Testing for regime switching in Singaporean business cycles

Abstract

We examine a Markov Switching model of Singaporean GDP using a combination of formal moment-based tests and informal graphical tests. The tests confirm that the Markov Switching model fits the data better than a linear, autoregressive alternative. The methods are extended to allow us to identify precisely which features of the data are better captured by the non-linear model. The methods described here allow model selection to be related to the intended use of the model.