Predictive Power of Markovian Models: Evidence from U.S. Recession Forecasting

Abstract

This paper brings new evidence of predicting the U.S. recessions through Markovian models. The Markovian models, including the Hidden Markov and Markov models, incorporate the temporal autocorrelation of binary recession indicators in a more traditional and natural way. Considering interest rates and spreads, stock prices, monetary aggregates, and output as the candidate predictors, we examine the out-of-sample performance of the Markovian models in predicting the recessions one to twelve months ahead. The Markovian models are superior to the probit models in detecting a recession and capturing the recession duration. We find the "one-month lag phenomenon" that the best Markovian model supported by statistical model selection procedures can always recognize the onset of a recession one month after it starts. In addition, the yield spread continues to serve as the most ecient predictor variable in explaining business cycles.