-B New Immigrants Old Disadvantage Patterns? Labour Market Integration of Recent Immigrants into Germany

I M I G 0 -B New Immigrants Old Disadvantage Patterns? Labour Market Integration of Recent Immigrants into Germany Irena Kogan ABSTRACT This paper examines the labour market integration of immigrants who have entered Germany since 0, and compares their situation with that of their predecessors. The analyses based on the cumulative micro-census data reveal that recent immigrants into Germany are on average bettereducated than their earlier counterparts, and some ethnic groups are even better- educated than the national average. Despite their high levels of formal education, these immigrants coming mostly from Eastern Europe, Africa and the Middle East face severe integration problems in the German labour market. Thus, after taking into account the value of human capital represented by these immigrants, their ethnic disadvantages appear to increase. This stands in sharp contrast with the disadvantages faced by classic immigrants who arrived in Germany during the 0s and 0s, for whom lack of human capital had been identified as the main obstacle to labour market integration. INTRODUCTION Dispatch:.. Journal: IMIG CE: Krishna Sarma Journal Name Manuscript No. Author Received: No. of pages: PE: Senthil doi:./j.-..000.x The growing body of empirical research in Germany has repeatedly demonstrated that the prospects of immigrants in the labour market have been weaker than native Germans. Immigrants are over-represented among the unemployed and in the lower strata of the occupational hierarchy (Seifert, ; Velling, ; Bender and Seifert, ; Szydlik, ; Kogan, 0, 0). This being said, the patterns as well University of Mannheim, Mannheim. Ó The Author Published by Blackwell Publishing Ltd., Journal Compilation Ó IOM 00 Garsington Road, Oxford OX DQ, UK, International Migration and 0 Main Street, Malden, MA 0, USA. ISSN 00-

Kogan as the underlying factors behind these disadvantages are somewhat different for the two major migratory waves after World War II. Initially, immigrants arrived within the framework of guest worker recruitment and subsequently accompanied by their family members, and thereafter beginning in the late 0s, immigrants arrived as refugees, asylum seekers, diaspora returnees and labour migrants. The main explanation for the disadvantages experienced by classic immigrants who arrived in Germany during the 0-0s within the framework of guest worker recruitment schemes from Italy, Spain, Greece, Turkey, Portugal and Yugoslavia is their negative selectivity (the fact that they arrived from economically depressed areas of their native countries and possessed little human capital, including education). Thus, once education has been controlled for, the general disadvantages faced by first- generation labour migrants in Germany appear to decrease (Kalter and Granato, 0). Recent immigrants coming mostly from Eastern Europe, Africa and the Middle East face no least severe problems when attempting to integrate into the German labour market, despite their comparatively high levels of human capital. Thus, their ethnic disadvantages appear to increase once one has taken the human capital characteristics of these immigrants into account (Kalter and Granato, 0). The aim of this paper is to directly compare the labour market situation of more recent immigrants to their predecessors while allowing the disaggregation of the former group to enable an adequate interpretation for the findings (Kalter and Granato, 0). Before examining possible reasons for the disadvantages faced by immigrants in the labour market, the next section will detail the recent migratory waves into Germany. The methodology of this study will be further discussed, to be followed by a presentation of the descriptive and analytical results. The paper concludes with a summary of the main findings of the study and their implications. RECENT IMMIGRATION TO GERMANY Germany s transformation into an immigrant-receiving society began when the post-war economic miracle resulted in the expanding recruitment of labour recruitment from foreign countries. Labour recruitment from Italy began in, from Spain and Greece in 0, from Turkey in, from Morocco in, from Portugal in, from Tunisia in, and from Yugoslavia in. Recruitment from these countries in the period between and brought a large number of lowly- Ó The Author Journal Compilation Ó IOM

xxxx skilled workers from Southern Europe (Kogan, 0; Kalter and Granato, 0). This phenomenon contributed to a massive increase in the number of foreign-born labour migrants in Federal republic of Germany (Rudolph, ). By, million relatively unskilled Gastarbeiter ( guest-workers ) laboured in Germany, the largest proportion of whom came from Turkey and Yugoslavia (Kalter and Granato, 0). The OPEC oil embargo () and the recession that followed forced the German government to curb their guest-worker recruitment policy, but immigration continued as the families of guest workers already in Germany came to be reunited with their families. Consequently, the immigration of the spouses and children of guest-workers accounted for 0-0 per cent of Germany s immigration between and. This trend fundamentally changed the demographic composition of the immigrants: labour force participation fell and the unemployment rate for immigrants was higher than that of native Germans for the first time since guest-worker recruitment began (Kalter and Granato, 0). Although the German government provided financial incentives for guest-workers to return home, these policies failed to entice return migration and consequently the average length of a guest-worker s sojourn in Germany increased. By the mid-0s immigration to Germany had assumed a more humanitarian nature. Refugees from behind the Iron Curtain, Yugoslavia and the Kurdish regions of Turkey and Iraq began to seek asylum in Germany (Kogan, 0, 0). Hence, the number of applications increased from less than 0,000 in to almost,000 in (OECD, 0). Only a minority of asylum applications appeared to be truly attributable to political persecution, thus acceptance rates were relatively low (e.g.,.% in ). After, the asylum law was amended and annual applications for asylum fell below 0,000. In the mid-00s, the biggest numbers of asylum claims came from the Middle East and Asia (Iraq, Turkey, Iran, Syria, Lebanon and Afghanistan), as well as from Serbia and the Russian Federation (Migrationsbericht, 0). In Germany introduced a settlement programme for Jewish immigrants from the former Soviet Union, granting them a quota refugee status (Cohen and Kogan, 0, 0). About 0,000 Jews and their family members have been re-settled to Germany to foster its Jewish life. Due to their special status, Jewish quota refugees enjoyed until recently very extensive integration support, far beyond what other immigrants groups are entitled to (e.g., unrestricted labour market rights and social Ó The Author Journal Compilation Ó IOM

