We estimate the impact of compulsory schooling on earnings using the changes in compulsory schooling laws for secondary schools in West German states during the period from 1948 to 1970. The German school system streams students into three separate types of secondary schools and the change in compulsory schooling laws affected students in the lowest level stream. There is a sizeable literature studying the returns to compulsory schooling for other countries, including the UK. These studies typically indicate returns in the range of 10 to 15 percent for every additional year an individual stays in school in response to a raising of the school leaving age. While our research design is very similar to many of these previous studies, we find very different estimates of the returns. We find no return to compulsory schooling in Germany in terms of higher wages. There are various possible explanations for the different returns in Germany, which we discuss in the paper. One possibility is that returns are low because of wage rigidity. Institutions like union wage setting may prevent the emergence of sufficient wage differentials. In order to investigate this possibility we also look at the employment effects of the change in compulsory schooling. If firms cannot differentiate wages sufficiently by skill, they should be more interested in hiring the more skilled workers (in this case the ones with more schooling). We find no evidence for this. We also find no evidence for any returns among the self-employed, who should not be subject to institutional wage rigidities. A second possibility is that an additional year of schooling matters less because of the existence of the apprenticeship training system in Germany. We also find no evidence for this explanation by looking at the returns for school leavers, who do not get any post-compulsory training or schooling. We conjecture that the different returns in Germany may be due to the fact that the basic skills most relevant for the labor market are learned earlier in Germany than in other countries. We present some evidence that the school leavers affected by the law change only tend to use very basic academic skills in their jobs, and that Germans typically score relatively well in international comparison tests on those types of skills. This is true in particular in comparison with the Anglo-Saxon countries, where returns to compulsory schooling are high. Hence, returns to compulsory schooling may be low in Germany because the school system is (or has been) rather successful in endowing students with the skills relevant for the labor market. Sample Means are appended. (Contains 11 tables, 3 figures and 19 footnotes.)