A Tale of Two Premiums: The Role of Hedgers and Speculators in Commodity Futures Markets

• DOI: http://doi.org/10.1111/jofi.12845
• Date: 01 February 2020
• Author: KE TANG,K. GEERT ROUWENHORST,WENJIN KANG
• Published date: 01 February 2020

THE JOURNAL OF FINANCE •VOL. LXXV, NO. 1 •FEBRUARY 2020

A Tale of Two Premiums: The Role of Hedgers

and Speculators in Commodity Futures Markets

WENJIN KANG, K. GEERT ROUWENHORST, and KE TANG∗

ABSTRACT

This paper studies the dynamic interaction between the net positions of traders and

risk premiums in commodity futures markets. Short-term position changes are driven

mainly by the liquidity demands of noncommercial traders, while long-term varia-

tion is driven primarily by the hedging demands of commercial traders. These two

components influence expected futures returns with opposite signs. The gains from

providing liquidity by commercials largely offset the premium they pay for obtaining

price insurance.

AN IMPORTANT QUESTION IN THE commodity futures literature concerns the in-

fluence of speculative activity on the price formation in futures markets. Ac-

cording to the traditional view of Keynes (1923) and Hicks (1939), the presence

of speculative capital facilitates risk-sharing for hedgers who seek insurance

against future price fluctuations. A central assumption of the theory of nor-

mal backwardation is that the hedging demand for futures is net short, and

∗Wenjin Kang is at the School of Finance, Shanghai University of Finance and Economics and

the Shanghai Institute of International Finance and Economics. Geert Rouwenhorst is at the Yale

School of Management and a partner at SummerHaven Investment Management. Ke Tangis at the

Institute of Economics, School of Social Sciences at Tsinghua University. We have benefited from

comments and suggestions from Hendrik Bessembinder; Geetesh Bhardwaj; Ing-haw Cheng; Gary

Gorton; Ravi Jagannathan; Andrei Kirilenko; Pete Kyle; Peng Liu; Anna Pavlova; Neil Pearson;

Tarun Ramadorai; Michel Robe; Matthew Spiegel; Marta Szymanowska; Dimitry Vayanos; two

anonymous referees; the Editor (Stefan Nagel); and seminar participants at 2013 National Bureau

of Economic Research (NBER) Meetings on the Economics of Commodity Markets, 2015 Western

Finance Association Annual Conference, 2015 China International Conference in Finance, First

Commodity Conference at Leibnitz University Hannover, United Nations Conference of Tradeand

Development (UNCTAD), Collegio Carlo Alberto, the CFTC, the Chinese University of Hong Kong,

the City University of Hong Kong, CKGSB, Notre Dame, Hong Kong Polytech University,JP Mor-

gan Commodity Centre at University of Colorado Denver,Luxembourg School of Finance, Pontifical

Catholic University of Chile, Shanghai JiaoTong University, Shanghai University of Finance and

Economics, SKKU Graduate School of Business (Seoul), University of Macau, Washington Univer-

sity St. Louis, and YaleUniversity. Kang acknowledges research support from Shanghai University

of Finance and Economics. Rouwenhorst acknowledges research support from the Yale School of

Management. Tang acknowledges financial support from the National Natural Science Fundation

of China (Grant No. 71973075) and National Science Fund for Distinguished Young Scholars of

China (Grant No. 71325007). SummerHaven Investment Management invests in, among other

things, commodity futures. The views expressed here are those of the authors and not necessarily

those of any affiliated institution.

DOI: 10.1111/jofi.12845

C2019 the American Finance Association

377

378 The Journal of Finance R

that hedgers induce speculators to absorb the risk of commodity price fluctu-

ations by setting futures prices at a discount relative to expected future spot

prices.

While the view that insurance provision is an important element of com-

modity futures trading is not controversial per se, there are several reasons

to believe that the Keynesian view is an incomplete description of the trading

motives of commodity futures markets participants. First, it is often not pos-

sible to draw a clear distinction between speculative and hedging activities.

For instance, commercial “hedgers” may decide to selectively hedge based on

their market views, in which case their positions can be thought of as having

both a hedging as well as a speculative component. Second, noncommercial

“speculators” form a heterogeneous group that includes hedge funds, money

managers, and index traders. These speculators’ motives to trade are likely

independent of accommodating commercial hedging demands. Consider, for

example, commodity trading advisors (CTAs) that actively pursue momentum-

style investment strategies,1or institutional investors that take exposure to

a commodity index for the purpose of diversification.2It seems unlikely that

these investment decisions originate simply from passively meeting the hedg-

ing demands of commercial market participants, as would be required to

fit the Keynesian story. A central focus of this paper is to explore the fac-

tors which influence position changes that are unrelated to hedging demands

and how these factors impact expected risk premiums in commodity futures

markets.

