The effects of investor attention on commodity futures markets

• Author: Ziying Li,Liyan Han,Libo Yin
• Published date: 01 October 2017
• Date: 01 October 2017
• DOI: http://doi.org/10.1002/fut.21853

Received: 4 March 2017

|

Accepted: 4 March 2017

DOI: 10.1002/fut.21853

RESEARCH ARTICLE

The effects of investor attention on commodity futures markets

Liyan Han

1

|

Ziying Li

1

|

Libo Yin

2

1

School of Economics and Management,

Beihang University, Beijing, China

2

School of Finance, Central University of

Finance and Economics, Beijing, China

Correspondence

Libo Yin, Associate Professor, School of

Finance, Central University of Finance and

Economics, No. 39 South College Road,

Haidian District, Beijing 100081, China.

Email: yinlibowsxbb@126.com

Funding information

National Natural Science Foundation of China,

Grant numbers: 71671193, 71371022;

Program for Innovation Research in the Central

University of Finance and Economics

This study utilizes the search volume for key terms on Google as a direct and timely

proxy for investor attention in order to examine how attention impacts commodity

futures prices, We provide significant evidence for attention’s influence on 13

commodity futures and the interaction between attention and returns, even after

controlling for important macroeconomic variables. We also examine the impact of

investor attention on market efficiency. Results show that rising attention, on one

hand, increases information efficiency and attenuates arbitrage opportunities,

whereas, on the other hand, decreases market efficiency by facilitating herd behavior.

1

|

INTRODUCTION

Commodity related institutional investments grew sharply from less than $20 billion in 2003 to more than $250 billion in 2010.

By the first quarter of 2016, managed futures funds had almost $333.9 billion of assets under management (AUM) according to

Barclays Hedge.

1

Such a rapid influx of money into commodity markets has the potential to dramatically affect commodity

futures trading volumes, prices, and volatilities as well as the performance of managed futures funds. The growing importance of

commodity market investments has led researchers to examine the factors that might impact the behavior of commodity futures

prices and conditions that influence the efficiency of market pricing.

Information efficiency plays a central role in modern financial theory. Fama (1970) argues that in an efficient market prices

fully and correctly reflect available information and change only in response to the arrival of new information.

2

This makes most

changes in prices unpredictable because the arrival of new information is unpredictable. Information efficiency depends on (at

least some) investors paying sufficient attention to the asset so they can respond quickly to news that impacts the market.

3

While

extreme returns, trading volume and sentiment for news and headlines are used as indirect proxies for investor attention,

4

an

increase in these proxies does not guarantee growing attention, unless investors actually are aware of the news of extreme returns

or turnovers. With the use of the internet penetrating throughout daily life, investors not only passively absorb information from

the news but also actively gather it from the Internet. The search volume on Google has been utilized as a direct, active, and

1

Available online at: http://www.barclayhedge.com/

2

Marsh and Rosenfeld (1986) show that prices may bounce between bid and ask prices even in the absence of the arrival of new information.

3

Grossman and Stiglitz (1980) have shown that perfectly efficient capital markets are impossible when information production is costly.

4

Barber and Odean (2008) provide a comprehensive review of these proxies. In the commodity futures market, evidence of the relationship between news

sentiment and futures return has been provided (Gao & Süss, 2015; Smales, 2014).

J Futures Markets. 2017;37:1031–1049. wileyonlinelibrary.com/journal/fut © 2017 Wiley Periodicals, Inc.

|

1031

timely proxy for investor attention, more specifically retail investor attention (Ben-Rephael, Da, & Israelsen, 2016), and has

proved influential in price efficiency in the equity market.

5

As the expansion of the market inevitably exposes commodity futures to significantly more attention, empiricists have begun

to investigate how investor attention impacts commodity prices. Peri, Vandone, and Baldi (2014) test the relationship between

Internet searches and noise trading in corn futures prices. Li, Ma, Wang, and Zhang (2015) uses the Google search volume index

(GSVI) to measure investor attention and investigate the relationships among the GSVI, different trader positions, and crude oil

prices. Still, evidence for the role of investor attention on commodity futures prices is limited in scope. Following the emerging

literature, we examine how investor attention impacts the behavior of commodity futures prices and ascertain the extent of

market efficiency for each futures with respect to investor attention.

