How do regional labor markets adjust to immigration? A dynamic analysis for post-war Germany

https://doi.org/10.1016/j.jinteco.2020.103416Get rights and content

Highlights

  • After World War II, eight million forced migrants arrived in West Germany.

  • We study how regional labor markets adjusted to this inflow.

  • We document empirically long-lasting adjustment dynamics after the inflow.

  • We develop a dynamic equilibrium model that closely fits empirical facts.

  • Unemployment differences and regional migration characterize adjustment process.

Abstract

This paper provides a comprehensive analysis of the dynamic labor market effects of one of the largest forced population movements in history, the mass inflow of eight million German expellees into West Germany after World War II. The expellee inflow was distributed very asymmetrically across two West German regions. We develop a dynamic equilibrium model that closely fits two decades of historical data on the regional unemployment differential and the regional migration rate. Both variables increase dramatically after the expellee inflow and decline only gradually over the next decade. The long-lasting adjustment process implies losses in the lifetime labor income of native workers that are not covered by conventional steady state analyses. Regional migration serves as an important adjustment margin for native workers to insure against local labor supply shocks.

Introduction

This paper studies how regional labor markets in West Germany adjusted to one of the largest forced population movements in history, the mass inflow of German expellees after World War II. The eight million expellees who had arrived in West Germany by the end of 1949–most of whom from the territories that Germany relinquished after the war–were distributed very unevenly across the country. The share of expellees in the population ranged from 3.1% in the federal state of Rhineland-Palatinate to 33.3% in Schleswig-Holstein. We exploit this large, unexpected, and highly uneven inflow of expellees to study how quickly and by what margins regional labor markets adjusted to the inflow.

Our paper contributes to an extensive literature that analyzes labor market adjustments to immigration.1 Over the past two decades, interest in the issue has been fueled by sharply rising numbers of international migrants and public concern about the consequences of immigration. The recent surge in forced migration and the growing inflow of refugees into Europe have only added to these concerns. Despite strong public interest, the burgeoning literature on immigration has hardly addressed two major aspects of the adjustment process (Borjas, 2014): the length of the adjustment process and the relative importance of different adjustment margins along the adjustment process.2 Our paper addresses these two aspects.

We start our analysis by presenting novel empirical facts on how regional labor markets in West Germany adjusted to the inflow of expellees, drawing on administrative data from 1939 to 1970. We derive these facts by contrasting the economic development of two stylized regions: a high-inflow region H that consists of what were known as the “refugee states” of Bavaria, Lower Saxony, and Schleswig-Holstein, and a low-inflow region L that consists of the remaining West German states. The two regions had similar unemployment and population growth rates before World War II. However, during and after the war, much more expellees fled or were transferred to region H than to region L, largely because region H is located much closer to the homelands of expellees than region L. We compare economic developments in the two regions before the inflow and document their subsequent relative development until 1970. We argue that this comparison is informative about the adjustment path because regional differences in initial inflow rates were large and exogenous to local labor market conditions.

Our empirical facts reveal that by 1950, the unemployment rate in region H was as high as 16.7% and exceeded the unemployment rate in region L by a factor of two and a half. Regional unemployment rates then gradually converged during the 1950s and eventually fell to around 1% by the early 1960s. The decrease and convergence in regional unemployment was accompanied by large regional migration flows. In 1950 alone, net migration from region H to L amounted to 1.9% of the population in region H.

These large migration flows diffused the labor market effect of the expellee inflow from region H to region L and re-equilibrated local regional labor markets. Consequently, a direct comparison of labor market outcomes between the two regions will underestimate, except on impact, the true causal effect of receiving high rather than low expellee inflows even if the initial distribution of expellees was exogenous to local labor market conditions. After all, such comparison would show little or no differences because, over time, the high inflow of expellees into region H affected not only region H itself but also region L.3

To interpret the empirical pattern, and isolate the causal effect of the expellee inflow along the adjustment path, we develop a parsimonious dynamic equilibrium model that accounts for regional migration and its effect on regional labor markets. The backbone of this model is a two-region search and matching model of unemployment where workers take recurrent and forward-looking migration decision between the two regional labor markets. In each region, a representative firm employs many workers and accumulates capital to produce output. The firm faces costs when it adjusts its workforce or capital stock.

