Chinese Management Studies

Impact of media on stock returns: an in-depth empirical study in China

Yahui Zhang Difang Wan Leiming Fu
Article information:
To cite this document:
Yahui Zhang Difang Wan Leiming Fu , (2013),"Impact of media on stock returns: an in-depth empirical
study in China", Chinese Management Studies, Vol. 7 Iss 4 pp. 586 - 603
Permanent link to this document:
http://dx.doi.org/10.1108/CMS-09-2013-0172
Downloaded by University of Southern Queensland At 22:38 01 April 2016 (PT)

Downloaded on: 01 April 2016, At: 22:38 (PT)

References: this document contains references to 32 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 370 times since 2013*
Users who downloaded this article also downloaded:
(2013),"Corporate accidents, media coverage, and stock market responses: Empirical study of the
Chinese listed firms", Chinese Management Studies, Vol. 7 Iss 4 pp. 617-630 http://dx.doi.org/10.1108/
CMS-09-2013-0171
(2013),"Ultimate ownership, institutionality, and capital structure: Empirical analyses of Chinese data",
Chinese Management Studies, Vol. 7 Iss 4 pp. 557-571 http://dx.doi.org/10.1108/CMS-09-2013-0175
(2011),"Abnormal stock returns, for the event firm and its rivals, following the event firm's
large one-day stock price drop", Managerial Finance, Vol. 37 Iss 2 pp. 151-172 http://
dx.doi.org/10.1108/03074351111103686

Access to this document was granted through an Emerald subscription provided by emerald-srm:405406 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for
Authors service information about how to choose which publication to write for and submission guidelines
are available for all. Please visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as well as
providing an extensive range of online products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee
on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive
preservation.

*Related content and download information correct at time of download.

 The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1750-614X.htm

CMS
7,4 Impact of media on stock returns:
an in-depth empirical study
in China
586
Yahui Zhang
Xi’an Jiaotong University, Xi’an, China and
China Merchants Bank, Shenzhen, China, and
Difang Wan and Leiming Fu
Downloaded by University of Southern Queensland At 22:38 01 April 2016 (PT)

Xi’an Jiaotong University, Xi’an, China

Abstract
Purpose – Media-effect refers to the phenomenon that stocks with no or low media coverage earn
higher returns than stocks with high coverage. This paper aims to explore the existence of media-effect
in China stock market and tests the two competing hypotheses explaining this phenomenon.
Design/methodology/approach – The authors construct a research sample based on a
media-coverage event: the publications of lists of the most wealthy Chinese individuals; in addition,
they identify the stocks of which listed firms are led by a controller who is recognized on the publicized
lists. This paper uses event study methodology to test the existence of media effect in China A-share
market. The authors employed propensity score matching (PSM) to construct a control group with
same number of non-listed stocks. Then compared the returns of the two portfolios to test the risk
premium hypothesis, and the abnormal trading volume and price reaction around the event date is
explored to test the over-attention underperformance hypothesis.
Findings – Sampled stocks show significantly negative abnormal returns within the event period,
but the matched control group formed by PSM shows no significant abnormal return, indicating that
the risk premium hypothesis is not supported. Covered stocks show significantly magnified trading
volume. The portfolio gains significant positive return before the event date but turns significantly
negative afterward, which is consistent with the over-attention underperformance hypothesis.
Originality/value – This paper offers insights into media-effect in China stock market and provides
empirical evidence explaining its existence.
Keywords Limited attention, Media-effect, Over-attention underperformance, PSM, Risk premium
Paper type Research paper

Introduction
Both theorists and practitioners have growing interest in how media affect asset prices.
There are two basic predictions regarding this question. Traditional finance theory
explains that asset price reflects market information; when the media disseminate new
information about changes in firm value, stock prices will rise or fall in accordance.
Many scholars confirm the prediction that stock price is sensitive to the tenor of media
reports (Tetlock, 2007, 2008). Behavioural finance theory provides the second prediction:
as investors have bounded rationality and limited attention, from thousands
of alternatives they tend to buy those stocks that attract their attention (Barber and
Chinese Management Studies Odean, 2008). Such attention causes buy-pressures and a probability that the stock
Vol. 7 No. 4, 2013
pp. 586-603
q Emerald Group Publishing Limited
1750-614X
The authors acknowledge financial support from the Chinese National Science Funds (Grant
DOI 10.1108/CMS-09-2013-0172 No. 70972101).
 price will rise. However, research about media and asset prices shows that, Impact of
disregarding good or bad news, highly covered stocks earn significantly lower media on
returns than do less-covered stocks. Various scholars have recognized and recorded this
phenomenon, called media-effect (Fang and Peress, 2009; Tetlock, 2011), which violates stock returns
both predictions above. First, despite being covered by reports of different tenors, the
media-covered stocks show identical return patterns. Second, highly covered stocks do
not outperform less-covered stocks as predicted. Thus, media-effect is considered an 587
anomaly in the financial markets.
Why does the media-effect occur? To answer this question, researchers have
suggested different hypotheses. Risk premium hypothesis attributes media-effect to the
compensation of uncovered stocks, not to the abnormal return of highly covered ones.
The behavioural hypothesis is based on limited attention, arguing that the media-effect
Downloaded by University of Southern Queensland At 22:38 01 April 2016 (PT)

