1. Introduction

The geographical proximity and the substantial pay gap between Germany and Poland create financial incentives for Polish nationals to seek employment in Germany. Polish workers are the second-largest group of foreign employees in Germany, with approximately 530,000 employed in 2023. They have low unemployment rates and contribute significantly to the social security system. However, Poland is currently experiencing a population decline, resulting in a growing shortage of (qualified) personnel for Polish employers. The economic situation in Poland has improved, with the country experiencing an upswing; wages are expected to rise, increasing incentives for Polish workers abroad to return-migrate (Leschke and Weiss 2023; Dziekońska 2024). In response to the rising labour demand, the Polish Government has launched return campaigns. It has also defended companies’ rights to use “wage competitiveness” to enhance mobility, rather than prioritizing improvements to the rights of Polish mobile workers. In this respect, a better understanding of the wage structure of Polish employees in Germany is needed to identify the groups most likely to return-migrate. From the German perspective, however, the key issue is how to retain foreign labour (Fuchs, Söhnlein and Weber 2021; Fuchs, Kropp and Matthes 2020).

This article makes several contributions to the existing literature. To the best of our knowledge, no other study has analysed Polish employees in Germany in such detail, with a specific focus on wages. We estimate an augmented Mincerian wage equation, drawing on a large administrative database covering workers subject to social security contributions in Germany. In addition to individual, job-related and firm-level characteristics, and unobserved firm heterogeneity (Card, Heining and Kline 2013; Brunow and Jost 2022), we focus on network- and migration-related characteristics to account for many aspects of the migration and return migration decisions (Miszczak 2012). We further differentiate between immigration groups based on year of arrival to control for the effects of policy change – specifically, the more restrictive legislation prior to 2012, in the aftermath of the eastern enlargement of the European Union (EU).¹ The effects are in line with expectations as regards individual, job-related and organizational characteristics. The evidence suggests negative network effects within occupations, particularly among lower-skilled workers who entered the German labour market from 2012 onwards. Labour market tightness is associated with significantly higher wages for Polish workers, providing wage incentives to remain in Germany.

The remainder of the article is structured as follows. Section 2 reviews related literature and identifies relevant aspects that are considered in our empirical work. Section 3 introduces the data and variables used in our study. A description of the wage structure of Polish employees in Poland and Germany is provided in section 4. Section 5 presents the estimates of an augmented Mincerian wage equation. These are discussed together with an analysis of subgroups in section 6. Section 7 concludes.

2. Immigration legislation, networks and related literature

After the Second World War, Germany developed long-standing immigration flows from various countries (Brunow and Jost 2021a). Until 1989, Germany was divided along the “Iron Curtain”, and migration between East and West Germany was almost impossible. However, while migration from Poland to West Germany was restricted, it was occasionally permitted – mainly for political reasons. Migration was also possible for the group known as Aussiedler (BVA 2023) – ethnic Germans living abroad who held non-German citizenship. Now referred to as Spätaussiedler, when they immigrate to Germany, they are usually granted German citizenship. Additionally, bilateral agreements between West Germany and Poland allowed certain temporary work contracts before 1989 (Hönekopp 1997). Accordingly, before 1989, there was a steady flow of emigrants from Poland to other countries, with West Germany as the main destination.

With the fall of the Iron Curtain in 1989 and Germany’s reunification, Polish workers were able to emigrate to both parts of Germany. Although improved economic conditions in Germany provided a financial incentive for emigration, Poland was not yet a member of the EU, which meant that emigration from Poland to Germany was generally limited to skilled and highly skilled workers, with some exceptions for unskilled workers (Fiałkowska and Piechowska 2016; Green 2013; Becker and Heller 2009). According to Miera (2008), after arriving in Germany on tourist visas, many Polish workers entered into informal arrangements, particularly in the cleaning, caretaking and construction sectors.

