There is a long tradition of believing that exchange rates cannot be forecast with economic fundamentals. The power of fundamental models to beat a random walk always been poor albeit trend and carry models have proven to be successful. A new research paper from the BIS, "When the walk is not random: commodity prices and exchange rates", suggests that currencies that are highly tied to commodity exports do not follow a random walk but are closely tied to the market prices of their exports.
There have been researchers who have tried to look at commodity prices and currencies but have not done the careful work of creating country-specific indices that reflect their exports and thus their terms of trade. When country specific indices are created there is a forecasting link at short-term horizons that extend out to two months. You can track these export weighted commodity price indices and be able to do better than a random walk.
More importantly, these forecasts do better than a carry model, a model that has exchange rate drift, or a model that focuses on risk appetites through the VIX index. Commodities are unique driver to currencies and the researchers show that the direction moves from commodities to currencies and not the reverse.
Oil exporting countries have their currencies tied to commodity price moves. Base metals or agricultural prices also influence currencies of those countries that have large exports related to these markets. The upheaval we have seen in currency markets is not some currency war or just an issue of rates and capital flows. It is also a response to the shock to commodities that we have seen with the end of the super-cycle. Now this may seem obvious to many traders, but the key to this paper is the focus on country specific indices. If you get the export weighting right, you have a good forecasting tool.
No comments:
Post a Comment