In light of these conceptual concerns, it is perhaps not surprising that em-

pirical support for the central prediction of the Keynesian view of commodity

futures markets has been weak (see Rouwenhorst and Tang (2012) for a litera-

ture review). In empirical work, it is common to construct a measure of hedging

pressure from the weekly reports on trader positions published by the Com-

modity Futures Trading Commission (CFTC). The test of the theory examines

whether variation in hedging pressure—measured as the net short position of

commercial traders—helps predict variation in expected futures risk premi-

ums. One possible reason for the failure to find a strong link is that the CFTC

trader classifications do not align with the distinction between speculative and

hedging positions. An alternative explanation suggested by Cheng, Kirilenko,

and Xiong (2015, CKX hereafter) and pursued in this paper emphasizes the

importance of motives to trade that are distinct from insurance provision.3

1Fung and Hsieh (1997,2001) analyze trend-following strategies by hedge funds, which have

become a major source of speculative capital over recent decades. Rouwenhorst and Tang (2012)

document that changes in speculative positions are positively correlated with relative returns in

commodity futures markets. Moskowitz, Ooi, and Pedersen (2012) find that speculators follow

time-series momentum strategies in many futures markets. Bhardwaj, Gorton, and Rouwenhorst

(2014) show that the returns of CTAs correlate with simple momentum and carry strategies in

stocks, currencies, and commodities.

2See, for example, Hamilton and Wu (2015) and Basak and Pavlova (2016).

3CKX (2015) show that during the financial crisis, financial traders such as hedge funds and

index investors reduced their net long positions in response to market distress. This was facilitated

by hedgers who reduced their net short positions as prices fell.

A Tale of Two Premiums in Commodity Futures Markets 379

We conjecture that much of the short-term trading by noncommercials leads

to variation in observed hedging pressure that is not driven by commercial

hedging motives. More importantly, such trading demands liquidity from their

commercial counterparts, for which commercials receive compensation in the

form of a risk premium.

To establish the liquidity provision channel, we examine futures returns fol-

lowing the weekly position changes reported by the CFTC. We find that active

trading decisions by noncommercials influence the price setting in commod-

ity futures markets. Our empirical strategy follows Kaniel, Saar, and Titman

(2008) by testing for the predictability of short-term returns following position

changes, and using the direction of this return predictability to infer who pro-

vides and who consumes liquidity in futures markets.4We find that during

the weeks following a position change, commodities that were bought by non-

commercials earn significantly lower returns than commodities that were sold

by them, while commodities that are purchased by commercial traders subse-

quently outperform those that are sold by them. Our empirical findings parallel

the prediction from microstructure theory5that liquidity providers (i.e., com-

mercials) trade as contrarians, while impatient traders (i.e., noncommercials)

consume liquidity and need to offer a price concession to incentivize risk-averse

market makers to take the other side of their trades.

The short-term underperformance of commodity futures sold by commercials

is opposite to the prediction of the Keynesian view,which associates an increase

in commercial selling (hedging) pressure with higher expected risk premiums.

We conjecture that variation in hedging pressure has two components: a short-

term component that is driven primarily by the liquidity demands of noncom-

mercials, and a longer-term component that stems from changes in the hedg-

ing demands of commercial market participants. The latter is relatively stable

over short-term horizons due to the slow evolution of underlying production

decisions in physical markets. When we control for the variation in hedging

pressure that is induced by short-term trading, the positive relationship be-

tween hedging pressure and expected futures risk premiums reemerges. This

leads to the second main finding of our paper, namely, that the expected excess

return to a commodity futures contract embeds two return premiums related

to position changes: a premium paid by commercials to noncommercials for ob-

taining price insurance, and a premium paid by noncommercials to commercial

traders for accommodating their short-term liquidity needs.

Tofurther support our hypothesis, we examine the sensitivity of liquidity pro-

vision to the risk environment. First, using option-implied commodity volatility

4Kaniel, Saar, and Titman (2008) study the dynamic relation between net individual investor

trading and short-horizon returns for a large cross-section of NYSE stocks, and show that the

demand for trade execution immediacy by institutions is met by liquidity provision by individual

investors and leads to predictable returns following their trades. In our context of commodity

futures markets, we test for the predictability of short-term returns following position changes

by commercial hedgers and noncommercial speculators, and we use such return predictability to

make inferences about who provides liquidity in these markets.

5See, for example, Grossman and Miller (1988) and Campbell, Grossman, and Wang (1993).