativee

Shared

Relativee

Shared

Relativee

2004 (as %)

 EU 15

3.6

35.5

1.1

18.6

0.1

6.2

 New Member States 12

0.4

4.7

0.7

11.2

0.6

47.2

 Europe outside EU 27

1.7

18.6

1.6

28.2

1.4

37.4

 Other Africa

0.8

8.0

0.7

8.3

0.0

1.0

 North Africa, Near/Middle East

1.8

16.2

0.5

6.4

0.1

1.7

 East and South Asia

1.0

10.9

0.1

1.5

0.1

4.5

 Latin America

0.3

3.0

1.7

23.1

0.0

1.0

 North America and Australia

0.3

3.0

0.1

2.7

0.0

0.9

 Share of non-natives smigrantsr

9.9

 

6.5

 

2.3

 

2008 (as %)

 EU 15

4.1

33.8

1.3

14.3

0.1

4.3

 New Member States 12

0.9

7.2

1.6

14.6

0.9

49.6

 Europe outside EU 27

1.9

19.3

2.3

27.3

1.8

34.9

 Other Africa

1.0

8.9

0.7

6.0

0.0

0.1

 North Africa, Near/Middle East

2.2

19.2

1.2

10.8

0.2

5.0

 East and South Asia

0.7

7.0

0.5

4.6

0.1

4.3

 Latin America

0.3

2.8

2.7

19.9

0.0

0.8

 North America and Australia

0.2

2.0

0.2

2.5

0.0

1.0

 Share of non-natives smigrantsr

11.3

 

10.5

 

3.1

 

AT, BE, DE excl. eastern Germany, DK, FI, FR, IE, LU, NL, SE, UK

b ES, IT, PT

c CZ, EE, HU, LT, LV, PL, RO, SI, SK, eastern Germany

d Share of group as a % of the total population

e Share of group as a % of the foreign population

Source: EU Labour Force Survey; own calculations

What is also known from migration literature is that migrants tend to settle in regions with a lower risk of unemployment. This probability is generally lower in more densely populated regions. The proportion of human capital is also larger in densely populated (agglomeration) regions, which raises problems of identification when the two variables are correlated with each other. A simple correlation between the proportion of foreign-born migrants and the human capital measure \(\ln h_{r}\) is 0.40, which provides first evidence supporting this hypothesis. After absorbing the region fixed effects, the correlation is even higher, namely 0.45.

Table 2 provides an overview of our main variables, some of them not presented in log form. Besides the total variation of the sample it also reports the variation after the fixed effects transformation has been performed. In the latter case no mean is reported, since it is zero. There are some regions with very low participation and employment levels, which in turn means a very high dependency ratio. On the other hand, in some regions over half of the population participates in the labour market and is employed. When we examine the transformed data set we find that changes in participation and unemployment rates occur. Focusing on nr + g + δ clearly shows that European regions do not grow or shrink much during the sample period in terms of population. Both, δ and g are fixed values and we use 0.08 for the sum, which is a common value emerging in the literature.
Table 2

Overview of main variables

 

Overall variation

Fixed effects transformed

Mean

SD

Min

Max

SD

Min

Max

\(\ln \left( GDP per capita_{r}\right)\)

3.024

0.543

0.899

4.395

0.086

−0.335

0.361

\(p_{r}\left( 1-u_{r}\right)\)

0.43

0.049

0.256

0.563

0.01

−0.036

0.032

\(\left( n_{r}+g+\delta \right)\)

0.05

0.003

0.074

0.091

0

−0.002

0.003

hr

0.192

0.074

0.059

0.413

0.016

−0.094

0.058

smigrantsr

0.081

0.068

0

0.453

0.013

−0.066

0.065

Fractionalization

1.673

0.207

1

1.96

0.057

−0.423

0.293

Neg. Herfindahl

0.641

0.175

0

0.845

0.055

−0.462

0.266

Polarization

0.673

0.137

0

1

0.069

−0.656

0.532

Some variables not in log form

Source: EU Labour Force Survey; own calculations

The average share of regional human capital is less than 20 %. A look at the upper right panel of Fig. 1 reveals fairly high values in northern Spain and rather low levels for instance in western Germany. We therefore test the validity of the data employed by comparing it with data on human capital as provided by Eurostat and national statistic offices. The Eurostat data suffer from the problem of missing values which is not the case for the values computed on the basis of the ELFS data. In Fig. 3 we therefore provide a comparison of the distributions of the human capital variables, which are available in both data sets, on the basis of a Kernel density plot. Whereas the absolute values of the Eurostat data are higher on average (left panel), the distribution of values within each region are almost identical (right panel). The correlation between both variables of the two data sources is 0.96. We therefore conclude that the measure based on the LFS data is reliable.
https://static-content.springer.com/image/art%3A10.1007%2Fs10663-012-9201-z/MediaObjects/10663_2012_9201_Fig3_HTML.gif
Fig. 3

