Search

The existence of a quasi-periodic cycle in growth rates of major economic indicators has become one of the more popular stylized facts in interpreting postwar macro-dynamics. In this paper, we examine the statistical record of sixteen OECD countries over the 1950-1970 period to determine the degree of regularity in movements of annual growth rates in both GNP and several disaggregated components. The WALLIS-MOORE test is employed to establish the significance of the deviation of growth rate fluctuations from those of a purely random series. In only 13 of the 90 series investigated were these deviations statistically significant and in only two countries, Germany and Finland, were more than one growth series significantly non-random. These results suggest that there is little statistical significance to the growth cycle hypothesis from an annual perspective and that postwar macro-dynamics generally fail to reflect in the growth rate domain much of the regularity observed in the more traditional business cycle of the prewar period. [ABSTRACT FROM AUTHOR]