The paper analyzes the employment effects of reduced working time when firms' endogenous responses of overtime and shoftwork are considered. The paper distinguishes between Keynesian and classical regimes as well as between interior and corner solution. The main conclusion is that it is only in very simple models that employment increases unambiguously. Under plausible assumptions, a reduction of working time is quite likely to be counterproductive as an employment-promoting measure. The reason is that relative factor costs or relative productivities on the margin are affected in such a way that firms substitute overtime and shiftwork for employment.