Gender-Specific Labor Market Performance of Mediterranean Immigrants in Germany and Hispanic Immigrants in the United States Compared.
This article examines labor market mobility of Hispanic immigrants in the United States and immigrants from recruitment countries in Germany. The data comes from two similar longitudinal samples conducted in Germany and the United States. This allows for a comparison of labor mobility between immigrants and natives in both countries. The comparison reveals a higher degree of mobility on all levels among immigrants in the United States. Over time, Hispanic immigrants were able to reduce the income gap compared to natives while this gap increased in Germany. This higher degree of mobility can be traced back to two factors: A higher human capital among Hispanic immigrants and a more liberal labor market regime in the United States.
Introduction: Different approaches towards the integration of immigrants in Germany and the United States
THE UNITED STATES AND GERMANY have very different attitudes towards immigration. The United States has a tradition of immigration which is still alive. Germany has 7.3 million foreigners living on its territory but has not officially declared itself a country of immigration. Therefore, no law exists that regulates legal immigration into Germany (Dittgen, 1998). There are laws for certain groups like ethnic Germans (Aussiedler), and for certain types of migration like family unification, but no general immigration law exists (Neuman, 1998). Before a change in law in 1991, becoming a German citizen was difficult and also expensive. This was a result of Germany's definition of citizenship that was based on descent (ius sanguinis). Since the change in law, immigrants are entitled to citizenship after having lived in Germany for 15 years. Starting in the year 2,000, immigrants can apply for citizenship after being in Germany for 8 years. Citizenship might not be granted if the applicant does not speak German, has a criminal record or depends on unemployment or social welfare benefits. The children of immigrants in the year 2,000 will receive dual citizenship. However, when they reach 23 years of age, they will have to decide which citzenship to keep.
Until the 1970s immigration in the United States was regarded as being the result of a self-selection process. This was based on the assumption that only the best and most highly motivated workers would come to the United States (Chiswick, 1978, 1980). Therefore, positive assumptions were made with regard to the integration of immigrants. It was assumed that immigrants would rapidly adapt to the requirements of the labor market and would achieve the average American income within 15 years (Chiswick, 1978, 1980). Later longitudinal studies reveal that the period in which immigrants reach the American mean is substantially longer (Borjas, 1985; Chiswick, 1986). This was caused by a change in the composition of immigrants. In 1960, three quarters of the foreign born in America were of European origin. Thirty years later, less than one-third were of European origin. The average level of education of immigrants, however, decreased not only as a result of the lower percentage of European immigrants but decreased a lso within certain immigrant groups. Whereas male Mexican immigrants who arrived before 1960 had on average completed 9 years of school, those Mexican men who arrived between 1987 and 1988 had only completed 5.3 years of school (Bean et al., 1994:86).
In comparison to immigration in the United States, Germany had a completely different pattern of selection. Between 1955 and 1973 immigrants were recruited by the State. The employment of foreign workers was considered to be a short-term solution to bridge periods of extreme labor shortage. Foreign workers received job contracts only for a specified period of time (Seifert 1995). This "guest-worker system" was intended to keep the foreign labor force flexible and adaptable to the demands of the German labor market. Thus, during the economic recession of 1966-67 the number of foreign workers employed decreased sharply and their number increased again quickly during the subsequent economic recovery (Hollifield, 1992). After the 1973 oil crisis, the recruitment of foreign labor was halted, though, the number of foreigners in Germany remained at a high level. Many of the foreign workers brought their families and prepared themselves for a long-term or even permanent stay. With the halt of the rotation of workers , temporary immigration became permanent. Despite increasing levels of unemployment and structural changes, the German industry remained unable to do without foreign workers. Foreign employees were concentrated in industrial sectors which were unattractive to German workers.
Due to the specific recruitment process, predominantly young men were selected. As permanent immigration was not planned, immigrant workers were restricted to jobs that did not require any specific qualification. Training would have been pointless for people who were to stay for only a short time. Therefore immigrants were concentrated in industrial mass production and heavy industry. Correspondingly, guest workers had on average low levels of education and were integrated into the lowest segment of the labor market. Furthermore, the German dual system of education and vocational training, which requires specific certificates for the advancement into higher sectors of the labor market, limited the mobility of the immigrants (Seifert, 1995, 1996a). With such conditions, an equal distribution in the employment hierarchy, even for the second generation, was not to be expected.
The United States has experienced various patterns in the process of integration of immigrants. Cubans in America tend to be economically successful within a short time after their arrival (Portes and Jensen, 1989). Haff of the refugees of the Mariel boat-lift were employed in firms owned by Cubans and 20 percent are self-employed (Portes and Rumbaut, 1990:92). Mexicans in America also have a high labor force participation rate (Rosenfeld and Tienda, 1997). With the exception of Japanese immigrants, Asian immigrants tend to have a lower household income than non-Hispanic Whites. However, on average they are in a more favorable position than Mexican immigrants (Lee and Edmonston, 1994). These differences in terms of economic success have been traced to differences in education, qualification, and resources. "The scarcity of human capital, characterizing immigrants of modest origin, makes raw physical power their principal marketable asset in the American labor market" (Portes and Rumbaut, 1990:84).
In Germany, differences by country of origin are not as pronounced as in the United States. This is a result of the recruitment process. Although unskilled workers were recruited in all countries of origin, nationality-specific differences can be seen. Turkish immigrants in particular tend to be employed in the lowest segment of the labor market than other immigrants, and Turks are more often unemployed (Bender and Seifert, 1996). This can be explained by different human capital. Turks have come from rural areas in higher proportions than other immigrants. Within Germany, Turks can more often be found in the north and the west, especially in the Ruhr area, which is an area with high unemployment. While 45 percent of the Italians live in economically strong states (Lander), such as Bavaria and Baden-Wiirttemberg, only 30 percent of the Turks live there.
Many female immigrants come to Germany as dependent family member. Therefore, they often do not have the necessary skills for the labor market and are thus concentrated in low-paid jobs.
The following comparisons of immigrant groups in the United States and Germany will be based on two theoretical approaches. On the one hand, it is assumed that the human capital of immigrant groups is relevant for their occupational careers. It is assumed that immigrants with higher language proficiency, longer on the-job training and higher levels of qualification will advance into more attractive occupational positions outside of industrial mass production. For the children of immigrants (i.e., the second generation) it can be assumed that German school certificates will offer better opportunities for occupational advancement. But the degree of mobility does not exclusively depend on individual characteristics. It is also possible that the disadvantaged position of immigrants results from the structure of the labor market. This hypothesis is reflected in the theory of labor market segmentation (Piore, 1978; Blossfeld and Mayer, 1988; Seifert, 1995, 1996b). The assumption of this theory is that several sect ors of the labor market are systematically closed to immigrants. If this holds true, no important intergenerational mobility can be expected. If labor markets are not segmented along ethnic lines, and career lines are not dependent on characteristics like ethnicity and gender, upward mobility can be expected with increasing human capital.
The analyses are based on longitudinal samples of the German Socio-Economic-Panel (GSOEP) for Germany and the Panel Study of Income Dynamics (PSID) for the United States. The GSOEP is a longitudinal sample, first conducted in 1984, which contains separate representative samples drawn from the central register for immigrants (Auslanderzentraliregister) for the five most important immigrant groups from the recruitment countries Turkey, former Yugoslavia, Italy, Greece and Spain. The total sample contains 3,000 immigrants and 9,000 German nationals. The sample has a similar structure as the PSID sample. There also exists an equivalent file of GSOEP and PSID data. As 97 percent of these immigrant groups were concentrated in the Old Federal States (Alte Bundeslander), only these states are analyzed here.
In the following, education, occupational and income positions of Mediterranean labor migrants from Turkey, fromer Yugoslavia, Italy, Greece and Spain in Germany will be compared to Hispanic immigrants in the United States. Mobility patterns will be analyzed in particular. The term Hispanics in this context includes all people of Hispanic ancestry. It also includes those who were born in the United States (second generation). In the PSID study, the second generation cannot be isolated, as no information on the birthplace of the parents is available. Those who denote themselves as Mexican-American belong to the second generation.
Education of Immigrants in the United States and Germany
First, the degree of education among immigrants of Hispanic decent in the United States and Mediterranean immigrants in Germany will be examined.' The possibilities of a direct comparison of education between Germany and the US are limited due to the different systems of education.  Differences in the educational structure of immigrants and natives in the two countries will be examined. During the period examined, the education variable in the PSID data set was changed. Therefore, no longitudinal observations are possible; data on education were available only for 1990. Education can be considered a stable feature; drastic changes are therefore not to be expected. Due to this stability, for Germany only the years 1990 and 1995 were taken into account instead of considering all five years (1990-1995).
In 1990, the educational structure of Hispanic men differed considerably from American men. Among Hispanic men, the share of those who attended school for a maximum of 8 years was above average (Table 1). Nevertheless, Hispanics cannot exclusively be found in the lower educational groups. At least 43 percent completed high school and further 8 percent attended college. The educational structure differs considerably among the various immigrant groups. Mexican men on average had the lowest educational degrees. Fifty-nine percent of them attended school for a maximum of 8 years. Educational degrees above average can be found for the group: "other Hispanic men," especially Cuban men. Fifty-eight percent of the Cubans had a high school diploma and further 10 percent had attended university.
The educational structure of American women is similar to American men except that women have a lower share of college degrees. Hispanic women also have a similar educational structure as Hispanic men. Among women, the same nation-specific differences can be seen as observed for men. A positive result can be seen for Mexican-American women of which 54 percent have completed high school compared to 46 percent among men of the same group. Moreover, the share of college degrees is higher among Mexican-American women. Since most of this group are born in the United States, this shows a high degree of intergenerational mobility among women.
Due to this educational structure, it can be expected that Mexican immigrants will be concentrated in the lowest segments of the labor market. Some Hispanic immigrants, however, should also be found in the higher segments of the labor market.
For German men, lower secondary school leaving certificates (Hauptschulabschlu[beta]) are the most common certificates (Table 2). Between 1990 and 1995 the share of those with upper secondary leaving certificates (Abitur) increased up to some 25 percent. The share of the other education certificates remained more or less on the same level as in 1990. In 1990, 20 percent of men from Mediterranean countries had no leaving certificates at all, and 44 percent had attended elementary school abroad. At least one quarter of the foreign men had attained a school leaving certificate in Germany. The majority of them gained a leaving certificate of a lower secondary school; higher certificates are scarce. Overall, no major changes in the educational level can be observed for immigrants from Mediterranean countries between 1990 and 1995. The share of those with an educational certificate increased while the group of those without a degree decreased.
In 1995, the educational structure of Turkish immigrants was more or less the same as that of all immigrants from Mediterranean countries. Among second generation men three quarters had a lower secondary degree in 1990. In 1995, the educational structure was more favorable. Then, 68 percent had a lower secondary school certificate, 22 percent had a intermediate school certificate, and 8 percent had a higher secondary degree.
Differences can be seen between German women and men, especially in terms of higher secondary school certificates. In 1990, the share of women with a higher secondary certificate was 14 percent compared with 22 percent of men. This share increased until 1995 for men as well as for women by 3 percent, so that the differences remained the same. German women however had a higher share of intermediate school certificates than their male counterparts. The share of those with lower secondary degrees was lower among women than among men.
Among Mediterranean women, the share of those without educational certificates is higher compared to men. In 1995, 25 percent of the women and 17 percent of the men did not have a degree. The share of secondary degrees attained abroad is lower among women compared to men. With regard to German degrees, no significant differences can be observed between Mediterranean women and men. The educational degree of Turkish women did not differ substantially from that of all immigrant women. Women of the second generation, however, were more successful than second generation men. This is especially true for intermediate school certificates which were gained by 30 percent of the women compared to 22 percent of men. Between 1990 and 1995 the share of women with intermediate and higher secondary school certificates increased significantly.
Labor force participation, and unemployment in Germany and the United States
In the following paragraph, it will be shown where immigrants are active in the labor market and to which degree they are affected by unemployment. The United States is a country of high employment. In 1990, 87 percent of American men at working age were gainfully employed (Table 3). The data for 1992 refer to the same individuals as in 1990; a decline in the labor force participation rate was to be expected since two birth cohorts had reached retirement age. In 1992, the labor force participation rate was 83 percent. American women were also active in the labor force to a high degree. In 1992, two out of three women were active in the labor force. Male Hispanic immigrants in the United States have higher labor force participation rates than American men. In contrast, the participation rate of Hispanic females was lower than that of American women. Fify-four percent of woman with Latin American backgrounds were active in the labor force. While the country of origin makes no differences for men, labor force p articipation differed considerably for the respective female groups. For instance, in 1992, the labor force participation rate among Cuban and Mexican women was below average. Back in 1990, however, these groups had participation rates above average. A clear increase in the labor force participation can be shown for Mexican-American women.
