Commodities As Portfolio Insurance

May 06, 2013

Misconceptions about contango and costs can leave a hole in an investor's portfolio.

There are two pieces of "conventional wisdom" on investing in commodities that seem to be held by many, if not the vast majority, of investors. The first relates to the belief that commodity investors are doomed to pay the costs of high and persistent contango. The second is that live mutual funds cannot track their benchmarks well due to costs.

When academic research began to be published on the benefits of commodities as an effective diversifier of the risks of both stocks and nominal return bonds, and products began to be developed and marketed, there was a large surge in demand for commodity futures contracts by investors.

Observers of this trend noted that it was accompanied by years of persistent contango in the markets for oil contracts (though not in any easily storable commodity, such as gold). When contango exists, commodity futures funds have to sell contracts at low near-month prices and buy into costly further-month contracts. Along with other costs, persistent contango can cause the returns of a fund to trail the returns of spot prices by large amounts.

One argument that was heard against investing in commodities futures was that the demand from hedgers permanently doomed the markets to a state of large contango. And that seemed to be the case for quite a while. As investor demand soared, assets in commodities futures rose sharply and contango in the oil market reached high levels.

However, while the hedging demand from investors doesn't seem to have lessened, the size of the contango has been coming down over the past several months. On March 27, spot NYMEX-traded WTI light sweet crude closed at 96.58. The May and June contracts were both trading below that, at 96.18 and 96.44, respectively, while the July contract was trading just slightly above, at 96.62. If we look further out, the futures contracts turn toward backwardation with the September, December and December 2014 contracts trading at 96.26, 95.03 and 90.48, respectively. While the index always uses the near month, a fund can choose which month's contracts to buy.

It's also important to note that the commodities indexes all assume contracts will be rolled over on five consecutive days in equal amounts, and a pure commodities index fund would do just that to make sure it tracks its index. However, a well-run fund could avoid the well-known costs of such a transparent strategy. Simply avoiding those days (because the active traders know indexers must trade and can exploit that) can improve returns.

In addition, there's evidence demonstrating that a shifting maturity approach to futures contracts (avoiding the most expensive months when contango is greatest or when backwardation smallest) can improve returns as well.

Since none of us has clear crystal balls, we can't know if the large contango we had experienced in recent years will return. However, as we've seen, the increased hedging demand from investors hasn't doomed the market to persistent high contango. In fact, the change in the situation may have occurred because of the relatively poor returns commodities have experienced since 2010. In 2011, the DJ-UBS Commodities Index lost 13.3 percent, in 2012 it lost 1.1 percent, and through February 2013, it had lost 1.8 percent.

If investors earn poor returns because of persistent large contango, it seems highly likely that money will leave the asset class, reducing or even eliminating the problem. Which leads us to the second issue: Can a fund overcome all costs to match its benchmark index? To see if that is the case, we'll take a look at the performance of two commodities funds: Pimco's PCRIX and DFA's DCMSX.

While looking at the two funds, it's important to understand that both are passive in terms of their approach to commodities exposure, maintaining the same exposure to various commodities as does its benchmark index. However, both are active in terms of investing the collateral they use to support the futures contracts. (Neither invests in three-month Treasury bills as the index assumes, and Pimco is more aggressive in taking credit risk with some of its choices.)

 

 

PCRIX also typically has a high allocation of TIPS with its collateral, while DFA uses short-term investment-grade bonds for its collateral. In addition, both funds avoid the dates that are used by the indexes when they roll their contracts over. And they both use a shifting maturity approach to determine the month they'll use to buy contracts.

We'll first look at DCMSX, which has an expense ratio of 0.35 percent. From inception in December 2010 through December 2012, the fund's annualized return was -0.8 percent, 1.7 percent better than the -2.5 percent return of the benchmark DJ-UBS index.

In the first two months of 2013, that trend continued, as the fund lost 1.6 percent and the index lost 1.8 percent. We now turn to PCRIX, which has a much higher hurdle to overcome than does DFA's fund since its expense ratio is 0.74 percent.

From inception in July 2002 through December 2012, the fund returned 9.5 percent, outperforming the DJ-UBS index by 4.5 percent a year. The fund's total return was 160.3 percent, 93.5 percent higher than the return of the index. In the first two months of 2013, the fund returned -1.5 percent, outperforming the index, which returned -1.8 percent, by 0.3 percent.

The evidence seems clear that a well-designed fund can not only replicate the returns of its benchmark, but to date, they have been able to significantly outperform the index even after expenses. Now let's turn to the right way to look at commodities by seeing how their addition to a portfolio impacted the performance of the portfolio.

The table below shows the results for the 10-year period 2003-2012. Note that during this period, the S&P 500 Index returned 7.1 percent and the DJ-UBS Commodities Index returned just 4.1 percent. Thus, one would think that in hindsight adding a 5 percent allocation of commodities and reducing the stock allocation by the same amount would not have been a good idea.

Portfolio

Annualized Return (%)

Annual Standard Deviation (%)

60% S&P 500/40% 5-Year Treasury

6.76

9.56

55% S&P 500/ 5% DJ-UBS Commodities Index/40% 5-Year Treasury

6.62

9.43

 

As you can see, the portfolios had almost identical returns with virtually the same level of volatility. The portfolio without commodities had both slightly higher returns and slightly higher volatility. As even during periods when the insurance that commodities have historically provided wasn't needed—as both stocks and bonds had good returns over the 2003-2012 period—the addition of commodities certainly didn't hurt the portfolio.

On the other hand, if the future brings supply shocks (like an oil embargo or a Middle East war that disrupts supply), or if we experience the high inflation many fear might be created by the Fed's easy monetary policy, history suggests that commodities would help the portfolio during such periods. And after all, that's why one should consider including commodities in the first place: as portfolio insurance.


 

Find your next ETF

Reset All