Distribution of human capital variables

The average proportion of migrants within regions is 8 % with a range between zero and over 45 %. The regions with the largest proportions of non-natives are Brussels, London and Luxembourg.

The structure of the correlation between the diversity measures and the log of GDP per capita is 0.5 for the fractionalization index, 0.4 for the Herfindahl-like index and about 0.1 for the polarization index. All three correlation structures vanish after the fixed effects transformation. The correlation drops to values between 0.02 and 0.07. The impact of the combination of migrants on income might be negligible. The first impression obtained by using bivariate correlation seems to suggest that immigration has a positive effect on regional income. However, does this picture remain when other effects are controlled for?

6 Regression analysis

In Sect. 3 we derived a regression model inspired by a neoclassical production function which reads as
$$ \begin{aligned} \ln \left( GDP per capita_{r}\right) &=\alpha _{0}+\alpha _{1}\ln \left( n_{r}+g+\delta \right) +\alpha _{2}\ln s_{kr}+\alpha _{3}\ln h_{r} \\ &\quad+\alpha_{4}\ln \left[ p_{r}\left( 1-u_{r}\right) \right] +\alpha _{5}\ln \left[ \left( 1-\tau _{r}\right) DIV_{r}\right] +\mu _{r}+\mu _{t}+\varepsilon _{r}. \end{aligned} $$

The theoretical approach of Brunow and Brenzel (2011) treats total factor productivity as constant across all regions. This strong assumption is hardly found in reality and therefore region-specific effects μr have to be taken into account. Region-specific effects capture all unmodelled time-invariant differences between regions, such as differences in technology. We compared the results obtained using fixed effects models with those obtained using OLS models. In any model the fixed effects model is superior to OLS. OLS estimates deliver inconsistent results. We therefore do not report OLS estimates. In all of the models region-specific fixed effects μr and time fixed effects μt are thus controlled for. Because of the complementary and substitutable relationship between the variables included in the regression model and others that are omitted, the explanatory variables are assumed to be correlated with μr. Then a fixed effects model is preferred over random effects models on the basis of theory. This result is confirmed when estimates using the random effects model are compared with the fixed effects estimates. We do not provide random effects estimates due to their inconsistency. We operationalize cultural diversity by using smigrantsr and the different diversity measures as introduced in Eqs. (2)–(4).

As already mentioned, the migration decision is made on the basis of wage differentials between the migrant’s potential host country and his or her home country. Because of the selection problem of potential migrants one might expect better performing regions to attract migrants more frequently, which in turn would increase the proportion of non-natives in that particular region. We partially overcome this problem by using a regional fixed effects model but also explicitly control for this endogeneity. Additionally, increases in productivity and thus in income in former centrally planned countries might be expected. This catch-up effect cannot be explained by the variables under consideration. We therefore also interact the time dummies with a dummy variable for Eastern European regions including eastern Germany (without Berlin) and add it to our empirical model. It emerges that these dummy variables are always highly significant and positive,7 providing evidence of this catch-up effect. Any estimates are efficient for arbitrary heteroscedasticity.

We estimate various models. The Base model does not control for cultural diversity issues. The Share model considers the proportion of all non-natives in the population, smigrantsr. If the proportion exhibits a positive sign, then there is a positive correlation between GDP per capita and the proportion of non-natives, as suggested in Fig. 2. Note that in the fixed effects analysis we cannot state that an increase in the number of migrants improves regional income because we do not test for causality. Models CD 1 to CD 3 control for the fragmentation of the non-natives in a particular region, employing the fractionalization index (CD 1), the Herfindahl-like index (CD 2) and finally the polarization index (CD 3) as outlined in Eqs. (2)–(4). These three models answer the question of whether there are additional gains (or losses) in GDP per capita the more fragmented the non-natives are with respect to their country of birth or if there is a tendency towards dominant groups.

