The time series literature reports two stylized facts about output dynamics in the United States. GNP growth is positively autocorrelated over short horizons and negatively autocorrelated over longer horizons, and GNP has an important trend reverting component which has a hump-shaped moving average representation. We consider whether various real business cycle (RBC) models are consistent with these stylized facts. ; We find that many RBC models have weak internal propagation mechanisms and must rely on exogenous factors to replicate both stylized facts. In particular, intertemporal substitution, capital accumulation, and capital adjustment costs do not provide strong sources of propagation. Models which incorporate employment lags or labor adjustment costs are partially successful. They endogenously generate roughly the right pattern of autocorrelation in output growth, and they generate a small humpshaped trend reverting component in output. However, they must rely on implausibly large transitory shocks to match the large trend reverting component found in U.S. data. Our results suggest that RBC theorists ought to devote further attention to the problem of modeling propagation mechanisms.