Kogan assistance). With regard to aid they are almost equivalent to ethnic German immigrants (Aussiedler), another group whose immigration has experienced a surge since the late 0s. In 0 alone, almost 0,000 Aussiedler entered Germany (OECD, 0). In, as a response to the migration wave of early 0s, the German government passed two major pieces of legislation aimed at reducing the numbers of asylum seekers and ethnic German immigration. Since immigration of ethnic Germans has been limited to those coming from the successor states of the former Soviet Union under certain quotas. Since ethnic Germans have had to pass a language test to obtain their status. These and other measures have resulted in a steady decline in the number of individuals entering Germany with the status of ethnic Germans, and rather in an increase of those entering the country as accompanying family members of Aussiedler. The decline in the share of ethnic Germans with language skills is said to increasingly hamper the socioeconomic integration of this group (OECD, 0). Accounting for the degree of integration of ethnic Germans has been limited until recently due to the fact that the German official data (e.g., the German microcensus up until 0) categorized immigrants solely by nationality, leaving ethnic German immigrants unrecognisable from other Germans, including naturalized immigrants. The present paper can thus be considered to be one of the first accounts of the labour market situation of ethnic German immigrants in comparison to other immigrant groups based on micro-census data. Another important source of migration to Germany in recent years has been the inflow of seasonal labourers and the labour migration of highly-skilled specialists. Seasonal workers come mainly from Central and Eastern Europe (above all Poland, Romania, Croatia, and Bosnia and Herzegovina, see Migrationsbericht, 0) for employment in agriculture, forestry, and hospitality industries. In addition, qualified nursing stuff for hospitals and nursing homes are in need, and these vacancies are often filled by immigrant workers. Although the formal recruitment policy has been abandoned since, persons generally exempt from the recruitment ban are scientists, executives and managers, foreign language teachers (native-speakers), chefs, chaplains, artists, models, professional athletes and trainers. In addition, foreign youth may work as an au pair for up to one year, or, for those enrolled in a university, work during the holidays and take part in internships related to their area of study. Furthermore, since August 00, foreign specialists in information technologies (IT) with relevant university degrees have been invited to work in the IT industry in Germany (Werner, 0). By September Ó The Author Journal Compilation Ó IOM

xxxx 0, more than,000 permits for such work, known as Green Cards, have been issued and since then foreigners in these occupations have been eligible for permanent residence in Germany. The most frequent nationalities among Green-Card holders are Indians, Chinese, Romanians, Poles, Russians as well as immigrants from other Eastern European countries (Migrationsbericht, 0). Permanent residence may also be granted to other highly-skilled workers, such as senior academics and researchers, top-level managers in business and industry. Self-employed foreigners may also emigrate to Germany if their business is of economic interest for the country. Highly-qualified immigrants continue to arrive from other EU- countries, with Italians, Frenchmen, Greeks, Dutchmen, Austrians, Britons and Spaniards dominating (Migrationsbericht, 0). THEORETICAL PERSPECTIVES AND EMPIRICAL EXPECTATIONS According to human capital and other economic theories dealing with accumulation and transferability of skills as well as migrant selection processes (Borjas,, ; Chiswick, ) the two waves of migrants to Germany should differ first of all in the degree of their selectivity both with regard to formal educational qualifications and intrinsic characteristics (e.g., motivation). From the discussion presented in the previous section becomes obvious that immigrants coming to Germany in the 0s have not been undergoing negative selection processes as have their predecessors of the guest worker wave, thus their higher levels of education can be attributed ipso facto. Furthermore, educational expansion took place in every country in the world, albeit to different degrees, therefore more recent immigrant cohorts should be better-educated then their predecessors, due to rising levels of education in the countries of origin. In aggregate, the higher levels of education of more recent immigrant cohorts may also be attributed to the altered composition of the migration flow from countries outside the EU-. Among such immigrants, those from Central and Eastern Europe often have had better schooling than other newcomers. It should be noted, however, that even high-level educational qualifications may be of little relevance for the host country if they are not adequately recognised by the host country and are not accompanied a by correspondingly high proficiency in the host country language. Highly qualified immigrants seem to face serious obstacles in the German Ó The Author Journal Compilation Ó IOM

