We analyze the impact of short-run economic fluctuations on age-specific mortality using Bayesian time series econometrics and contribute to the debate on the procyclicality of mortality. For the first time, we examine the differing consequences of economic changes for all individual age classes. We employ a recently developed model to set up structural VARs of a latent mortality variable and of unemployment and GDP growth as main business cycle indicators. We find that young adults noticeably differ from the rest of the population. They exhibit increased mortality in a recession, whereas most of the other age classes between childhood and old age react with lower mortality to increased unemployment or decreased GDP growth. In order to avoid that opposed effects may cancel each other, our findings suggest to differentiate closely between particular age classes, especially in the age range of young adults. The results for the U.S. in the period 1956–2004 are confirmed by an international comparison with France and Japan. Long- term changes in the relationship between macroeconomic conditions and mortality are investigated with data since 1933.