To calibrate the model to historical data, we exploit both the cross-section and time dimension of the data. We show that the calibrated model's adjustment dynamics after the asymmetric historical expellee inflow closely fit our empirical facts. Using the calibrated model, we find that it takes regional labor markets about a decade to absorb the expellee inflow. Regional migration played a crucial role in the adjustment process. By 1965, more than one-third of the initial increase in region H's population was absorbed through migration to the low-inflow region L. The model also suggests that migration was more important than capital accumulation in narrowing the gap between regional capital-population ratios, created by the asymmetric expellee inflow.

The adjustment process of the native population differs strongly from the process of the population as a whole. The employment probability of native workers decreases in the first few quarters of the adjustment process, whereas the employment probability of the average worker increases monotonically throughout. The negative effect of the expellee inflow on native employment is largest ten quarters after the arrival of the expellees. When measured at that time, native employment decreases by 4.33 workers for every ten expellees who arrive in region H. Of those 4.33 native workers, 1.30 leave the labor force, 2.42 enter the unemployment pool, and 0.61 leave region H for region L.

Our calibrated model also allows us to make inference on income losses of natives, for which, unfortunately, we have no data. We find that the expellee inflow decreased the expected discounted lifetime labor income of the average native worker in West Germany by 1.51%. The short-term decline, measured by per-period labor income, is much larger and reaches 5.91% nine quarters after the shock. Regional migration is an important adjustment margin for native workers to ensure against income losses arising from local expellee inflows. The short-run wage elasticities, which in our calibrated model are between −0.17 and −0.25, are in line with empirical estimates that we produce based on a large-scale survey of employees' earnings in trade and industry. All these results survive a battery of checks, in which we check for the–potentially confounding–influence of pre-existing differences between regional labor markets.

Related literature. Our finding of prolonged adjustment processes after the expellee inflow complements a nascent literature that studies the dynamic wage effects of immigration. Cohen-Goldner and Paserman (2011) analyze the impact of the inflow of more than one million Soviet Jews into Israel after the collapse of the Soviet Union. The authors show that the initially negative wage effect of the inflow dies out after five to seven years. Edo (2020) studies the sudden inflow of repatriates from Algeria to France in 1962. He finds that wages fell between 1962 and 1968 and then returned to their pre-shock level in 1976. Cohen-Goldner and Paserman (2011) and Edo, 2020 exploit variation in the migrant share across labor market segments, assuming that these segments are isolated from each other. In contrast, our structural approach directly accounts for movements between regional labor market segments and therefore allows us to quantify the role of regional migration as an adjustment margin.

In recent work, Colas (2018) considers a dynamic equilibrium model of local US labor markets to study dynamic wage and income effects of immigration. He finds that the wage effect of immigration is cut in half ten years after the shock, as natives move away from high inflow locations. In a related setting, Monras (2020a) studies local labor market adjustments in the US, including internal migration and its effect on life-time income, induced by the Great Recession.4 When treating this event as a permanent change in local productivity, he finds that adjustments in most locations are completed within a decade. An important difference between our paper and both Colas (2018) and Monras (2020a) is the calibration strategy. While Colas (2018) and Monras (2020a) calibrate their models to moments of the distribution of workers obtained from cross-sectional or panel data, we calibrate the model directly to the adjustment dynamics following the expellee inflow. We also consider a natural experiment involving a one-time inflow of exceptional size.

Braun and Mahmoud (2014) is, to the best of our knowledge, the only other study that analyzes the labor market effects of our specific episode of forced migration.5 The authors demonstrate that expellees had a substantial negative effect on native employment in 1950. In contrast to our paper, Braun and Mahmoud (2014) focus on the short-term effect of the expellee inflow and do not quantify the relative importance of different margins through which the West German economy adjusted over time to the expellee inflow.

Another related literature studies the link between immigration and subsequent internal migration of native workers. Some studies find that native workers indeed respond to immigration by moving out to other areas (see, for example, Filer, 1992, Borjas, 2006 or Boustan et al., 2010), while other studies find no such effect (see, for example, Card and DiNardo, 2000, Card, 2001 or Kritz and Gurak, 2001). However, this literature generally abstracts from adjustment dynamics.

