Updated Asset Class Correlations and the Commodity Controversy

Richard Ferri's company Portfolio Solutions website has a handy updated 14 page downloadable pdf titled Asset Class Correlations, which updates many of the major asset class correlation numbers up to the fall of 2006. He indicates an intention to do quarterly updates so it promises to be a useful on-going tool. Ferri is the author of the fine book All About Asset Allocation, which I previously reviewed here. The document contains an appendix with a very good explanation of correlation and its part in asset allocation (it looks like it was lifted from the book). Ferri's site also includes, for free, documents that track the long-term evolution of investment indicators, and one of his books called Serious Money, which I have on my list to review.

One of the interesting and controversial parts of the above paper is that Ferri finds that having a commodity holding in a portfolio reduces returns. He doesn't address whether it reduces volatility ... though the chart of returns appears to have less variation in the portfolio that includes the commodity. This is directly contrary to the findings of Roger Gibson in his book Asset Allocation which I reviewed here. What explains this difference I cannot tell - possibly a different time interval for back-testing the "with "and "without" portfolios, the different commodity index each used (GSCI for Gibson, CRB for Ferri) or something else, I cannot tell. Ferri himself says the topic is subject to on-going debate. For now, I've got a commodity holding in my portfolio - DJP, the iPath Dow-Jones-AIG Commodity Index - and it will stay there till next year's portfolio rebalancing at least.

Update June 26: Zephyr Associates has also compiled a series of asset class correlations, oriented to US investors and investments. There is a downloadable pdf here. It is interesting that the correlation of equity returns of every foreign country with the US has increased in the last ten years to the 0.7/0.8 level, which is quite high and lessens the benefits of diversification. Is this a sign of a world economy that is ever more inter-connected and inter-dependent and is thus permanent, or is it just a period that will pass? Correlations of other asset classes with equity like real estate and bonds are much lower and in some cases, widening, making the diversification benefit even greater. The correlations of the overall US market with various equity sectors is also seen to be much lower and again in some cases, widening. Is industry sector asset allocation a more promising strategy? After all, one can hold the entire market split up in different ways; it doesn't necessarily have to be on a geographic country basis. btw, thanks to Canadian Capitalist for the link to Zephyr Associates for his blog posting on the new Invest Skeptically blog where I found the posting with the link to Zephyr.