Monday, March 1, 2010

The Problem with Commodities ETFs

Last week, Hard Assets Investor published another one of my articles. In it, I compared the correlations between several energy ETFs and the prices of commodities they are tracking. The takeaway message is that, for the most part, ETFs do a bad job tracking commodities prices. There are a few reasons for this.
The first, and primary reason why the correlation is less than perfect, is that the front-month futures contract for every commodity is constantly changing. Since most ETFs hold only the front-month contract, this means that sometime before expiration, the fund must sell its holding of one (soon to expire) contract and then purchase another (soon to be front-month) contract. There is almost always a significant discrepency between the two prices, particularly if it's a commodity that doesn't deliver every month. Take a look at the futures chains on the CME group website for crude oil, gold, and corn. As of writing, the difference between the front month and second month contract is about 40¢, $1, and 10¢, respectively. That may not seem like much, but that price difference represents a pure loss in the value of the fund with every roll that takes place. Just to prevent declines, the fund would, on average, have to be increasing in value between 1 and 5% a month. That means that, depending on the commodity and the fund, a monthly percentage increase, if not sufficiently high, may actually mean you're losing money.
The second problem with ETFs is that they create a secondary market on top of what is already a volatile market in its own right. And, while the underlying instruments are the same for both, (futures contracts) the factors driving the supply and demand may be drastically different. For example, while natural gas might be moving up on strong demand resulting from a particular weather forecast, if the primary participants are energy hedgers, no such demand would exist for UNG, the world's biggest natural gas etf. As such, UNG might theoretically follow the price of gas with something of a lag, rather than actually track the price in real time.
Don't take my word for it, take a look at this graph comparing oil price movements with the price of USO, the world's biggest crude oil ETF, over the last four years:
Now, you might be saying "those charts seem to follow each other reasonably well" but that's the problem. Reasonably well isn't good enough. An investor buying a share or shares in a commodity ETF that puportedly tracks the price of a commodity wants her investment to actually do just that, not merely track the price "reasonably well".
It's true that there are some ETFs, especially those dealing with precious metals, that eliminate the problem of the roll yield altogether simply by buying and hoarding the commodity in question and storing it in a vault somewhere; no futures contracts required. While certainly effective in mitigating the problem, this only works for a very specific subset of commodities that do not have a cliff for their usefulness. Other than those few cases, in general, the moral is that if you really, truly want exposure to commodities, don't waste your time with an ETF. As they stand, very few of them can be relied upon to achieve their goals. Though that may change in the future, for now, your best bet is get into the futures market directly.