Data show an emigration-to-return-migration ratio of about 18 in the 1990s, declining steadily to about 2.5 in 2010, before rising again to around 4 in 2014.² Since the COVID-19 pandemic, flows have been relatively balanced, with the ratio ranging from approximately 1.1 to 1.5. Following Poland’s integration in the EU and its access to the free movement of workers, the German labour market has been fully open to all Polish workers since 2012, without legal restrictions. As a result, the share of unskilled Polish workers (without recognized vocational training) in Germany has increased substantially. Notably, between 2010 and 2023, the number of people commuting from Poland to Germany for work rose from approximately 3,900 to 94,000 (Seibert 2024). Today, Polish nationals comprise 1 per cent of Germany’s population and 1.5 per cent of total employment, indicating strong integration into the labour market and low unemployment rates.

From a theoretical perspective, (international) migration occurs when individuals decide to leave their current country of residence to live and work in another country. The motives for emigration can be roughly grouped into forced and voluntary movements. In this article, we consider voluntary movements, particularly economically motivated emigration. Potential migrants compare wages in their country of origin with wages in the potential destination country, under conditions of uncertainty, and decide to emigrate if the expected real wage abroad exceeds the real wage at home, taking migration costs into account (Borjas 2024). Therefore, a better understanding of immigrants’ wage structures in the destination country is essential to understanding migration decisions. Empirical studies often focus on the wage gap between natives and immigrants to identify differences in wage-setting behaviour and potential discrimination (e.g. Brunow and Jost 2021a and 2021b and 2022; Lehmer and Ludsteck 2011 and 2015). From the emigrant’s perspective, such a pay gap does not influence the migration decision, which is based on the wage differential with their country of origin. Other price differentials, such as the cost of living, are also taken into consideration.

An augmented Mincerian wage equation provides a robust approach to analyse characteristics that lead to wage differentials between individuals in general (Heckman, Lochner and Todd 2006). Furthermore, individual characteristics and aggregated characteristics – such as those at the firm or regional level – help explain individual wage differentials. Dostie et al. (2023), as well as Lehmer and Ludsteck (2015), Brixy, Brunow and Ochsen (2022) and Brunow and Jost (2021a and 2022), provide evidence that firm-level characteristics significantly account for wage differences between employees across firms.

Lee (1966) highlights the role of push-and-pull factors, such as differences in job opportunities, levels of unemployment and amenities in origin and destination countries, which are frequently cited as determinants in individual migration decisions. Brunow, Nijkamp and Poot (2015) review several economic migration theories, examining the drivers of migration and the resulting impact on both origin and destination economies, which ultimately influence the push-and-pull factors and incentives to migrate. Iranzo and Peri (2009) present a theoretical and simulated case to show that the emigration of better-skilled workers from Eastern European countries leads to an increase in welfare for both the Western European destination countries and the origin countries. The movement of workers in a particular segment of the labour market may lead to higher competition in the host country, whereas the labour market in the country of origin becomes tighter (i.e. unemployment falls). Thus, expected wage reactions might be either positive or negative for native workers in the destination country (Amior and Manning 2021; Andersson, Eriksson and Scocco 2019; Brunow, Nijkamp and Poot 2015) and in the country of origin (Acuna 2023). Additionally, the individual incentives of an immigrant to stay abroad or to return-migrate depend on wage differentials (Dustmann 2003). These incentives are also influenced by decisions to invest in host-country-specific human capital (Adda, Dustmann and Görlach 2022; Wahba 2022), by acquiring language skills and making efforts to assimilate, among others (Diehl, Fischer-Neumann and Mühlau 2016). It may be assumed that, over time, immigrants tend to increase their investments in host-country-specific human capital, which may allow them to access higher wages.

Employers’ wage-setting behaviour is influenced by regional labour market conditions. The so-called “wage curve” describes the inverse relationship between the regional unemployment rate and regional and individual wages (Blanchflower and Oswald 2008; Blien et al. 2024; Brunow, Lösch and Okhrin 2022). We consider whether the wages of Polish workers increase when the German labour market tightens, which would indicate that immigrants benefit from regional tightness. Lower unemployment rates in the host region may be seen as another pull factor, according to Lee (1966). Rokicki et al. (2021) identify a Polish wage curve whereby Polish regions with lower levels of unemployment enjoy higher wages, leading to a reduction in the incentive to emigrate. For the period 1998 to 2007, Dustmann, Frattini and Rosso (2015) show that Polish emigration led to a slight increase in the wages of middle- and high-skilled workers, and a potential income loss for less-skilled workers in Poland, which again changes (group-specific) incentives to emigrate.

