International exchange rate dynamics and purchasing power parity

Abstract

This thesis provides evidence in favour of the long-run validity of Purchasing Power Parity (PPP) using primarily a linear error correction framework. Through an examination of PPP where proportionality and symmetry are implicitly imposed, it is shown that a selection of twelve EU real exchange rates is stationary on a univariate basis. The contribution here is based on the reconciliation of unit root test outcomes across univariate and panel tests. Following this analysis, the Johansen cointegration procedure is employed to examine whether long-run equilibrium relationships can be identified in systems of real exchange rates. The implications of results found are set out in terms of regional exchange rate policy co-ordination, exchange rate regime appropriateness, and monetary integration. By focussing on interdependent regions that were affected by a major financial shock (Europe: EMS crisis; Latin America: Mexican crisis; South East Asia: 1997 crisis), the real exchange rate dynamics are compared in pre- and post-crisis scenarios.This thesis also presents evidence in favour of PPP by examining the less restrictive scenario where neither proportionality nor symmetry is imposed. Given the fact that most developed economies have highly integrated goods and capital markets and liberalised capital accounts, the failure to find evidence for PPP in previous studies may be due to the exclusion of factors that might reflect the behaviour of capital markets and their influence on the exchange rate. To test this, the traditional nominal exchange rate and domestic/foreign price based system is augmented with an interest rate component. In a tripolar specification, the joint test of PPP and Uncovered Interest Parity (UIP) is found to hold in a system comprising Germany, Denmark and the UK, suggesting well-integrated goods and capital markets and the long-run convergence evident suggests that Denmark and the UK might be suitable for membership of the euro area. This convergence appears to be stronger when short-term interest rates are used as opposed to long-term rates (perhaps since they are not subject to distortions such as taxation and maturity levels). Furthermore, long-rates have been associated recently with an inversion of the yield curve, while evidence to support the yield curve in non-crisis times is mixed. Finally, multivariate and panel cointegration procedures are employed to provide evidence for the suitability of potential future euro area entrants from Central and Eastern Europe in tri-variate systems comprising the euro nominal exchange rate and two price series.