CYCLICAL PROPERTIES OF A REAL BUSINESS-CYCLE MODEL

This paper tests the well-known real business cycle model of Kydland and Prescott (1988), using spectral methods for linear filters. Model spectra, coherencies, phase shifts, and correlations are tested against US post-war data using both asymptotic and small-sample distributions. Compared with the model, the data have shorter business cycles, smaller co-movements of different macro variables, and less of a leading role for output.