Epstein (2008) highlights network and herd effects as key drivers of migration to specific regions. Network effects tend to be weaker in rural than in urban regions, owing to the lower density of potential employees and the lower quality of social capital or technologies available (Dietz 1999; Longhi et al. 2002). Regional economic networks are often described as connections between firms, their interaction possibilities and their shared information and knowledge base (Miszczak 2012). However, a link also exists between firms, their employees and potential migrants. The strength of connections between different groups of workers influences both migration prospects and employment opportunities. The most substantial connections are found among ethnic groups and women (Glitz 2017; Kracke and Klug 2021). Glitz (2017) and Ioannides and Loury (2004) analyse co-workers. When two workers collaborate, the likelihood that they will recommend each other increases, particularly if one of them becomes unemployed. Family, friends and ethnic groups can also provide information hubs (Cappellari and Tatsiramos 2010; Goel and Lang 2010; Loury 2006), influencing the migration decision. Andersson, Larsson and Öner (2021) provide evidence for Sweden that ethnic enclaves are associated with higher self-employment rates when many individuals of the same ethnicity are self-employed. However, they also find a negative effect on self-employment incentives as the size of the enclave increases, potentially because people can find employment within their ethnic group. An influx of workers into a specific region can lead to either lower or higher wages, depending on the qualification structure, diversity and competition among workers, and on the segment of the labour market affected (Suedekum, Wolf and Blien 2014; Piyapromdee 2021).

In 2016, the Polish company Work Service conducted a survey among Polish individuals regarding their incentives to emigrate (Work Service 2016).³ The most frequently cited incentives were higher wages abroad, the lack of suitable jobs in Poland, and better prospects for professional development in other countries. Most respondents who planned to emigrate intended to work in the construction industry, and 12 per cent wanted to run their own business abroad. Other motivations included the opportunity to travel and live abroad, improved living standards and the desire to be united with family or friends. Better social conditions and a more favourable tax system were also mentioned. The greatest disincentive to emigration was attachment to family and friends living in Poland. Another reason to stay was limited foreign language proficiency. Lastly, holding a desirable job in Poland was strongly associated with decisions to remain.

Based on theoretical arguments and the literature review, we expect the wages of Polish workers in Germany to be higher when they reside and work in Germany relatively longer, have higher skill levels and arrived prior to gaining full access to the German labour market in 2012. Furthermore, based on the wage curve literature, we expect wages to be higher in regions with lower unemployment rates. Lastly, wages may vary depending on the strength and direction of network effects, which are likely to differ across regions.

3. Data

We make use of the Integrated Employment Biographies (IEB), a database provided by the German Institute for Employment Research (IAB). This administrative database covers information on all workers in Germany who are subject to social security contributions, are marginally employed⁴ or are unemployed. It excludes civil servants and self-employed workers, representing a slight limitation for our research: Polish firms operating in Germany often employ Polish workers registered in Poland, and we know that some Polish workers in Germany are self-employed. Although we lack data on the extent of self-employment, since the IEB covers approximately 92 per cent of the German labour market, we consider the employment structure of Polish workers subject to social security contributions in Germany to be sufficient for our research hypotheses. We use a 10 per cent sample of the IEB, considering full-time employees in a cross-sectional analysis for 2018. Marginal and part-time employees are excluded owing to the absence of information on hours worked and the fact that wage is recorded on a daily basis. Including these groups would introduce too much heterogeneity, compromising the validity of our wage analysis. Although there was a steady influx of Polish nationals to Germany before 1989, our dataset includes only 503 individuals who entered Germany during that period. We explain this low number by the fact that (Spät)Aussiedler were granted German citizenship and, therefore, cannot be identified as immigrants in the data. Accordingly, we restrict our analysis to Polish immigrants who arrived from 1991 onwards. In total, our sample comprises more than 64,000 Polish nationals employed in Germany, including border commuters.

