«SOEPpapers on Multidisciplinary Panel Data Research Sebastian Gundel Heiko Peters Assimilation and Cohort Effects for German Immigrants Berlin, ...»
Panel Data Research
Assimilation and Cohort Effects for German Immigrants
Berlin, November 2007
SOEPpapers on Multidisciplinary Panel Data Research
at DIW Berlin
This series presents research findings based either directly on data from the German SocioEconomic Panel Study (SOEP) or using SOEP data as part of an internationally comparable
data set (e.g. CNEF, ECHP, LIS, LWS, CHER/PACO). SOEP is a truly multidisciplinary
household panel study covering a wide range of social and behavioral sciences: economics, sociology, psychology, survey methodology, econometrics and applied statistics, educational science, political science, public health, behavioral genetics, demography, geography, and sport science.
The decision to publish a submission in SOEPpapers is made by a board of editors chosen by the DIW Berlin to represent the wide range of disciplines covered by SOEP. There is no external referee process and papers are either accepted or rejected without revision. Papers appear in this series as works in progress and may also appear elsewhere. They often represent preliminary studies and are circulated to encourage discussion. Citation of such a paper should account for its provisional character. A revised version may be requested from the author directly.
Any opinions expressed in this series are those of the author(s) and not those of DIW Berlin.
Research disseminated by DIW Berlin may include views on public policy issues, but the institute itself takes no institutional policy positions.
The SOEPpapers are available at http://www.diw.de/soeppapers
Georg Meran (Vice President DIW Berlin) Gert G. Wagner (Social Sciences) Joachim R. Frick (Empirical Economics) Jürgen Schupp (Sociology) Conchita D’Ambrosio (Public Economics) Christoph Breuer (Sport Science, DIW Research Professor) Anita I. Drever (Geography) Elke Holst (Gender Studies) Frieder R. Lang (Psychology, DIW Research Professor) Jörg-Peter Schräpler (Survey Methodology) C. Katharina Spieß (Educational Science) Martin Spieß (Survey Methodology) Alan S. Zuckerman (Political Science, DIW Research Professor) ISSN: 1864-6689 German Socio-Economic Panel Study (SOEP) DIW Berlin Mohrenstrasse 58 10117 Berlin, Germany Contact: Uta Rahmann | email@example.com Assimilation and Cohort Effects for German Immigrants Sebastian Gundela and Heiko Petersb This Version: October 2007 Abstract Demographic change and the rising demand for highly qualified labor in Germany attracts notice to the analysis of immigration. In addition, the patternof immigration changed markedly during the past decades. Therefore we use the latest data of the German Socioeconomic Panel up to the year 2006 in order to investigate the economic performance of immigrants. We perform regressions of three pooled cross sections (1986, 1996, 2006) to estimate assimilation and quality of immigrants as reflected by their earnings. Further we take the heterogeneity of immigrants into account by separating them by country of origin. The rising wage inequality in Germany since the mid nineties will also be considered. We find a negative wage gap and a yearly assimilation rate of 2.3 percent. Due to a changing immigration pattern the cohort quality is declining.
Keywords: Assimilation, immigrants, cohort quality, Germany JEL Classification: J31, J61, C21
a University of Muenster. Address for correspondence: Center of Applied Economics, Am Stadtgraben 9, 48143 Muenster, Email: firstname.lastname@example.org, phone: +49 251 8322976.
b University of Muenster. Institute for Economic Education.
1. Introduction Globalization brought a deterioration of the economic position of low qualified workers and an increased need for human capital. In combination with the demographic development of many industrialized countries, immigration policy is becoming an important instrument for economic growth. The goal should be to attract and successfully integrate highly qualified “white collar” workers and try to retain the “blue collar” workers.