Our primary contributions can be summarized as follows. First, we confirm attention’s influence on the returns of a range of

commodity futures categories including: grains, softs, metals, and energy, unlike existing research that focuses on a specific

commodity category. We consider the search queries for certain keywords as investor attention toward the futures contract.

Using cocoa as an example, our study focuses on the search volume of “cocoa,”“cocoa futures,”“cocoa cost,”“cocoa price,”

“cocoa production,”and “cocoa supply.”The returns of 13 futures contracts that we examine have a statistically significant

relation with investor attention. The significant interaction between attention and futures returns indicates that they are mutually

influential, which is consistent with the theory that attention represents the information flow to investors. The effectiveness

remains when controlling for four macroeconomic variables: VIX, Economic Policy Uncertainty (EPU), the Aruoba-Diebold-

Scotti (ADS) business conditions index, and the Treasury bill-Eurodollar futures (TED) spread, which implies that the

information contained in investor attention is not included in macroeconomic sources.

Second, our study sheds some new light on the findings in the literature concerning the exact role of investor attention from a

perspective of market efficiency. Consistent with the evidence that searches in Google Trends directly represent retail investor

attention (Ben-Rephael et al., 2016), our study highlights that the investor attention serves as an indicator for informational

efficiency, but the exact role of investor attention in market efficiency remains elusive. According to Grossman and Stiglitz

(1980), more information leads to more informative prices, that is, a more informationally efficient market. If increased attention

leads to more information absorbed by the market, then increased attention should improve market efficiency (Vozlyublennaia,

2014). However, some researchers argue that more investor attention can create extra noise and, consequently, reduce market

efficiency (Da, Engelberg, & Gao, 2015). The results show that rising attention increases information efficiency and attenuates

arbitrage opportunities but conversely decreases market efficiency by facilitating herd behavior. The specific effect depends on

the futures contracts and the relative magnitude of the conflicting influences.

Third, although there is an extensive body of theoretical and empirical literature that examines the cointegration of spot and

futures prices to ascertain the efficiency of commodity futures markets (i.e., Kellard et al., 1999; Saberi, Stoorvogel, & Sannuti,

2011; Switzer & El-Khoury, 2007), research that examines a broad swath of commodities to compare market efficiency among

different futures markets is limited. Our study shows that metals futures are the most efficient of the four commodity categories

examined while grains futures are the least efficient. Briefly stated, the ranking of commodity futures contracts from strongest to

weakest in terms of market efficiency are metals > energy > softs > grains. Moreover, the out-of-sample results suggest time-

varying market efficiency for some commodity futures contracts. Specifically, our proxy for investor attention predicted cocoa

and copper futures with more accuracy than historical models for some periods, but the predictive power vanished in recent

years, coincident with the financialization of the futures markets (Cheng & Xiong, 2014). Some researchers argue that

financialization has increased trading in commodity futures markets and enhanced market efficiency. However, natural gas and

cotton futures show a decline in measured market efficiency during the period from 2010 to 2012.

In addition, our study finds both discrepant results among futures contracts that belong to the same category and similarity

among contracts from different categories. We posit investors’heterogeneous preferences in trading commodity futures as a

possible explanation, and this heterogeneity is not limited by the contract’s category. For example, while copper and silver

belong to metals futures, the evidence suggests that investors trading copper futures tend to take into account information over 3

consecutive days, while silver futures market participants tend to trade on information shortly after it arrives. Meanwhile, cocoa,

natural gas, and gold futures display similar interactions with attention. Of course, an alternate explanation is that market

participants do not behave differently but the evidence is limited. One way to further investigate this hypothesis, for example, is

comparing results of CBOE wheat futures and CME Kansas wheat futures. Yet this paper does not provide additional evidence

for this possibility.

5

The existing literature includes the studies of Aouadi, Arouri, and Teulon (2013), Da et al. (2011, 2015), Hou and Ding (2015), Siganos (2013),

Vozlyublennaia (2014), and so forth.

1032

|

HAN ET AL.