The data include individual characteristics, such as age and sex, educational and vocational attainment; job-related characteristics, such as occupation and the task level; and information on industry and region. We used data on employment and unemployment spells to construct indicators related to individual worker performance. Seasonal Polish workers are not eligible for unemployment benefits during periods of work interruption in Germany, when they typically reside in Poland. In such cases, the workers temporarily disappear from the data. We address this issue using another control variable. A unique firm identifier exists for each employee, allowing us to examine individual wage differentials in relation to firm characteristics (Dostie et al. 2023; Brunow and Jost 2021a and 2022). Regional classification is based on NUTS3 regions,⁵ which we aggregate into 151 regional labour markets, following Eckey, Kosfeld and Türck (2006). According to a classification based on accessibility and population density,⁶ each region is classified into one of three categories: urban, semi-urban or peripheral region. Lastly, as outlined in table 1, we consider a set of characteristics at various levels of aggregation to account for individual wage differences.

4. The employment structure of Polish employees

In this section, we examine the employment structures of Polish employees in Poland and Germany and compare them with those of German employees. Table 2 presents distributions by sex, age, level of education and occupation. Polish employees in both countries are younger than their German counterparts. Institutional differences between the German and Polish education systems exist and are taken into consideration in this analysis. For example, university qualifications are more prevalent among employees in Poland, whereas vocational training is more common in Germany. Fewer Polish employees in Germany hold a vocational qualification than their German counterparts, and they are less likely to have a university qualification. Occupational distribution varies across all three groups. The shares of Polish employees in Germany in craft and related trades and in elementary occupations are significantly higher than among the two other groups. Conversely, the proportion of managers and professionals is considerably lower among Polish employees in Germany than in the non-migrating groups.

Figure 1 shows the regional distribution of Polish employees in Germany. Panel A shows the share of Polish employees in total employment. It becomes evident that the eastern German border regions particularly benefit from immigration and commuting from Poland. Higher shares are also observed in Bavaria (around Munich), other economic centres around Frankfurt–Mannheim, and the peripheral regions of the northwest. Panel B presents the share of Polish employees among all foreign employees. Interestingly, the overall proportion of foreign employees in eastern Germany is low, which results in a relatively high share of Polish employees. Given the comparatively weaker economic structure in East Germany,⁷ we expect that many Polish employees are seasonal workers in agriculture, tourism and construction. Lastly, panel C displays the distribution of all Polish employees across Germany. It indicates an uneven distribution, suggesting a tendency towards regional concentration and the possible presence of network effects.

Figure 1. Distribution of Polish employees in Germany, 2018 (percentages)

Source: Our own calculations based on statistics of the German Federal Emploment Agency; maps created with RegioGraph 10.

Table 3 presents monthly gross salaries by industry. Notably, Polish employees in Germany earn, on average, almost twice as much as employees in Poland, with the gap ranging from 1.4 times the wage in mining and quarrying to 3.1 times in education. However, the wages of Polish employees in Germany are considerably lower than those of their German counterparts: Germans earn 12 per cent more in agriculture, forestry and fishing and 64 per cent more in manufacturing. A significant portion of this gap can be attributed to the fact that half of the Polish employees in Germany perform unskilled tasks – even though they may hold vocational degrees. Nevertheless, most of the unskilled jobs in Germany offer Polish workers higher wages than equivalent jobs in Poland.

Lastly, figure 2 reports the 2018 wage distribution in Poland by voivodeships (the country’s administrative regions). Wages were higher in the industrial centres and especially in the Warsaw region. The lowest value was recorded in the Warmińsko-Mazurskie region. In most regions, average wages were less than €1,100. Thus, there is a clear financial incentive to work in Germany (see table 3).

Figure 2. Average monthly gross salaries in Poland by region, 2018

Notes: Regional classification based on NUTS3 regions (NUTS 2021).

Source: Statistics Poland; German Stock Market.