comes from return reversal after over-attention, the sudden growth of media coverage on
one special event or within a specific time period. Abnormal attention makes investors
overreact to reported stocks, but then reversal follows short-time buy pressures.
Although researchers have thoroughly discussed attention-driven buying behaviour
(Barber and Odean, 2008; Seasholes and Wu, 2007), they have paid insufficient attention
to the reversal process, which is our focus. Using a unique data set of three ranking lists
from the China stock market, we contribute to this research by exploring the
media-effect in China and testing the two competing hypotheses explaining why it
occurs. We find that from the day the rank is released, a negative 1.06 percent market
adjusted abnormal return occurs within the 11-day event window around the event date,
while the abnormal return from a matched non-listed sample was not significant.
Further tests provide mixed evidence regarding the over-attention underperformance
hypothesis but no evidence supporting the risk premium hypothesis.
A key problem in testing the two competing hypotheses is how to distinguish whether
new information content causes the market to react or whether the attention effect causes
price movement. To avoid this problem, we chose listed firms whose controller appeared
on the list of most wealthy. Although Western business has compiled such lists for many
years, they are new to China. Forbes provided China’s first in 1999, Hurun Richest 100, but
it gained little attention until the last decade. Since then, the list has flourished in China’s
business world. Although the Chinese entrepreneurs used to “hide” how wealthy they are,
it is now publicized widely on a list that varies greatly from year to year; the newly richest
join the list and formerly richest drop away, especially after ChiNext opened in 2009 and
became the most important new billionaire-making machine. This great variation makes
the annual release of such lists an important news event that grabs attention throughout
China. Most major portals provide special reports following the yearly announcements,
and these reports are hot topics throughout society. Many entrepreneurs on the lists
become well-known “stars” as well.
This environment allows us to test the information-based and attention-based
hypotheses. For our research, we chose firm controllers who were featured on the three
most popular wealthy lists in China – Hurun Richest 100, Forbes China Rich List, and
New Fortune Richest 500 – for three reasons. First, those lists are the most popular and
influential in mainland China. The media release them to great fanfare that catches
investors’ attention, making it unnecessary to measure investor attention quantitatively.
Second, the financial information used to rank entrepreneurs is publicly available, and
the lists do not convey firm-value information that is new to the market. When investors
 CMS view the entrepreneurs named on the list, their attention is directed to the stock,
7,4 so we are assured that price fluctuations are caused by mere attention rather than
adjustment according to new information, which makes our results “cleaner”. Third,
rich lists are released on specific dates, which make it possible for us to directly explore
the reported stocks’ short-time price reaction to publication, rather than comparing
portfolio returns as related researchers have commonly done (Fang and Peress, 2009).
588 Besides, we must emphasize that the “investor attention” we discuss in this paper is
a potential chain effect from entrepreneurs’ appearance on the rich list to investor
attention on the stock of their listed firms. The list presents assets, company name and
industry along with the entrepreneur’s information. Usually the news reports include
the entrepreneur’s name with the company name (such as Nine Dragons Paper
Zhang Yin, Huang Guangyu from Gome), and frequently the online news includes a
Downloaded by University of Southern Queensland At 22:38 01 April 2016 (PT)

hyperlink to the real-time stock performance. Thus, we believe this chain effect is possible
and valid. After ChiNext was opened, the broadened channel of going public tightened
the connection between the stock of the listed firm and the entrepreneur on the rich list.
The remainder of the paper is organized as follows. We review related research and
develop our hypotheses in the related literature and hypotheses development section.
We discuss the sample selection procedures and the methodology used to test the
hypotheses in the research methodology section. We present and discuss empirical
results in analysis of results, and provide implications and limitations of the study in the
discussion.

Related literature and hypotheses development

1. Literature review on how media affects asset prices
Media affects the security market by attracting investor attention (Shiller, 2000).
Recently investor attention theory has become popular for demonstrating this
mechanism. The basic logic of investor attention theory is that information impacts the
stock market through investor behavior; thus stock prices show the effects only when
investors notice the information (Barber and Odean, 2008). Investor attention is the
premise and objective condition of market reaction. However, limitations of time and
professional skills make investor attention a scarce resource (Kahneman, 1973).
Thus, modern society’s well-developed news media work as an attention-allocation
mechanism, with the hidden proposition that media hype can gain investors’ maximum
attention. Thereby media may play an important role in the stock market (Huberman
and Regev, 2001). Tetlock (2011) found that investors overreact to even previously
reported stale information, which directly confirms the possibility that media guide
investor attention. So, considering that attention is limited, is attracting investor
attention always a good thing? Research shows that this is not necessarily the case.
Ritter (1991) found that IPO stocks with common attention usually have high
short-time returns followed by long-run underperformance. Fang and Peress (2009)
found that highly covered stocks have lower returns the next month.
Although many scholars have proved that media affect asset prices and that broad
dissemination of information has a greater impact than the quantity or quality of
press-generated information (Bushee, 2010; Engelberg and Parsons, 2011), Fang and
Peress (2009) were the first to identify the media-effect, arguing that media reports can
predict stock returns. Their sample included all companies listed on the NYSE and
500 randomly selected companies listed on the NASDAQ between January 1, 1993
 and December 31, 2002. They collected the number of newspaper articles about Impact of
a stock from the four most influential daily newspapers with nationwide circulation to media on
proxy for the stock’s overall media exposure. By constructing a zero-investment
portfolio longing non-covered stocks and shorting highly covered ones, they found that stock returns
stocks with no media coverage earned higher returns than stocks with high media
coverage. The results were significant even after they controlled for well-known risk
factors. They further confirmed that media-effect is different from some 589
well-documented time series patterns of stock returns, such as short-run momentum
and long-run reversal, and is a unique anomaly in the stock market.