The fixed-effects regression analysis provides first results, which are presented in Table 3. A look at the F test of the estimation results shows that the explanatory variables of our models are jointly significant. Also, the fixed effects are jointly significant such that OLS yields inconsistent results. Let us take a first look at the evidence. An increase in the participation and employment rates is positive and highly significant. Of course, the lower the dependency ratio, the higher the sum of wage payments is and this in turn makes a region relatively wealthier. In the context of demographic change the participation rate will decline during the transition period, when the proportion of the elderly as a percentage of the total population is relatively larger, which lowers regional income.
Table 3

Fixed Effects Regression on GDP per capita for EU regions

\(\ln \left( GDP per\,capita_{r}\right)\)

Base

Share

CD 1

CD 2

CD 3

\(\ln \left[ p_{r}\left( 1-u_{r}\right) \right]\)

0.433***

0.431***

0.449***

0.431***

0.403**

 

(0.163)

(0.163)

(0.156)

(0.159)

(0.156)

\(\ln \left( n_{r}+g+\delta \right)\)

−0.001

−0.001

−0.001

−0.001

−0.002

 

(0.002)

(0.002)

(0.002)

(0.002)

(0.002)

\(lag \ln h_{r}\)

0.138***

0.134***

0.131***

0.136***

0.134***

 

(0.031)

(0.032)

(0.029)

(0.03)

(0.03)

\(lag \ln s_{kr}\)

0.018

0.021

0.016

0.02

0.021

 

(0.029)

(0.029)

(0.028)

(0.028)

(0.028)

smigrantsr

 

0.103

0.093

0.103

0.076

  

(0.099)

(0.106)

(0.104)

(0.106)

Diversity1

  

0.097**

0.086*

0.107**

   

(0.041)

(0.049)

(0.044)

Region fixed effects, time fixed effects, East European Countries* time fixed effects

F

152.4***

144.6***

130.6***

128.8***

128.6***

RMSE

0.034

0.034

0.034

0.034

0.034

r2

0.841

0.841

0.845

0.843

0.845

Within R2

0.841

0.841

0.845

0.843

0.845

Overall R2

0.051

0.062

0.128

0.105

0.058

Between R2

0.055

0.071

0.18

0.142

0.067

Valid cases

666

666

666

666

666

No. of regions

156

156

156

156

156

Robust s.e. in (); * p < .1; ** p < .05; *** p < .01; 1 Cultural diversity measures are the Herfindahl index for CD 1, the fractionalization index for CD 2, and the polarization index for CD 3

A change in nr + g + δ has no effect on GDP per capita. There is also no significant effect when the endogeneity of migrants is controlled for. Fischer (2011) suggests a value of 0.05 for g + δ. If we use this value we still obtain insignificant estimates for the nr + g + δ term in the fixed effects and the instrumental variable fixed effects models. Despite the insignificance of the \(\ln \left( n_{r}+g+\delta \right)\) term, the remaining estimates are unaffected with regard to its value and significance such that our conclusion does not depend on the definition of g + δ. The insignificance might be because of the short time period and little variation in regional population growth. In migration literature it is frequently shown that younger people are more likely to migrate, which possibly raises the natural population growth rate. Again, because of the short time period in our study such an effect will not affect the short-run perspective.

As expected, an increasing stock of human capital improves regional performance and thus GDP per capita. The elasticity is about 13.4 %. A larger stock of human capital boosts regional income. The proportion of capital investment clearly does not influence regional income. We will discuss the impact of human capital and capital investment in more detail later, when the endogeneity of foreigners is explicitly controlled for.

Bearing Fig. 2 in mind, it is somewhat surprising that the proportion of immigrants has an insignificant impact on regional income, although one might expect the proportion to have a positive or negative impact, depending on the net effect of gains and losses from immigration. However, once we control for region and time fixed effects and other well established variables, the possible effect of changes in the composition of natives and foreign-born migrants disappears. If there is only a labour supply effect of foreign-born people which influences participation and unemployment, then this effect is already captured by \(p_{r}\left( 1-u_{r}\right).\) Interestingly, the diversity variables among foreign-born people are all highly significant and positive, indicating that a culturally diverse labour supply matters and raises GDP per capita.

