Quarterly data on industrial production and deflated gross national product in the United States from 1947 through 1985 are decomposed into independent nonstationary trend and stationary cycle components using Kalman filtering and smoothing techniques. Estimates of the model parameters imply that at least half of the quarterly innovation in U. S. economic activity can be attributed to a stationary cyclical component that persists over periods of time as long as five years. This finding is inconsistent with the hypothesis that most of the apparent variation in U. S. economic activity can be attributed to a nonstationary trend component.