Fundamentalist exchange rate trading performances

The question on the predictive power of different exchange rate equilibrium models is one of the oldest and most intriguing in international finance and assets trading. On one side there is evidence – starting from Meese and Rogoff (1983) to Cheung, Chinn and Pascual (2005) - about the poor performance of fundamental models such as purchasing power parity (PPP) or covered interest parity (CIP). On the other side, practitioner asset managers often trade with a very simple rule - carry trade (CT) – assuming that currencies with high interest rates will appreciate against those with low interest rates, which goes exactly the opposite way than what predicted by CIP. Many studies, for example Gyntelberg and Remolona (2007) or McGuire P. et al. (2007) revealed a good performance of CT models in recent years, but some authors underlined that returns exhibit a strong (left) asymmetry in their distribution - Gagnon and Chaboud (2007) - and are negative in periods of high volatility, such as market crises – Brunnermeier, Nagel and Pedersen (2009), Cairns, Ho and McCauley (2007). In the light of these new findings and relying on recent evidence - by De Zwart et al. (2009), Jordà and Taylor (2009) - that fundamentals have informative power we reconsider the main fundamental relations on exchange rates comparing the trading performance obtained going long the under-valued currency and short the over-valued one. Our sample is made of G-10 currencies plus 8 currencies of minor/emerging countries with deliverable forward contracts. We tested the performance of simple trading rules based on mispricing signals from three different exchange rate fundamental relations: PPP, UIP in terms of real interest rates, GDP growth differentials through the so called Taylor Rule (TR). Models are estimated with monthly data and trading performances evaluated in terms of return, volatility and Sharp ratio over the entire sample and portofolios made of selections of currencies. Henceforth the comparison between fundamentalist trading strategies takes into account differences across the sample, potentially linked to the different nature of the three equilibrium relations employed. Results are supportive for Purchasing Power Parity and for the Taylor Rule.