SOEPpapers on Multidisciplinary Panel Data Research 113
To study the development of wage inequality is important for the economic performance as well as for the development of employment. First, I estimate the remuneration to personal characteristics for Germans and immigrants across the wage distribution using quantile regression. My database is the German socio-economic panel for the period 1984-2006. I find a higher inequality between skill groups for Germans relative to immigrants. The returns to skill for the highest educational attainment are higher for Germans across the wage distribution compared to immigrants. But within-group inequality for the group with the highest educational attainment is higher for immigrants. Both groups have concave experience-earnings profiles. One more year of work experience increases the wage more for Germans. Secondly I use the decomposition method of Melly (2006). Decomposition methods are suitable to get further insights into the question as to whether or not the observable differences in the distribution are caused by the difference in the composition or differences in the estimated coefficients. Immigrants have a negative wage gap relative to Germans. The wage gap rises across the distribution and is due to a rising discrimination of immigrants across the wage distribution for the years 1992 and 2006. For the year 1984 the characteristic effect is responsible for the wage gap. Inequality rises for both groups between the year 1992 and 2006. The increase is much stronger for Immigrants. The coefficient effect is mainly responsible for the wage increase across time for both groups.