Common idiosyncratic volatility and returns: From an investment horizon perspective

• DOI: http://doi.org/10.1002/ijfe.1668
• Author: Zhi Su,Tengjia Shu,Libo Yin
• Date: 01 January 2019
• Published date: 01 January 2019

RESEARCH ARTICLE

Common idiosyncratic volatility and returns: From an

investment horizon perspective

Libo Yin

1

| Tengjia Shu

2

| Zhi Su

3

1

School of Finance, Central University of

Finance and Economics, Beijing, China

2

Tippie College of Business, The

University of Iowa, Iowa City, Iowa

3

School of Statistics and Mathematics,

Central University of Finance and

Economics, Beijing, China

Correspondence

Libo Yin, School of Finance, Central

University of Finance and Economics, No.

39 South College Road, Haidian District,

Beijing 100081, China.

Email: yinlibowsxbb@126.com;

0020130053@cufe.edu.cn

Funding information

National Natural Science Foundation of

China, Grant/Award Numbers: 71871234,

71671193 and 71473279; Program for

Innovation Research in Central University

of Finance and Economics; National

Social Science Foundation of China,

Grant/Award Number: 15ZDC024

Abstract

“Idiosyncratic risk–return puzzle”is conflicting and confusing. It becomes

more complex by the introduction of the common idiosyncratic volatility

(CIV). We shed new light on the issue from the perspective of heterogeneity

of investors with different investment horizons. We study the “CIV puzzle”

with Chinese A‐Share market evidence and further contribute by employing

the wavelet multiresolution analysis framework to decompose the overall

CIV into timescales that refer to risks in different terms. We apply these

timescales in Fama–Macbeth regressions to investigate their pricing effects.

The results suggest that the “CIV puzzle”is an investment horizon specifica-

tion problem. The relationship between returns and common idiosyncratic risk

is negative in the short run, positive in the intermediate run, and then negative

again in the longer run. We also contribute to the international empirical

evidence with an in‐depth analysis of the Chinese stock market over the period

1999–2016. The results are robust across different specifications of the

CIV. Our findings have important implications for portfolio and risk

management.

KEYWORDS

Chinese stock market, idi osyncratic risk–return puzzle, idiosyncratic volatility, investmenthorizon

perspective, wavelet multiresolution analysis

1|INTRODUCTION

The relationship between idiosyncratic volatility (IVOL)

and expected returns has received considerable attention

recently because of the suggestion that IVOL matters

after all (Campbell, Lettau, Malkiel, & Xu, 2001). How-

ever, current empirical evidence on the existence of idio-

syncratic risk premia in expected returns has not only

produced mixed results but also been quite limited to

developed markets.

One strand of the research literature focuses on the

relationship between IVOL and cross‐sectional stock

returns, which are highly contentious and are still

intensely debated in the literature. Some literature, on

the other hand, argue that the relationship between

returns and idiosyncratic risk is not significant (Bali,

Cakici, Yan, & Zhang, 2005; Fink, Fink, & He, 2012;

Jiang & Lee, 2006; Wei & Zhang, 2005). Xu and Malkiel

(2003) argue that investors with poorly diversified portfo-

lios will require extra compensations for holding stocks

exposed to greater idiosyncratic risk. This supports

Merton's (1987) theory in that excess return and IVOL

are positively related when asset diversification is limited.

Other return models incorporating investor under‐diver-

sification obtain similar results (Barberis & Huang,

2008; Boyle, Garlappi, Uppal, & Wang, 2012).

However, Ang, Hodrick, Xing, and Zhang (2006,

2009) demonstrate that idiosyncratic risk is negatively

Received: 22 January 2018 Revised: 10 July 2018 Accepted: 9 September 2018

DOI: 10.1002/ijfe.1668

370 © 2018 John Wiley & Sons, Ltd. Int J Fin Econ. 2019;24:370–390.wileyonlinelibrary.com/journal/ijfe

priced in the cross section of stock returns with evidence

from 23 international countries. It triggers the intense

debate of “IVOL puzzle”: The negative connection

between idiosyncratic risk and equity return seems to be

quite abnormal and goes against Merton's (1987) theory.

Related literature attributes the “IV puzzle”to monthly

stock return reversals (Huang, Liu, Rhee, & Zhang,

2011), the specification of data frequency, the weighting

scheme of calculating the average portfolio return, the

breakpoints to sort portfolios' quintiles (Bali & Cakici,

2008), sample selection (Bali & Cakici, 2008; Fu, 2009),

the estimation method (Fu, 2009), and mispricing

(Zhong, 2017). Recently, Herskovic, Kelly, Lustig, and

Van Nieuwerburgh (2016) construct a common factor

for IVOL (common idiosyncratic volatility [CIV]) and ver-

ify its negative link with the cross section of stock return.

