Business-Cycle Phases and Their Transitional Dynamics

Abstract

This article examines differences in expansionary and contractionary phases of the business cycle. By extending the nonlinear Markov-switching estimation method of Hamilton to incorporate time-varying probabilities of transitions between the phases, the marginal benefits oi the timevarying transition probability Markov-switching model are highlighted. Using this technique, ! document the high correlation between the evolution of the phases inferred from the model and traditional reference cycles for monthly output data. Many of the economic variables rhat determine the time-varying probabilities help to predict turning points. The predictive power of standard leading indicators is evaluated and compared.

Regime switching model of US crude oil and stock market prices: 1859 to 2013

“…Several studies successfully use MS models to analyze aggregate output and business cycles (e.g., Hamilton 1989;Diebold, et al 1994;Durland and McCurdy 1994;Filardo 1994, Ghysels 1994Kim and Yoo 1995;and Filardo and Gordon 1998). Numerous studies also utilize MS models in the context of stock market returns (e.g.…”

A classifying procedure for signalling turning points

“…However, the transition matrix, (e.g. Filardo (1994) allows it to vary), is constant and it works as a linear operator. For our method, it is crucial that one estimates the model for horizon l, and elicits information from y t on J t+l .…”

Twenty-five years of progress, problems, and conflicting evidence in econometric forecasting. What about the next 25 years?

“…In leading indicator work, it was first used by Lahiri and Wang (1994) to predict cyclical turning points. Filardo (1994) extended their work to permit time varying transition probabilities, which appeared to improve forecasts marginally. Diebold and Rudebusch (1996) provided another early analysis.…”