As a result of the compressed earnings distribution in Germany compared to other industrial, emerging and developing countries and due to the beneficial social security system, Germany attracts less qualified immigrants (see Borjas (1987), pp. 532-534 and Gernandt; Pfeiffer (2006), p. 1). After the huge influx of guest workers between 1961 and 1973, due to a strong demand for relative unskilled workers, Germany became an immigration country like Australia, Canada and the United States. During the first oil crisis the hiring of guest workers ended with a recruitment ban and a period mainly dominated by family reunification began.
Since 1989 the fall of the “Iron Curtain” defined the picture of immigration. During that time a large remigration of ethnic Germans and immigration from former communist countries of Eastern Europe took place and changed the pattern of immigration origin markedly (see Zimmermann et al. (2006), pp. 16-29).
Today, Germany faces a high level of low qualified unemployed and the need for a positive immigration balance in order to soften the ageing of the German population. Particularly the need for high qualified workers is of crucial importance. According to the OECD study on migration even the highly qualified immigrants have a higher probability of becoming unemployed than Germans obtaining the same educational level (see OECD (2007)). This highlights the importance of labour market efficiency as reflected by the possibility to absorb the new immigrants and successfully overcome demographic change. The intensity, the arrival, the quality and the assimilation of immigrants differ strongly between economies and are important parameters for assessing the success of immigration. If the new immigrants are not successfully integrated they are an additional burden for the public welfare system.
For the first time, Chiswick (1978) analyzed the assimilation of immigrants as reflected by their earnings using the 1970 Census for the United States. Chiswick indicated an initial wage gap of immigrants relative to natives of 17 per cent. 10-15 years after immigration the earnings of the immigrants equal the natives’ and start to exceed them with each year of labour experience in the host country. Thus the assimilation rate corresponds to the variable years since migration, sometimes called years of residence. The theoretical explanations are twofold: Initially, immigrants earn less than natives because they have lower country specific human capital like labour market customs, language skills and business practices. Each subsequent year that immigrants build up country specific skills is rewarded by the labour market. Therefore they experience steeper earnings growth compared to natives which leads to a narrowed gap. The argument explaining the immigrants’ wages exceeding the natives’ ones derives from selection. The decision to emigrate and start a new life allows the immigrants to become a positive selected group that is more ambitious to work harder and
longer (see Borjas (1994), pp. 1671f.; Carliner (1980), pp. 88f. and Chiswick (1978), pp. 899Chisiwick’s seminal article was followed by a huge literature (Borjas (1985, 1987, 1989, 1995); Chiswick (1986) and LaLonde; Topel (1990, 1991)) on the assimilation hypothesis for immigrants in the United States.
Different empirical approaches to the study of immigrants’ assimilation can be found in the literature. Chiswick (1978) used a cross-sectional standard OLS estimation of the human capital earnings function1 to estimate the assimilation of immigrants in the USA.2 A correlation of the year since migration variable and the wage of immigrants could be caused by either age (assimilation) or cohorts effects (quality change). If the less successful have a higher propensity to remigrate or the most recent immigration cohorts have a lower unobserved ability relevant for the labor market, the simple OLS cross-section estimation of the assimilation is biased upwards. To account for this bias, the cohort effects have to be taken into consideration. Because of identification problems the separation of age and cohorts effects is not possible with single cross section estimation. Cohort or longitudinal data are necessary for this kind of analysis. Borjas (1985, 1995) uses different cross-sections in his estimations. Assuming the same period effect for immigrants and natives, it is possible to identify age and cohort effects simultaneously. This approach poses problems due to selective emigration, changes in the composition of the samples over time and the difficulty of disentangling longitudinal changes and period effects (Chiswick et. al (2002), p. 1). Instead of using different cross-sections with the restriction of the same period effects, panel data techniques could be used. Borjas (1989) and Hu (2000) estimated a system of natives and immigrants human capital functions using longitudinal data. The main problem using panel data techniques is panel attrition. Panel attrition will bias the estimation results if the probability to leave the sample is systematically linked to labor market developments.