However, as was mentioned earlier, the proportion of foreign-born migrants is highly endogenous. We therefore estimate the same models but treat the proportion of migrants as an endogenous variable. From migration literature we know that network effects of migrants exist. Additionally, if a region and its neighbouring regions already accommodate a larger proportion of foreigners, then this region might be attractive to new migrants, because it is an established immigration/ destination area. We therefore add two instruments to explain the current proportion of migrants: first, the proportion in the previous period as an internal instrument. Second, we define an average proportion of migrants in the surrounding regions as an external instrument, again using the time-lagged value. When computing this instrument we use a distance-based weighting matrix to give nearby regions a higher weight compared to regions further away. Let smigrantsk be the proportion of migrants in region k ≠ r and let 0 ≤ wkr < 1 be the weight for the extent to which smigrantsk contributes to the external instrument \(\hat{s}_{migrants}^{r}\) for region r. We apply the distance-decay function8 suggested by Niebuhr (2001) to relate the distance between k and r negatively to wkr. Then, the instrument is computed as
$$ \hat{s}_{migrants}^{r}=\sum_{k\neq r}^{R}w_{kr}s_{migrants}^{k} ,\qquad \sum_{k\neq r}^{R}w_{kr}=1. $$
The instrument can be understood as a kind of migrant potential. It captures the effects of established immigration regions. If, for instance, a region has a relatively large proportion of immigrants but the neighbouring regions have hardly any, this region might not be particularly attractive for further immigrants compared to a region whose neighbouring regions also have a large number of immigrants.
The estimates of the instrumental variable approach are presented in Table 4 and the parameters are derived employing GMM.9 All parameters are jointly significant, as reported by the F test. The Hausman specification test compares the IV estimates with the estimates of the fixed effects model presented in Table 3. The Hausman test is valid as the basis of homoscedasticity and is therefore performed under this assumption. The tests indicate that the IV fixed effects models should be preferred over the fixed effects model. A general problem in IV regressions is that of under- and over-identification. We therefore provide the Sargan and Hansen J test for overidentification and the Kleibergen-Paap LM statistics of underidentification (weak instruments). The instruments are strong enough as confirmed by the Kleibergen-Paap test. The tests for overidentification are insignificant, indicating that our instruments are not correlated with the error term of the regression. This is the relevant assumption for the validity of the chosen instruments. Other test statistics which are not presented here are in line with the reported statistics. Thus, the IV estimation is preferred over the fixed effects model.
Table 4

Fixed Effects Instrumental Variable Regression on GDP per capita for EU regions

\(\ln \left( GDP_{r} per\,capita\right)\)

IV Share

IV CD 1

IV CD 2

IV CD 3

\(\ln \left[ p_{r}\left( 1-u_{r}\right) \right]\)

0.336**

0.352**

0.342**

0.295**

 

(0.145)

(0.147)

(0.145)

(0.137)

\(\ln \left( n_{r}+g+\delta \right)\)

0.001

0.001

0.001

0.001

 

(0.002)

(0.002)

(0.002)

(0.002)

\(lag \ln h_{r}\)

0.081***

0.077**

0.081***

0.080***

 

(0.031)

(0.03)

(0.03)

(0.028)

\(lag \ln s_{kr}\)

0.068**

0.070**

0.072**

0.066**

 

(0.029)

(0.03)

(0.029)

(0.026)

smigrantsr

1.061*

1.175**

1.124**

0.882*

 

(0.55)

(0.562)

(0.56)

(0.488)

Diversity1

 

0.055

0.097**

0.128***

  

(0.039)

(0.043)

(0.04)

Region fixed effects, time fixed effects, East European countries * time fixed effects

F

158.4***

133.7***

136.1***

163.4***

RMSE

0.033

0.034

0.033

0.032

Within R2

0.837

0.835

0.839

0.85

Valid cases

496

496

496

496

No. of regions

145

145

145

145

Sargan test value

0.004

0.013

0.035

0.002

Hansen J test value

0.006

0.016

0.042

0.003

Kleibergen-Paap

8.9**

9.7***

8.9***

9.2***

Robust s.e. in (); * p < .1; ** p < .05; *** p < .01; 1 Cultural diversity measures are the Herfindahl index for IV CD 1, the fractionalization index for IV CD 2, and the polarization index for IV CD 3. Sargan test valid for the assumption of homoscedasticity

Table 5

Countries included and number of regions per year

 

 

