From the July 01, 2011 issue of Futures Magazine • Subscribe!

Avoiding correlation surprises

Commodity traders need to prepare themselves for the worst-case scenario. Your worst drawdown is always the one ahead of you. The markets are unforgiving, especially with leveraged vehicles. One can look at 2008 as well as 2001, 1998 and1994 for periods with massive disruptions and tremendous volatility. When you are convinced of the rarity or the impossibility of a six-sigma event (six standard deviation move thought to be statistically impossible under a normal distribution assumption), you are open to risk of ruin.

These six-sigma events in the markets seem to be occurring in greater frequency. The goal of commodity traders is to produce positive returns while managing their risk. Too often traders delude themselves into thinking correlation ratios are written in stone. What may be proven as a concept does not prove certainty; there is no certainty in the markets. The irony is that by studying correlation to reduce risk, traders can expose themselves to greater risk through making faulty assumptions.

There are well known assumptions in various markets regarding correlations that supplement this delusion. We assume a perfect negative correlation between gold and the U.S. dollar; gold and the Swiss Franc generally trend in the same direction. There is the correlation between the Canadian, Australian and New Zealand dollars to commodities. There is the correlation between the Canadian dollar to oil prices, as it is a major exporter. Conversely, there is the Japanese yen, which is at risk because of rising oil prices as it imports most of its oil.

These correlations are real and knowing them can be helpful, but they are not law. As much as we rely on these correlations in asset allocation, we must realize they can change and, in extremely volatile markets, break down.


The correlation between crude oil and the Canadian dollar from 2006 to 2009 exemplifies this. The markets had a correlation of approximately 0.8 over long periods of time; however, in January of 2008 there was a major divergence (see "Side by side," above).

comments powered by Disqus