Compared to the United States the labor force participation rate in Germany is low. Three quarters of German men and half of the women were active in the labor force (Table 4). Parallel to Hispanic immigrants in America, in 1990 men from the recruitment countries had a higher labor force participation rate than German men. But this has changed. In 1995, only two out of three men from Mediterranean countries were active in the labor market. Among Turkish men this decline was especially high; the participation rate declined from 76 percent to 61 percent. The participation rate of second generation men was also below average. But within this group some still have not finished their education or occupational training. This can be seen in the increasing participation rate during 1990 to 1992. Among women, the labor force participation rate declined only insignificantly. The participation rate of Mediterranean women (44 percent) was only slightly below that German woman (47 percent). Turkish women, however, had a participation rate below average. The participation rate of second generation women was only slightly below average.
The period 1990 to 1995 analyzed here is characterized by different economic developments in Germany and the United States. In America, the employment growth was higher than in Germany. This can be seen clearly in the unemployment rates, which were lower in the United States. In 1995, the average unemployment rate in the United States was 5.6 percent. Some groups, however, had unemployment rates above average. Thirty percent of young African-Americans, for example, were unemployed. Residents of Hispanic background, too, were unemployed above average (Werner 1997a). The unemployment rate varies considerably by season. Different seasons, in which the interviews were held, might distort the comparison between Germany and America. In order to avoid this problem, unemployment here refers to at least one period of unemployment during the year before the interview. Both data sets contain monthly data on unemployment, which allows us to reconstruct the employment history in the year before the interview. In this con text a person is defined as unemployed if he or she was unemployed at least for a month during the 12 month preceding the interview.
In the United States, men were more often affected by unemployment than women. This holds true for American women as well as for Hispanic women with the exception of Mexican women. In 1992, 12 percent of the American male labor force was unemployed in the preceding year, for females the share was 8 percent. In 1990, Hispanic immigrants were especially affected by unemployment (19 percent). This share decreased to 15 percent in 1992. Especially for Cubans, important deviations from the average can be seen in both years. While in 1990 their unemployment was above average (29 percent), it was below average (9 percent) in 1992. Cubans entered the United States in large numbers during the late 1980s. By 1990 they were still not integrated in the labor market, as can be seen in the high unemployment rate. However, the low unemployment rate in 1992 indicates that unemployment is a temporary experience for most of the immigrants.
The unemployment rate for females with Hispanic background was above average. But in contrast to Hispanic men, their unemployment rate still increased. Mexican women were unemployed above average while among Cuban women only 3 percent were unemployed in 1992.
In Germany, gender-related differences in terms of unemployment are minor. However, Mediterranean immigrants were more often unemployed than Germans. In 1990, the unemployment rate among immigrants was twice as high as among German men. This gap became even larger. In 1995, 22 percent of the immigrant men were unemployed for one month within one year. Within the German labor force this share increased from 6 percent to 9 percent. The Turkish labor force in particular was unemployed above average. In 1995, one in four Turkish men was unemployed. Moreover, second generation immigrants also had a high level of unemployment. Unemployment rates for foreign women were similar to those of foreign men. Among women, the group of Turkish descent and the second generation were affected by high unemployment. In 1995, both groups of women experienced a 31 percent unemployment rate. The transition from school to work in 1995 was more difficult for second generation women than for men, whereas these women had lower unemplo yment rates in 1990.
The employment situation of the foreign labor force deteriorated in Germany during the 1990s. Turkish immigrants and the second generation were affected in particular. In the United States, unemployment was higher among Hispanic immigrants than among Americans.
Greater fluctualtions in unemployment rates can be observed in the U.S. labor market. Groups which had high unemployment rates in 1990 had low rates in 1992 and vice versa.
Occupational positions of immigrants in Germany and the United States
The following section highlights the occupational positions of immigrants in Germany and the United States. 
The United States can be described as a so-called service society. In 1992, only one quarter of American men held a blue collar position. More than half of the labor force were employed in white collar positions, and one in five men were self-employed (Table 5).
In the short time between 1990 and 1992 the share of lower white collar positions declined considerably among all immigrant groups, whereas the share of middle and high level white collar positions increased. This development indicates that the American labor market is not ethnically segmented. Immigrants are mobile to a high degree. Moreover, they are able to improve their occupational positions within a short period of time.
American women scarcely work in blue collar positions. In 1990 as well as in 1992 only 8 percent were employed in these occupations. Hispanic women, especially those from Mexico, worked more often in blue collar positions. However, Hispanic women can be found more often in white collar positions than men. Their share in low level white collar positions was about the same as that of American women. Among the respective groups different patterns of mobility can be observed. While the share of low level white collar positions increased among Mexican-American women, it decreased among Mexican and Cuban women. The latter can be found more often in middle and high level white collar positions now. For immigrant women a high mobility can be observed. This mobility indicates that there were no indications for a gender-related segmentation of the labor market that would disfavor women.
In 1990 male Hispanic immigrants were employed above average as unskilled and semi-skilled workers. This share declined until 1992 when only one out of five Hispanic males were employed as unskilled and semi-skilled workers. Mexican immigrants were concentrated in unskilled and semiskilled positions above average. Their share in these positions declined, however, from 38 percent in 1990 to 30 percent in 1992.
The employment profile of immigrants from the recruitment countries in Germany differs more from that of the German labor force than do Hispanics compared with Americans in the United States. The distribution of occupational positions of foreign workers still reflects the fact that the recruitment process focussed on jobs for which only a low level of qualification was required. In 1990, 55 percent of immigrant men worked as unskilled or semi-skilled workers (Table 6). This share declined until 1995, when 48 percent of foreign men were employed in such positions. The share of skilled workers increased only slightly. In 1995, foreign men were more often employed in white collar positions than in 1990. Their share (13 percent) is still very low compared to German men (42 percent) in these positions. German men were also more often self-employed than Mediterranean immigrants.