Because of the high panel data attrition of the immigrants in the German socioeconomic panel (GSOEP) for the period 1986-2006 we do not use panel data techniques.3 Instead, we use a similar empirical framework as Borjas (1985, 1995) to identify assimilation and cohort effects of immigrants in Germany simultaneously.
For Germany, Pischke (1992) analyzed the assimilation of guestworkers using data from the GSOEP 1984-1989 and different estimation methods. Pooled cross-section and panel estimates yield similar results indicating no assimilation. After controlling for different variables, the assimilation effect is insignificant and negative, indicating lower earnings growth for immigrants than for Germans. Dustmann (1993) differs between temporary and permanent migrants using the expected length of stay variable to disentangle. Estimating a
See Mincer (1974) for a theoretical derivation of the human captial earnings function.
Carlinger (1980) and De Freitas (1980) use also a simple cross section estimation of the assimilation of immigrants in the USA.
A balanced panel would contain 1288 (1015 natives and 273 immigrants) individuals over the period of 23 years. After correcting for missing values and our data selection we would have only 207 (180 natives and 27 immigrants) individuals.
cross-section of the 1984 wave of the GSOEP, Dustmann found a positive but insignificant yearly assimilation rate. The absent significant assimilation is explained with the low incentive of temporary migrants to invest into host country specific human capital. Licht and Steiner (1994) perform fixed effects panel estimates based on the first six waves of the GSOEP. They found no evidence of the assimilation hypothesis. Schmidt (1997) compared the assimilation of ethnic Germans and German guest workers and couldn’t find a stable pattern. His analysis was based on cross-sections from two different sources, from which one was the GSOEP wave 1984. Bauer et al. (2005) couldn’t support the assimilation hypothesis either. However, Constant and Massey (2005) found a positive assimilation rate that equates the earnings of immigrants and natives after 17 to 23 years analyzing the GSOEP waves from 1984-1997. Most recently, Fertig and Schurer (2007) analyzed the assimilation hypothesis using a longitudinal panel data set of 21 waves lasting from 1984 to 2004. They estimated the assimilation for four different entry cohorts in order to control for immigrant heterogeneity using fixed effects models. They found positive assimilation rates for two cohorts lasting 9 and 16 years, respectively. Summing up the existing literature, there is no consistent picture about the catching up process. However a large initial earnings differential in Germany exists.
Further there is no unambiguous evidence about the development of cohort quality so far.
This paper uses the latest data of the GSOEP to estimate the earnings equations in a pooled cross section of the years 1986, 1996 and 2006 and thus incorporates the development of the changing immigration pattern of the past 15 years. Further we want to account for the changing wage structure that occurred in Germany at the end of the 20th century (see Gernandt; Pfeiffer (2006) and Peters (2007)).
The remainder of this paper is organized as follows. The next section describes the underlying data and some stylized facts. We will exhibit the wage differential between immigrants and Germans for each survey year and different immigrant cohorts. The changing and groupspecific differently affecting wage structure will be accounted for by using two different wage deflators. The empirical model will be compiled in section 3. After discussing the different model specifications and the results, section 4 concludes with a short summary.
2. Data and Summary Statistics
For the analysis we use the GSOEP, collected for the period 1984 through 2006. The GSOEP is a yearly conducted household survey.4 The data we are using are limited to men, aged 18working at least 1820 hours per year with an hourly wage higher than 1€ and living in West Germany5. The individual real hourly wages are calculated deflating the nominal yearly labor earnings by the harmonized consumer price index based on the year 2000 and divided
See Wagner et al. (2007) and Haisken-DeNew/Frick (2005) for a detailed description of the GSOEP.
The immigration sample D for the period 1984-1993 is restricted to West Germany due to few observations in the eastern states of Germany. See Pannenberg et al. (2005), p. 180.
by yearly hours worked. We are focusing on three cross sections, derived from the latest one in year 2006, 1996 and 1986, respectively. After correction for individuals exhibiting missing
values the 2006 wave consists of 2576 natives and 431 immigrants (1996: 1916, 688; 1986: