Parity Theorem Revisited: An ARDL Bound Test with Non-Parity Factors

Abstract

The research question addressed in this paper is, do inflation and interest rate differences across two major economies fully drive the long-run exchange rate changes if controls for non-parity factors are embedded? Exchange rate behaviour research is once again an interesting topic given the availability of powerful econometric approaches to resolve unsolved issues. We re-examine the exchange rate behaviour of the US economy, applying a more appropriate econometric model using 55 years of quarterly data. The model explains 96% of variation in exchange rates, which testifies to the model’s appropriateness. The error correction estimate indicates a time-to-equilibrium of 0.139 per quarter; that is, full adjustment takes seven quarters. Tests indicate evidence of a long-run relationship among the exchange rate, prices, and interest rates. The coefficients on both parity factors (prices and interest rates) are statistically significant with correct theory-suggested signs. These findings constitute strong evidence in support of parity and non-parity theorems while confirming that the US currency behaviour over 1960–2014 is consistent with parity and non-parity theories.