Among Turkish men the decrease in blue collar positions was stronger than among all immigrants. In 1990, they were still above average in unskilled and semiskilled positions. In 1995, the share of Turkish men in these positions reached the average level of immigrant men. Turkish men can now be found in skilled worker positions above average. In white collar positions they were still underrepresented. Considerable occupational mobility can be seen within the second generation. While in 1990, almost every second man was employed as unskilled or semi-skilled worker, this share declined to 32 percent in 1995. Blue collar jobs remained, however, the most important employment area for this group. At least 42 percent were employed as skilled workers (1995). In white collar positions the second generation is still underrepresented.
In 1990, women from the recruitment countries were still more often employed as unskilled or semi-skilled workers than men. Their numbers declined considerably. In 1995, 64 percent of the immigrant women were employed as unskilled or semi-skilled workers. Only small numbers of immigrant women work in blue collar positions; their share in white collar positions, however, is higher than that of immigrant men. The share of those in middle and higher positions was twice as high as for men. Turkish women in 1990 were represented in higher shares in unskilled or semi-skilled positions than immigrant women on average. In contrast to men their numbers remained above the average of all female immigrants. Fewer women of the second generation, however, worked in unskilled and semi-skilled positions compared to second generation men. In 1995, only one in four second generation women worked in these positions. In 1990, 15 percent of second generation women worked in middle and higher level white collar positions, but in 1995 this share increased to 39 percent. A further 35 percent were employed in low level white collar positions.
Between 1990 and 1992, the occupational mobility among Hispanic immigrants in the United States was higher than among immigrants from the recruitment countries in Germany. In the United States more group-specific differences can be observed than in Germany.  Only for the second generation in Germany are gender-specific differences clear: Women can predominantly be found in white collar jobs while men were able to establish themselves in skilled worker positions.
Income positions and dynamics among immigrants in Germany and the United States
Before the income development among immigrants in Germany and America will be dealt with, some basic developments which are relevant for the interpretation of the data will be discussed.
On the basis of an average national gross income, adjusted for inflation and purchasing power parities, it can be seen that the increase in income was low in the United States. In 1983 the average hourly income was at $16.10 (U.S.) and by 1992 it had reached $17.20 (U.S.). In West Germany, the hourly average income increased from $12.40 to $16.20 (U.S.) (Werner, 1997b:590). This means that the increase in employment in this period was based on slow increasing wages in the United States.
The distribution of income in Germany differs and from the United States. In America, the income inequality is higher than in Germany. Whereas the income inequality in the United States increased between 1979 and 1993, this was not the case for Germany (Steiner and Wagner, 1997). In the United States a contrary development can be seen in the lower and the higher income positions. While in the highest income decile, i.e., those 10 percent of the American employees with the highest income, the wages and salaries increased by 11 percent, the income in the lowest decile decreased by 4 percent (Werner, 1997b:590).
In the following paragraph, it will be analyzed in which income positions immigrants are concentrated. First, the gross mean income will be analyzed. Then by means of a regression model, it will be examined which factors are relevant for the height of the income. No conversion into one currency or adjustment for inflation will be used. These transformations are not necessary because the differences in the income will be analyzed between locals and immigrants in both countries and not between the two countries.
At first sight, differences in hourly wages and salaries between Hispanic and American men are obvious in 1990 (Table 7). The occupational mobility of Hispanic men between 1990 and 1992 had a positive impact on their income. In this period, Hispanic men had higher wage increases than American men. In 1990, Hispanic men earned 73 percent of the average income of Americans, and in 1992 they earned 87 percent. Parallel to the income mobility, different income developments can be observed for the various immigrant groups. In 1990, Mexicans and Cubans gained wages clearly below the average. While the income of Cubans increased substantially and was above average by 1992, the wages of Mexican men remained at about $9.80 (U.S.) and below the average of Hispanic men.
In 1990, income differences between Hispanic and American women were less significant than among men. This was not caused by higher earnings of Hispanic women but by the low income of American women. Between 1990 and 1992, Hispanic women were able to narrow the gap to the average wage level of American women. In 1990, they earned on average 80 percent of the income of American women; in 1992 they reached 94 percent. For women considerable group differences can be observed. Cuban women (as well as Cuban men) gained wages below average in 1990 but reached the average income level in 1992. Mexican women had an increase in earnings from $6.60 to $8.20 (U.S.). However, this was still the lowest income of all groups.
In Germany, the income differences between immigrants and the local labor force were lower than in the United States. In 1990, immigrant men earned 18.90 DM on average. This is the equivalent of 86 percent of the average income of German men (Table 8). While Hispanic immigrants in the United States were able to narrow the income gap, this was not the case for the foreign labor force in Germany. Here, the income differences even increased. In 1995, foreign workers gained only an equivalent of 80 percent of the average German earnings. Turkish men gained earnings above the average in 1990 but their income development was not stable. Therefore, in 1995 their income was below the average of all Mediterranean immigrants. The earnings of the second generation, too, were below average. Their higher level of qualification had no impact on their wages so far. They are still at the beginning of their job career and consequently have not reached well-paid positions yet. Between 1990 and 1992 a substantial increase of i ncome can be observed. In the years 1993 to 1995, too, the level of income increased. However, this increase was not as substantial as in the previous years due to high pressures on the labor market.
Women from the recruitment countries were not able to improve their income position compared to German women. In 1990, they gained an equivalent of 84 percent of the average German woman's income. In 1995, they gained 83 percent. The earnings of Turkish women are more or less on the same level of all foreign women. Women of the second generation had an average income per hour of 11.60 DM, which is very low. Until 1995 their wages increased but their average earning of 15.70 DM per hour was lower than that of men. This was to be expected due to their lower level of education and qualifications.
The high degree of occupational mobility among Hispanic immigrants in the United States had an impact on their wage level. Considering the occupational change in Germany, a wage adjustment could have been expected. However this was not the case, the income gap even widened. The reasons for this can partially be seen in the unfavorable economic development. Especially those branches where immigrants work were especially shaken by the economic crisis and therefore the increase in wages remained low. The degree of occupational mobility was obviously not strong enough to compensate for income losses in other sectors, or they did not pay off with regard to wages.
