Rouwenhorst: Correlation In Commodities Up

August 07, 2012

The Yale professor and commodity index builder says a rise in commodities correlation is real, but is no deterrent to the asset class' diversification power.


Geert Rouwenhorst is probably best known as one of the first to hail the virtues of commodities as a bona fide asset class helping fuel the proliferation of commodities indexes and the ETFs that have followed them in the past decade or so. The Yale School of Management professor and deputy director for the International Financial Center at the university is also a partner at SummerHaven Investment Management, and the mastermind behind some of the benchmarks underlying United States Commodity Funds' ETFs, including the United States Commodity Index Fund (NYSEArca: USCI).

In anticipation of his keynote speech at IndexUniverse's upcoming Inside Commodities conference in Chicago, he caught up with IndexUniverse Correspondent Cinthia Murphy, and acknowledged the challenges of creating a catch-all broad commodities benchmark. But he was also quick to say that despite rising correlation between various commodities due to the increasing acceptance of commodities as an asset class, they remain strong diversification tools that should be found in every investor's portfolio.


Murphy: Your work has been key to the rise of commodities ETFs. They have in fact been one of the fastest-growing ETF asset classes recently. But some argue that there’s no such thing as commodities beta. How do you define commodities beta?

Rouwenhorst: Beta is a measure of the sensitivity of an investment to broad market movements. In equity markets, which is where the notion of a beta was first used, “the market” usually means a broadly diversified portfolio that includes many stocks, usually weighted by market capitalization. The average stock beta is 1. Stocks with a beta above 1 would on average move more than 1-for-1 with the market, whereas stocks with betas below 1 would have less sensitivity to market movements.

A similar calculation can, in principle, be performed for commodities markets. How sensitive are individual commodities to the overall commodity market? In commodity markets, there is less agreement on what the appropriate broad market benchmark is.

Murphy: What are some of the challenges with creating a catch-all commodities index that would be viewed as a broad benchmark akin to the way the S&P 500 is viewed for equities?

Rouwenhorst: First, most commodity indexes are constructed from futures prices, and therefore exclude commodities for which there is an important spot market, but no futures market. Next, there is no natural weighting scheme equivalent to market capitalization in commodity futures, because for every long there is a short in futures: The overall net market cap of any futures contract is zero. This is why commodity indexes have looked elsewhere to determine weights. The earliest indexes used a combination of world production and liquidity and other factors to determine the index weights.

Murphy: So, how does one capture true commodities beta?

Rouwenhorst: Of course, betas cannot be observed in practice, and have to be estimated. In this estimation, I would choose an index that is well-diversified to calculate betas. Or alternatively, always be clear and report the index that is used to calculate beta. If a majority of investors use a particular index as their benchmark, then that could be an alternative as well.

Sometimes people take commodities beta to mean the sensitivity of commodities (as an asset class) to the overall stock market. This would be more like the “equity beta” of commodities. Beta defined this way has historically been low, and during some periods even negative. In recent years, the correlation between commodities and equities has increased, and the equity beta of commodities has turned positive. Some researchers have attributed this to the “financialization” of commodities.

Murphy: How much correlation is there between commodities?

Rouwenhorst: They are net positively correlated to each other. We’ve seen typical commodities correlation of about 10 percent in the past 40-year history, but in the last three years, that correlation has jumped to 50 percent.


Find your next ETF

Reset All