In this paper we show that the Markov switching model is a relevant statist
ical alternative to the classical martingale model for exchange rates. By e
xtending the standard Markov switching model we decisively reject the marti
ngale model. Moreover, the model generates autocorrelations and linear stru
ctures in Line with what is observed in reality. Subsequently, we test whet
her this model can explain chartist profits. We find that the extended Mark
ov switching model is able to explain the profitability of a simple MA-30 r
ule. Finally, we decompose the profitability of the MA-30 rule into a linea
r and nonlinear part. We find that, although the implied linear structure o
f the Markov model explains a substantial part of the profitability, part o
f the profits of the MA-30 rule can be attributed to the specific nonlinear
ities implicit in the Markov model. (C) 2001 Elsevier Science Ltd. All righ
ts reserved. JEL classification: F31.