Manacled short sellers and return premium: New evidence

• Published date: 01 March 2022
• Author: Xiao‐Ming Li
• Date: 01 March 2022
• DOI: http://doi.org/10.1111/eufm.12309

Eur Financ Manag. 2022;28:403–432. wileyonlinelibrary.com/journal/eufm © 2021 John Wiley & Sons Ltd.

|

403

DOI: 10.1111/eufm.12309

ORIGINAL ARTICLE

Manacled short sellers and return premium:

New evidence

Xiao‐Ming Li

School of Economics and Finance,

Massey University, Auckland,

New Zealand

Correspondence

Xiao‐Ming Li, School of Economics and

Finance, Massey University, Private Bag

102904, North Shore, Auckland 0745,

New Zealand.

Email: x.n.li@massey.ac.nz

Abstract

Investigating the short‐selling regulation of the Hong

Kong market, we document that shortable stocks, on

average, earn significantly higher returns than non‐

shortable stocks. However, loadings of stocks/portfo-

lios on the shortable minus non‐shortable misvaluation

factor SMN predict a significant negative return pre-

mium in the cross‐section of returns. We measure SMN

by applying both value‐and return‐weighted methods

with various time lags. We propose a behavioural

model to rationalize our results. The model shows that,

if investors are overconfident regarding short‐selling

regulatory factor signals, it is possible to detect a po-

sitive average/abnormal return but a negative future

return premium on SMN.

KEYWORDS

Hong Kong market, mispricing, overconfidence bias, short‐sale

regulation

JEL CLASSIFICATION

G02, G10, G12, G28

EUROPEAN

FINANCIAL MANAGEMENT

I thank an anonymous referee, the editor of EFM John Doukas, the editor at ACADEMICWORD David Levy and the

participants at the Asian Finance Association 2015 Annual Meeting, for their comments and suggestions that have

greatly improved this paper. The usual declaimer applies.

1|INTRODUCTION

In conventional wisdom, which ignores the issues of investors' behavioural biases, there al-

ready have been debates regarding the price effects of short‐sale constraints. On one hand,

Miller (1977) shows that asset prices will be upward biased, since only optimistic views are

reflected in prices. On the other hand, Diamond and Verrecchia (1987) predict that security

prices are not biased upward, since risk‐neutral rational investors correctly account for excess

demand by optimistic investors.

Empirical studies have largely found evidence consistent with the theoretical prediction of

Miller (1977). Specifically, securities with more binding short‐sale constraints proxied by in-

stitutional ownership (Asquith et al., 2005), higher short‐sale cost measured by rebate rates

(Drechsler & Drechsler, 2016), or greater short‐sale risks measured by the variation in security

loan fees (Engelberg et al., 2018) tend to earn lower expected returns in the near future.

Recently, Chu et al. (2020) show that pilot stocks (with short‐sale constraints loosened) earn

higher returns during the pilot period than during the nonpilot period (when the stocks become

more constrained for short sales). In contrast, however, Bai et al. (2017) document evidence

inconsistent with Miller's (1977) prediction. The authors find that stocks off the officially

designated shorting list (treated as non‐shortable) earn significantly higher average returns

than stocks on the list (treated as shortable). The authors attribute this finding to three types of

risk entailed by non‐shortable stocks: overvaluation risk, constraint‐induced liquidity risk and

constraint‐induced information risk. Although reporting conflicting results, the above‐cited

empirical studies share a commonality, in that they are conducted basically within the ra-

tionality framework.

Unlike the rationality framework, behavioural finance has two building blocks: limits to

arbitrage and psychology (Barberis & Thaler, 2003). Going along with the behavioural ap-

proach, we ask whether the price effects of short‐sale constraints would be influenced by

investor behavioural biases. If so, what might be the consequent results in terms of expected

returns, abnormal returns and predicted return premiums? Whereas short‐sale constraints are

well known for their misvaluation effects, several studies have shown that investor behavioural

biases can also cause common misvaluation across all stocks in the market (Chang et al., 2007;

Hirshleifer & Jiang, 2010, hereafter, HJ). In these studies, the central link between behavioural

biases and misvaluation comovement is economic factors that do not involve the government

sector. The theoretical underpinnings of the studies' focus on economic factors include irra-

tional shifts of investors towards certain stock characteristics (Barberis & Shleifer, 2003) and

the overconfidence of investors interpreting economic (firms') fundamentals (Daniel

et al., 2001). The results demonstrate that systematic misvaluation related to economic factors

has a strong return‐predictive power beyond that of conventional risk factors.

It has become a consensus that governments can create or worsen market frictions. Gov-

ernment policy/management/regulation is an important piece of information that will influ-

ence investors' decisions. For example, the short‐selling regulation of the Stock Exchange of

Hong Kong (SEHK) determines a stock's short‐selling status, thereby adding one more char-

acteristic (i.e., shortability) to its existing ones (size, book‐to‐market, etc.). This additional

characteristic, resulting from the short‐selling regulation, is classified in our study as a reg-

ulatory factor. A regulatory factor, as an economic factor, must possess payoff signals. Thus, in

providing theoretical underpinnings for our empirical results, we propose a behavioural model

that adds, as a novelty, the short‐selling regulatory factor to HJ's model, which only allows for

economic factors. However, unlike an economic factor that mainly conveys private signals (HJ),

404

|

EUROPEAN

FINANCIAL MANAGEMENT

LI