Determinants of the level of income -- Two multivariate models
Which factors are relevant for the amount of earnings? Wage levels are considered to be a good indicator for the level of the integration of immigrants. By means of a regression model determinants of the amount of (log) earnings will be analyzed. Two models will be calculated. The first model gives the factors for the total labor force and the second only for immigrants.
For the United States the model with the total male labor force showed a weak correlation with nationality (Hispanics versus Americans). The income of Hispanic males was slightly lower than that of American men. For women almost no differences were found between Hispanic and American women (Table 9). Therefore, no far-reaching wage discrimination can be assumed for the United States.
As expected, a higher degree of education had a positive effect on the earnings of workers in general. In terms of occupational position and industry, wage differences are apparent. Membership in a trade union had a positive impact on the wage level. Those who worked in the same firm for a longer period of time gained higher incomes. Moreover, a higher age had a positive impact on the income of men but not for women. For women a slightly negative effect can be observed, that is, younger women gained higher incomes than older ones. The explained variance in the model covering the total labor force was higher for females in the total model than for males. Therefore these factors have a different explanation power depending on gender.
The model that covers Hispanics only shows interesting deviations from the model for the total labor force. Immigrants with higher education do not gain higher wages to the same degree as was observed in the model for the total labor force. While self-employment for Hispanic men was linked with higher wages, this was not the case for Hispanic women. The duration of employment in a firm had a higher correlation with income than was the case in the total model. For women this factor was more important than for men.
Within the group of Hispanics clear income differences can be seen. Especially among Mexicans and Cuban women, a strong negative deviation can be seen. For the latter, the correlation would be less negative if instead of 1990 the year 1992 would have been chosen for the regression model. The results would have been different due to the high mobility of this group. Other factors, which can not be measured here, such as language proficiency, social network strength etc., are relevant factors for the level of income.
For Germany no indications for wage discrimination of immigrants can be found. When all other factors are controlled, immigrants, especially women, gained even higher wages than Germans did (Table 10). This surprising result has to do with the specific working conditions of immigrants. They tend to work more often in night shifts, and are more often paid on the basis of piecework rates. Since they predominately work in big firms, this factor pays off. Another surprising result is the negative correlation for some educational groups. This is a result of the peculiarity of the group of those without an educational degree. On average, older men with a high share of self-employment and incomes above average belong to this group. Since this group was chosen as reference group, a negative correlation was given for other educational groups.
In contrast to the United States, where longer duration of employment in the firm had a positive impact on the wages of women but not a higher age, in Germany a positive correlation with age was found. Duration of employment only had a slight impact on the earnings. In contrast to the United States the explained variance above was higher for men (33 percent) than for women (20 percent).
In the regression model for Mediterranean men the explained variance is higher than in the model for the total labor force. For women, however, the explained variance for the immigrant model is lower than for the labor force. The factors have a different explanation power for men and women. Especially for foreign men the impact of a higher education is low. As in the model for the total labor force, men without an educational degree had incomes above average. Since they are the reference group, there is a negative correlation for other educational groups. Apart from that, the number of those in the higher educational groups is still small. Most of them belong to the second generation and were at the beginning of their careers and therefore have on average low wages.
As expected, those groups in higher occupational positions gain higher wages and salaries. While self-employed males gained high incomes, this was not the case for females. For women no significant income differences by industry can be seen. Men, however, gained a higher income in social and public services in contrast to construction, wholesale, retail, traffic, and consumer oriented services. A higher age is generally linked with higher income. Compared to other immigrants, no major differences can be noted for the group of Turkish employees.
The analysis clearly shows that higher qualifications pay off for immigrants in the American labor market as well as in the German labor market. Immigrants do not receive lower wages than Americans and Germans respectively. Their lower wage level is caused by different levels of education. The income mobility among immigrants was higher on the U.S. labor market than on the German labor market. In the United States, the income gap between Hispanics and Americans narrowed whereas the income gap between Mediterranean immigrants and Germans increased.
The different patterns of occupational and income mobility among immigrants in the United States and Germany must be seen, first of all, in the context of the different economic development in these countries. While unemployment increased in Germany, new jobs were created in the United States. The fact that immigrants in Germany had a lower income growth rate than Germans indicates a different income development in the segments of the labor market where immigrants were concentrated. Werner (1997b) showed in his study that wages declined in the lowest segment of the U.S. labor market since the beginning of the 1980s. In this specific segment, Hispanics are concentrated. Therefore, one could have assumed that this decline in wages would affect Hispanic immigrants in particular. However, the immigrants were able to compensate this negative development with a high degree of occupational mobility. This higher degree of mobility must be explained by the different educational structure of the immigrants and the dif ferent structure of the labor markets. On average, it is common for immigrants from the recruitment countries in Germany to have low educational degrees. The second generation, which obtained German degrees, also on a middle and higher level, is still at the beginning of their occupational careers. Therefore they have not reached well-paid positions yet. Among Hispanics in the United States a concentration of immigrants in the lower educational groups can be seen. To a higher degree than in Germany there was a severe share in higher educational groups. The different structures of the U.S. and German labor markets had effects on occupational mobility. The German labor market is regulated to a higher degree than the U.S. labor market. Most of the qualified jobs require certificates that immigrants normally do not have. The German dual system of occupational vocation training is a disadvantage for immigrants who were not educated within this system.
Immigrants in the U.S. labor market started in low occupational positions, which might be an effect of a limited transferability of skills. However, when they have been in the U.S. labor market long enough, they acquire new skills and then they are able move into positions with better pay. In the U.S. labour market it is easier for immigrants to make use of their skills and require new ones than in the German labor market. Therefore education and skills of immigrants pay off better in the U.S. labor market than in the German labor market. Foreign labor in Germany is tied in the lowest segments of the labor market. Only for the second generation is an occupational advancement observed.
