Do commodity price changes predict currency exchange rate fluctuations for commodity-exporting countries? In their March 2016 paper entitled “When the Walk is Not Random: Commodity Prices and Exchange Rates”, Emanuel Kohlscheen, Fernando Avalos and Andreas Schrimpf analyze relationships between commodity prices and exporter exchange rates. They first construct daily commodity export price indexes tailored to 11 commodity-exporting countries (Australia, Brazil, Canada, Chile, Colombia, Malaysia, Mexico, Norway, Peru, Russia, South Africa), encompassing 83 commodities (26 metal, 36 agricultural, 11 livestock, 10 energy). They then relate index levels to daily currency exchange rates by country. Using daily UN Comtrade statistics, commodity prices and currency exchange rates in U.S. dollars and Japanese yen as available during January 2004 (Malaysia starts in August 2005, and Russia starts in February 2009) through February 2015, they find that:
- Overall, contemporaneous linear regressions with daily data show that a 10% increase (decrease) in prices of a country’s commodity exports accompanies on average a 2.1% appreciation (depreciation) of the respective. This relationship varies considerably in strength across countries, explaining more than 23% of the exchange rate variation versus the U.S. dollar for Australia and Canada but only about 3% for Peru.
- For in-sample predictive linear regressions, monthly data produce the most robust predictions across countries, exhibiting statistical significance for 10 of 11 countries (excepting South Africa).
- Predictions from out-of-sample linear regressions based on daily data and a rolling five-year estimation window outperform a random walk and a random walk with drift for 10 of 11 countries (excepting Norway) during January 2009 through February 2015. However:
- Results generally weaken for measurement intervals longer than one day.
- Inserting lags in commodity price indexes to ensure time for data collection materially weakens predictive power, suggesting that currency exchange markets are fairly efficient.
- Findings are generally robust to:
- Use of Japanese yen rather than U.S. dollar exchange rates.
- Accounting for changes in country interest rates.
- Accounting for changes in risk aversion as indicated by the S&P 500 Implied Volatility Index (VIX).
In summary, evidence indicates that commodity prices do predict currency exchange rates for commodity exporters, but a fairly efficient exchange market confounds simple exploitation of the predictive power.
Cautions regarding findings include:
- The study does not investigate any currency trading strategies.
- Relationships between export prices and currency exchange rates may not be linear (as assumed in applying linear regressions).