The empirical results confirm the findings of Ang et al.

(2006, 2009) and further generate the “CIV puzzle”: for

one thing, though Bali and Cakici (2008) argue that the

puzzle is insignificant in monthly evidence, the 1‐month

CIV is still negatively priced to a significant degree

(Herskovic et al., 2016); for another, Fu (2009) challenges

Ang et al. (2006) on firm size for stocks with IVOL as

high as 40% accounting for merely 9% of the total market

capitalization, but CIV solves such concern.

However, the sample above is restricted to developed

countries. Evidence in emerging markets on this puzzle

is very limited with only a few exceptions. Nartea, Wu,

and Liu (2013) present evidence that the negative pricing

effect of IVOL is also observable in Chinese markets and

contrasts with that of Nartea, Ward, and Yao (2011) who

discounted the presence of a negative IVOL effect in the

five largest emerging markets of Southeast Asia (Singa-

pore, Malaysia, Thailand, Philippines, and Indonesia).

This gap in the literature is surprising. Emerging markets

are ideal to examine the pricing of idiosyncratic risk,

because investor under‐diversification is quite prevalent

in these relatively small, concentrated, and often illiquid

markets.

In particular, Mainland China's stock market is worthy

examining for its unique features and background. First,

China has been the world's largest emerging and transi-

tional economy with competitive growth rates and enor-

mous investment opportunities. A‐Shares has been

gaining status and included in MSCI in 2018; however,

the pricing of A‐Shares remains largely unknown. Second,

several market characteristics make the study of China's

stock market illuminating. For example, there are vast dif-

ferences between A‐Share market and develop ones in

terms of investor protection, short‐sales prohibition, dom-

inance of individual investors, quality of reporting stan-

dards, intensity of private information acquisition and

dissemination, and ownership structure (Firth, Rui, &

Wu, 2011; Lemmon & Lins, 2003; Meng, Li, Jiang, & Chan,

2017). These differences are likely to have impact on firm

valuation, stock moving patterns, and idiosyncratic risks,

contrasting with those documented in developed markets.

Su, Shu, and Yin (2018) investigate the pricing effect of

CIV in China's A‐Share market and confirm the findings

of Herskovic et al. (2016). The study on A‐Shares amplifies

the “CIV puzzle”with developing market evidence, sug-

gesting that the puzzle is not limited to the U.S. market

but also holds for China.

However, the “CIV puzzle”still remains unsettled.

Herskovic et al. (2016) rationalize the “CIV puzzle”with

a heterogeneous agent model and calibrate that the stocks

with high IVOL raises the average household's marginal

utility thus are lower expected in returns. But there's still

little practical implication. Su et al. (2018) study on China's

A‐Shares while leaving the “CIV puzzle”unexplained.

Actually, in terms of “IV puzzle,”several studies have

attempted to explain it (see, e.g., Angelidis & Tessaromatis,

2009; Guo & Savickas, 2006, 2008; Kang, Lee, & Sim, 2014;

Kearney & Poti, 2008; Lee, 2015; Nath & Brooks, 2015;

Vozlyublennaia, 2013), but none appears totally convinc-

ing when adding China's market unique characteristics

into consideration. For example, because strict price‐limit

rule and short‐selling restrictions are conducted in China's

stock market, the prices of stocks exposed to high‐volatility

risk is unlikely to crash in relatively short horizons. Equiv-

alently, the short‐term reversal of price is unlikely to hap-

pen. Gu, Kang, and Xu (2016) are inspired by such

unique features of China's market and find that IVOL

had a negative effect on price with high limits of arbitrage

while contributing little to solve the puzzle. What's more,

there are notable differences in construction method and

economic implication between IVOL and CIV, so explana-

tions or solutions to the “IV puzzle”are not applicable to

CIV. In general, the “CIV puzzle”still remains confusing,

even with the mentioned China's unique characteristics

considered. Our paper, in this sense, contributes to the lit-

erature by shedding light on a new explanation for the

“CIV puzzle.”

The heterogeneity of investors offers a unique per-

spective in explaining the puzzle. On one hand, the exis-

tence of heterogeneous investors can explain fat tails and

volatility clusters commonly present in most markets

(Balli, Pericoli, & Pierucci, 2016; Gil‐Bazo, Moreno, &

Tapia, 2007; Lévy, Lévy, & Solomon, 2000; McMillan &

Speight, 2006). On the other hand, focusing on the short

investment horizon while ignoring the relatively long

ones is prone to generate biased risk loadings, which

are dependent largely on time interval and systematic risk

(e.g., Levhari & Levy, 1977). Malagon, Moreno, and

Rodríguez (2015) propose a hypothesis that the heteroge-

neity of market players is reflected by heterogeneous

YIN ET AL.371