we consider a threshold time series model in order to test the ppp hypothesis with brazilian effective real exchange rate dataset in the long run. by following keenan (1985), tsay (1986) and chan (1990), we test brazilian dataset for several types of nonlinearities. so, after apply hansen's test to infer about the number of regimes, we apply the more recent methodology of self-exciting threshold autoregressive (setar) model to point out some threshold to which a signal of turning point could be given in the states of the exchange rate dynamics. all the tests suggest that the brazilian real exchange rate is highly nonlinear. the skeleton of the setar models fitted shows that ppp hypothesis is supported in the long run in spite the deviations from short run.