5. Regression analysis

The analysis that follows identifies characteristics relevant to the wage-setting and earnings of Polish workers in Germany. We perform a detailed subgroup analysis to identify immigration group differences caused by different immigration legislations. We consider the characteristics outlined in table 1, estimate a Mincerian wage equation using ordinary least squares (OLS) regressions and compute robust standard errors. We report the results of the control variables of this regression in table 4, while table 5 shows the results of the focus variables. The first columns show the reference model, while the second columns show the results for an interaction model in which the effect of the network-related characteristics may differ depending on the regional type. The third and fourth columns present the respective estimates for German workers for comparison. Although the difference between Polish and German workers is not the focus of our analysis, providing this comparison alongside a survey regarding emigration incentives in Poland may help us better understand the migration decisions of Polish workers.

The estimates for the control variables in table 4 are robust, although some group-specific differences exist. There is almost no coefficient variation between the reference and interaction models, indicating no strong correlation between additional regressors and the reference variables. The effects are generally greater for German employees, indicating a wider wage dispersion between German employees relative to Polish employees. This also indicates relatively higher returns to characteristics for Germans. Brunow and Jost (2021b) highlight that flatter experience curves are related to unskilled tasks and that the accumulation of German-specific knowledge leads to wage convergence.

A detailed analysis of the results indicates that, relative to unskilled labour, wages are higher for individuals performing skilled tasks or working as specialists or experts. We also observe a wage premium for those holding a vocational training certificate or university degree, occupying a leadership position or having acquired further qualifications. There is a slight positive age effect, although the age–experience curve is flatter for Polish employees than for their German counterparts. After controlling for other characteristics, Polish women earn approximately 7.4 per cent less than Polish men, whereas German women face a pay gap of more than 15 per cent. The longer an individual has been registered as unemployed, the greater their wage loss relative to someone rarely unemployed. Polish employees who arrived in Germany during the period of immigration restrictions (1991 to 2011) earn 10 per cent more than those who arrived afterwards. This is conditional on other characteristics of individual experience, such as task profile, leadership position, vocational attainment and occupational selectivity, and provides further insight into the human capital endowments of this group. This wage advantage may reflect better German language proficiency, greater German labour market experience and stronger negotiating power vis-à-vis employers. As regards seasonal work, wages tend to decrease as the share of time without observations in German data increases (i.e. the person was neither registered as employed nor unemployed), suggesting limited learning and human capital accumulation. In addition, Polish workers earn higher wages when employed in larger firms and firms with greater human capital. Lastly, unobserved firm-level heterogeneity is positively correlated with individual wages, suggesting that better-performing firms tend to pay higher wages to both German and Polish workers.

The estimation results of the focus variables are presented in table 5. Column 2 shows that the interaction effects on the regional distribution of Polish employees across occupations are statistically insignificant, indicating no specific regional wage-setting differences within occupations. Employment in East Germany is associated with approximately 1.6 per cent lower wages. Wages also respond to variations in the regional occupation-specific unemployment rate: the higher the unemployment rate, the lower the wages. At the regional level, an increase in the proportion of human capital within an occupation is associated with higher wages for Polish workers. We also find a positive association between the share of foreigners within the occupation and region and wage levels. Given that this effect is conditional on the regional occupation-specific unemployment rate, this estimate may reflect labour shortages in Germany and suggest that employers may offer higher wages to attract immigrant workers to certain regions.

Regarding potential occupation-specific clustering, there is no evidence that a higher concentration of employees in an occupation within a given region is associated with higher wages. However, when a specific occupation accounts for a larger share of total regional employment, wages tend to increase. This suggests that the occupation holds relative importance in the region, in line with the hypothesis of regional clustering. Analysis of the network-related characteristics yields a surprising result: a negative association between the number of Polish employees in a region and wage levels. This contradicts the migration incentives mentioned earlier, suggesting that this wage-decreasing effect is specific to Poland. There is regional heterogeneity in this regard: the negative network effect is strongest in urban regions (–0.02), less pronounced in semi-urban regions (–0.012) and weakest in peripheral regions (–0.007). We observe an additional wage reduction – regardless of regional type – when the number of Polish workers employed in a region within an occupation increases. These negative effects relate to network opportunities specific to Poland. They may correlate with other semi-urban-related variables and characteristics of the occupation-specific regional environment. Excluding these Poland-specific variables does not alter the results reported thus far, indicating no correlation with other variables and suggesting a potential bias in the overall estimates. Interaction and mediating effects between the sets of characteristics do not yield further insights. More detailed data – potentially collected through a survey – would be needed to explore this issue further.

