We show that there is a strong, positive correlation across countries and industries between the standard deviation of the seasonal component and the standard deviation of the nonseasonal component of aggregate variables. After documenting this stylized fact, we discuss possible explanations and develop a model that generates our empirical finding. The main feature of the model is that firms endogenously choose their degree of technological flexibility as a function of the amounts of seasonal and nonseasonal variation in demand. Although this model is intended to be illustrative, we find evidence supporting one of its key empirical implications.