2. Competing hypotheses explaining media-effect and hypotheses development

Previous researchers have suggested three explanations for the media-effect, from
Downloaded by University of Southern Queensland At 22:38 01 April 2016 (PT)

liquidity, risk premium, and investor recognition.

2.1 Limits to arbitrage hypothesis. Understanding limits to arbitrage is important in
explaining stock market anomalies (Hong and Stein, 2007). Limits to arbitrage include
short-sell constraints, liquidity constraints, and the risks from uncertainty that
irrational investors bring. Given investor rationality, if the no-media premium reflects
mispricing, then profit-motivated traders will exploit and thereby eliminate the
mispricing. Media-effects will persist only if market frictions are severe enough to
prevent arbitrageurs from exploiting it. Fang and Peress (2009) provided support for this
hypothesis. They found that media-effect is strong among small stocks and stocks with
high bid-ask spreads. Similarly, Vega (2006) found that transaction cost (or limits to
arbitrage) explains why post-earnings announcement drift (PEAD) persists, although
the reason may be that announcements with unexpectedly high (low) earnings make
investing in these firms more (less) risky, and the drift may be explained as a risk
premium. The weakness of this hypothesis is that it may help explain the persistence of
media-effect, but it fails to explain its existence. Thus, this hypothesis is not tested in this
research.
2.2 Risk premium hypothesis. Risk premium hypothesis focuses on the two kinds of
information risk in the stock market. Merton (1987) posited the investor recognition
hypothesis, that information incompleteness causes investors to be unaware of all
securities in the market. Easley et al. (2002) pointed out that information is
asymmetrically distributed among informed and uninformed investors. Based on these
two information allocation characters in the stock market, news media broaden investor
recognition by disseminating information to a broader audience, while stocks with no
media coverage or lower coverage have limited recognition and thus have higher risk.
From this view, risk premium hypothesis posits that stocks with lower investor
recognition must offer higher returns to compensate their holders for the fact that
incomplete or asymmetric information causes them to be imperfectly diversified.
Fang and Peress (2009) proved that risk premium explains the media-effect,
and impediments-to-trade only explains its persistence. They used analyst coverage and
fraction of individual ownership to indicate low information dissemination and
idiosyncratic volatility and idiosyncratic volatility per investor to indicate high
idiosyncratic risk. Their results provided support for the risk premium hypothesis.
Media-effect is more pronounced among stocks with high individual ownership and low
analyst following. This indicates that traditional information transmission
channels (such as analyst coverage) cover such stocks poorly and that media enhance
 CMS information dissemination. Media-effect grows stronger as actual idiosyncratic risk
7,4 increases. Among one-third of most risky stocks, the return of the zero-investment
portfolio is almost 1 percent, which is economically significant. Based on the above
analysis and findings, we propose the first hypothesis:
H1. Stocks not reported by the media earn a significant higher return than media
covered stocks after the news.
590
2.3 Over-attention underperformance hypothesis. This hypothesis, based on investors’
overreaction to stocks that grab their attention, has roots in cognitive psychology.
Kahneman (1973) pointed out that attention is a scarce cognitive resource. Shiller (1999)
also put attention on a critical position when summarizing market inefficiency based
on behavioural factors. He explained that investors’ decision on investment type is
Downloaded by University of Southern Queensland At 22:38 01 April 2016 (PT)