Kogan labour market, which is known for its pronounced segmentation along professional lines. Access to occupations is largely provided by certificates acquired within the German system of vocational training (including tertiary education) (Kalter and Granato, 0). Recent studies (Esser, 0) have shown anew that unsatisfactory language proficiency among highly- qualified immigrants is not associated with any meaningful labour market gains, which in the end might lead to even more pronounced ethnic penalties. Penalties associated with immigrant status that remain after controlling for education and other socio-demographic characteristics can also be attributed to their pursuit of labour market opportunities and investments different from the local population. Immigrants may decide not to invest in acquiring the human capital relevant in the host country, or long searches for higher-status jobs if job-search costs are very high or if they are unlikely to be resident in the country long enough to enjoy its rate of return (Kalter and Kogan, 0). Instead, they often choose jobs with immediate financial returns, (Dustmann, 00) which eventually results in their over-representation in low-status positions even without apparent discrimination from employers. Even in the case of no apparent return intentions migrants may end up in low-status employment after having sought and tried suitable jobs for a long time without any apparent success. In the long-run, immigrants may become trapped in the secondary labour market characterised by low-status, poorly-paid, dangerous or unpleasant jobs, as mobility between the primary and secondary labour markets is quite limited (Doeringer and Piore, ; Piore, ) particularly in Germany, a country characterised by a strong insider-outsider divide. Additional hurdles to immigrants integration into the labour market could be present in the German legal regulations, such as the so-called German priority law. According to this law, German nationals are preferred over immigrants for any job opening, unless a special justification for employing a foreigner could be provided by an employer. Furthermore, immigrants from the EU- enjoy preferential treatment over other immigrants according to this law, so a certain queuing in the labour market entry is legally institutionalized in Germany. Furthermore, restrictive citizenship laws make it difficult to immigrants, older or newer, to reach better positions in the German labour market (Kogan, 0). Due to their higher levels of education, we may expect recent cohorts of immigrants to be more ambitious with regard to the prestige of their Ó The Author Journal Compilation Ó IOM

xxxx jobs. However, such ambitions might be marred by greater difficulties and hence higher unemployment risks. For the older immigrant cohorts, finding employment may also have been a problem, but to a lesser extent than for more recent immigrants, due to their targeting of less prestigious jobs, which might be easier to enter. Due to the more privileged legal status of Western immigrants in Germany and an easier transferability of skills across industrialized countries, we expect better labour market prospects for this group of immigrants, irrespective of the timing of migration. Finally, based on the German state s efforts to integrate Aussiedler, to officially recognise their education and labour market experiences from abroad and to provide them with language and training courses (Kogan, 0), this group should be among those immigrants displaying the most favourable labour market status, enjoying returns to education comparable to those of the native-born Germans. To what extent more and less recent immigrants, as well as nationalethnic groups of the most recent immigrant cohort differ in their human capital endowments, their present standings in the labour market and returns on their qualifications will be analysed in the empirical analyses presented below. RESEARCH METHODOLOGY To examine how newer immigration cohorts fare compared to classic migrants in the labour market we rely on the data from the German micro-census (a German equivalent to labour force surveys used in other countries) for the years, 00 and 0. The German Micro-census is an annual survey of per cent of all German households, and is also a rotating panel, in which the whole sample is substituted quadrennially. We analyse the Microcensus Scientific Use File, which is a factually anonymized 0 per cent subsample of the original microcensus. This data source is appropriate for the aims of our analyses due to its provision of detailed information on education and labour market integration, and because its large sample sizes allow for distinction between the relevant immigrant groups. A particular advantage of including the microcensus for the year 0 is that for the first time, one could properly identify ethnic Germans and compare them to the rest of immigrants with this dataset. This implies that with the data prior to the year 0, it is impossible to analyse ethnic German immigrants, and the information we have for this group stems only from the 0-dataset. Ó The Author Journal Compilation Ó IOM

Kogan Our analyses are limited to respondents aged - years old. Immigrants are defined by country of birth or nationality and year of migration. We consider immigrants those who arrived to Germany after age. For comparison of the labour market situation of old and new immigrants, we distinguish four groups: immigrants coming from EU- countries, the US, Canada, Australia and New Zealand, and immigrants from the rest of the world. In addition we explore the labour market status of ethnic German immigrants. In each of the groups we differentiate those arriving prior to and those who came after 0, and compare them to native Germans. For more detailed analyses of the new immigrants, we differentiate the following immigrant groups: Italians, Iberians (immigrants from Spain and Portugal), ex-yugoslavs, immigrants from the former Soviet Union (FSU), Poles, immigrants from the rest of Eastern Europe, Turks, immigrants from Africa, Asia, Latin and Southern America, and the residual category of other immigrants. Ethnic Germans arriving since 0 are also included in these analyses. All these groups are compared to immigrants from EU- countries, the USA, Canada, Australia and New Zealand. To analyse the avoidance of unemployment and inactivity among immigrants and to compare those figures with that of native Germans, a Heckman probit model for a binary dependent variable is fitted. In terms of analysis of the occupational standing of immigrants, a multinomial logistic regression is applied with the following categories of the dependent variable distinguished: () salaried and intermediate occupations, () petty bourgeoisie (small employers and own account workers), () skilled manual workers and finally () routine non-manual and unskilled workers (reference category). We deviated from other contributions in this edited volume by using the European Socio-economic Classification of occupations (ESeC) (Rose and Harrison 0) instead of EGP for the identification of occupational destinations. This is predicated on the fact that it was impossible to construct EGP with the micro-census data for the year 0. Thus we applied a reduced version of ESeC class schema, according to which salaried and intermediate occupations correspond to EGP classes I, II and IIIa and ESeC classes I, II and III; petty bourgeoisie to EGP classes IVa,b,c and ESeC classes IV and V; skilled manual workers EGP V, VI and ESeC VI and VIII; routine non-manual and unskilled workers EGP IIIb and VII and ESeC VII and IX. There are certain discrepancies in the distributions of occupational positions among immigrants and natives according to EGP and ESeC, but the overall trend is reproduced similarly by both classifications. Ó The Author Journal Compilation Ó IOM

