This proposal builds upon the seminal work of Engel & West (JPE, 2005) on the relationship between FX rates (FXRs) & fundamentals (FMs). They deal with the long-standing puzzle in international economics, which is the difficulty o...
ver más
¿Tienes un proyecto y buscas un partner? Gracias a nuestro motor inteligente podemos recomendarte los mejores socios y ponerte en contacto con ellos. Te lo explicamos en este video
Fecha límite de participación
Sin fecha límite de participación.
Descripción del proyecto
This proposal builds upon the seminal work of Engel & West (JPE, 2005) on the relationship between FX rates (FXRs) & fundamentals (FMs). They deal with the long-standing puzzle in international economics, which is the difficulty of linking floating FXRs to macroeconomic FMs. In this project, we take a new line of attack on the question of co-movement between FXRs & FMs as well as between FXRs. We work with the class of asset-pricing models of Engel & West (EW). We attempt to empirically verify their theoretical conclusion that large discount factors account for random walk behavior in FXRs. We also deal with the following: if the FX models imply RW behavior, so that their changes are unpredictable, how then can we validate the models? The study investigates different forecasting horizons & frequencies of the 6 FX majors & differentials of Macro-FMs relative to US. The project develops in three phases; first we explore the nature & direction of FXRs-FMs causality. Then a new linear/nonlinear causality-based model selection methodology is introduced & applied to forecasting. Finally, the causality-based criteria are compared to well-known methods & forecast measures. We extend the various linear multivariate models of EW towards nonlinear ones with time-varying parameters, structural breaks, regime switches & Bayesian estimation. The recent evidence on causality is based on the linear parametric Granger test, although it has low power against nonlinear alternatives. The nonparametric test by Hiemstra-Jones (1994), which is a modified version of the Baek-Brock’s (1992), is regarded a test for nonlinear dynamic causal relationship. We employ a multi-step empirical methodology as in Bekiros & Diks (2008), using multivariate VAR/VECM/GARCH filtering combined with linear & nonlinear causality tests. Improved knowledge of the direction & nature of causalities will expand the information set available to policymakers for decision-making.