Spillovers and Directional Predictability with a Cross‐Quantilogram Analysis: The Case of U.S. and Chinese Agricultural Futures

• Date: 01 December 2016
• Author: Huayun Jiang,Neda Todorova,Jen‐Je Su,Eduardo Roca
• DOI: http://doi.org/10.1002/fut.21779
• Published date: 01 December 2016

Spillovers and Directional Predictability

with a Cross-Quantilogram Analysis:

The Case of U.S. and Chinese

Agricultural Futures

Huayun Jiang, Jen-Je Su, Neda Todorova,*and Eduardo Roca

This paper examines the daily, overnight, intraday, and rolling return spillovers of four key

agricultural commodities—soybeans, wheat, corn, and sugar, between the U.S. and Chinese

futures markets via a newly developed quantile dependence measure called quantilogram. The

results reveal significant bi-directional dependence between the two markets across commodi-

ties which is greater in extreme quantiles and moderately stronger from the United States

to China. These findings offer valuable insights into investors’ behavior, market integration,

dissimilarity, and market efficiency in both countries. ©2016 Wiley Periodicals, Inc. Jrl Fut

Mark 36:1231–1255, 2016

1. INTRODUCTION

The United States and China are two of the world’s biggest players and trading partners

in relation to agricultural commodities. The U.S. agricultural futures market, i.e., Chicago

Board of Trade (CBOT) and Inter Continental Exchange (ICE), is the world’s most active

while the Chinese market, i.e., Dalian Commodities Exchange (DCE) and Zhengzhou

Commodities Exchange (ZCE), is the world’s fastest growing (Acworth, 2013). Hence,

the interaction between these two markets is of significant importance as it would have

an impact on other markets (Christofoletti et al., 2012; Han et al., 2013; Hernandez

et al., 2014). Applying MGARCH, VAR, VECM, Granger Causality, Wavelet, and impulse

response functions1to daily closing prices, the existing literature consistently demonstrates

greater price discovery in the United States compared to China. However, there is a lack

Huayun Jiang is a PhD candidate at the Griffith Business School, Griffith University,Australia. Jen-Je Su is a

Senior Lecturer in Economics at the Griffith Business School, Griffith University,Australia. Neda Todorova

is a Senior Lecturer in Finance at the Griffith Business School, Griffith University,Australia. Eduardo Roca

is a Professor in Finance at the Griffith Business School, Griffith University,Australia. We thank the Editor,

Robert Webb, and two anonymous referees for their insightful comments which helped us to improve our

paper substantially. We also would like to thank Han et al. (2014) for making their codes available and

acknowledge the comments from participants of the 44th Australian Conference of Economists, the 27th

Australasian Banking and Finance Conference, the Griffith University Finance PhD Peer Collaboration

Symposium 2015, and the Brownbag Seminar Series of the University of Queensland.

JEL Classification: C12, C13, C22, Q11, Q14

*Correspondence author,Griffith Business School, Griffith University, 170 Kessels Road, Nathan, Queensland

4111, Australia. Tel: +61 7 3735 7219, Fax:+61 7 3735 3719, e-mail: n.todorova@griffith.edu.au

Received January 2015; Accepted December 2015

1Section 2 provides a detailed literature review.

The Journal of Futures Markets, Vol. 36, No.12, 1231–1255 (2016)

©2016 Wiley Periodicals, Inc.

Published online 14 March 2016 in Wiley Online Library (wileyonlinelibrary.com).

DOI: 10.1002/fut.21779

1232 Jiang et al.

of studies that systematically investigate the mechanism and dynamics of the spillovers

between these two markets. Our study addresses this important knowledge gap.

This paper investigates the return spillovers between the U.S. and Chinese agricultural

futures markets in relation to the four major commodities—soybeans, wheat, corn, and sugar,

based on a newly developed methodology called quantilogram. Given the close connection

between the United States and China in terms of trade and investment, we expect a significant

bilateral dependence between the two markets. However, since the U.S. futures market

is globally well-established and easily accessed by different types of investors whereas the

Chinese futures market is more restricted to foreign participation (Talpsepp, 2011; Wang,

2012), we conjecture a stronger return spillover from the United States to China than vice

versa.

This study makes three contributions to the empirical literature. First, the cross-

quantilogram introduced by Han et al. (2014) offers the advantage of providing a com-

prehensive examination of return spillovers and directional predictability between the U.S.

and Chinese futures markets as it allows the analysis of the relationship between the two

markets at different quantiles, rather than just at the median. The recent Global Financial

Crisis (GFC) and its aftermath have highlighted, once again, the imperative of understand-

ing and modeling relationships between markets at the extremes of the return distributions

rather than just at the center. Although there are other methods that can be used to analyze

relationships between variables at different quantiles, for example, quantile regressions, the

cross-quantilogram has advantages over them. As demonstrated by Han et al. (2014), the

cross-quantilogram can detect the magnitude, duration, and direction of the relationship si-

multaneously, which cannot be done by means of a quantile regression. One can also select

arbitrary quantiles for both time series, rather than preset ones. Furthermore, the stationary

bootstrap is used to construct critical values and allows the use of large lags. The cross-

quantilogram has been applied only to the equity market by Han et al. (2014). Thus, this

paper is the first to apply this newly introduced and advantageous methodology in studying

commodity markets.

Second, unlike existing research, this study analyzes the daily interaction between the

U.S. and Chinese agricultural futures markets based on trading and non-trading parts of

the day with an adjustment on cross-country nonsynchronous trading hours. Motivated by

the extensive body of literature on market microstructure in relation to the stock market,

which suggests that the market behaves differently during trading and non-trading periods

(Barclay and Hendershott, 2008; Hong and Wang, 2000), in this paper, daily (close-to-close)

returns are decomposed into overnight (close-to-open) and intraday (open-to-close) returns

components. To the best of the authors’ knowledge, the extant literature on agricultural

futures is primarily based on daily (close-to-close) returns. Only one study (Fung et al., 2013)

is found to have decomposed daily returns into overnight and intraday returns. However, this

study differs from ours as it mainly focuses on the contemporaneous and lead–lag dynamic

return transmissions in the conditional mean using a standard approach.

Third, different from other studies, our paper tests whether the movements in returns

in each of the two countries are indeed driven more bilaterally rather than by their own

historical returns. This is done by performing an autocorrelation analysis of each country’s

returns using the auto-quantilogram. Moreover, a rolling cross-quantilogram based on a

rolling window of 2 years is examined to identify the stability and dynamics of the observed

spillovers found in the entire sample.

With return series obtained from daily closing prices, the cross-quantilogram results

show a significant short-lived (1 day) positive directional predictability for soybean, wheat,

corn, and sugar futures from the United States to China across various quantiles. The trans-

mission from China to the United States is also short-lived (1 day) for each of the four