xxxx Due to important gender contrasts in labour market patterns, for both native Germans and immigrants, all statistical models have been run separately by gender. Independent variables include education coded against the International Standard Classification of Education (ISCED) in the following categories: () primary education or less, () lower secondary, () secondary vocational, () secondary general, () post-secondary, () tertiary vocational (short), and () tertiary academic. Individuals with missing information on education have been assigned to a separate category. Marital status of the respondents was represented by a number of dummy variables distinguishing between single respondents, single parents, married persons without children and married persons with children. For the Heckman selection equation, this variable was substituted by a dummy variable differentiating between married individuals and the rest. Furthermore, persons with dependent children in age up to years old and those aged - were accounted for. Age and age squared were also included to account for another demographic characteristic relevant for labour market inclusion. In the models pertaining to immigrants, we included years since migration (YSM) as a variable. The variable was derived from the year in which the interview took place minus the year of respondent s arrival. In addition, we control for the year of survey (, 00 or 0, the latter being a reference category) and regional variation differentiating between Eastern and Western federal states. FINDINGS Descriptive evidence In line with expectations concerning the selection of immigrants to Germany, it is evident that more recent immigrants are better educated than the less recent ones. Improvements in the immigrants quality are evident for both newcomers from EU- or other western countries, and for those coming from the rest of the world. It is also apparent that among more recent immigrants from the EU- the proportion possessing tertiary education was even higher than for native Germans. A larger percentage was, however, also endowed with the lowest qualifications as compared to the German charter population. Among new non-western immigrants the proportion with academic tertiary qualifications was as high as among Germans, the proportion of loweducated was, however, also considerably higher. The pattern of educa- Ó The Author Journal Compilation Ó IOM

Kogan Colour online, B&W in print 0% 0% 0% % % 0% FIGURE EDUCATIONAL ATTAINMENT OF THE GERMAN CHARTER POPULATION AND IMMIGRANTS Germans Western new Non-West new Western old Non western old Ethnic Germans Source: German micro-census 0. Ter. long Ter. short Post-sec. Sec. Gen. Sec. Voc Low-sec. Primary tional attainment among ethnic Germans was somewhere between the more recent immigrants, and older immigrants, which reflects the heterogeneous character of this group. About per cent of ethnic Germans possessed any type of tertiary education, but at the same time more than per cent surveyed had lower-secondary education only. Since the educational patterns for men and women appeared quite similar, they are not shown here by gender, but available from the author upon request. Among the recent immigrants, the most educated groups (i.e. the groups with the high proportion of tertiary educated) were immigrants from the FSU, Latin America, Western and Northern Europe. Turkey continued to send much less-educated immigrants than any other country, and as compared to the German charter population. Somewhat higher was the proportion of poorly-educated among those coming from Southern Europe and the countries once constituting Yugoslavia. A peculiarity of new immigrants from Asia is their relative over-representation of both highly- and lowly-skilled individuals. Gender-specific employment patterns for various immigrant groups compared to native Germans are shown in Table. It is immediately evident that all immigrant groups have higher unemployment rates than native Germans. Moreover, unemployment seems to be somewhat higher among newer immigrants than among less recent ones. Interestingly, Ó The Author Journal Compilation Ó IOM

xxxx TABLE EMPLOYMENT PATTERNS, BY ILO DEFINITION, OF IMMIGRANTS AND NATIVE-BORN IN GERMANY, BY GENDER Employed Unemployed Student Inactive Men Women Men Women Men Women Men Women Germans 0....0.... Ethnic Germans...0..0... Western old........0 Non-Western old........ Western new........ Non-Western new....0.... EU+OECD......0.0. Italy 0........ Iberia........ Ex-Yugoslavia 0........ FSU.......0. Poland.0...0.... Other EE..0...... Turkey.0...0.... Africa........ Asia........0 Latin America and other........ Source: German microcensus 0. inactivity is much less pronounced among newer male immigrants, which could at least partially be attributed to their younger age and labour orientation of their migration. Ethnic Germans, both men and women, are also less likely to be inactive. Among women, labour force participation patterns do not differ much between migration cohorts, inactivity being higher among those coming outside the EU- or other western industrialised countries. Older immigrant cohorts supplied fewer students, whereas a larger proportion of newer immigrants come to Germany to study. From the descriptive results, it is hardly visible that ethnic Germans are doing better in the German labour market in terms of gainful employment compared to other immigrants. Their situation is somewhere between the immigrants from Western and non-western countries. Apparently, intensive efforts to integrate them into the German labour market have not yet resulted in better outcomes. If one scans employment patterns among specific groups of more recent immigrants, one notices extremely high unemployment among both male and female immigrants from the former Soviet Union (see also Cohen and Kogan, 0, 0), and among male immigrants coming from Africa and Turkey. This occurs despite the fact that apart from Turkish Ó The Author Journal Compilation Ó IOM

