Wednesday, March 21, 2007

Stocks and commodities positively correlated?

There has been growing acceptance lately for including commodities in the average portfolio because of their inverse correlation to stocks. In simple terms, this means that one asset class (commodities, for example) tends to well in periods where the other does poorly.

This idea has become one of the more widely-held notions among investors and investment managers. But does it stand up to closer examination?

Steve Saville has written an article for Safehaven entitled, "Stocks, Commodities, and Inflation", that tackles this subject. Here is an exerpt from that piece:

Continuing with the "sometimes things are different" theme, during the current cycle there has clearly been a change in the relationship between stocks and commodities. Stocks and commodities have traditionally been either inversely correlated or uncorrelated asset classes, and as a result there has been a general tendency for professional money managers to use commodity-related investments to hedge their equity exposure.

However, the following chart shows that over the past seven years there has been a strong positive correlation between the S&P500 Index (SPX) and the Industrial Metals Index (GYX). The GYX has clearly been the superior investment in that it fell by a much smaller percentage during 2000-2002 and rose by a much greater percentage thereafter, but the SPX and the GYX have essentially become plays on the same underlying global growth trend. Therefore, the idea that commodities in general and the industrial metals in particular can be used to offset stock market exposure needs to be binned, or at least re-considered.

So in Steve's view, at least one group of commodities (the base or "industrial" metals) seems to be positively correlated with US stocks. His conclusion is that the commonly espoused view of negative correlation between stocks and commodities needs to reexamined.

Saville is not the first to voice this opinion. Last year, Societe Generale and Legal & General both came out with reports that warned of lower returns from commodities and an end to the low correlation environment between commodities and stocks.

From the June 2006 report:

Both Société Générale and Legal & General said that the traditional view that commodities were a viable alternative investment because of their low correlation to equities and bonds was no longer valid.

That was because commodity prices had moved in tandem with equity markets and therefore had a closer correlation to equities. That in turn would have a negative impact on asset allocation towards the sector.

What's interesting about this principle of gearing investment decisions around reports of low/high correlation, is that it seems inevitably voided by the bandwagon effect.

In other words, as more investment managers spot low or high correlation between assets and then act to take advantage of this relationship, the relationship changes. Instead of diverging, asset prices may actually begin to converge and move higher or lower together.

It seems an interesting example of the idea that observing an event/phenomenon is enough to change it. Or am I misusing that scientific principle by incorrectly applying it to this aspect of investment decision-making?

Update: For more on this theme, see FT Alphaville's recent entry, "Uncorrelated assets are now correlated".

No comments: