Model based clustering of economic time series

Markov switching models provide a flexible way of dealing with parameter changes in sequentially observed data, like time series data, by assuming that the parameter change is driven by a hidden Markov chain. Many applications are to be found in economics, where the business cycle causes switches between regimes. Challenging issues in the econometric estimation of these model is the estimation of the number of regimes, and dealing with identifiability problems. At the IFAS, these models are applied in cooperation with the Austrian National Bank to various data sets from empirical economics (bank lending data, industrial production).