A further test of the influence of leading indicators on the probability of US business cycle phase shifts

In earlier work Filardo used a variable transition probability Markov regime-switching model to investigate the usefulness of a number of leading indicators in anticipating phase changes in the business cycle. Filardo used Industrial Production as a proxy for the business cycle and found the leading indicators were not only statistically significant determinants of the transition probabilities but that they also substantially improved the dating of the business cycle. However, Industrial Production is a very narrow proxy for the business cycle and represents a relatively small and decreasing component of economic activity. Here a broader, more comprehensive proxy for the business cycle is employed to test the usefulness of leading indicators in forecasting the likelihood of future business cycle phase shifts. Specifically, the Economic Cycle Research Institute's (ECRI) coincident composite index is employed as a summative measure of the business cycle. Then, following Diebold et al., the transition probability parameters are allowed to vary. In particular, the ECRI leading and long leading indexes are used as putative determinants of these transition probabilities.