Kogan immigrants we are dealing with relatively skilled individuals. As aforementioned, newer immigrants to Germany are largely found among students, which holds true for almost all nationalities except for migrants arriving from Southern Europe and the former Yugoslavia. Inactivity appears to be a dominant pattern for female immigrants from Turkey, Africa and Asia. Among male immigrants, somewhat higher proportions of inactivity occur among immigrants from the former Yugoslavia and Turkey. Not only are immigrants disadvantaged when regarding their employment patterns, their occupational attainment also lags behind those of native Germans, as it is evident from Figures a (for men) and b (for women). Only for few groups of recent immigrants was the proportion of those employed in salaried professions similar to that of native Germans. These groups encompass western- and northern-europeans, followed by Latin Americans. Nonetheless, for these groups and even more so for recent immigrants from Africa, Turkey, the former Yugoslavia and FSU we observe high proportions employed in unskilled jobs. A similar pattern of over-representation in the lower echelons of the occupational hierarchy is evident for earlier cohorts of immigrants and ethnic German immigrants. It is also evident that male immigrants from Western and Northern Europe, Africa, Asia and Latin America are under-represented among skilled manual workers, whereas for the rest of immigrants, skilled manual work is accessible to the same degree as among native Germans. Self-employment has become a venue of professional integration only for selected groups of immigrants, that is, older cohorts of male immigrants from Western countries as well as newcomers from Italy and Asia. Employment patterns: Results of the multivariate analyses The aforementioned descriptive findings may appear somewhat misleading, as they do not take into account the basic socio-demographic characteristics of the analysed population. For example, the fact that immigrants from the earlier cohorts are on average somewhat older than those who arrived recently may have been a critical factor. To control for the characteristics of the sample and to assess the differentiating role education might have played for labour market inclusion, we ran a series of multivariate analyses. Results of the analyses for employment and labour force participation using the Heckman probit selection model are shown in Table. Conditional on labour force participation and net of demographic characteristics (age and family status), year of observation Ó The Author Journal Compilation Ó IOM

xxxx a) 0% 0% 0% 0% 0% 0% % % % % 0% b) 0% 0% 0% 0% 0% 0% % % % % 0% FIGURE A: ESEC CLASS POSITIONS OF THE MALE GERMAN CHARTER POPULATION AND IMMIGRANTS. B: ESEC CLASS POSITIONS OF THE FEMALE GERMAN CHARTER POP- ULATION AND IMMIGRANTS Salariat Petty bourgeosie Skilled manual Unskilled Salariat Petty bourgeosie Skilled manual Unskilled Source: German micro-census 0. Ó The Author Journal Compilation Ó IOM

Kogan TABLE HECKMAN PROBIT MODEL OF AVOIDING UNEMPLOYMENT (EMPLOYMENT EQUA- TION) CONDITIONAL ON LABOUR FORCE PARTICIPATION (SELECTION EQUATION), BY GENDER Employment equation Men Selection Employment equation Women Selection Model Model Model Model Model Model West new )0. )0.** )0.** )0.** )0.** )0.** Non West new )0.** )0.0** )0.0** )0.** )0.0** )0.0** West old )0.** )0.0** 0.** )0.** 0.0 0.** Non West old )0.0** )0.** 0.0** )0.** )0.** )0.0** Ethnic Germans )0.0** )0.** )0.0 )0.** )0.** 0.** Age 0.0** 0.0** 0.** 0.0** 0.0** 0.** Age squared )0.0** )0.00** )0.00** )0.00** )0.00** )0.00** Married, no kid 0.** 0.** )0.0** )0.0 Married, kid 0.** 0.** )0.** )0.** Single parent 0.0 0.0** )0.** )0.** Other )0.0** )0.0** )0.** )0.** Married 0.** )0.** Child below )0.** )0.** Child 0.0** )0.** Year 0.** 0.** )0.** 0.** 0.** )0.** Year 00 0.** 0.** )0.** 0.** 0.** )0.** East )0.** )0.** )0.0** )0.** )0.** 0.** Low secondary 0.** 0.** 0.** 0.** Secondary voc. 0.** 0.0** 0.** 0.** Sec. gen. (Abitur) 0.** 0.** 0.** 0.** Post-secondary 0.** 0.** 0.** 0.** Tertiary short 0.**.0** 0.** 0.** Tertiary long 0.**.** 0.**.** Education 0.** 0.** 0.** 0.** missing Constant 0.** )0.0** ).** 0.** 0.** ).0** Rho )0.0** )0.** )0.** )0.** No, observations No, censored 0 Wald Chi DF Source: German microcensus, 00, 0. Note: *p<0.; **p<0.0. and region of residence (see model for men and women), all immigrants show greater propensities for unemployment. The disadvantages of newer immigrants are larger than earlier cohorts, whereas immigrants from outside the EU fare worse than westerners. The disadvantages of ethnic Germans are smaller than among non-western immigrants, but Ó The Author Journal Compilation Ó IOM

xxxx larger than among immigrants from Western countries. Male immigrants suffer larger penalties than women. Once we control for education (see model ), the native-immigrant gaps are somewhat reduced, but to a varying degree for older and newer immigrant cohorts. Education or, more precisely, the lack of relevant education seems to explain the greater part of ethnic disadvantages among earlier cohorts of immigrants. For more recent immigrants as well as for ethnic German immigrants, education has proven to be less of a factor in determining the immigrants lack of employment prospects as compared to the natives, and this has been particularly true for women. In Figures a (for women) and b (for men), we examine the ethnic penalties on employment for various groups of recent immigrants compared to immigrants coming from Western and Northern Europe, i.e. immigrants who enjoy treatment almost equal to those of the charter population. The gross ethnicity effects (without controlling for education) are compared there to the net effects (controlling for education), and the results are consistent to those presented in Table. For the majority of immigrant groups, education hardly explains the extent of ethnic penalties. In case of immigrants from the FSU and Latin America, ethnic penalties further increase once education is taken into account. For more recent immigrants from Italy, Portugal and Spain, advantages over native Germans with regard to employment further increase once education is taken into account, which suggests a clearly employment-oriented nature of migration for these groups. The best employment prospects, even better than among comparable native Germans, are observed among immigrants from Southern and Eastern Europe (except for Poland and the FSU), and the worst are found for immigrants from the former Soviet Union, Africa, Turkey and Asia. The next question we ask is: to what extent foreign education has been devaluated for immigrants, and which groups suffer the biggest loss in their host country. In Table it is once again confirmed that newer immigrants undergo somewhat larger losses in the transferability of their educational qualifications. This is not surprising, since older immigrant cohorts might have had a bigger chance to acquire some education in Germany. Furthermore, among men from non-western countries, practically every educational certificate is of a lower value than a comparable German certificate. Overall, the mostly devaluated seem to be educational qualifications from tertiary and vocational education, which is not surprising for Germany, known for the relevance of educational creden- Ó The Author Journal Compilation Ó IOM