(*.) Humboldt-Universitat, Lehrstuhl Bevolkerungswissenschaft, Unter den Linden, 6 D-10099 Berlin, Germany
(1.) The following analyses are based on longitudinal samples of the German Socio-Economic-Panel (GSOEP) for Germany and the Panel Study of Income Dynamics (PSID) for the U.S. The GSOEP is a longitudinal sample, first conducted in 1984, which contains separate representative samples for the five most important immigrant groups from the recruitment countries Turkey, former Yugoslavia, Italy, Greece and Spain As 97 percent of these immigrant groups were concentrated in the Old Federal States (Alte Bundeslander), only these states were analyzed here. In 1990, a sub-sample for the population of Hispanic origin was integrated in the PSID data set for the first time. This sample was restricted to 382 counties in the United States with the highest concentration of Hispanic immigrants. More than 90 percent of immigrants with Hispanic backgrounds resided in these counties. Due to a complicated system of data processing of the PSID data, the last available year that was completed was 1992. Therefore, for the comparison between Germany and the United States longitudinal data sets were used for the years 1990 to 1992. Moreover, for Germany a second data set was used for some analyses for the years 1993 to 1995 since the economic background changed in this period.
(2.) For details on the German and American education systems see Faist 1995b.
(3.) The PSID data set had no comparable variable for the occupational position as available in the GSOEP data. But there are detailed data on occupation, which also contain information on the position in a firm's hierarchy. From these sources of information, a similar variable as given in the GSOEP was constructed. But in this particular point, GSOEP and PSID are not fully comparable.
(4.) See Bender, Seifert 1996.
(5.) However, this model is not specific enough to assess wage discrimination.
BEAN, F.D., J. CHAPA, R.R. BERG and K.A. SOWARDS
1994 "Educational and Sociodemographic Incorporation Among Hispanic Immigrants to the United States". Pp. 73-100 in: B. Edmonston and J. Passel (eds), Immigration and Ethnicity. The Integration of America's Newest Arrivals, London: University Press of America.
BENDER, S. and W. SEtFERT
1996 "Zuwanderer auf dem Arbeitsmarkt: Natinalitaten- und kohortenspezifische Unterschiede." Zeitschrift fur Soziologie 25: 454-76.
BLOSSFELD, H.P. and K.U. MAYER
1988 "Arbeitsmarktsegmentation in der Bundesrepublik Deutschland. Eine empirische Uberprufung von Segmentationstheorien aus der Perspektive des Lebenslaufs." Kolner Zeitschrift fur Soziologie und Sozialpsychologie 40: 262-83.
1994 "Immigrant Skills and Ethnic Spillovers." Journal of Population Economics 7: 99-118.
1997 Migration als Problem offener Gesellschaften. Globalisierung und sozialer Wandel in Westeuropa und in Nordamerika. Opladen: Leske and Budrich.
1978 "The Effect of Americanization on the Earnings of Foreign-Born Men." Journal of Political Economy 86: 897-922.
1980 An Analysis of the Economic Progress and Impact of Immigrants. Washington: Department of Labor.
1986 "Is the New Immigration Less Skilled Than the Old?" Journal of Labor Economics 4: 168-92.
1998 "Power, Perceotions, and the Politics of Immigration and Welfare." Pp. 9-28 in: H. Kurthen, I. Fijalkowski and G. Wagner (eds), Immigration, Citizenship, and the Welfare State in Germany and the United States. Part B, London: JAI Press.
1998 "Volk Natio or Nation of Immigrants? The current debate about Immigration in Germany and the United States in Comparative Perspective." Pp. 107-39 in: H. Kurthen, J. Fijalkowski and G. Wagner (eds), Immigration, Citizenship, and the Welfare State in Germany and the United States. Part B, London: JAI Press.
1992 Immigrants, Markets, and States. The Political Economy of Postwar Europe. Cambridge: Harvard University Press.
LEE, S.M. and Barry EDMONSTON
1994 "The Socioeconomic Status and Integration of Asian Immigrants." Pp. 101-38 in: B. Edmonston and J. Passel (eds), Immigration and Ethnicity. The Integration of America's Newest Arrivals, London: University Press of America.
MARTIN, P. and E. MIDGLEY
1994 "Immigration to the United States: Journey to an Uncertain Destination." Population Bulletin, 49: 1-47.
1998 "Nationality Law in the United States and the Federal Republic of Germany: Structure and Current Problems." Pp. 247-97 in: P. Schuck and R. ? (eds), Paths to Inclusion. The Integration of Immigrants in the United States and Germany, New York: Berghahn.
1978 "Lernprozesse, Mobilitatsketten und Arbeitsmarksegmente." Pp. 67-98 in: W. Sengenberger (ed.), Der gespaltene Arbeitsmarkt, New York: Campus.
PORTES, A. and L. JENSEN
1989 "The Enclave and the Entrants: Patterns of Ethnic Enterprise in Miami before and after Mariel." American Sociological Review 54: 929-49.
PORTES, Al. and R.G. RUMBAUT
1990 Immigrant America. A Portrait. Los Angeles: University of California Press.
ROSENFELD, M.J. and M. TIENDA
1997 "Labor Market Implications of Mexican Migration: Economies of Scale, Innovation and Entrepreneurship." Pp. 201-27 in: F.D. Bean, R.O. de Ia Garza, B.R. Roberts and S. Weintraub (eds), At the Crossroads. Mexican Migration and U.S. Policy, London: Rowman & Littfield Publishers.
1998 "Context of Immigrant Incorporation: Locating Dimensions of Opportunities and Constraints in the United States and Germany." Pp. 91-106 in: H. Kurthen, I. Fijalkowski and G. Wagner (eds), Immigration Citizenship, and the Welfare State in Germany and the United States. Part B, London: JAI Press.
1995 Die Mobilitat der Migranten. Die berufliche, okonomische and soziale Stellung auslandischer Arbeitnehmer in der Bundesrepublik. Berlin: Edition Sigma.
1996a "Berufliche, okonomische und soziale Mobilitat von Arbeitsmigranten zwischen 1984 und 1993." Pp. 240-63 in: W. Zapf, J. Schupp and R. Habich (eds), Lebenslagen im Wandel: Sozialberichterstattung im Langsschnitt, New York: Campus.
1996b "Neue Zuwanderergruppen auf dem westdeutschen Arbeitsmarkt." Soziale Welt 47:180-201.
1997a "Beschaftigungswunder USA. Viel Licht und viel Schatten." IAB Kurzbericht 1:1-7.
1997b "Die Arbeitsmarktentwicklung in den USA - Lehren fur uns?" Mitteilungen aus der Arbeitsmarkt- und Berufsforschung 30: 585-600.
