Out-of-sample exchange rate predictability in emerging markets: Fundamentals versus technical analysis

Abstract

We provide an in-depth analysis of the predictive ability of models with fundamentals and technical indicators for fourteen emerging market currencies. Our findings suggest that the forecasts from the symmetric Taylor rule as well as from a predictive regression exploiting the informational content of the momentum indicator are statistically superior to those of the random walk and other competing models. We combine the forecasts from the two best performing models via simple techniques and assess the economic significance of the out-of-sample forecasts using a trading strategy based on the sign of the predicted currency returns. Our economic significance results demonstrate that the symmetric Taylor rule, momentum and combination forecasts generate the largest net-of-transactions costs and risk-adjusted returns. © 2019 Elsevier B.V.