Kogan (a) 0.0 0. 0. 0. 0. 0.00 0. 0. 0. 0. 0.0 0.0 (b) 0.0 0. 0. 0. 0. 0.00 0. 0. 0. 0. 0.0 0.0 Italy Italy Iberia Iberia Yugoslavia Yugoslavia FSU FSU FIGURE A: EFFECTS OF ETHNICITY FOR RECENT IMMIGRANTS IN COMPARISON WITH IMMI- GRANTS COMING FROM WESTERN AND NORTHERN EUROPE, FROM THE HECKMAN PROBIT MODEL OF ACTIVITY AND EMPLOYMENT PATTERNS, MEN. B: EFFECTS OF ETHNICITY FOR RECENT IMMIGRANTS IN COMPARISON WITH IMMIGRANTS COMING FROM WESTERN AND NORTHERN EUROPE, FROM THE HECKMAN PROBIT MODEL OF ACTIVITY AND EMPLOYMENT PATTERNS, WOMEN Poland Poland Other EE Other EE Source: German micro-census, 00, 0. tials, preferably German credentials, for the labour market allocation. A single group, for which education in the country of origin appears to be not totally lost, was the group of ethnic German immigrants. Indeed, Turkey Turkey Africa Africa Asia Without Education With Education Asia America America Ethnic Germans Without Education With Education Ethnic Germans Ó The Author Journal Compilation Ó IOM

xxxx TABLE DIFFERENTIAL RETURNS ON EDUCATION FOR VARIOUS MIGRANT GROUPS (INTERACTION TERMS OF EDUCATION AND MIGRANT STATUS) FROM THE HECK- MAN PROBIT MODEL OF AVOIDING UNEMPLOYMENT (EMPLOYMENT EQUATION) CONDITIONAL ON LABOUR FORCE PARTICIPATION (SELECTION EQUATION), BY GENDER Main effect Interaction effect of education with migrant status for West new Nonwest new West old Nonwest old Ethnic Germans Men Low secondary 0.** )0. )0.** )0.0 )0.0 0.0 Secondary vocational 0.** )0.** )0.** )0.** )0.** )0. Secondary gen. (Abitur) 0.** )0.* )0.** )0.** )0.** 0. Post-secondary 0.** )0. )0.** )0.** )0.** )0. Tertiary short.0** )0.** )0.0** )0.** )0.** )0. Tertiary.0** )0.** )0.** )0. )0.** )0. Education missing 0.** Women Low secondary 0.** )0. )0.** )0.** )0.0 )0. Secondary vocational 0.** )0.** )0.** )0.** )0.** )0.** Secondary gen. (Abitur) 0.** )0. )0.** )0.* )0.** )0. Post-secondary 0.** )0.** )0.** )0.** )0. )0. Tertiary short 0.** )0.0** )0.** )0. )0.0 )0. Tertiary 0.** )0.** )0.** )0.** )0.** )0.0** Education missing 0.0** Source: German microcensus, 00, 0. Note: *p<0.; **p<0.0. among men no educational discount is experienced by ethnic German immigrants, whereas for women only those with university education and vocational training are significantly disadvantaged. Apparently efforts to recognise education among Aussiedler and to provide re-training pay off by their somewhat better usability in the host country. Going back to the selection equation from model, in Table, one also notices significant differences in the overall participation rates between natives and immigrants, which are, however, no larger than the employment gaps. Low labour market participation is particularly pronounced for new non-western immigrants, both men and women, but also for newcomers from western countries. For earlier cohorts of male immigrants we do not observe lower rates of participation; on the contrary, these people seem to participate in the labour market to the same or an even larger extent than comparable natives. The same holds true for ethnic German immigrants. Ó The Author Journal Compilation Ó IOM

Kogan The fact that the q-coefficient is significant in all Heckman probit models indicates that there are indeed important selection processes taking place. Accounting for selection somewhat reduces the gaps between natives and immigrants in employment, which is also emerges from comparison of the coefficients from the presented model and the probit model for employment without a selection correction (the model is not shown, but available from the author upon request). Selection effects suggest that if immigrants increased their participation rate, their employment prospects vis-a-vis native Germans would actually be smaller. This implies that those who are not currently participating in the labour market can show better employment performances than currently active taking into account observable socio-demographic and educational characteristics. The existence of negative selection, which was more pronounced when modelling recent immigrants (not shown here), may be an indication of the fact that some persons with personal characteristics favourable for employment, are in fact not economically active at all, which might be attributed to their status as refugees or asylum seekers. Occupational attainment: Results of the multivariate analyses Not only are immigrants in Germany disadvantaged regarding employment, they also occupy lower ranks in the occupational hierarchy, which we already referred to in our presentation of the descriptive results. In the current section, results of the multinomial logistic regression of class attainment measured against the ESeC class schema collapsed into categories (as shown in the methodological part), as are shown in Table for all immigrants, and in Table for recent immigrants. In accordance with the earlier results pertaining to the employment patterns, here once again we observe larger occupational disadvantages for non-western immigrants disregarding which occupational category we are looking at. Occupational disadvantages seem to be more pronounced for earlier cohorts of immigrants when it comes to the contrast between the salaried and unskilled occupations, but lower when self-employment and skilled manual work are juxtaposed with unskilled work. These findings are more or less the same for both genders and confirm our hypotheses on the milder disadvantages for newer immigrants in accessing more prestigious jobs, but with higher risks of unemployment. Controlling for education, we observe significant reductions of ethnic penalties in all occupations (as opposed to unskilled work) for earlier cohorts of immigrants. For more recent immigrants the same can be said only for the contrast of skilled manual labour, but not for the salar- Ó The Author Journal Compilation Ó IOM

