The Long-Run Equilibrium of Foreign Exchange Markets after the Subprime Crisis

Cointegration is a statistical concept that has always been associated with the economic concept of equilibrium. This is a static relationship commonly called long-run equilibrium between two or more financial time series. When this dependency exists, the mean-reversion logic holds: cointegrated time series, as stated in ‘Granger’s representation theorem’, cannot drift too far apart from equilibrium, because an error correction term, represented in practice by various economic mechanisms, will act to restore such equilibrium. Thus, a long-term relationship implies the that an error correction term in the model automatically exists and this error correction term captures every short-run dynamics of the system, by quantifying deviations from the long-run, therefore it contains information regarding the future movement of one variable based on past prices. According to this fundamental property and since Granger’s work and recommendation, cointegration has been widely used in the foreign exchange market to test for market efficiency. This linkage is of particular interest for investors and traders looking at any possible arbitrage opportunity. However, relatively recent studies suggested that cointegration is not suitable for testing market efficiency and consequently market predictability at any period. Our study uses the multivariate Johansen’s approach to investigate the presence of co-movements among some of the most important exchange rates after the subprime crisis. Our main purpose is not to provide investors with advice on trading or portfolio diversification but to look at whether the recent subprime financial crisis has modified the long-run equilibrium in Foreign Exchange markets.