Our paper also contributes to an emerging literature that studies the effect of immigration within search and matching models. Ortega (2000) studies a two-country model, in which unemployed workers decide where to search for a job. Chassamboulli and Palivos (2014) analyze the effects of US skill-biased immigration, while Liu (2010) and Chassamboulli and Peri (2015) focus on illegal immigration into the US. Battisti et al. (2018) analyze the welfare effects of immigration on low-skilled and high-skilled natives in 20 countries. Our work differs from these papers in three main respects. First, we focus on the adjustment dynamics triggered by immigration rather than on the steady-state effects of immigration. Second, we study the role of regional migration within a country as an adjustment margin to immigration. Third, we calibrate key model parameters using data from a natural experiment.

This paper proceeds as follows. Section 2 provides background on our historical setting, and Section 3 derives empirical facts on how regional labor markets in West Germany adjusted to the expellee inflow. Section 4 develops the dynamic equilibrium model that we use to analyze the historical data. Section 5 explains the calibration of model parameters and assesses the model fit. Section 6 contains our main results on the channels through which regional labor markets adjusted to the expellee inflow and the associated income effects for native workers. Finally, Section 7 concludes.

Section snippets

Historical background and nature of the expellee inflow

Below, we shall refer to those territories east of the present-day eastern border of Germany that were part of the German Reich before World War I as eastern territories (see Fig. 1 for an overview of Germany's territorial losses between 1919 and 1945). We shall refer to the Federal Republic of Germany as West Germany and to the German Democratic Republic as East Germany (again, see Fig. 1).

Empirical facts on regional development

We now present novel empirical facts on how regional labor markets in West Germany adjusted to the large and asymmetric inflow of expellees. We derive these facts by comparing the demographic and economic developments of two stylized regions, a high-inflow region H and a low-inflow region L, between 1939 and 1970. We thus only exploit the variation in the expellee inflow across regions, in line with recent recommendations in Dustmann et al. (2016).8

A dynamic model of regional labor markets

The backbone of our model is the textbook Diamond-Mortensen-Pissarides search and matching model of unemployment. We depart from a version of this model with endogenous labor force participation and extend it in two directions that are motivated by the particular historical episode that we study.16

Model fit and capital versus migration as adjustment margins

We choose the parameters of the model in three steps. In the parametrization step, we set initial conditions and expellee inflow rates to historical values and a first set of parameters to values conventional in the literature. In the second step, we calibrate a second set of parameters by targeting steady-state values of endogenous variables. In the third step, we calibrate the remaining eight parameters by minimizing the distance between the model's adjustment dynamics after the expellee

Main results on the labor market effects of the expellee inflow

We have shown that our parsimonious model explains the empirical facts surprisingly well. In this section, we hence use this model to address our key research questions. Subsection 6.1 discusses how quickly and by what labor margins regions H and L adjust to the expellee inflow, and Subsection 6.2 quantifies the effect of the inflow on native labor income. Appendix D shows that our main results are robust to various changes in the model calibration. In particular, the robustness checks allow

Discussion and conclusion

This paper has analyzed how regional labor markets in West Germany adjusted to the inflow of eight million expellees after World War II. Three key findings emerge. First, it took regional labor markets more than one decade to absorb the expellee inflow. Second, the adjustment process was characterized by large differences in regional unemployment rates and strong migration from the high- to the low-inflow region. Third, the large and long-lasting adjustment dynamics in regional labor markets

Declarations of interest

None.

References (89)

  • G.H. Hanson et al.

    Labor-market adjustment in open economies

    J. Int. Econ.

    (2002)
  • J. Lafortune et al.

    More hands, more power? Estimating the impact of immigration on output and technology choices using early 20th century US agriculture

    J. Int. Econ.

    (2015)
  • X. Liu

    On the macroeconomic and welfare effects of illegal immigration

    J. Econ. Dyn. Control.

    (2010)
  • D. Lkhagvasuren

    Big locational unemployment differences despite high labor mobility

    J. Monet. Econ.

    (2012)
  • A. Aydemir et al.

    Attenuation bias in measuring the wage impact of immigration

    J. Labor Econ.

    (2011)
  • O. Bargain et al.

    Comparing labor supply elasticities in Europe and the United States

    J. Hum. Resour.

    (2014)
  • G. Basso et al.

    Immigrants, labor market dynamics and adjustment to shocks in the euro area

    IMF Econ. Rev.

    (2019)
  • M. Battisti et al.

    Immigration, search, and redistribution: a quantitative assessment of native welfare

    J. Eur. Econ. Assoc.

    (2018)
  • T. Bauer et al.

    The economic integration of forced migrants: evidence for post-war Germany

    Econ. J.

    (2013)
  • R. Beyer et al.

