Do Commodity Prices Affect the Australian Real Effective Exchange Rate: A Real-Time Quantile-Regression Analysis
45 Pages Posted: 16 Oct 2018
Date Written: September 21, 2018
Abstract
I use a real-time quantile-regression approach to analyze whether commodity prices have predictive value for movements of the Australian real effective exchange rate. To do so, I use a modified version of Frankel’s (1986, 2008, 2014) and Frankel and Rose’s (2010) model of commodity price determination using real interest rates and real exchange rates. In-sample analyses and inclusion of the GSCI (sub-)indices in out-of-sample models indicate some predictive value for forecasters that have an asymmetric loss function. These results are partly confirmed when compared with different benchmark models using an out-of-sample R2 statistic and in terms of Fair-Shiller regressions. However, these findings depend on the studied forecasting technique and the loss function that is assumed.
Keywords: Quantile Regression, Exchange Rates, Commodity Prices, Forecasting
JEL Classification: C53, E17, Q02
Suggested Citation: Suggested Citation