based on constant alternation between attention and inattention, and even the crash of
financial markets is an attention effect. Because investors’ attention changes with time,
uncoordinated large attention suddenly pouring into the market causes major crashes.
Limitation of attention and media-reported stocks save investors search costs; they
tend to buy only the stocks that grab their attention.
Barber and Odean (2008) tested attention’s impact on different types of investors and
found that attention-grabbing stocks experience significant buy-pressure from
individual investors who lack the professional skills to process information and thus
rely on media-provided public information. Seasholes and Wu (2007) used a similar
strategy and found that when stocks hit the upper limit in China’s stock market,
investors turned attention to those stocks. They were then likely to buy in, which caused
them losses.
Rao et al. (2010) further developed this hypothesis and argued that highly covered
stocks show returns that indicate overreaction. Then the returns negatively reverse,
similar to the long-run underperformance of IPO stocks described by Ritter (1991), only
more rapidly. Seasholes and Wu (2007) documented a significant reversal after stocks
hit the upper limit, which is probably the reason why attention-driven buyers suffer
losses. Barber and Odean (2008) referred to this reversal as “disappointing subsequent
returns”, for the buy-pressure cast by attention-driven buying is not supported by
upward change of stock value. Although this hypothesis seems feasible theoretically,
we must further test stock returns because of direct reaction to media news. Based on
the above analysis and researches, we come to the following hypothesis:
H2. After the news, media-covered stocks earn a significant lower return than
stocks not covered by the media.
The difference between the H1 and the H2 is that, although they both agree there is a
media-effect that media-covered stocks underperform not-covered stocks subsequently,
the reason is different. The H1 states that media-effect comes from the higher return of
not-covered stocks, i.e. the risk premium; while the second states that it comes from the
lower return of media-covered ones, i.e. the reversal after overreaction. Specifically,
to describe the over-attention underperformance more clearly, we further developed
H2a and H2b:
H2a. The trading volume of media-covered stock is significantly enlarged after the
news is released.
 H2b. Attention-driven buy pressure causes an upward price reaction of Impact of
media-covered stock and will soon reverse. media on
H2a and H2b describe the over-attention underperformance from the price reaction stock returns
and trading volume, respectively.

Data and methodology 591

1. Data
The core of our data is a hand-collected list of listed firms whose top managers were on
Hurun Richest 100, Forbes China Rich List, or New Fortune Richest 500 between 2003
and 2010. Hurun Richest 100 began in 1999. Before 2003, it named only the 50 richest
people. To keep the sample selection consistent, we included only the lists after 2003
Downloaded by University of Southern Queensland At 22:38 01 April 2016 (PT)

for all three. For the same reason, we included only the 100 richest people listed by New
Fortune Richest 500. We acquired the lists from the Annual Summary of Financial
Topics of Sina (web site: http://finance.sina.com.cn/zt/index.html). Sina’s financial
sector is one of the biggest and best financial portals in mainland China, and during
our sample period it collected the lists in our sample and provided in-depth yearly
reports on the event. We did not acquire the lists directly from the production agencies
because New Fortune magazine composes the New Fortune Richest 500. They do not
provide the rich list for past years as the other two institutions do. We thoroughly
compared the lists of Hurun Richest 100 and Forbes China Rich List in our sample and
the lists derived from the web site of Hurun Report and Forbes China, respectively, and
found no difference. The stock price data and firms’ financial data are from
Genius Financial Database, and we used STATA 10.0 software to analyze the data.
Because our sample included only firms that were listed on Shanghai or Shenzhen
Stock Exchanges, and because we were expecting investors to transfer their attention
from the entrepreneurs on the list to the stock of their listed firms, we had to put the
entrepreneurs on the lists into a one-to-one correspondence with listed firms to assure a
strong-enough connection between the entrepreneur and the listed firm. Building this
correspondence presented several difficulties. First, not all entrepreneurs on the list were
from listed firms or firms listed in mainland China (some were from firms listed in
Hong Kong, Singapore, or US stock exchanges). Second, for those from listed firms in
mainland China, some were from the same firm and some controlled more than one
listed firm.
We used strict screening criteria to ensure that entrepreneurs on the list had control
power over the sample firms. First, we excluded unlisted firms, firms listed in Hong
Kong, and abroad China. Compared with stock markets abroad or in Hong Kong, China’s
stock market has different requirements in information disclosure, policy environment,
and market freedom. The compositions of investors as well as the maturity of investors
are also different, so we excluded stocks listed outside mainland China. Second, for those
from the same firm appearing on the same list in the same year, we counted only the one
with the highest title. This occurred because we did not weigh the sample. That is, we did
not use how many times managers from the same listed firm were on the list as an
indicator, as Malmendier and Tate (2009) did to proxy for the competitiveness of CEO
award winners, thus we avoid counting one firm repeatedly. Third, for a group of listed
firms controlled by the same entrepreneur on the list, we included only the earliest listed
firm[1]. After eliminating the stocks with missing company data, we had 89 A-share
listed firms in the sample that had controllers on the three rich lists, for 487 times.
 CMS All sample firms were private enterprises coming from nine of 13 first-level blocks
according to the China Security Regulatory Commission Industry Classification.
7,4 Figure 1 shows the industry distribution of sample firms by year.

