Markov Switching Causality and the Money-Output Relationship

The empirical relationship between money and output is one of the most studied issues in macroeconomics, and a large literature has examined the causal links between monetary variables and output. One puzzle from this literature is that the results of causality tests appear to be sensitive with respect to the sample period that one considers. As a way of overcoming this problem, we propose a new method for analysing causal links that allows for changes in these links over the sample period. Our method is based on a VAR model with time-varying parameters. We model the time-variation in the parameters as reflecting changes in causality, and assume that these changes are stochastic and governed by an unobservable finite Markov chain. One important advantage of our method relative to alternative methods is that it allows for arbitrarily many changes in causal links during the sample and enables the identification of sample points at which causality has changed. When applied to US data, we obtain results that allow us to reconcile previous puzzling differences in the outcome of standard Granger causality tests.