Abstract:
A recent innovation in modeling exchange rates has been the use of nonlinear techniques such as threshold autoregressive models and its smooth transition variants. This paper investigates the smooth transition autoregressive (STAR) modeling strategy in an application to real exchange rates. The key findings are as follows. First, using the methodology advocated by Teräsvirta (1994), we find evidence of nonlinear dynamics for several of the spot dollar real exchange rates using monthly data on five of the G7 countries. However, once estimated, we find that the STAR specification is appropriate for only one of the three exchange rate series indicated to be an ESTAR process. Moreover, using simulations, we show that the underlying methodology used to detect nonlinearities in the data exhibit substantial size biases, which we attribute to influential observations. We also investigate an alternative nonlinear specification and find that we can model the dollar-sterling and the dollar-lira real exchange rates better as an open-loop TAR process instead of a SETAR process.