xxxx TABLE ESEC MULTINOMIAL LOGISTIC MODEL, BY GENDER Salariat and intermediate Petty bourgeoisie -edu Skilled manual -edu -educ. -educ. -educ. Men West new )0.** )0.** )0.** )0.** )0.** )0.** Non West new ).** ).0** ).** ).** )0.** )0.** West old ).0** )0.** )0.** )0.0 )0.** )0.0** Non West old ).** ).** ).0** )0.** )0.** )0.0** Ethnic Germans ).** ).** ).0** )0.** )0.** )0.** Pseudo R 0. Women West new )0.** )0.** )0. )0. )0.** )0.* Non West new ).** ).** ).0** ).0** )0.** )0.** West old )0.** )0.** )0. )0.0 0.** 0.** Non West old ).0** ).** ).0** )0.** 0.** 0.** Ethnic Germans ).** ).** ).** ).** 0.00 0.0 Pseudo R 0. Source: German micro-census, 00, 0. Note: *p<0.; **p<0.0; Unskilled and low-level routine non-manual is a reference category. Variables included into the equation: age, age squared, family status, observation year, region (dummy for East Germany), education (see Table for similar application) iat and self-employment. The disadvantages for new immigrants indeed further increase once we control for education in the case of salaried workers. These findings imply different returns on education for native Germans and education obtained abroad by immigrants. To estimate access to various occupations eschewing any assumption of equal returns on education, we ran the models allowing for the interaction of education and membership in various immigrants cohorts. To ease the interpretation of results we plot predicted probabilities of access to salaried and unskilled work (other destinations are not shown but available upon request) by education, holding all other covariates at their means (see Figure a and b). The figures suggest that the higher the level of education, the larger the gaps between natives and immigrants. More educated immigrants, and above all those coming from outside the EU, are more likely than comparable natives to be employed in unskilled sector. This is true for both men and women. The order of immigrants with regard to their employment chances in unskilled sector is also more or less similar for both genders. Natives have the lowest probability of being employed in Ó The Author Journal Compilation Ó IOM

Kogan TABLE ESEC MULTINOMIAL LOGISTIC MODEL, BY GENDER Salariat and intermediate Petty bourgeoisie Skilled manual -educ. -educ. -educ. Men Italy ).0 ** )0.0** )0. 0.0 0.0 0. Iberia ).** )0.** )0. )0. 0.** 0.** Yugoslavia ). ** )0.** ).** ).** 0.** 0.* FSU )0.** )0.** )0.* )0.** 0.* 0. Poland )0. )0. 0.** 0.0** 0.** 0.** Other EE 0.** )0.0 )0. )0. 0.** 0.0** Turkey ). ).** )0.0** )0.** )0.** )0.* Africa ).**** ).** ).0** ).** )0.** )0.** Asia )0.** )0.** )0.* )0.* )0.** )0.0** America 0. )0.0 )0. )0. 0.0 0. Ethnic Germans ).** )0.** )0. )0. 0.** 0.** YSM )0.0** )0.0 )0.0 0.00 0.0** 0.0** Pseudo R 0. Women Italy ).** ).** )0. )0. )0. )0. Iberia ).** )0.** ). ). )0. )0. Yugoslavia ).** )0.** ).** ).** )0.** )0.** FSU )0.** )0.** )0. )0. )0. )0. Poland )0.0** )0.** 0. 0. )0.0 )0.0 Other EE )0.** )0.** )0. )0. )0.0 )0.0 Turkey ).** ).** ). ). 0.00 0.0 Africa ).0** ).** )0. 0.0 )0. )0. Asia ).0** ).00** )0. )0. )0. )0.0 America )0.** )0.** )0. ).0 ).** ).** Ethnic Germans )0.** )0.** ).** ).** 0.0 0.0 YSM 0.00 0.0** )0.0 )0.0 0.0 0.0 Pseudo R 0. Source: German micro-census, 00, 0. Note: *p<0.; **p<0.0; Unskilled and low-level routine non-manual is a reference category. Variables included into the equation: age, age squared, family status, observation year, region (dummy for East Germany), education (see Table for similar application). unskilled jobs, followed by EU immigrants from earlier cohorts, followed by more recent EU immigrants. The least favourable chances for higher- status jobs and stronger representation in the unskilled sector fall on immigrants from outside EU and other western countries, from both cohorts. Ethnic German immigrants are in-between, resembling immigrants from outside EU- more than Western immigrants. For this group we notice particular disadvantages experienced by tertiaryeducated immigrants in their ability to attain a salaried position (a phenomenon particularly pronounced among women). Highly-educated Ó The Author Journal Compilation Ó IOM