The difference in parameters between the reference and interaction models shows minimal change in the effects of the control variables and their interpretation. Accordingly, there is no significant correlation between the network characteristics and these control variables. Given that the interaction model differentiates the strength of potential network effects by the type of region in which Polish workers are employed, it seems evident that network effects are not influenced by differences in the qualification structure, age or sex of Polish workers. Rather, the strength of network effects on wages depends on the region of employment, not on the individual characteristics of Polish workers.

6. Discussion and subgroup analysis

Compared with Germans workers, Polish employees show lower returns on their characteristics. The Oaxaca–Blinder (OB) decomposition is commonly used to characterize the explained and unexplained components of the pay gap. The clearly lower returns for Polish workers point to a pronounced coefficient effect, which may be interpreted as discrimination.⁸ However, the coefficient effect should be interpreted carefully, as the identification of slope parameters in the OLS regression – which are needed to calculate the coefficient effect – is not independent of the estimate of the constant term. The constant term relates to the reference group in the regression and the coefficient effects measure the deviation relative to the average wage of that group. Based on the interaction models, the predicted average gross daily wages for the reference groups are approximately €53.78 for Germans and €61.56 for Polish employees.⁹ Several descriptive statistics confirm relatively higher wages, particularly among unskilled Polish workers. This suggests that the wage distribution among Polish workers is more equal than among German workers, and wage differentials based on individual characteristics are smaller. The wage profile is flatter for Polish employees, confirming the findings of Brunow and Jost (2021b). This evidence on the expected wage structure of German employees is consistent with the results of the survey conducted among Polish workers in Poland. The majority of Polish respondents indicated that the wages of nationals in a potential destination country are not a major concern, provided that their own wage in that country is sufficiently higher than the equivalent wage in Poland.

Network-related characteristics indicate a relative disadvantage in contexts of specific regional competition. Accordingly, the advice for a potential migrant considering employment in Germany would be to identify a region and, within that region, an occupation with a low concentration of Polish workers. However, this conclusion may be relevant for some but not all subgroups.

Table 6 considers several subgroups of Polish workers in Germany. The first column reports the reference model discussed above for comparison. The second column considers only skilled workers, specialists and experts, while the third column examines unskilled employees. Based on sex-specific occupational differences, the fourth and fifth columns restrict the sample to women and men, respectively. Lastly, the sixth and seventh columns present the results for Polish employees who first started working in Germany between 1989 and 2011, and those who arrived from 2012 onwards.

Relative to urban areas, wages in peripheral regions are significantly lower for unskilled workers, men and individuals who immigrated to Germany from 2012 onwards. Furthermore, these groups are more exposed to the unemployment rate and, to some extent, the share of foreign workers in the region and occupation. This suggests higher competition and potentially lower wages. Skilled workers and migrants who arrived in Germany before 2012 benefit most from a regional occupation-specific concentration, indicating that their skills are in demand and rewarded with higher wages. However, their wages are relatively lower when competition in the region and occupation is higher owing to the presence of other Polish workers. For skilled workers, the negative effect is strongest in semi-urban regions (–0.422), waning in urban areas (–0.257) and weakest in peripheral regions (–0.137), suggesting that potential skilled worker shortages lead to a less pronounced wage penalty among skilled workers. For workers who arrived in Germany before 2012, we observe a negative effect in urban areas (–0.158), but there is no further region-specific pattern, as the interaction effects for semi-urban regions and rural areas are insignificant. Nevertheless, the fact that the effect remains negative for these two groups, indicates the predominance of competition among Polish employees.