2. Calculation of CARs
We employed standard event study methodology to examine the abnormal returns of
592 sample stocks around the media report date. The date on which the list was publicly
released served as the event date. For event window, we considered the 11 trading days
surrounding the award release date (days [2 5, þ 5]). We used market-adjusted-return
model to calculate the normal return, for the parameter estimation in other methods is
unstable and the high volatility of stock prices in China may affect test efficiency.
Considering the co-integration of Shanghai and Shenzhen markets, we used the
Downloaded by University of Southern Queensland At 22:38 01 April 2016 (PT)

arithmetic mean of the return of Shanghai Index for A-Shares (000002) and return of
Shenzhen Index for A-Shares (399107) as market return, following prior work.
Let Ri,t designate the observed arithmetic return for security i on day t. Define ARi,t
as the excess return for security i on day t. For every security, the excess return for
each day in the event period is calculated using the following procedures:
Ri;t ¼ lnðP i;t Þ 2 lnðP i;t21 Þ ð1Þ
Ri;t ¼ Rm;t 2 1i;t ð2Þ
ARi;t ¼ 1i;t ¼ Rm;t 2 Ri;t ð3Þ

100

80

60

40

20

0
2003 2004 2005 2006 2007 2008 2009 2010

A C D E G H J L M
Notes: This figure shows the sample distribution by year and by industry;
the CSRC industry classification is as follows: A – agriculture, C –
manufacturing, D – electric power, gas and water production and supply;
E – construction, G – information technology, H – wholesale and retrial
Figure 1. trade, J – real estate, L – communication and cultural industry, M –
Sample firms by year
comprehensive
 where Pi,t2 1 is the closing price of security i on day t 2 1 and Pi,t is the closing date of Impact of
that security on day t, and Rm,t is the market return on day t. The average abnormal
return of N stocks on day t is calculated as:
media on
stock returns
1X N
ARt ¼ 1i;t ð4Þ
N t¼1
593
And the cumulative abnormal return for stock i from day t1 to day t2 is:
X
t2
CARi ðt1 ; t 2 Þ ¼ ARi;t ð5Þ
t¼t 1
Downloaded by University of Southern Queensland At 22:38 01 April 2016 (PT)

We removed the stocks that missed more than two consecutive trading days within the
event period. For the very small percentage of stocks missing return data for one day in
the event period, we averaged the succeeding day’s return over the missing day and the
succeeding day, as Peterson (1989) recommended.

3. Calculation of abnormal trading volume

Following Barber and Odean (2008), we used volume change in the event period
relative to the base period to capture abnormal trading behavior. Using the daily
average of trading volume one month (20 trading days) prior to the event period as
benchmark, we calculated the abnormal volume for stock i on day t as:
X
t21
VOLi;d
VOLi ¼ ð6Þ
d¼t220
20

VOLi;t 2 VOLi;t
AVOLi;t ¼ ð7Þ
VOLi;t
where VOLi,d is the trading volume of stock i on day d prior to the event period
(d is between t 2 20 and t 2 1); VOLi is the average daily trading volume of stick i the
month prior to the event period; VOLi,t is the trading volume of stock i on day t; and
AVOLi,t is the abnormal volume of stock i on day t.

4. Propensity score matching method

To test the risk premium hypothesis, we used propensity-score matching (PSM)
method (Abadie and Imbens, 2007) to construct a control group with the same number
of non-listed stocks. Matching algorithm is a sampling method used mostly in control
group construction. Sampling the control group with the most similar background to
the treated group reduces the disturbance of confounding factors and yields better test
efficiency (Rosenbaum and Rubin, 1985). Stock index option research commonly uses
stocks matching, where matching stocks can control other factors’ impact on
cross-sectional volatility of stock price (Bae, 2004). An example of such research is
Harris (1989), who picks 500 matching non-index stocks for S&P 500 index stocks with
beta, price, firm size, and trading frequency. Research regarding media and asset prices
usually chooses similar stocks based on book-to-market ratio and firm size (Chan, 2003).
Following Malmendier and Tate (2009), we used asset, market capitalization, ROA,
 CMS financial leverage, sales growth, Tobin’s Q and past month return to conduct
7,4 propensity score matching, and we also matched year and industry. These seven
indicators reflect firm size, profitability, capital structure, growth, risk, and price level,
respectively.

Analysis of the results

594 1. Media report and stock return
Table I presents how media reports affected stock return within the event window.
Overall, reported stocks had negative abnormal return in the 11-day period, which is
not consistent with Fang and Peress’s (2009) risk premium hypothesis. Although two
days have positive abnormal returns before the event day as Barber and Odean (2008)
predicted, the positive abnormal return is not significant. The overall trend of stock
Downloaded by University of Southern Queensland At 22:38 01 April 2016 (PT)

return is basically consistent with H2, proposing media-effect comes from the lower
return of media-covered stocks.
Concretely speaking, the abnormal return on event day is 2 0.05 percent with
t-value equal to 2 0.05 percent, which is not significant. As for event window,
consistent with prior research, reported stocks have a significant 2 0.97 percent
(t-value is 2 3.27) abnormal return one week after the event date ([0, þ 6]), while
abnormal return one week before the event date ([2 5, 0]) is not significant. On one
hand, event study results show that media reports directly and negatively affect
reported stocks and that the media-effect exists in the China stock market. On the other
hand, such results broaden existing knowledge on media-effect by showing that the
lower return of reported stocks at least partially caused the media-effect.