xxxx ) b) 0 0 0 0 0 0 0 0 0 0 0 0 0 Men Primary Low sec. Sec. voc Sec. gen Post-sec. Tert.short Tertiary Native-born Germans West new Non West new West old Non West old Ethnic Germans Men Primary Low sec. Sec. voc Sec. gen Post-sec. Tert.short Tertiary Native-born Germans West new Non West new West old Non West old Ethnic Germans FIGURE A: ACCESS TO THE SALARIAT BY EDUCATION (PREDICTED PROBABILITIES FROM MULTINOMIAL LOGISTIC REGRESSION WITH INTERACTION TERMS FOR EDUCATION AND IMMIGRANT GROUP). B: ACCESS TO THE UNSKILLED OCCUPATIONS BY EDU- CATION (PREDICTED PROBABILITIES FROM MULTINOMIAL LOGISTIC REGRESSION WITH INTERACTION TERMS FOR EDUCATION AND IMMIGRANT GROUP) 0 0 0 0 0 0 Primary Low sec. Sec. voc Sec. gen Post-sec. Tert.short Tertiary Primary Low sec. Sec. voc Sec. gen Post-sec. Tert.short Tertiary Aussiedler have disproportionally higher chances of ending in unskilled occupations, with women standing out once again. Different returns on education for various immigrant groups are also statistically significant (results are not shown but available upon request). In the final analyses we compare how newer immigrants fare with regard to occupational status as compared to immigrants from Northern and Western Europe. Results are shown in Table, in which the access to the salaried and intermediate positions, petty bourgeoisie and skilled manual labour is contrasted to unskilled employment and low-level routine non-manual jobs. Each contrast is shown once without and once 0 0 0 0 0 0 0 0 Women Native-born Germans West new Non West ne West old Non West old Ethnic Germ Women Native-born Germans West new Non West new West old Non West old Ethnic Germans Source: German micro-census, 00, 0. Note: For reference see Table. All other covariates are at their means. Ó The Author Journal Compilation Ó IOM

Kogan with controlling for education. In addition to the ethnic effects we include the coefficient for YSM. Consistent with our earlier findings new immigrants are mostly disadvantaged when it comes to entering salaried jobs. The lowest log odds of entering this class position are exhibited by Turks and immigrants from Africa (both men and women), male immigrants from ex-yugoslavia and southern European countries. The least disadvantaged (as compared to immigrants from western and northern Europe) appear to be immigrants from Latin America, Eastern European countries (except for Poland) and FSU. Ethnic penalties on entry into salaried positions (as opposed to unskilled occupations) among immigrants (both male and female) from southern Europe, Turkey and ex-yugoslavia, as well as among recent cohorts of ethnic Germans are at least partially explained by their lack of education. This means that the penalties decrease once education is taken into account. The same holds true for women coming from Africa and Asia, and men coming from Poland. For immigrants from the FSU, male immigrants from Africa and Asia, female immigrants from Poland, Eastern Europe and Latin America, employment penalties became even more pronounced once education is taken into account. The lowest odds for opening small businesses (compared to entering unskilled labour) are observed among male immigrants from the former Yugoslavia and Africa, or female immigrants from the former Yugoslavia and ethnic Germans. Education is less powerful in explaining ethnic differences with regard to self-employment than it is the case with the entry into a salaried job. Entry into skilled manual employment as opposed to unskilled labour is another important mobility venue for male immigrants. In Germany, the best chances of avoiding unskilled positions are found among newcomers from Poland and Eastern Europe, the lowest chances are reserved for immigrants from Africa, Asia and Turkey. With longer tenures in the host country, the chances for skilled manual work (as opposed to unskilled jobs) increase among male immigrants. This is not the case for those entering salaried professions. It seems people coming more recently have better prospects of entering salaried professions (as opposed to unskilled work), which might be related to the increased demand for high-qualified professionals in Germany in very recent years. Among women, one notices the positive effect of YSM for Ó The Author Journal Compilation Ó IOM

xxxx entry into salaried professions as opposed to jobs in unskilled manual sector. CONCLUDING REMARKS Until recently the main explanation for immigrants inferior placement in the German labour market has been their lack of resources relevant to succeed in the rigid and highly credentialist labour market of Germany (Kalter and Granato, 0; OECD, 0). Indeed, as found elsewhere and again shown in this paper, education plays a very important role for the success of immigrants in Germany. This means that for immigrants with lower education who largely arrived in 0-0s and were later on accompanied by their family members, lack of education appeared to hinder their professional mobility. For more recent cohorts of immigrants, that is, those arriving after 0s, who on average are better educated than their predecessors and in some cases even more qualified than natives, education is largely discounted in the labour market and is practically useless for access to suitable occupations. In accordance with our hypothesis, more recent cohorts of immigrants appear to be more ambitious in their professional expectations, and indeed manage to enter higher-level jobs as compared to their counterparts from earlier immigrant waves. Being able to access better-quality jobs, these immigrants suffer, however, from larger penalties upon entering work. For the older immigrant cohorts, gaining employment is also a problem, albeit less so than for more recent immigrants, but they target clearly less prestigious occupations, which are somewhat easier to attain. Western immigrants in Germany clearly have better labour market prospects than non-western immigrants, irrespective of the timing of their migration. Apparently, more privileged legal status and easier transferability of qualifications pay off in the labour market. Although we expected a successful integration of ethnic German immigrants, our predictions were substantiated only partially. Aussiedler do fare somewhat better than non-western immigrants in Germany in almost all indicators of labour market integration measured in this study, but they fail to match Western immigrants in this respect. Education obtained abroad is not discounted among ethnic Germans to the same degree as among other immigrants, but when it comes to the access to highly qualified positions, German employers seem to be sceptical about suitability of Aussiedler, even those with high-level qualifications. Ó The Author Journal Compilation Ó IOM