ASSESSMENT OF MODELS TO FORECAST EXCHANGE RATES: THE QUETZAL-U.S. DOLLAR EXCHANGE RATE

Based on Cheung, Chinn and García-Pascual (2004) and Meese and Rogoff (1983), the forecasting performance of a wide variety of theoretical and empirical exchange rate models (PPP, UIP, flexible and sticky-price monetary models, portfolio balance, and a BEER model) is tested against the random walk specification to determine their assessment in predicting the quetzal-U.S. dollar nominal exchange rate. Such models are estimated by applying a recursive regression methodology to quarterly data for the period 1995-2009. Estimations are performed based on an innovative trend-gap data disaggregation methodology, and an error-correction specification to contrast short vs. long run prediction performance, which is evaluated up to eight periods ahead for all model specifications. Different from previous results, forecasts provided by most specifications in the very short run (up to 2 quarters ahead), particularly the BEER specification, consistently outperform those obtained from the random walk model.