    Labour market adjustments and migration in Europe and the United States: how different?

    Econ. Policy

    (2015)
  • O.J. Blanchard et al.

    Regional evolutions

    Brook. Pap. Econ. Act.

    (1992)
  • G.J. Borjas

    Self-selection and the earnings of immigrants

    Am. Econ. Rev.

    (1987)
  • G.J. Borjas

    The labor demand curve is downward sloping: reexamining the impact of immigration on the labor market

    Q. J. Econ.

    (2003)
  • G.J. Borjas

    Native internal migration and the labor market impact of immigration

    J. Hum. Resour.

    (2006)
  • G.J. Borjas

    The analytics of the wage effect of immigration

    IZA J. Migr.

    (2013)
  • G.J. Borjas

    Immigration Economics

    (2014)
  • G.J. Borjas

    The wage impact of the Marielitos: a reappraisal

    ILR Rev.

    (2017)
  • G.J. Borjas et al.

    The labour market consequences of refugee supply shocks

    Econ. Policy

    (2017)
  • L.P. Boustan et al.

    The effect of internal migration on local labor markets: American cities during the great depression

    J. Labor Econ.

    (2010)
  • S. Braun et al.

    The employment effects of immigration: evidence from the mass arrival of German expellees in post-war Germany

    J. Econ. Hist.

    (2014)
  • S. Braun et al.

    Local labor markets and the persistence of population shocks: evidence from West Germany, 1939–70

    J. Econ. Geogr.

    (2020)
  • K.B. Burchardi et al.

    The economic impact of social ties: evidence from German reunification

    Q. J. Econ.

    (2013)
  • M.C. Burda

    Factor reallocation in eastern Germany after reunification

    Am. Econ. Rev.

    (2006)
  • D. Card

    Immigrant inflows, native outflows, and the local labor market impacts of higher immigration

    J. Labor Econ.

    (2001)
  • D. Card et al.

    Do immigrant inflows Lead to native outflows?

    Am. Econ. Rev.

    (2000)
  • A. Chassamboulli et al.

    A search-equilibrium approach to the effects of immigration on labor market outcomes

    Int. Econ. Rev.

    (2014)
  • A. Chevalier et al.

    Immigrant Voters, Taxation and the Size of the Welfare State

    (2018)
  • D. Chiquiar et al.

    International migration, self-selection, and the distribution of wages: evidence from Mexico and the United States

    J. Polit. Econ.

    (2005)
  • M. Colas

    Dynamic Responses to Immigration

    (2018)
  • I. Connor

    Refugees and Expellees in Post-War Germany

    (2007)
  • M. Dao et al.

    Regional labor market adjustment in the United States: trend and cycle

    Rev. Econ. Stat.

    (2017)
  • Deutscher Städtetag

    Statistisches Jahrbuch Deutscher Gemeinden 1949

    (1949)
  • C. Dustmann et al.

    How do Industries and firms respond to changes in local labor supply?

    J. Labor Econ.

    (2015)
  • C. Dustmann et al.

    The effect of immigration along the distribution of wages

    Rev. Econ. Stud.

    (2013)
  • Cited by (0)

    We would like to thank Nicola Fuchs-Schündeln and Jochen Streb for their extensive comments and suggestions. The paper has also benefited from comments by Klaus Adam, Almut Balleer, Toman Barsbai, Michael Binder, Marcus Böhme, George Borjas, Michael Burda, Larry Christiano, Wouter Den Haan, Eva Moreno-Galbis, Fabio Ghironi, Peter Haan, Michael Krause, Winfried Königer, Wolfgang Lechthaler, Jochen Mankart, Christian Merkl, Gianmarco Ottaviano, Horst Raff, Morten Ravn, Michael Reiter, Dennis Snower, Nikolai Stähler, Ignat Stepanok, Jan Stuhler, the editor Guido Lorenzoni, two anonymous referees, and various seminar and conference participants. Richard Franke and Anika Martin provided excellent research assistance. The research in this paper was funded by Deutsche Forschungsgemeinschaft, Germany (grant no. BR 4979/1-1). We also gratefully acknowledge financial support from the Leibniz-Gemeinschaft, Germany ("Worker Flows, Match Quality and Productivity - Evidence from European Micro Data"). The opinions expressed in this paper are those of the authors and do not necessarily reflect the views of the Deutsche Bundesbank, the European Central Bank or the Eurosystem. Any remaining errors are our own.

    View full text