Descriptive analysis and FE

IV FE

Data source

NUTS

2004

2005

2006

2007

2008

2005

2006

2007

2008

Variable sk

AT

1

3

3

3

3

3

3

3

3

3

EU

BE

2

11

11

11

11

11

11

11

11

11

EU

CZ

2

8

8

8

8

8

8

8

8

8

EU

DE

1

16

16

16

16

16

16

16

16

16

EU

DK

1

1

1

1

1

1

1

1

1

1

EU

EE

1

1

1

1

1

 

1

1

1

 

EU

ES

2

16

16

16

16

16

16

16

16

16

CE

FI

2

5

5

5

5

5

5

5

5

5

EU

FR

2

22

22

21

21

22

22

21

21

21

CE

HU

2

7

7

7

7

7

7

7

7

7

EU

IE

2

2

2

  

2

2

   

EU

IT

2

  

21

21

21

  

21

21

EU

LT

1

1

1

1

1

1

1

1

1

1

EU

LU

1

1

1

1

1

1

1

1

1

1

CE

LV

1

 

1

1

1

1

 

1

1

1

EU

NL

1

1

1

1

1

1

1

1

1

1

EU

PT

2

5

5

5

5

5

5

5

5

5

EU

RO

2

 

8

8

   

8

  

EU

SE

2

8

8

8

  

8

8

  

EU

SI

2

2

2

2

2

2

2

2

2

2

EU

SK

2

4

4

4

4

4

4

4

4

4

EU

UK

1

12

12

12

  

12

12

  

CE

Number of regions within each country included in the descriptive statistics and regression analysis

Data Source: CE Cambridge Econometrics, EU Eurostat

The overall picture of the estimates concerning variables which do not relate to cultural issues is unaffected by the instrumental variable approach. The lagged value of the human capital variable is now lower, indicating that the parameter was upwardly biased when the endogeneity of foreigners was not explicitly controlled for. As already mentioned in the descriptive section, the proportion of high-skilled workers and the proportion of foreigners are correlated. The proportion of foreigners was downwardly biased in the pure fixed effects model. Once we control for endogeneity, smigrantsr is no longer downwardly biased and the human capital measure is no longer upwardly biased.

With respect to content, regions offer higher incomes the more human-capital-intensive their production is. Rural regions within the EU typically do not attract much human capital because of a lack of relevant employment opportunities. Then, persistent regional income disparities are expected to be present and are a constant, long-term outcome within the EU as a result of the unequal distribution of human capital.

The parameter estimates for capital investment also become significant when the endogeneity of foreigners is controlled for. An increase in investment raises GDP per capita in the future, which is in line with the prediction of the theoretical model. Because human capital, capital formation and migration patterns are interrelated variables and the selection mechanism of potential migrants depends on income expectations, taking into consideration the endogeneity of the proportion of the foreign-born population improves the consistency of the parameter estimates of the interrelated variables in the model.

The estimate of the proportion of migrants is positive and significant when its endogeneity is explicitly controlled for. Note that the effect on GDP per capita is rather small since the proportion of migrants does not enter the regression model in log form. An increase in the proportion of migrants by 1 percentage point yields an increase in GDP per capita10 of about 0.012 for the estimates in columns IV CD 1 and IV CD 2. It is lower for IV CD 3 (0.009). This estimate is lower than that obtained by Ozgen et al. (2010), who report a value of 0.1 based on meta-analytic evidence. In our case, a 1 % increase in the proportion of foreigners in a region means a fairly large inflow of migrants at NUTS 2 level. The overall effect of immigration in EU regions is thus positive but small.

So far we have considered immigration, but what about the migrants’ region of origin, especially when it is one of the regions in our sample? First, when it is mainly employed workers that migrate, then the dependency ratio of outflow-regions will rise. This effect is captured in the \(\ln p_{r}\left( 1-u_{r}\right)\) term. An outflow of workers then results in a loss of regional income. If, however, unemployed or economically inactive people leave, then the dependency ratio will decline and the impact on regional GDP per capita will be positive. Thus, depending on the migrants’ employment status, the outflow-regions do not necessarily deteriorate. The study by Basile et al. (2010) reviews related literature. They work out that eastern European regions face a reduction in unemployment because of the outflow of people. However, they also show in their review that there is no equalisation of the unemployment level. Etzo (2011) concludes that wage differentials and unemployment levels are push factors that influence the decision for out-migration in the case of Italy. Based on the evidence from existing literature we might conclude that regions gain from out-migration.

