This paper constructs a consistent set of quarterly Japanese data for the 1960-2002 sample period and compares properties of the Japanese and U.S. business cycles. We document some important differences in the adjustment of labor input between the two countries. In Japan most of the adjustment is in hours per worker of males and females and also in employment of females. In the U.S. most of the adjustment is in employment of both males and females. We formulate, estimate and analyze a model that makes the distinction between the intensive and extensive margin and allows for gender differerences in labor supply. A weak empirical correlation between hours per worker and employment in Japanese data is a puzzle for our theory.