Cointegration tests of purchasing power parity

Im et al. (Unpublished working paper, 2008) develop cointegration tests using stationary instrumental variables. Their tests avoid the need to simulate critical values for the cointegration estimations, especially problematic in the presence of a nuisance parameter. Likewise, bootstrapping errors is unnecessary. Using an updated version of the Taylor (Rev Econ Stat 84(1):139–150, 2002) data set, the Im et al. (Unpublished working paper, 2008) approach is applied to two well-known, single equation cointegration methods to test for purchasing power parity. The estimations with instruments provide evidence of purchasing power parity (PPP) for more than half of the countries studied; but the empirical results, hence conclusions regarding PPP sometimes differ with the choice of instrument.