Testing for regional cycles: a Markov-switching approach

This paper presents a method for testing whether business cycle expansions and contractions at different levels of aggregation, such as regional and national cycles, occur at the same time. The procedure is to fit markov-switching models to the time-series and then apply a bootstrap test of equality to the residuals. Applied to US employment data, the test shows significant differences between fluctuations in US national employment and those in some US states. Application to European data reveals significant differences between employment fluctuations at the aggregate EMU level and those in Germany and France.