The differential labor market developments in Germany and in the United States are often cited to support the hypothesis of a trade-off between more jobs for unskilled workers on the one hand, and a less equal earnings distribution on the other. In contrast to the United States, unemploment of unskilled German workers increased substantially since the early 1980's while the distribution of earnings changed little (see, e.g., Gottschalk and Smeeding 1997, Steiner and Wagner 1998). Given that the German economy is affected by technological change and international competition to a similar extent as the U.S. economy, these differential labor market developments are usually explained by the much greater importance of institutional factors in Germany which are hypothesized to lead to a rigid wage structure. These factors include effective wage floors set by collective bargaining agreements, unions' "solidaristic wage policy" aiming at uniform relative wage increases, and income support schemes characterized by high earnings replacement ratios (see, e.g., Abraham and Houseman 1995, Siebert 1997). However, not all observers seem convinced that these factors have contributed significantly to the high unemployment rates of unskilled workers in Europe, and in Germany in particular (see, e.g., Nickell 1997 for a pessimistic view).Although the relationship between (changes in) the employment of unskilled workers and relative wages has been analyzed in a number of empirical studies, no consensus view seems to have emerged so far. For a small cross-section of developed market economies, OECD (1996) reports a significant positive relationship between unemployment rates and a measure of earnings inequality; this correlation disappears, however, when first differences of unemployment rates rather than their levels are considered. In a study covering a larger number of countries, Blau and Kahn (1996) find that employment ratios of low-skilled workers are lower in countries with a more compressed earnings distribution than in the U.S., but this correlation, too, seems far from conclusive, to say the least. After comparing the development of relative unemployment rates and wages for various skill groups in a number of OECD countries, Nickell and Bell (1995: 46) conclude that "there seems to be no evidence that the unemployment rate effects are any more severe in countries where the wage effects are small". On the basis of a comparative study including the U.S., Canada and France, Card, Kramarz and Lemieux (1996) find little evidence for the hypothesis that the more compressed earnings distribution in France, which hardly changed during the observation period, generated significantly different employment trends than in the other two countries characterized by a higher degree of wage flexibility.On the basis of a similar methodology as employed by these authors, Krueger and Pischke (1997) and Beissinger and Möller (1998) find for Germany that there is no significant correlation between the change in the employment-to-population ratio disaggregated by age-education cells and the respective wage in the base period. Other studies for Germany, which rely on the estimation of standard partial-equilibrium labor demand models, tend to find a significant negative relationship between the relative employment of unskilled labor and relative wages. However, estimated substitution elasticities between unskilled and skilled labor vary a great deal between the various studies, depending on the economic sector analyzed, the time period, the way skills are measured and the specification of the production technology. The same is also true regarding the estimated effects of skill-biased technological change on the relative demand for labor.This paper builds on and extends previous work for Germany by Steiner and Wagner (1997) who find a rather low substitution elasticity between unskilled and skilled male labor for the whole manufacturing sector of about -0.3 and a trend decline in the skills ratio of about 3% per year. Given these estimates, the authors conclude that even reductions in the relative earnings of unskilled workers on a scale observed for the U.S. labor market would not have been sufficient to bring employment of unskilled workers in West German manufacturing back to previous levels. In this paper, we extend their analysis for the manufacturing sector to the whole German economy and analyze the economic factors which have contributed to the dramatic decline of the employment share of unskilled labor in German manufacturing, in particular the role played by the relatively rigid earnings structure. In the next section, we present some stylized facts on relative earnings and employment trends in the German economy since the mid-1970's. The econometric model of substitution between unskilled and skilled labor and its relation to the development of relative earnings is set out in section 3. Estimation results are presented and discussed in section 4, and section 5 concludes.