The Cross Section of Commodity Returns
What’s the most useful way to look at commodity prices? A team of Smith School researchers sought to answer that question by examining commodity futures returns over a 42-year period.
Jan 29, 2018

The Cross Section of Commodity Returns

As Featured In 
Management Science

Explaining the Cross Section of Commodity Returns

What’s the most useful way to look at commodity prices?

Some say it’s best to consider the storage and convenience yield — the benefit of holding the asset itself, rather than the contract or derivative. That inkling, and the hedging motives of producers, consumers and speculators has sparked various theories about the behavior of commodity futures prices and has spearheaded efforts to understand the economic drivers of commodity risk premia.

Still, questions remained unanswered: Which asset pricing models are capable of reconciling the cross-sectional properties of commodity futures returns, at both the portfolio and individual commodity levels? What is a possible explanation for the positive average returns of the commodity carry and momentum factors? How are the commodity pricing factors related to innovations in economic variables? How should one interpret the behavior of the commodity factors over the different stages of the business cycle?

In new research, a team from the University of Maryland’s Robert H. Smith School of Business sought to answer those questions.

The research was authored by Gurdip Bakshi, the Smith School Dean’s Professor of Finance; Xiaohui Gao Bakshi, visiting professor of finance at the Smith School, and Alberto Rossi, assistant professor of finance at the Smith School.  

The researchers examined commodity futures returns over a 42-year period, beginning in 1970. They studied baseline portfolios, managed portfolios and individual commodities, and found that a three-factor model — driven by an average factor, a carry factor, and a momentum factor — outperformed the nested one- and two-factor counterparts in capturing the cross-section of commodity returns. And they drafted an analysis, centering on an economic explanation of the commodity factors.

The carry strategy, they found, performs poorly when global stock volatility increases, and innovations in global equity volatility can price portfolios sorted on carry. “The economic intuition is that investors dislike these adverse changes in the investment opportunity set and require positive average returns to carry as compensation,” the researchers say.

But they had a harder time explaining commodity momentum, because no traditional macroeconomic variable appeared to satisfactorily explain the returns on such strategy. The researchers did find, however, that the momentum strategy performed well when aggregate speculative activity increases, and innovations in aggregate speculative activity can price portfolios sorted on momentum.

Because commodity speculators are known to follow momentum strategies, the researchers viewed the finding as evidence that whenever more agents in the market behave as momentum traders, momentum strategies self-perpetuate. Commodities whose prices have been increasing continue to increase in price, while the commodities whose prices have been decreasing continue to decrease in price.

Among their other findings, the researchers found three “worth emphasizing.”

- First, innovations in global equity volatility cannot price momentum portfolios, and innovations in speculative activity cannot price carry portfolios.

- Second, the authors’ measure of speculative activity is specific to commodity markets and is unrelated to many proxies that capture the availability of arbitrage capital and liquidity constructed from the equity and bond markets.

- And third, neither carry nor momentum portfolios can be priced by innovations in many other economic variables.

Read more: Understanding the Sources of Risk Underlying the Cross-Section of Commodity Returns is featured in Management Science.

About the Author(s)


Professor Bakshi's research interests include asset pricing, international finance, term structure of interest rates, default risk, and pricing of derivative securities. His works have been published in theAmerican Economic Review, Journal of Business, Journal of Finance, Journal of Financial Economics, Management Science, Journal of Financial and Quantitative Analysis, and Review of Financial Studies. His recent work focuses on studying the valuation structures of technology stocks, risk premiums, investor irrationality, probability of stock market crashes, and credit risks.

Xiaohui Gao Bakshi holds a B.S. in Mathematics from Fudan University, M.S. in Mathematics from the University of Florida, and Ph.D. from the University of Florida (granted in 2008). Professor Gao Bakshi is interested in conducting research in empirical corporate finance and empirical asset pricing. One of her current work features explanations for the recent drought in Initial Public Offerings in the US, in particular, where the paucity of new issues could be traced to economies of scope and regulatory overreach (e.g., Sarbanes-Oxley). She has written on the link between individual investor trading behavior and gambling tendencies of investors in a natural experiment setting.


Alberto Rossi is an Assistant Professor of Finance at the Smith School of Business, University of Maryland at College Park. His research interests include theoretical and empirical asset pricing, portfolio choice and financial econometrics. His recent work concentrates on networks, institutional investors' performance, and the risk-return trade-off in financial markets. He also studies stock return predictability and commodity markets.

More in


How to Optimize Executive Incentives
Leaders can lose their jobs for simply having the right skills at the wrong time, which can complicate negotiations for executive incentive packages.
Jan 16, 2018
Slow and Steady Avoids the Crash
Wall Street traders can keep the market from crashing by implementing a slow-and-steady trading strategy.
Nov 27, 2017
Giving Power to the People
Virtually all director elections in the U.S. are uncontested and a single “for” vote is enough to elect a director. Do votes in these routine elections even matter? Yes, according to new research.
Sep 28, 2017