Does purchasing power parity hold after all? Evidence from a robust test

Abstract

The paper tests the purchasing power parity (PPP) hypothesis using the Brazilian (versus US) real exchange rate from 1855 to 1990. The novelty of the approach pursued here is methodological. Instead of relying on traditional unit roots and cointegration tests found elsewhere, a robust unit root test recently proposed by Hasan and Koenker is applied. Contrary to traditional least squares based tests, the robust test used here is more powerful under non-Gaussian disturbances. The more conventional ADF, PP and KPSS tests are also applied to the series in question. Contrary to most studies that employ longer samples, the unit root hypothesis cannot be rejected, thus, weakening the validity of PPP as a long-run concept. The inability to reject the unit root hypothesis with more than a century of data poses serious questions for the profession's consensus that PPP holds in the long run.