With this theorem in mind, it would make sense to evaluate any portfolio commodity strategy in light of the time horizon it is best suited for. Commodity futures is inherently a short-term game. Contracts expire and rolling them over comes at a substantial cost. Inflationary expectations are hardly a consideration for commodity futures trading mavens. Instead, their fodder, is weather forecasts, tomorrow's crop reports and new gold deposit rumors. Playing this game well is more than a full time job even with today's best technology on your side. This is why few individual investors and only a handful of successful ones play the game.
Attempting to extend this game to the longer side are automated trading platforms and commodity ETNs. The goal of any automated trading program is to take emotion out of the picture, while, hopefully, being on the right side of a trade more often than not, after expenses. As tempting as this approach may seem, getting it to work for any significant stretch of time is mission impossible even without the added expenses of the trading platform itself.
Commodity ETNs generally engage in some form of automated futures trading, as well. However, most of the larger ones, like PowerShares DB Agriculture Fund (DBA), limit such trading to rolling over contracts nearing expiration into ones with further out expiration dates. This approach works reasonably well under normal market conditions of backwardation to cover expenses and preserve principal. However, it breaks down under contango, or even under backwardation, when expenses exceed implied interest rates. Thus, under a deflationary scenario, commodity bullish ETNs are no better bets than cash in the mid to longer term.
The dual chart of DBA and the Deutsche Bank Diversified Agriculture Index, which it is suppose to be tracking, below it confirms that the slight upward bias of the index since January 2009 is negated by tracking error and expenses associated with DBA. Here, it is important to note that the recent peaks in both of these charts are not systemic and can be easily tracked back to the individual commodities that experienced strength at the time. The January 2010 peak, marked in red on the charts is associated with the strength in Cocoa, as also seen on the July Cocoa futures chart. The most recent, should we say current, peak, on the other hand, is prominent on the September Wheat Futures chart, as well.
To be more exact, the Diversified Agriculture Index and DBA ETN appear to peak a couple of weeks after the leading commodity. It doesn't matter whether the leading commodity peaked due to alleged cornering of the market by Anthony Ward and company or poor crops in the Ivory Coast, as was the case with the cocoa induced peak. It also doesn't matter if the current wheat peak is due to fires and unusually high summer temperatures in Russia, or because of the export limitations that this condition is producing. In every case, the result is the same, easing of commodity price levels once the situation works itself out and that it always does in longer term. In the case of wheat, world inventories are more than enough to prevent further price increases and Colorado has just reported the best wheat harvest in its entire history!
My expectations of the economy, looking forward for the next several years, being strictly deflationary (just look carefully at the employment trends of the past two years, if you don't agree), I took the opportunity to exit DBA on Thursday, September 2, 2010 at $26.39 / share. I bought in on January 19, 2009 at $24.66 / share for a total gain of just 7%.
P.S. When I told Vitaliy Katsenelson of Investment Management Associates at a friendly game of poker on Friday night (we don't play for money) that my deflationary views lead me to selling DBA, he was surprised and exclaimed: "So, why did you buy farmland?" Before you ask the same, allow me to explain that my long term view (say, a decade out) is inflationary and in the meantime, current income from farmland should carry it quite nicely. This is quite different from the situation with DBA, which has no potential to either generate current income or appreciate in a deflationary environment. Let me also add that as an investment, farmland is quite different from other real estate. For example, I would not touch any residential rental real estate right now and for the foreseeable future, but that's a subject for an entirely different article.
E. Jake Berzon