Pairs trading across Mainland China and Hong Kong stock markets

• Author: Hanxiong Zhang,Andrew Urquhart
• Date: 01 April 2019
• DOI: http://doi.org/10.1002/ijfe.1687
• Published date: 01 April 2019

RESEARCH ARTICLE

Pairs trading across Mainland China and Hong Kong stock

markets

Hanxiong Zhang

1

| Andrew Urquhart

2

1

Lincoln International Business School,

University of Lincoln, Lincoln, UK

2

Southampton Business School,

University of Southampton, Southampton,

UK

Correspondence

Andrew Urquhart, Southampton Business

School, University of Southampton,

Southampton SO17 1BJ, UK.

Email: a.j.urquhart@soton.ac.uk

JEL Classification: G10; G11

Abstract

Motivated by the rationale that market inefficiency arises from a combination

of less than fully rational demand and limits to arbitrage, this paper investi-

gates the profitability of pairs trading across Mainland China and Hong Kong

on highly liquid large‐cap and midcap stocks from January 1996 to July 2017.

We have three main findings. First, we find that pairs trading constrained

within each market generates no significant abnormal returns. However, if

investors can trade across Mainland China and Hong Kong, pairs trading is

profitable after adjusting for risk and transaction costs, where the annualized

abnormal return is 9% over the full sample. Second, by using a rolling‐window

regression, we find that the profitability of the strategy is time‐varying. The

bootstrap simulations suggest that the decline in profitability of the strategy

since 2012 is due to random chance rather than poor ability of identifying

mispriced stocks. However, the vast majority of profitable periods reflect the

strategy's ability to choose profitable stocks rather than random chance. Third,

the profitability of the strategy is somewhat sensitive to market conditions,

most notably, the strategy is more profitable during longer term market turbu-

lence. Overall, our empirical findings are consistent with the Adaptive Market

Hypothesis in that the integration of financial markets and market conditions

determine the level of market efficiency.

KEYWORDS

Adaptive Market Hypothesis, bootstrap simulation, cross‐border trading, market conditions, market

efficiency, pairs trading

1|INTRODUCTION

Motivated by the theoretical hypothesis and empirical

findings that market inefficiency arises from some combi-

nation of less than fully rational demand and limits to

arbitrage (Baker, Bradley, & Taliaferro, 2014; Jacobs &

Weber, 2015), this paper studies pairs trading from three

perspectives: First, we study whether pairs trading across

Mainland China and Hong Kong stock markets is more

profitable than trading in the individual market by using

highly liquid stocks; second, whether the profitability of

pairs trading declines since the higher integration

between Mainland China and Hong Kong stock markets;

and third, what drives this profitability and whether this

profitability is time‐varying.

Pairs trading is a popular statistical arbitrage strategy

used by hedge funds and investment banks since 1987

(Gatev, Goetzmann, & Rouwenhorst, 2006). Essentially,

the strategy finds two stocks whose prices move together

over time, and if the pair of prices diverge wide enough,

investors buy the declining stock and sell the increasing

stock simultaneously to take the advantage of the

Received: 4 December 2017 Revised: 29 July 2018 Accepted: 10 September 2018

DOI: 10.1002/ijfe.1687

698 © 2018 John Wiley & Sons, Ltd. Int J Fin Econ. 2019;24:698–726.wileyonlinelibrary.com/journal/ijfe

mispriced stocks while maintaining a level of market

neutrality. The rationale behind pairs trading is to profit

from mean‐reversion forces that eliminate short‐term

price deviations in favour of long‐term historical pricing

relationships. There have been a number of papers that

have examined the profitability of pairs trading. Gatev

et al. (2006) find that pairs trading yields statistically sig-

nificant monthly excess return in the order of 0.9% and a

risk‐adjusted return of 0.76% before transaction costs in

the U.S. stock market by using the CRSP stocks from

1962 to 2002. After considering market impact, proxied

by bid–ask spread, the monthly profits drop to 0.19% to

0.38%. To rule out the concern that their findings are data

mining, Gatev et al. (2006) conduct an out‐of‐sample and

show that pairs trading remains profitable. Using CRSP

data from 1962 to 2009, Do and Faff (2010) find that the

profitability of pairs trading peaks in 1970s and 1980s

and declines since 1990s. Do and Faff (2012) find that

the strategy performs strongly during periods of market

turbulence, and after adjusting for commissions, market

impact, and short‐selling fees, they find that the results

of Gatev et al. (2006) lose its profitability. Nevertheless,

pairs trading remains profitable in a fairly small number

of refined versions and at much diminished levels. Gatev

et al. (2006) and Do and Faff (2010, 2012) focus on the

highly liquid U.S. stocks by including stocks in the CRSP

daily files that are alive throughout the 18‐month period

(12‐month formation period and 6‐month trading period).

