Technology shocks and the business cycle: An empirical investigation

This paper uses maximum likelihood to estimate a prototypical real business cycle model under several different assumptions regarding the stochastic process governing technological change. The estimation approach permits the use of formal statistical hypothesis tests to help discriminate between competing specifications; the approach also allows each specification to be evaluated based on its ability to generate accurate out-of-sample forecasts. Overall, the results indicate that the data prefer a version of the model in which technology shocks are extremely persistent but still trend stationary. (C) 2001 Elsevier Science B.V. All rights reserved.