Day AR (%) t-statistic , 0 samples $0 samples

25 2 0.064 2 0.517 270 217

24 2 0.040 2 0.335 251 236
23 2 0.263 2 2.272 * * 272 215
22 0.141 1.253 245 242
21 0.134 1.114 240 247
0 2 0.051 2 0.417 282 205
1 0.009 0.073 262 225
2 2 0.350 2 2.955 * * * 266 221
3 2 0.224 2 1.727 * 261 226
4 2 0.298 2 2.523 * * 279 208
5 2 0.052 2 0.480 259 228
[21, 0] 0.083 0.475 234 253
[0, 1] 2 0.041 2 0.224 252 235
[25, 0] 2 0.143 2 0.470 256 231
[0, 5] 2 0.966 2 3.270 * * * 278 209
[21, 1] 0.093 0.425 244 243
[22, 2] 2 0.117 2 0.414 249 238
[23, 3] 2 0.604 2 1.732 * 257 230
[24, 4] 2 0.941 2 2.469 * * 261 226
Table I. [25, 5] 2 1.058 2 2.577 * * 267 220
Abnormal and
cumulative abnormal Notes: Abnormal return is significantly different with zero under *10, * *5 and * * *1 percent levels of
return within event significance; this table shows the event-study result using release day as event date; we used t-test to
window examine whether the abnormal return of sample firms is significantly different with zero
 The abnormal return two days before the event day is 0.14 and 0.13 percent (t-values are Impact of
1.253 and 1.114, respectively), which is positive but not significant. According to Barber media on
and Odean (2008), stocks that grab investors’ attention face attention-driven
buy-pressure, and media report is one of the three measures of investor attention. So, stock returns
if the over-attention underperformance hypothesis holds, significant buy-pressure will
be detected within the event window, which may result in higher return. The event
study results show that positive abnormal return exists, but it is not significant. 595
Notice also that although positive abnormal return is not significant, it is detected two
days before the event day. We attribute this to the market’s advance reaction to the
information. Prior research has shown that information leak is common in the China
stock market (Zhang and Zhu, 2003); thus abnormal returns show anticipated
movements before the event date (Haw, 2006). The event in this research is
Downloaded by University of Southern Queensland At 22:38 01 April 2016 (PT)

press-released public information rather than formal information disclosure. Because

the information lacks economic value, it may be inappropriate to call it information leak.
However, corresponding magazines publish all three rich lists and the releases are
arresting news events, so many individuals desire to know about it ahead of time, and
the market may learn about it in advance. We will further test whether this abnormal
return is caused by attention-drive buying, as theory predicts.

2. Risk premium hypothesis

Using media coverage, Fang and Peress (2009) divided the stocks into three groups and
constructed a zero-investment portfolio longing non-covered stocks and shorting
highly covered stocks. They found significant positive returns coming from the
premium of non-covered stocks and non-significant returns from highly covered
stocks, thus supporting the risk premium hypothesis. However, we used event-study to
test media-effect, which made it impossible for us to follow this strategy. For similar
comparison, we took the sample firms as media-covered group (treated group), and
constructed non-covered group (control group) from all other stocks unexposed to the
media-coverage event. Then we tested the returns of both portfolios. Although the rich
list ranks only assets of individual entrepreneurs, firm character and stock price surely
affect firm size and controllers’ assets, especially for listed firms.
Table II reports the logit regression result of the seven indicators on whether the
firm is on the rich list. The dependent variable is getting-on-the-list dummy; if the
firm’s top manager has been on the rich list it equals 1, otherwise it equals 0. Results
show that the seven indicators we chose to describe the financial and return character
of the firm significantly impacted whether the controller of the firm got on the rich list.
This indicates that sampling matched control group based on these seven indicators
can effectively control other factors’ impact on the attention effect.
To test the return of both portfolios one week after the event date, we examined two
different models on factors known to affect the cross-section of returns: CAPM model
and the Fama and French (1993) three-factor model. Table III contains the results of
both regressions. Results show that the media-covered portfolio exhibits significant
negative return one week after the event date, which cannot be absorbed by traditional
risk factors, while the return of the not-covered portfolio is not significant. The market
adjusted alpha of media-covered portfolio is 2 0.759 with a t-value of 2 3.02, and the
Fama-French three-factor adjusted alpha is 2 0.834 with a t-value of 2 3.34. This
shows that media-effect is directly caused by the negative return of media-covered
 CMS
Matching variables Coefficient Z-statistic
7,4
Assets 2 0.001 2 5.19 * * *
Market capitalization 1.172 10.16 * * *
ROA 0.062 7.57 * * *
Financial leverage 0.863 2.21 * *
596 Sales growth 0.001 4.53 * * *
Tobin’s Q 2 0.235 2 2.51 * *
Past month return 0.007 1.64 *
Intercept 2 31.099 2 11.96 * * *
Year dummy controlled
Industry dummy controlled
Adj. R 2 0.132
Downloaded by University of Southern Queensland At 22:38 01 April 2016 (PT)