Educational Degrees in the United States, 1990 (in percent) Elementary High High High College School School School with School and (0 to 8 without Diploma Vocational years) Diploma Training Men American 4 14 20 34 27 Hispanic 31 18 18 25 8 Mexican 59 11 17 11 2 Mexican-American 24 24 18 28 6 Cuban 24 7 22 36 10 Others 16 23 12 42 7 Women American 3 13 30 34 20 Hispanic 34 19 16 26 5 Mexican 54 12 16 18 1 Mexican-American 18 20 27 27 8 Cuban 22 9 26 34 9 Others 16 27 16 37 3 Data source: PSID longitudinal sample 1990-92. Educational Degrees in Germany (West), 1990-95 (in percent) Without Elementary Secondary Lower Inter- Certificate School School Secondary mediate abroad abroad school school Men 1990 Germans (West) 4 - - 54 19 Medit. immigrants 20 44 12 18 5 Turks 20 37 20 20 3 Second generation 0 0 1 75 19 Men 1995 Germans (West) 3 - - 52 19 Medit. immigrants 17 42 12 20 7 Turks 15 35 17 24 5 Second generation 0 0 2 68 22 Women 1990 Germans (West) 4 - - 54 27 Medit. immigrants 26 43 7 18 6 Turks 26 44 8 16 6 Second generation 0 3 1 70 22 Women 1995 Germans (West) 4 - - 50 28 Medit. immigrants 25 40 8 17 8 Turks 25 37 10 19 8 Second generation 0 2 1 61 30 Higher Secondary school Men 1990 Germans (West) 22 Medit. immigrants 1 Turks 1 Second generation 4 Men 1995 Germans (West) 25 Medit. immigrants 2 Turks 3 Second generation 8 Women 1990 Germans (West) 14 Medit. immigrants 1 Turks 0 Second generation 3 Women 1995 Germans (West) 17 Medit. immigrants 2 Turks 2 Second generation 8 Source: GSOEP (longitudinal samples 1990-92 and 1993-95). Labor Force Participation and Unemployment in the United States (in percent) Men Woemn 1990 1992 1990 1992 Labor force participation rate  American 87 83 69 66 Hispanic 91 85 54 54 Mexican 93 83 57 46 Mexican-American 89 83 53 62 Cuban 92 87 55 42 Others 88 89 51 58 Unemployment  American 11 12 8 8 Hispanic 19 15 12 15 Mexican 20 17 15 23 Mexican-American 21 15 7 13 Cuban 29 9 8 3 Others 14 13 8 13
(1.)Those at working age (16 through 64) only.
(2.)At least one month unemployed in the year before the interview.
Data source: PSID longitudinal sample 1990-92.
Labor Force Participation and Unemployment in Germany (West), 1990-95 (in percent) Men Women Men Women 1990 1992 1990 1992 1993 1995 1993 1995 Labor force participation rate  Germans (West) 75 75 49 51 74 73 51 47 Medit. immigrants 78 80 46 48 69 66 45 44 Turks 76 79 33 36 63 61 33 37 Second generation 57 68 44 45 53 57 34 42 Unemployment  Germans (West) 6 6 7 7 7 9 7 10 Medit. immigrants 12 13 12 11 15 22 12 21 Turks 15 13 18 15 15 25 14 31 Second generation 16 14 11 11 11 23 12 31
(1.)Those at working age (16 through 64) only.
(2.)At least one month unemployed in the year before the interview.
Source: GSOEP (longitudinal samples 1990-92 and 1993-95).
Occupational Position, United States, 1990-92 (in percent) Unskilled/ Skilled White collar White collar Self- semi-skilled workers low level middle and employed workers high level Men 1990 American 10 16 19 37 18 Hispanic 26 19 26 18 11 Mexican 38 16 26 13 7 Mexican-American 26 20 23 21 10 Cuban 8 19 28 31 14 Others 15 22 29 19 15 Men 1992 American 9 16 18 37 20 Hispanic 20 18 20 28 14 Mexican 30 18 23 20 9 Mexican-American 16 18 14 35 17 Cuban 7 13 10 62 9 Others 17 22 25 16 20 Women 1990 American 5 3 45 37 10 Hispanic 20 3 45 22 9 Mexican 36 3 39 19 3 Mexican-American 6 6 55 26 7 Cuban 14 4 42 32 8 Others 11 2 48 20 19 Women 1992 American 6 2 44 38 11 Hispanic 15 2 44 31 8 Mexican 32 3 27 33 6 Mexican-American 8 2 57 29 4 Cuban 8 2 31 53 6 Others 5 1 49 24 21 Data source: PSID longitudinal sample 1990-92. Occupational Position Germany (West), 1990-95 (in percent) Unskilled/ Skilled White collar White collar Self- semi-skilled workers low level middle and employed workers high level Men 1990 Germans (West) 11 28 3 34 12 Medit. immigrants 55 34 2 3 6 Turks 62 29 1 2 7 Second generation 49 38 2 9 2 Men 1992 Germans (West) 10 28 5 35 12 Medit. immigrants 51 35 3 4 8 Turks 57 29 2 2 10 Second generation 42 39 4 10 5 Women 1990 Germans (West) 21 4 19 41 8 Medit. immigrants 71 8 11 6 3 Turks 75 6 12 5 2 Second generation 33 19 30 15 3 Women 1992 Germans (West) 18 5 23 40 8 Medit. immigrants 65 8 15 8 5 Turks 68 6 14 10 3 Second generation 30 9 39 17 6 Men 1993 Germany (West) 11 26 5 34 12 Medit. immigrants 51 35 3 7 5 Turkey 57 34 2 3 4 Second generation 34 43 4 17 3 Men 1995 Germans (West) 9 25 5 37 13 Medit. immigrants 48 35 6 7 5 Turks 48 39 7 3 4 Second generation 32 42 7 15 4 Women 1993 Germans (West) 20 4 22 38 10 Medit. immigrants 71 3 16 8 2 Turks 77 2 14 6 1 Second generation 27 4 45 22 3 Women 1995 Germans (West) 17 3 20 42 11 Medit. immigrants 64 2 17 15 2 Turks 70 1 13 15 1 Second generation 25 1 35 39 1 The colunms do not add up to 100% for the German population, because civil servants (Beamte) were excluded. Source: GSOEP (longitudinal samples 1990-92 and 1993-95). Average Earnings per Hour in U.S. dollars, United States, 1990-92 Men Women 1990 1992 1990 1992 American 14.50 15.90 9.80 10.80 Hispanic 10.60 13.90 7.80 10.20 Mexican 8.80 9.80 6.60 8.20 Mexican-American 11.20 15.20 7.50 8.30 Cuban 9.70 16.20 8.20 10.80 Others 13.00 16.70 9.50 16.50 Data source: PSID longitudinal sample 1990-92. Average Earnings per Hour in Deutschmark (DM), Germany (West), 1990-95 Men Women Men Women 1990 1992 1990 1992 1993 1995 1993 1995 Germans (West) 22.10 24.90 16.60 18.40 26.00 28.50 19.00 21.40 Medit. immigrants 