When a stock in a pair is delisted from CRSP, Gatev et al.

(2006) close the position in that pair, using the delisting

return, or the last available price. Gatev et al. (2006) find

that more than 90% of stocks in the pairs trading come

from the top five size deciles using CRSP breakpoints.

Do and Faff (2012) excludes stocks in the bottom size dec-

ile using New York Stock Exchange breakpoints, and

those with closing prices in the 1‐year formation period

less than $5. Broussard and Vaihekoski (2012) extend

the research to the illiquid Finnish market and find the

strategy to be profitable even allowing for a 1‐day delay

in the trade execution after the trading signal. Jacobs

and Weber (2015) find that the strategy is persistently

profitable in 35 international stock markets and that the

strategy is quite successful for pairs that are hard to arbi-

trage or less visible. Using the constituents of FTSE All‐

Share Index from 1979 to 2012, Bowen and Hutchinson

(2016) find that the pairs trading performs well in market

turmoil and argue that the abnormal returns in the

United Kingdom can be accounted for by a combination

of time‐varying risk exposures and transaction costs.

A number of different methodologies have been

employed to take advantage of pairs trading. For exam-

ple, Elliott, Van Der Hoek, and Malcolm (2005) propose

a Gaussian‐Markov chain model for the spread. Tourin

and Yan (2013) propose a dynamic pairs trading strategy

using the stochastic control approach. In the spirit of

Vidyamurthy's (2004) cointegration method, Li, Chui,

and Li (2014) find that the pairs trading yields an average

annualized excess return of about 17.6% using 38 firms

dual‐listed on A‐shares market in Mainland China and

H‐shares in Hong Kong during the period 2009–2013,

whereas Marshall, Nguyen, and Visaltanachoti (2013)

find that the pairs trading is profitable and support the

Adaptive Market Hypothesis (AMH) by using two

extreme liquid S&P500 ETFs. Xie, Liew, Wu, and Zou

(2016) find that the copula method is superior to distance

method by investigating 89 U.S. stocks in the utility

industry with a sample from 2003 to 2012. Rad, Low,

and Faff (2016) investigate the performance of the dis-

tance method, cointegration method, and copula method

on the U.S. stock market from 1962 to 2014 and find that

all three methods show significant profitability; however,

the distance method outperforms the other methods

slightly after adjusting for transaction costs.

1

Profitable pairs trading, however, is a challenge to

Efficient Market Hypothesis (EMH), which in the weak

form states that prices already reflect all information that

can be derived by examining past market trading data

such as the history of past prices and trading volume. If

prices are predictable and profits could be made by using

historical data, arbitrage would eliminate these profits in

an efficiently operating market. Therefore, there should

be no predictability in security prices. However, success-

ful pairs trading employs past data to predict future prices

and therefore violates the EMH. From a theoretical per-

spective, Grossman and Stiglitz (1980) argue that a per-

fectly efficient market is impossible if prices correctly

reflect all available information, and no one would have

any motivation to acquire costly information. Campbell,

Lo, and MacKinlay (1997) suggest that market efficiency

is not an all‐or‐nothing absolute condition but a relative

notion. As a modification to EMH, Lo (2004) proposes

the AMH. The AMH extends the EMH view of the mar-

ket to argue that learning, competition and evolutionary

selection pressures govern the forces that drive prices to

their efficient levels. The AMH provides a number of

practical implications within finance. First, the risk pre-

mium varies over time according to the stock market

environment and the demographics of investors in that

environment. The second implication is that arbitrage

opportunities do exist from time to time in the market.

Thus, from an evolutionary viewpoint, active liquid

financial markets imply that profit opportunities must

exist. However, as they are exploited, they disappear.

1

For a recent review of the literature on the pairs trading, see Krauss

(2017).

ZHANG AND URQUHART 699