Table II.
Deciding factors of Notes: Significant at: *10, * *5 and * * *1 percent levels; the dependent variable is getting-on-the-list
getting on the rich list dummy; if the firm is within the on-the-list sample it equals to 1; otherwise it equals to 0

Media-covered portfolio Not-covered portfolio

CAPM FF three-factor CAPM FF three-factor

Intercept 20.759 * * * 20.834 * * * 2 0.518 20.517

(23.02) (23.36) (2 1.39) (21.43)
MP 0.626 * * * 0.639 * * * 0.568 * * * 0.597 * * *
(10.83) (11.41) (6.62) (7.33)
SMB 0.298 * 0.528 * *
(1.80) (2.19)
HML 0.187 0.002
(1.26) (0.01)
Table III. Observations 24 24 24 24
Regression of portfolio Adj. R 2 0.835 0.849 0.651 0.693
returns one week after the
event period Notes: Statistically significant at: *10, * *5 and * * *1 percent levels; t-value is in parentheses

stocks, as H2 proposed, rather than the risk premium of not-covered ones, and that the
risk premium hypothesis is not supported, as H1 proposed.
We emphasize that Fang and Peress (2009) used monthly data to sort the stocks by
media coverage and to test the next-month returns, while we defined a specific event that
caught media attention and used daily data to conduct similar tests. To test whether
media-effect holds in a longer time-span, besides testing the returns one week after the
event date ([1, 5]), we also tested the returns one week after the event period ([6, 10]) and
one month after that ([11, 30]). We detected no abnormal return within the latter two
periods, which suggests that the media-effect we discovered lasted for only one week
after the event day and diminished after that (these results are available from the author
by request). Still, it is reasonable that former research captured media-effect in a longer
time-span than ours, because they covered more reports and more types of content.

3. Over-attention underperformance hypothesis

The over-attention underperformance hypothesis explains that the media-effect is
caused by the reversal following attention and overreaction toward media reports.
 Testing this hypothesis includes two aspects: first whether media reports catch investors’ Impact of
attention; and second whether reversal occurs afterward. To test the attention-effect, media on
Barber and Odean (2008) and Seasholes and Wu (2007) both targeted attention-driven
buying. They examined the buy-sell imbalance to investigate how attention-grabbing stock returns
effects impact investor behavior. Because we lacked trading data for individual stocks,
following Schmitz (2007), we used abnormal trading volume to proxy for investor
attention. To test the reversal, we followed the calendar-time portfolio strategy Seasholes 597
and Wu (2007) employed to test price pressure brought by demand shocks. H2a and H2b
test the attention effect and the return reversal, respectively.
Table IV reports the abnormal trading volume within the event period. Using the
average daily trading volume 20 days before the event period as baseline, we found that
trading volume was enlarged before the event date, and it continued to increase
Downloaded by University of Southern Queensland At 22:38 01 April 2016 (PT)

significantly after the event date. The largest daily trading volume after the event date was
130 percent of the baseline with a t-vale of 4.013, while the smallest was still 110 percent of
the baseline with a t-value of 2.098. The significant change in trading volume was
consistent with the over-attention underperformance hypothesis as H2a states, which
showed that media report on the rich list release grabbed investors’ attention and
influenced their trading behavior. Table V shows the return of the calendar-time portfolio.
Results show that, overall, trading before the event date was profitable. Buying sample
stocks five days before the event date and selling it one day before had a significant 0.91
percent profit with t-value of 2.794, while buying the day before the event date and selling
it the next day had no significant profit. After the event date, buying the sample stocks on
the event day and selling it the next suffered a significant 0.29 percent loss (t-value is
21.955) while buying it one day after the event day and selling it four days after suffered a
significant 0.72 percent loss with t-value of 22.480. The trend of the profit change was in
accordance with the predictions of the over-attention underperformance hypothesis, with
the return of sample stocks changing from positive to negative as H2b proposes.
Combining the trading volume performance in Table IV and the return performance
in Table V, we can see that the volume increase appeared mainly after the event date,

Date AVOL t-statistic , 0 samples $0 samples

25 0.060 1.390 278 182

24 0.053 1.384 287 173
23 0.093 1.896 * 269 191
22 0.044 1.075 281 179
21 0.183 4.060 * * * 245 215
0 0.109 2.773 * * * 267 193
1 0.106 2.444 * * 263 197
2 0.097 2.098 * * 281 179
3 0.301 4.013 * * * 241 219
4 0.262 4.381 * * * 268 192
5 0.290 4.167 * * * 279 181
Notes: Statistically significant at: *10, * *5 and * * *1 percent levels; we used trading volume data Table IV.
20 days earlier to calculate the baseline of abnormal volume; some samples had no trading data for two Abnormal trading
or more consecutive days, so they were eliminated in this part, yielding 20 samples fewer than the volume within the event
whole sample period
 CMS
Holding period Daily price Return over holding
7,4 From To in days reaction (%) period (%) t-statistics

