It was with considerable interest and not a little chagrin that I read Joel Chernoff's May 15, page 3 article
Hot commodities attracting attention and skepticism. Like many quoted in the article, I, too, was perplexed by Ibbotson's findings. That is, I was perplexed until I read the study itself. There I found the analytical flaw that overstated the relative returns to commodities and so biased the entire study's conclusions.
The Ibbotson study compares investment returns in commodities to those in Treasury bills, bonds, equities and several other asset classes. It assumes the investor buys each asset class outright to capture its return, except commodities. With them, the study assumes the investor makes his or her purchase by putting up 100% of the notional value of the commodities in collateral kept in Treasury bills. The total return calculation on commodities then adds the interest earned on the Treasury bill collateral to any gains on the commodities themselves. Almost half the 12.6% annual historical return the study claims for commodities investing comes from the collateral investment in bills, which Ibbotson measures at a 6.1% annual return. None of the other assets tested in the study get to earn from two asset classes simultaneously. That is hardly an unbiased comparison. If the returns on the other asset classes were calculated in the same way, relative risk-reward tradeoffs would look very different, and the study would have very different conclusions, a lot less favorable to commodities investing.
Some might argue Ibbotson had no choice, that the only way investors can buy commodities is with such collateral. It is true that practically collateral is the only way to buy commodities, but that does not preclude Ibbotson from making a fair comparison with other asset classes. After all, just because an investor can buy other assets outright does not mean he or she must buy them that way. Most investors have the option to use a collateral approach with margin, just as with commodities. Had Ibbotson used comparable approaches to the purchases of all asset classes, it would have had a less biased study and arrived at more useful conclusions.
Once it becomes clear how the analysis is done, the Ibbotson findings are not at all surprising. What is surprising is that Ibbotson chose to do things in the way that it did and then claimed to be surprised by findings that seem in retrospect to have been foreordained.
partner, senior economist and market strategist
Lord, Abbett & Co.
Jersey City, N.J.