A comparison of labor market segmentation: Germany and the United States

This dissertation explores several topics related to international comparisons of labor market segmentation. A segmented labor market is one in which mobility barriers prevent certain workers from obtaining the job of their choice. Some theorists have asserted that women are segmented into certain occupations. If true, this "crowding" would lead to lower wages than are warranted by the incentives and demands of the occupations into which women are segregated. Previous empirical tests of this hypothesis have produced conflicting results. An international comparison offers a unique opportunity to examine the effects of differing gender roles and labor market institutions on the relation between occupational segregation and wages. Germany is an ideal subject country for this study because German women's labor market behavior is unique among industrial countries. The United States is the reference country. The German Socio-Economic Panel data are used as well as the Panel Study of Income Dynamics. The relationship between the gender composition of a worker's occupation and wages is described in detail. The techniques that have been previously applied to the U.S. are reproduced and applied to Germany. The presentation of the degree and kind of occupational segregation is more comprehensive than that in the previous literature of international comparisons. This adds the result that there is more extreme occupational segregation in the U.S. while still confirming previous research showing more aggregate occupational segregation in Germany. The main result is that there is a smaller independent effect of occupational segregation on wages in Germany than in the U.S. even though Germany has a greater gross correlation of the degree of segregation and wages. A more general comparative analysis of labor market segmentation is also developed. The empirical methods are evaluated to determine which would be suitable for international comparisons. The endogenous switching model is chosen because it does not rely on the judgment of the researcher in defining segments. However, the model does not produce labor market segments for Germany that fit expectations based on prior research. The implications of this unexpected finding are explored.