Exchange rate dynamics and economic fundamentals: Nonparametric forecasts and risk valuation

Liu Liu, Fordham University


This dissertation has examined the forecasting performance of a nonparametric kernel regression model under Taylor-rule economic fundamentals for both point and interval density of exchange rate dynamics, as compared with a set of conventional parametric exchange rate models. Based upon the density forecasts of returns for five floating exchange rates of OECD countries, a simple risk valuation approach is employed in the context of risk management in international investment portfolio. It aims to answer the questions whether the nonparametric method can perform better than the parametric methods with fundamental-based exchange rate models in predicting foreign currency downside risks, and whether economic fundamentals have more predictive power in nonparametric estimation. The empirical results of this dissertation provide evidence that the nonparametric exchange rate model under Taylor-rule fundamentals has greater predictive power than the conventional empirical models on the level of point forecasts, while the interval density forecasts and its applications in risk analyses have mixed results for different exchange rates. The results also demonstrate that more economic fundamentals have significant impact on exchange rate dynamics through nonparametric estimation.

Subject Area


Recommended Citation

Liu, Liu, "Exchange rate dynamics and economic fundamentals: Nonparametric forecasts and risk valuation" (2012). ETD Collection for Fordham University. AAI3543379.