Real-Business-Cycle Models and the Forecastable Movements in Output, Hours, and Consumption

This 1996 American Economic Review paper by Julio Rotemberg and Michael Woodford tests standard real-business-cycle (RBC) models against the forecastable – rather than total – movements in U.S. output, consumption, and hours, arguing that forecastable movements are the essence of what makes these variables “cyclical.” Using a three-variable VAR on postwar U.S. data, they find that the actual variance of forecastable output movements over a 12-quarter horizon is about one hundred times larger than the variance a calibrated single-shock RBC model predicts, and that actual forecastable changes in output, consumption, and hours are all strongly, positively correlated with each other. The standard RBC model, by contrast, implies that some of these correlations must be negative: because a positive technology shock typically means the current capital stock lies below its new steady-state level, the model’s intertemporal-substitution logic forces consumption and hours to be forecast to move in opposite directions, and – for standard parameter values – forecastable increases in output must be far more strongly tied to expected movements in consumption than to the roughly one-for-one movement with hours seen in the data. The authors show that varying preference parameters (labor-supply elasticity, intertemporal substitution, capital’s income share) can reverse some of these wrong-signed correlations, but each fix available within “plausible” parameter ranges creates or worsens a different failure – either reversing the hours-output sign at the cost of counterfactually volatile consumption, or vice versa – so that no single calibration reproduces both the sign pattern and the magnitude of the data’s comovements. They conclude that a standard stochastic-growth model driven solely by permanent technology shocks cannot account for the business cycle so defined, and suggest nominal-rigidity (“sticky price or wage”) models are, at least intuitively, better positioned to explain why output, consumption, and hours would all be expected to rise and fall together.

Julio J. Rotemberg | Macro Paper Warehouse

Real-Business-Cycle Models and the Forecastable Movements in Output, Hours, and Consumption