An Econometric Characterization of Business Cycle Dynamics with Factor Structure and Regime Switching

This dissertation proposes a dynamic factor model with regime switching as an empirical characterization of business cycles. The approach integrates the idea of comovements among macroeconomic variables and asymmetries of business cycle expansions and contractions. The first is captured with an unobservable dynamic factor and the second by allowing the factor to switch regimes. The model is fully estimated by maximizing its likelihood function and the estimation methods are designed to make inferences about the unobserved nonlinear factor and the latent Markov state. The empirical results indicate that the combination of a dynamic factor with Markov switches leads to a successful representation of the sample data relative to extant literature along several dimensions. The results corroborate previous evidence about asymmetries of business cycle phases as regards the differences in the internal dynamics of expansions and contractions and reestablish a strong correlation between the NBER business cycle dates and inferred probabilities for monthly and quarterly data. The extracted coincident indexes are strikingly similar to the Department of Commerce coincident indicator, and all recessions are well characterized for both quarterly and monthly analyses. The proposed model also provides a more rigorous and timely approach for dating business cycle turning points than traditional methods. Our approach is based on a probabilistic framework that can be used in real time to assess the state of the economy and can be replicated consistently at any time. The model performance is evaluated in terms of its ability to predict business cycle turning points in an out-of-sample exercise, for both quarterly and monthly data, using revised and real time data sets. We find that the model is successful in characterizing business cycle phases out-of-sample.