Is idiosyncratic asymmetry priced in commodity futures?

• Published date: 01 September 2023
• Author: Yufeng Han,Xuan Mo,Zhi Su,Yifeng Zhu
• Date: 01 September 2023
• DOI: http://doi.org/10.1111/jfir.12339

Received: 25 February 2022

|

Accepted: 4 May 2023

DOI: 10.1111/jfir.12339

ORIGINAL ARTICLE

Is idiosyncratic asymmetry priced in commodity

futures?

Yufeng Han

1

|Xuan Mo

2

|Zhi Su

3

|Yifeng Zhu

4

1

Belk College of Business, University of North

Carolina at Charlotte, Charlotte,

North Carolina, USA

2

School of Economics and Management &

Lab for Low‐Carbon Intelligent Governance,

Beihang University, Beijing, China

3

School of Statistics and Mathematics,

Central University of Finance and Economics,

Beijing, China

4

School of Finance, Central University of

Finance and Economics, Beijing, China

Correspondence

Yufeng Han, Belk College of Business,

University of North Carolina at Charlotte,

9201 University City Blvd., Charlotte,

NC 28223‐0001, USA.

Email: yhan15@uncc.edu

Funding information

National Natural Science Foundation of

China, Grant/Award Numbers: 72173145,

72134002 & 72172033

Abstract

In this article, we use a recently introduced asymmetry

measure, IE, to measure the idiosyncratic asymmetry of

commodity futures returns and find that idiosyncratic

asymmetry negatively and significantly predicts commodity

futures returns cross sectionally. Furthermore, we find that

a long–short trading strategy based on idiosyncratic

asymmetry generates significant abnormal returns, which

cannot be explained by traditional risk factors in commod-

ity futures and persists up to 12 months. Moreover,

idiosyncratic asymmetry appears to be a priced factor in

commodity futures with significant risk premium. Finally,

we confirm that IE is better at capturing the pricing effect

of idiosyncratic asymmetry than the traditional skewness

measure.

JEL CLASSIFICATION

G11, G12, G13

1|INTRODUCTION

Idiosyncratic asymmetry and its asset pricing implications have received much attention in investment theory and

practice. Tversky and Kahneman (1992), Mitton and Vorkink (2007), Barberis and Huang (2008), and Han et al.

(2022) theoretically demonstrate that idiosyncratic asymmetry negatively affects the cross section of expected

returns. Boyer et al. (2010), Annaert et al. (2013), Tavakoli Baghdadabad and Mallik (2018), and Bali et al. (2020),

among others, have provided substantial evidence in stock markets to support this negative relation.

J Financ Res. 2023;46:875–898. wileyonlinelibrary.com/journal/JFIR

|

875

This is an open access article under the terms of the Creative Commons Attribution‐NonCommercial‐NoDerivs License, which

permits use and distribution in any medium, provided the original work is properly cited, the use is non‐commercial and no

modifications or adaptations are made.

© 2023 The Authors. Journal of Financial Research published by Wiley Periodicals LLC on behalf of The Southern Finance

Association and the Southwestern Finance Association.

The traditional way to measure idiosyncratic asymmetry is the third‐order moment coefficient of skewness.

However, idiosyncratic skewness reflects the degree of asymmetry only to a certain point. To measure idiosyncratic

asymmetry more efficiently, Jiang et al. (2020) propose a distribution‐based measurement, IE, which is defined as

the difference between upside and downside return probabilities. Jiang et al. (2020) show that IE can provide

additional information beyond skewness. For a distribution with zero skewness, for example, it may still be

asymmetrically distributed, which can be captured by IE. Moreover, following cumulative prospect theory, Jiang

et al. (2020) derive a negative relation between IE and expected stock returns

1

and provide evidence that IE is more

powerful in predicting the cross section of stock returns than skewness.

2

In this article, we extend the work of Jiang et al. (2020) to commodity futures by examining whether

idiosyncratic asymmetry measured by IE can also predict the cross‐section of commodity futures returns. The

commodity futures market differs from the stock market in several important aspects, which make it nontrivial to

extend the findings in the stock market to the commodity futures market. First, it is dominated by institutional

investors who are presumably more rational and less subject to behavioral biases. Second, although both long and

short positions can be easily established in the commodity futures market with low trading costs, both sides are

subject to mark‐to‐market daily. Third, commodity futures prices are greatly affected by supply and demand

relations, in particular, seasonal variations between demand and supply. Fourth, perhaps because of the issues

mentioned and other issues, commodity futures returns are even more volatile than stock returns, which makes it

much harder to predict returns in the commodity futures markets.

Recently, Fernandez‐Perez et al. (2018) document a negative relation between total asymmetry as measured by

skewness and commodity futures returns. However, whether this negative relation is due to the idiosyncratic

component or systematic component remains unknown. This is important because the aforementioned theoretical

papers are based on idiosyncratic asymmetry instead of total asymmetry. Thus, it is necessary to investigate

whether the negative relation between total asymmetry and future return is mainly driven by idiosyncratic

asymmetry in the commodity futures market.

Additionally, Daskalaki et al. (2014) find that very few risk factors can be priced in commodity futures. They

conduct a comprehensive study to investigate whether macro factors, equity‐motivated tradable factors,

commodity‐specific factors, and latent factors from principal component analysis can be priced in commodity

futures. They find that none of the factors are priced. Therefore, it is important to exam whether idiosyncratic

asymmetry (IE) is a priced factor in commodity futures. If it is priced, it would make IE a rare candidate for any factor

models aiming to explain commodity futures returns.

In sum, in this article we seek to answer the following questions: (1) Can idiosyncratic asymmetry (IE) predict

the cross‐section of commodity futures returns? (2) Is the total asymmetry pricing effect mainly driven by the

idiosyncratic asymmetry (IE) in the commodity futures market? (3) Is idiosyncratic asymmetry a priced factor in the

commodity futures market?

First, we use both Fama–MacBeth (1973) regressions and portfolio sorts to investigate the pricing effect of

idiosyncratic asymmetry (IE) on a cross‐section of commodity futures returns. We compare the results of IE with

those of traditional measures and find that IE is better at capturing the negative relation between idiosyncratic

asymmetry and expected returns. Based on the Fama–MacBeth (1973) regression results, we find that a 1

percentage point increase in idiosyncratic asymmetry (IE) reduces the expected returns of commodities by 2.63

basis points. This supports the prediction provided by the model of Jiang et al. (2020). It is also consistent with the

theories of Mitton and Vorkink (2007), Barberis and Huang (2008), and Han et al. (2022). Our finding is robust when

adding additional control variables for gambling preference, liquidity, and financial crisis, as well as using alternative

methods to roll over the commodity futures contracts.

1

An analytical proof is provided in Jiang et al. (2020).

2

Evidence based on the US and Chinese stock markets are documented in Jiang et al. (2020) and Chen et al. (2022), respectively.

876

|

JOURNAL OF FINANCIAL RESEARCH