Title: Point and density exchange rate forecasts using yield curve factors in a time-varying framework
Authors: Anastasia Allayioti - University of Warwick (United Kingdom) [presenting]
Abstract: Many papers have offered several explanations for the lack of correlation between exchange rate movements and macroeconomic fundamentals. Recent research attributes this predictive failure to the use of inappropriate proxies for market expectations of future fundamentals. A separate literature has stressed how the presence of structural instabilities might hinder the forecasting ability of many exchange rate models. Taking into account evidence suggesting that the term structure of interest rates reflects expectations of market participants about future economic activity, we compare the predictive ability of specifications augmented with yield factors relative to the performance of models that use traditional macro-fundamentals and financial predictors. Given that the examined sample period includes a period that was characterized by negative interest rates and considerably narrower yield spreads, the adopted framework explicitly accounts for the existence of an effective lower bound on nominal interest rates. Following a comprehensive in-sample evaluation, we examine the relevance of the competing models in generating accurate point and density forecasts within a set-up that explicitly allows for time-evolving dynamics in both the slope and volatility parameters. The out-of-sample evaluation involves numerous statistical criteria and an assessment of the ability of the models under examination to generate economic value.