Focusing on the diversity issue reveals that fragmentation among foreigners does not matter with regard to the Herfindahl-like index (IV CD 1). The composition becomes significant for the fractionalization index and the polarization index. As was shown in the literature review, the results depend strongly on the measures applied. The estimate of the polarization index is positive, which indicates that a culturally diverse region gains when there is a tendency towards two dominant foreign-born groups. Then, a balanced blend of foreign-born people belonging to one of the two groups seems to raise GDP per capita. According to Ottaviano and Peri (2005) different groups of foreigners provide heterogeneous products, possess different skills and possibly select into distinct jobs and tasks that suit them best. Labour resources are then distributed among jobs and tasks where they offer the highest returns (Peri and Sparber 2009). A mixture of all cultures should then be favourable, as is suggested by the estimate of the fractionalisation index (IV CD 2). Based on the results of the fractionalisation and the polarization indices we conclude that regions benefit when there is a tendency towards equalisation in group size but also when there is a tendency towards dominant groups. The estimates appear to contradict each another; but only at first sight. The reason for the results obtained is that the proportion of people from the NMS 12 and from Europe outside EU 27 rose in the European regions observed during the sample period. This increase changes the relative composition of the diversity indices. At the same time the foreign-born people from the EU 15 lost some of their relative weight, whereas the relative composition of migrants from any other region remained more or less constant. These changes in the relative composition increase overall diversity that is captured by the fractionalisation index. In addition, as the proportion of foreign-born people from regions other than those in the EU 15, NMS 12 and Europe outside the EU 27 remains constant over time, and the composition of migrants from EU 15, NMS 12 and outside EU 27 has become more balanced, the polarization index signalises the emergence of dominant groups. Network effects between foreign-born migrants and their respective home countries could be one reason for the tendency to establish dominant groups. However, the number of different cultural backgrounds within each group could be high, since we are only able to distinguish between 8 distinct world regions of origin.

The question arises as to which regions benefit from polarization. The answer is easy—every region would gain from polarization. The logic behind this result is that we estimate fixed effects models. The fixed effects model examines a change in the polarization index within a region and this identifies the parameter estimate. Then, negative effects of cultural diversity decrease, the more similar the migrants’ cultural backgrounds are. An advantage of homogeneity among migrants could be interpreted as a factor that can reduce ethnic conflict. From the perspective of the native population the presence of dominant groups might also reduce integration costs, for example by reducing language barriers more easily, by providing migrant-specific public goods, or perhaps by reducing the fear of contact, thereby facilitating the integration of foreign-born migrants into the labour market. A tendency towards dominant groups could also be a result of migrants’ networks with their respective home countries abroad and indicates that they attract further migrants. This attraction of new migrants can be seen as an advantage.

To be more precise, the dominant groups are people from the EU 15, the NMS 12 and European countries outside the EU 27. The overall proportion of the three groups together account for approximately half of the foreign-born population in European regions. This provides clear evidence of the advantage of European integration and common labour markets. Since the effect of the labour supply side and the respective increases in employment and participation rates is already captured by the \(p_{r}\left( 1-u_{r}\right)\) term, the results obtained by smigrantsr and the diversity measures relate to productivity effects. Additionally, in terms of diversity the other half of the foreign-born population, those from outside Europe, also contributes to productivity gains. This effect is captured by the fractionalisation index. Hence the conclusion we draw is that it is not exclusively intra—EU migration and diversity which matters but the overall diversity of foreign-born people present in European regions.

In the literature review various ways in which cultural diversity matters are outlined. One of the branches of literature focuses on innovation abilities (Niebuhr 2010; Ozgen et al. 2011). We therefore tested various models that take the innovation issue into account. First, we added the proportion of high-skilled migrants to the model. Second, we interacted this proportion with the human capital measure. In this case an increase in the proportion of foreign-born high-skilled workers might strengthen human capital effects. Third, we focused on the diversity of foreign-born human capital. In all cases we obtained insignificant results, even after controlling for the endogeneity of migrants. One reason for this unexpected result could be that the proportion of foreigners in the total regional stock of human capital is relatively small, so the individual impact of this group of migrants vanishes with regard to average regional GDP per capita.

Our evidence supports the hypothesis that international migration into the EU and migration between European countries increase GDP per capita on average. A regional macro-model is limited, however, and cannot separate effects for different groups within the labour force. It is still possible that a sub-group of the native population or former migrants may have disadvantages when new migrants enter the regional labour market. The impact of immigration on specific groups of the labour market can be found in Nathan (2011) and Suedekum et al. (2009) for the UK and Germany, respectively. Both studies show negative effects when foreigners and natives compete for specific jobs. Given our data and modelling approach we cannot separate effects for sub-groups of natives and migrants.

