Cyclical dynamics of industrial production and employment: Markov chain-based estimates and test

This paper characterizes the business cycle as a recurring Markov chain for a broad set of developed and developing countries. The objective is to understand differences in cyclical phenomena across a broad range of countries based on the behavior of two key economic times series - industrial production and employment. The Markov chain approach is a parsimonious approach that allows us to examine the cyclical dynamics of different economic time series using limited judgment on the issue. Time homogeneity and time dependence tests are implemented to determine the stationarity and dependence properties of the series. Univariate processes for industrial production and employment growth are estimated individually and a composite indicator that combines information on these series is also constructed. Tests of equality of the estimated Markov chains across countries are also implemented to identify similarities and differences in the cyclical dynamics of the relevant series.