18.90 21.30 14.00 15.00 20.30 22.70 15.50 17.70 Turks 20.00 22.00 14.00 14.30 19.70 21.30 15.30 17.60 Second generation 15.50 19.80 11.60 12.80 17.90 20.50 13.30 15.70 Source: GSOEP (longitudinal samples 1990-92 and 1993-95). Determinants of Income, Beta-Coefficient (Regression), United States Total labor force Men Women Educational degree High School without diploma -.044 -.020 High School with diploma .103 [**] .064 [*] High School plus occupational training .152 [***] .183 [***] College .310 [***] .324 [***] Occupational position Skilled worker .082 [***] .052 [***] White collar low level .304 [***] .126 [***] White collar middle and high level .360 [***] .344 [***] Self-employed .330 [***] .119 [***] Industry Agriculture -.055 [**] -.052 [***] Constructions -.015 a Wholesale, retail, traffic -.108 -.186 [***] Production oriented services -.045 .004 Consumer oriented services -.133 [***] -.200 [***] Public and social services -.113 [***] -.105 [***] Trade union membership .106 [***] .076 [***] Age .149 [***] -.031 [*] Duration of employment in firm .128 [***] .266 [***] Hispanic -.049 [*] -.015 Mexican a a Mexican-American a a Cuban a a r2 .230 .381 Durbin-Watson 1.991 2.080 Hispanics Only Men Women Educational degree High School without diploma .151 [***] -.015 High School with diploma .049 .051 High School plus occupational training -.002 .219 [***] College .064 .210 [***] Occupational position Skilled worker .077 [*] .082 [*] White collar low level .354 [***] .168 [**] White collar middle and high level .414 [***] .266 [***] Self-employed .492 [***] .054 Industry Agriculture -.088 [*] -.045 Constructions -.029 a Wholesale, retail, traffic -.162 [**] -.139 [**] Production oriented services -.144 [***] .069 Consumer oriented services -.286 [***] -.122 [**] Public and social services -.186 [**] -.077 Trade union membership .290 [***] .070 [**] Age .057 -.092 [**] Duration of employment in firm .185 [***] .349 [***] Hispanic a a Mexican -.255 [***] -.277 [***] Mexican-American -.051 -.111 [***] Cuban -.176 [***] -.272 [***] r2 .447 .433 Durbin-Watson 1.985 2.432
(a.)Variable not included in the model.
(***.)significant at p [less than or equal to] .001,
(**.)significant at p [less than or equal to] .01,
(*.)significant at p [less than or equal to] .05.
Reference categories: without educational degree, unskilled and semi-skilled workers, manufacturing, for the total model: native work force, for immigrants: other Hispanics.
Data Source: PSID longitudinal sample 1990-92.
Determinants of Income, Beta-Coefficient (Regression), Germany (West) Total labor force Men Educational degree Primary degree without occupational training -.171 [***] Primary degree with occupational training -.076 [***] Secondary degree without occupational training -.025 Higher degrees .114 [***] Occupational position Skilled workers .066 [**] White collar low level -.006 White collar middle and high level .232 [***] Civil servants .031 Self-employed .116 [***] Industry Constructions -.068 [***] Wholesale, retail, traffic -.134 [***] Production oriented services -.016 Consumer oriented services -.345 [***] Public and social services .028 Age .143 [***] Duration of employment in firm .086 [***] Immigrants .043 [*] Turks a r2 .327 Durbin-Watson 1.993 Mediterranean immigrants Women Men Educational degree Primary degree without occupational training -.050 -.141 [***] Primary degree with occupational training .069 [*] -.088 [*] Secondary degree without occupational training .045 -.026 Higher degrees .160 -.021 Occupational position Skilled workers .031 .156 [**] White collar low level .024 .009 White collar middle and high level .274 [***] .140 [***] Civil servants .201 [***] a Self-employed -.017 .291 [***] Industry Constructions a -.142 [***] Wholesale, retail, traffic -.097 [***] -.159 [***] Production oriented services .029 -.001 Consumer oriented services -.068 [**] -.185 [***] Public and social services .016 .216 [***] Age .117 [***] .097 [***] Duration of employment in firm .018 .092 [**] Immigrants .081 [*] a Turks a .068 [*] r2 .197 .351 Durbin-Watson 2.074 2.374 Women Educational degree Primary degree without occupational training .094 Primary degree with occupational training .117 [**] Secondary degree without occupational training .083 Higher degrees .088 Occupational position Skilled workers .027 White collar low level .046 White collar middle and high level .267 [***] Civil servants a Self-employed -.044 Industry Constructions a Wholesale, retail, traffic .052 Production oriented services -.050 Consumer oriented services -.033 Public and social services .071 Age .196 [**] Duration of employment in firm .108 Immigrants a Turks .023 r2 .159 Durbin-Watson 2.381
(a.)Variable not included in the model.
(***.)significant at p [less than or equal to] .001.
(**.)significant at p [less than or equal to] .01.
(*.)significant at p [less than or equal to] .05. Reference categories: without educational degree, unskilled and semi-skilled workers, manufacturing, for the total model: native work force, for immigrants: former Yugoslavia, Italy, Greece and Spain.
Source: GSOEP (longitudinal samples 1990-92 and 1993-95).
|Printer friendly Cite/link Email Feedback|
|Publication:||International Journal of Comparative Sociology|
|Date:||Nov 1, 1999|
|Previous Article:||Does Economic Adjustment Affect the Legitimacy of Democracies?|
|Next Article:||Africa in Chaos.|