7 Conclusion

This paper investigates the impact of a culturally diverse population on regional income for EU regions including parts of the New Member States. We adopt a neoclassical approach inspired by the Mankiw-Romer-Weil (1992) model and augment this model by allowing labour heterogeneity with respect to the cultural background as suggested by Ottaviano and Peri (2005). The research question is whether a culturally diverse labour force has positive or negative productivity effects on regional GDP per capita. While controlling for regional and time fixed effects and the endogeneity of the proportion of foreign-born migrants, our estimation results suggest that per capita GDP increases when a region becomes more culturally diverse in two respects. First, we find evidence that an increase in the proportion of foreigners raises regional GDP per capita. Second, our estimates suggest that the overall diversity among foreign-born people and the tendency towards dominant groups reduces the costs of integration and thus promotes higher income. Obviously, both channels capture positive and negative effects of cultural diversity. The positive impact of dominant groups might be due to a rough classification of the cultural backgrounds and the change in the relative composition of dominant groups that are already present. It is argued that cultural similarities exist within such a rough classification which reduce the costs of integration or the costs of an efficient provision of migrant-specific public goods. However, the data also reveal that diversity in the sense of a broader mixture of people with distinct regions of origin enhances regional income. This finding is in line with the theoretical idea of a broader supply of different skills, knowledge and abilities due to differences in culture.

In the discussion surrounding demographic change, immigration is seen as a way to countervail labour shortages that may occur in the future. Compared to a scenario of no immigration, the inflow of workers compensates for labour shortages and therefore contributes positively to GDP per capita. Without immigration the participation rate and GDP per capita will decline. We provide evidence that regional income increases as a result of immigration due to productivity gains which are not captured in related variables such as the participation and unemployment rates. This leads us to conclude that there are positive income effects on average. When looking at migration within the EU, it is also necessary to consider the effects that may occur in the regions of origin due to the outflows. Depending on the labour market status of migrants before they leave, regional income might increase or decline after they depart. Then, an EU-wide policy should aim mainly to encourage people to migrate when the migrants’ region of origin can be expected to gain from the outflow. This reduces regional disparities to a certain extent. However, our findings also confirm that the costs of integration decrease in the presence of dominant foreign-born groups. We also find support for the hypothesis made by Ottaviano and Peri (2005) of a positive diversity effect among foreign-born people. Our results suggest that there are gains from immigration due to migrant-specific skills that increase productivity and thus income in addition to an increase of the stock of labour.

Footnotes
1

School enrolment rates or the number of students in higher education are frequently applied to represent investment in new human capital.

 
2

Table 5 in the "Appendix" reports the number of regions contained for each country and year.

 
3

In Table 5 we provide information, whether the data source for computing the investment rate is Cambridge Econometrics or Eurostat.

 
4

This is especially the case for Germany.

 
5

We still include respondents over the age of 65 who are active in the labour market.

 
6

The exception to this are the Baltic countries, where due to the Soviet history of these countries many Russians reside there.

 
7

The reference year for the dummy estimates is 2004.

 
8

The weight is computed as follows: \(w_{kr}=\exp \left( -\phi d_{kr}\right) ; \phi =-\ln\left[ \left( 1-\gamma \right) /D\right]\). The distance-decay parameter is γ = 0.5, which means that the proportion of migrants in the average neighbouring region (D = 167.24 km) enters with half of the original value. We standardize the weights so that the sum across all regions equals 1. In this case a weighted average of the proportion of migrants in all of the surrounding regions results.

 
9

We use the STATA xtivreg2 package, provided by Schaffer (2010).

 
10

With regard to immigration, the change in income is defined as \(dy/y=\beta \ast d\left( s_{migrants}^{r}\right),\) ceteris paribus.

 

Acknowledgments

This research is funded by the NORFACE research funding agency and part ofthe MIDI-REDIE sub-project, which we gratefully acknowledge. We would liketo thank two anonymous referees, Peter Huber (Wifo Vienna), the participantsof the 4th Wifo workshop in Vienna and the MIDI-REDIE workshop inTartu/Estonia for helpful comments and suggestions on earlier versions of this work.

Copyright information

© Springer Science+Business Media New York 2012