This paper empirically investigates the ability of a real business-cycle model with nonseparabilities in consumption and leisure and external habits both in consumption and leisure to fit the postwar US data. The results indicate a strong but fast-dying habit in leisure, and a somewhat weaker but more persistent habit in consumption. Intratemporal nonseparabilities in consumption and leisure play an important role in driving the response of real variables to a productivity shock. Adding capital adjustment costs to the model with nonseparabilities in consumption and leisure and external habits both in consumption and leisure changes the responses of real variables to a productivity shock, however, in a way similar to that documented for the models with capital adjustment costs and habit formation in consumption. The estimated persistence of the productivity shock is quite modest, which may be the factor that drives a procyclical response of hours worked to the positive productivity shock even when habit in consumption is strong.