Oil ETF vs. Commodity Indexes
With the introduction of an ETF based on the price of Brent Crude Oil launched in London, and similar products on the way in the United States, we compared the simulated historical performance of an oil ETF to the Goldman Sachs Commodity Index and the DowJones AIG Commodities Index.
While much more information about the GSCI and DJAIG indexes is available on our website, for the purpose of this analysis, the key points to keep in mind are as follows. The GSCI contains twenty-four different commodities, and weights them on the basis of the average quantity of each that is produced in a given year. The DJAIG believes that pure production weighting underestimates the relative importance of storable commodities (e.g., gold) to the world economy, and overestimates the importance of non-storable commodities (e.g., live cattle). Hence, it uses a weighting scheme based on a combination of production volume and the volume of futures contract trading for each of the nineteen commodities it contains. This difference in weighting schemes leads to one key difference between the GSCI and DJAIG: the weight of energy commodities in the former is almost 80%, while in the latter they are only 33%.
The following table shows the average annual return and standard deviation data in U.S. dollars for the three alternative commodity vehicles, between 1991 (when the DJAIG's simulated history begins) and the end of 2004.
The next table shows the correlation of returns between these three alternatives over the same 1991 - 2004 period.
The next table shows the correlation of real returns between the different commodity alternatives and U.S. investment grade bonds, commercial property, and equity, over the 1991-2004 period.
|U.S. Comm. Prop.||(.03)||.07||.14|
Based on this analysis, we aren't too enthusiastic about the oil ETF. As you can see, investing in it instead of one of the broader commodity indexes would have generated about the same return, with significantly more risk and somewhat lower correlation with U.S. commercial property and U.S. equity. While, like all statistics based on historical financial returns, these inevitably contain estimation errors, the difference between oil's standard deviation and those on the GSCI and DJAIG is very large. We suspect that it more than offsets any potential diversification benefit from oil versus the other two alternatives in a portfolio. To be sure, we will test this in our upcoming portfolio rebalancing. However, for now our preference for the Dow Jones AIG index remains unchanged.
While on the subject of commodity index products, we should also mention two others that sometimes appear in the media. What was formerly called the "Rogers Raw Materials Index" is now known as the "Rogers International Commodities Index", or RICIX. It includes 34 different commodities, with energy products (crude, heating oil, unleaded gasoline, and natural gas) receiving a fixed 44% weighting. To our knowledge, the only way to invest in the RICIX is through a privately placed security, which means higher costs, plus giving up the liquidity of a publicly traded product. We do not believe the RICIX offers benefits that justify these additional costs. If and when a retail product is introduced that tracks the RICIX, we will reconsider our opinion of it.
The other product is the Commodity Research Bureau Index, which, since June, 2005 has been known as the Reuters Jeffries CRB Index. This index has been around since 1957, but has gone through many changes to both the commodities it includes, and its weighting scheme. The most recent change to the latter came this past summer, which raised energy's weight in the overall index from 17.6% to 33%. While a futures contract on the index trades in the United States, there are no retail products that track it. Between 1991 and 2004 (the same period we used to compare the other indexes), the CRB index delivered average annual real U.S. dollar returns of 1.47%, with a standard deviation of 8.26%. Its correlation of returns with those on other indexes were as follows: GSCI, .63, DJAIG, .80, Brent Crude, .29, U.S. domestic bonds, (.03), U.S. commercial property, .22, and U.S. equity, .18. However, since the index weights were considerably changed in June, these statistics are not a valid estimate of the future performance of the RJCRB index.
Some Follow Ups from Previous Articles
We regularly receive comments from readers on the articles we write, some of which offer interesting pieces of information that we wish we had included. Here are some of those:
oThe minimum investment for a single foreign currency certificate of deposit at Everbank is only $10,000. The $20,000 minimum applies to their index CDs that track baskets of different currencies.
oOn the same subject, there is a futures contract on the trade-weighted U.S. dollar index, known as the USDX contract. Thus far, there is not a retail fund that is based on these contracts.
oOur article on equity volatility as an asset class drew a question from a European reader, who wanted to know how closely the VSTOXX index (which tracks implied volatility on the Dow Jones STOXX European Shares Index) is related to the VIX (which tracks the implied volatility on the S&P 500). We ran the numbers, and found that between January, 1999 and December, 2004 the correlation of the returns on these two indexes was .84. That makes them pretty close substitutes. Unfortunately, we have yet to see any fund products for retail investors based on either of these two indexes. For now, if you want to invest in this asset class, you have to purchase, and continuously rollover, the futures contracts on them.
oFollowing our article on private equity, another reader called our attention to a new product from Van Kampen funds (in the United States) which will track the IPOX-30 Index. The IPOX is a product developed by Josef Schuster, who was formerly on the faculty of the London School of Economics. It tracks the performance of initial public offerings for 1,000 days after they are launched. The IPOX-30 includes the 30 largest companies in the IPOX, based on their market capitalization (for more information on the index, see www.ipoxschuster.com). While we find the concept of the index intriguing, we have some concerns about the Van Kampen product. First, while its expenses are relatively low (at 32 basis points), it carries a hefty 4.95% front end sale load (commission). Second, it is clear that the fund may not always be able to buy new issues at their IPO price; hence, the performance of the fund may significantly diverge from the performance of the underlying index. In our view, it would be wise to wait and see how the fund performs before considering its use to gain exposure to venture capital as an asset class.
Active Versus Passive Investing Results in Canada
Standard and Poor's is gradually expanding their Index Versus Active Funds Scorecard (SPIVA) reports to geographies outside the United States. The most recent report covers Canada, and raises some interesting issues. As in other small countries, there are a few large companies in Canada (e.g., Nortel) that account for a very large percentage of the overall market index. To minimize the risk this presents for investors (as Nortel's recent history makes clear), in many smaller countries the local equity market index comes in two forms: one based on pure market capitalization, and another that limits ("caps") the maximum weight of any single company. On the one hand, this logic makes sense, if one believes that markets are not always efficient, and as a result companies can become substantially over and undervalued. However, it leaves one with the question about the appropriate level for the cap on any company's weight in the index. On the other hand, an efficient market purist would reject this approach, and say that market capitalization weighting is the right approach to use. They might note that if Microsoft had been established in Vancouver instead of Redmond, Washington, a cap would have deprived Canadian investors of a lot of upside returns, even though Microsoft would have dominated the Canadian market index.
The SPIVA data make this trade-off clear. Over the past five years, 39 percent of active managers of Canadian equity (pure) funds (i.e., those that primarily invest in Canadian companies) outperformed the uncapped Canadian equity index, while only 8.7% of these managers outperformed the S&P/TSX Capped Composite Index. On balance our view on the capped versus uncapped issue is a pragmatic one. In so far as small countries are likely to have more densely connected social networks, their local markets may therefore be more prone to herding, and therefore to more severe over and undervaluations. At the same time, investors will tend to differ in terms of their degree of home bias, or the extent to which they resist diversifying their equity holdings internationally. This leads us to conclude that if one has a portfolio that is significantly diversified internationally, one should probably invest in funds that track the uncapped index. While this can lead to a heavy weight on a few companies in the context of the local market, in terms of this investor's overall portfolio this will not be the case. On the other hand, an investor who is less diversified internationally should probably prefer products that track a capped index.