Did Leaving the Gold Standard Tame the Business Cycle' Evidence from NBER Reference Dates & Real GNP

Abstract

Cover and Pecorino (2005) demonstrate, using monthly NBER reference dates, that, the March 1933 departure from the gold standard is the most probable breakpoint ushering in an era of longer expansions, both absolutely and relative to recessions that follow. Cover and Pecorino view this finding as a challenge to real business cycle theory (RBCT). However, RBCT, along with most of recent macroeconomic research, has focused on the alternative concept of growth cycles - periods when the economy's production is above or below trend - rather than absolute increases or decreases in economic activity. Furthermore, empirical evaluation of RBCT has focused on the statistical properties of aggregate time series, rather than focusing exclusively on business cycle durations. Using HP-filtered quarterly real GNP, we demonstrate that tests of growth cycle durations still imply a most probable breakpoint close to 1933. However, we also test for structural breaks in the volatility of real GNP growth rates and deviations from HP-filter trends. These tests suggest that the most probable structural break is considerably later than 1933; perhaps as late as the 1950s. This conclusion is robust to the consideration of the alternative pre-1929 real GNP series constructed by Romer (1989) and Balke and Gordon (1989).