In general, we can conclude that Polish employees do not benefit from potential network effects among Polish immigrants. Those who immigrated before 2012 were subject to strict restrictions for entering the German labour market; it is, therefore, not surprising that they are less negatively affected by the presence of other Polish workers in their region and occupation. However, those who had full access to the German labour market from 2012 onwards are negatively affected by a spatial concentration of Polish employees. Regional occupation-specific labour market performance also matters, especially for unskilled workers, men and immigrants arriving after 2011, who are more strongly affected by higher unemployment rates. However, they partially benefit from higher shares of foreigners. This provides a strong indication that these regions and occupations are characterized by labour shortages associated with higher wages and are, therefore, attractive potential destinations for immigration.

Given the differences in returns to characteristics between German and Polish workers, it is evident that the Polish wage profiles are flatter and returns to education and skill-specific characteristics, for example, are lower. A statistical test on differences in parameters supports this finding. Performing the OB decomposition, therefore, suggests a pronounced coefficient effect,¹⁰ and thus unequal returns to characteristics (results not shown).

7. Conclusion

The geographical proximity between Germany and Poland and the substantial pay gap between the two countries creates financial incentives for Polish workers to seek employment in Germany. In the German labour market, Polish workers are the second-largest foreign employment group, accounting for over half a million employees. Both Poland and Germany face a population decline and labour shortages, which may lead to wage increases in both countries. Higher wages in Poland may increase the incentive for Polish workers in Germany to return-migrate, even if German wages remain relatively higher. To better understand the wage structure of Polish employees in Germany and to identify the groups most likely to return-migrate, we have considered an augmented Mincerian wage equation using a large administrative database covering workers subject to social security contributions in Germany. We focus on network- and migration-related characteristics, as these are relevant not just for the decision to emigrate to Germany but also for the decision to return-migrate to Poland. Our results show that the returns to individual characteristics – such as age, sex and education – follow the expected direction. However, the effects for Polish workers are generally weaker than those for Germans, suggesting that Polish wage profiles are flatter and returns to education and skill-specific characteristics, for example, are lower.

We have further provided evidence that wages are negatively affected by the size of the Polish workforce in a region. If the size of the Polish workforce in the same occupation increases, wages are relatively lower. We have interpreted this as indicating a relatively strong competition effect among Polish workers. A detailed subgroup analysis reveals that Polish workers who moved to Germany before 2012, when the German labour market was not yet fully open, are less affected by the size of the Polish enclave. They were subject to strict legislation and were therefore relatively better skilled, often holding at least a vocational training degree. Since 2012, less-skilled workers have made up a substantial share of the influx of Polish immigrants to Germany. Thus, the opening of the labour market and the change in policy have led to a shift in the qualification structure of immigrants. Our findings also suggest that a better labour market performance – characterized by lower unemployment rates – leads to substantive wage increases for Polish employees in Germany, indicating that tight German labour markets offer higher wages not just to natives but also to foreigners. Although we have observed considerable variation in characteristics, our data comprise a cross-section of individuals. Consequently, only static effects can be analysed directly; dynamic effects may differ. As we have shown, since 2012, following the unconditional opening up of the German labour market, the share of unskilled Polish workers has increased substantially. This shift in the workforce composition impacts the wage structure of Polish workers in Germany. A subsequent dynamic change in the composition of the Polish workforce would be likely to lead to a dynamic adjustment of the wage structure and of the overall impact of Polish workers on the German economy. Our results show the expected signs, identified using between-individual variation.

This brings us to the conclusion that German employers seeking to retain their Polish employees may consider aligning wages with those offered to German nationals to offset incentives for return migration, particularly if the Polish labour market catches up. For Polish individuals seeking to work in Germany in a specific occupation, the best advice would be to identify a region with a low concentration of Polish competition, as this may result in relatively higher wages. A complete comparison of Polish wages in Poland with the wage structure of Polish workers in Germany may provide additional insights in future research.

Competing interests

The authors declare that they have no competing interests.