Although an inverse relationship between sickness absence and unemployment has been documented in a number of studies using either quarterly or annual data from different countries with varying institutional frameworks, it is not yet clear whether this empirical regularity is due to changes in the individual costs of absence when unemployment increases (incentive effect) or, alternatively, to changes in the composition of the workforce over the business cycle (selection effect). In order to provide evidence to evaluate the relative importance of both effects we first investigate the effects of changes in the unemployment benefit entitlement system with monthly absence data for East and West Germany for the years 1991-2004. Second, we analyze the impact of differences in the costs of unemployment on the annual absence rates of workers in different sickness insurance funds using state-level annual absence rates for the years 1993-2004. We find clear evidence in favor of an incentive effect.