Friday, September 12, 2008

I am telling below what the oil price will be a year from now on, so read on

No, not really; I was just trying to trick you into having a look at this blog:) Joking aside, while they are far too humble to make such a bold statement, the authors of a recent paper attempt to do just that and have found an easy-to-use way. In Can Exchange Rates Forecast Commodity Prices (NBER working paper 2008-02), Yu-chin Chen, Ken Rogoff and Barbara Rossi show that commodity currencies, i.e. currencies of commodity exporters, can forecast the price of the country's major commodity export. Moreover, these commodity currencies do a good job in predicting overall price movements. The intuition is, according to the authors, that exchange rates are asset prices that embody expectations of future movements in macroeconomic fundamentals, specifically ones that will directly affect the exchange rates.

However, before you develop a strategy based on these results, here are a few "use at your own risk" disclaimers. First, the countries in the study are all major commodity exporters, with commodity exporters representing at least a quarter of their export earnings. Second, they are small countries, so can not affect the price of most of the countries they are exporting. Third, they all have floating, market-based exchange rates. So, rule number 1: This exercise will not work for any country/currency. Fourth, you actually need to know some forecasting and time series to model the relationship, so rule number 2: don't try to do this at home unless you know your metrics; buy your starving PhD student friend dinner so that she will do it for you. Fifth, we are talking about monthly data here, so if the the depreciation in the rand this week does not reflect to a fall in the price of gold the following week, do not blame me (or the authors for that matter). I would turn part of my apartment as shrine to for any economist able to capture that volatility and forecast exchange rates at those frequencies. Finally, rule number 5: This is not for daily trading, traders beware.

Despite the disclaimers above, the results of the paper, which are robust to the specification used (I have tried quite a few), are exciting, not only for practitioners, but also from a theoretical point of view: They validate the modeling of agents adjusting their behavior to expected future events, which is even though much criticized as unreal, has been the bread and butter of international macro for the past 30 years.

Coming to my claim in the title: Doing this exercise for oil is tricky because some major oil exporters have pegged currencies, others are not liked by the US, still others have presidents that do not like the US. While affinity to the US is not modelled in the paper, the currencies of those countries are not determined in international markets, either. Despite these challenges, the authors list predicting oil prices as a future avenue of research.

For a non-technical summary of the table, written by the authors themselves, see

No comments: