Measuring major and minor cycles in univariate economic time series

The coexistence of cycles with different periods complicates the assessment of the current macroeconomic conditions. In order to overcome this problem, a modeling of multiple autoregressive processes in a univariate time series is presented. In the proposed model, individual autoregressive processes are assumed to be mutually uncorrelated, and the number of individual autoregressive processes is determined using information criteria. Simulation results show that the proposed procedure is sufficiently applicable for measuring major and minor cycles. Empirical applications suggest the usefulness and limitations of the proposed method.