The first section is a study on the Thai Baht using an advanced regime-identification tool by Hamilton (1989) on the 1998-2005 monthly data, nesting the special case of a monetary model with emphasis on the real interest rate differential. Three scenarios post-crisis for the Thai Baht were identified by the nonlinear detection mechanism where the unobservable states are assumed to follow the Markov chain of events where past history does not matter. The states identified are described as panic, calm, and favourable markets for the currency. Furthermore, using the Markov-switching software developed by Krolzig (1998), it is possible to provide weak evidence that the nominal exchange rate moved to restore equilibrium in the fundamentals but not vice versa. The second section tested the productivity bias for the Thai quarterly data from 1993-2005, using Johansen cointegration method and found no evidence. The final section has the whole of the Thai economy specified and exogenous shocks sent to see the effects on the real exchange rate. In particular, a surge in productivity in the fully-specified economy causes the real exchange rate to appreciate, confirming the evidence for the productivity bias for the Thai Baht.