Buy on Sell on 4 0.229 0.913 2.794 * * *

t25 t21
Buy on Sell on t 1 0.181 0.181 1.325
598 t21
Buy on t Sell on 1 20.290 20.290 2 1.955 *
tþ1
Buy on Sell on 4 20.182 20.724 2 2.480 * *
tþ1 tþ5
Table V.
Price reaction around Notes: Statistically significant at: *10, * *5 and * * *1 percent levels; we assumed transactions took
Downloaded by University of Southern Queensland At 22:38 01 April 2016 (PT)

the event date place at closing prices

while trading stocks in this period suffered a loss. Thus, the attention-driven buying and
the price pressure it brought were not significant here, as predicted by the over-attention
underperformance hypothesis. This means that, although H2a and H2b hold,
respectively, combined they are not telling the exact story as the over-attention
underperformance hypothesis proposes; instead, the high volume dropping process was
highlighted. The market reaction within the whole event period reflects that investors
overreacted to the information that caught their attention. The abnormal return of
covered-stocks diminished after temporarily dropping, showing that the stock price was
back to its fundamentals. Assuming that the market pre-reacts, investors’ attention on
the sample stocks within the whole event period should go from low to high. We
attributed this to the competition between disagreement and attention-effect, as Quan
and Wu (2010) suggested. They found that investor attention and trading volume had a
U-shaped relationship: at low-attention level, disagreement made abnormal trading
volume less obvious; at high-attention level, attention-effect beat disagreement, and the
trading volume of the stock rose significantly. We assume that some “smart-traders,” as
Seasholes and Wu (2007) described, bought stocks before the attention-grabbing event
or just when it happened, and sold them when the word was out. Overall, our results
support the over-attention underperformance hypothesis but find attention-driven
buying to be less significant.

4. Robustness check
From 1999 to 2009, about 19 entrepreneurs from the Hurun Richest 100 were arrested,
investigated, or sentenced. They have been called “questionable billionaires.” Although
the Hurun Research Institute declared that only 1.4 percent of entrepreneurs on their list
were “questionable,” a rate no greater than that of developed countries, still the
significant abnormal return we found may be from anticipation that the entrepreneurs
and their companies would soon be more severely supervised. To guarantee that our
results were not driven by samples with “questionable” controllers, we excluded the
abnormal returns from such firms and retested them. We identified the questionable
entrepreneurs mainly from the 1999 Hurun Special Report. We searched by hand for
entrepreneurs’ names and stock names on Baidu, China’s biggest search engine. Table VI
shows the results of our robustness check on the abnormal returns, which do not differ
from the results presented in Table I.
 Impact of
Day AR (%) t-statistic ,0 samples $0 samples
media on
25 2 0.036 2 0.283 264 215 stock returns
24 2 0.020 2 0.165 245 234
23 2 0.258 2 2.224 * * 268 211
22 0.166 1.467 240 239
21 0.146 1.197 235 244 599
0 2 0.020 2 0.166 277 202
1 2 0.006 2 0.043 259 220
2 2 0.366 2 3.067 * * * 263 216
3 2 0.231 2 1.753 * 257 222
4 2 0.295 2 2.475 * * 274 205
5 2 0.059 2 0.537 255 224
Downloaded by University of Southern Queensland At 22:38 01 April 2016 (PT)

[21, 0] 0.125 0.714 229 250

[0, 1] 2 0.026 2 0.139 246 233
[25, 0] 2 0.022 2 0.073 251 228
[0, 5] 2 0.976 2 3.262 * * * 273 206
[21, 1] 0.121 0.547 238 241
[22, 2] 2 0.081 2 0.284 244 235
[23, 3] 2 0.568 2 1.625 252 227
[24, 4] 2 0.884 2 2.313 * * 255 224
[25, 5] 2 0.978 2 2.367 * * 260 219
Notes: Abnormal return is significantly different with zero under *10, * *5 and * * *1 percent levels of
significance; this table shows the event-study result using release day as event date; we used t-test to
examine whether the abnormal return of sample firms is significantly different with zero; we excluded Table VI.
the samples with questionable controllers Robustness check

Unsurprisingly, the results are the same as the previous ones because the controller was
“questionable” in only two stocks that had been on the lists seven times, accounting for
1.4 percent of the sample. Beside statistical evidence, potential endogeneity is not
problematic for two reasons. Theoretically, it is quite possible that because the
entrepreneurs on the rich lists are “famous,” their business scandals are more
eye-catching. From the socio-environmental perspective, the high rate of governmental
actions against firm owners might make it appear that this endogeneity is inherent in
China’s current socioeconomic environment, but the low ratio of “questionable” samples
shows that it is not necessarily the case.

Conclusions
Theories based on limited attention take investor attention as a scarce resource, and
suggest that investor attention itself can dramatically change stock prices (Shiller, 1999).
Attention-allocating media reports attract investors to buy publicized stocks. However,
empirical research has found the media-effect that reported stocks have lower returns
than non-reported stocks. Our objective in this study is to explain the media-effect. For
our research sample, we select firms whose controllers appeared on the three most
influential lists of wealthy individuals in China – Hurun Richest 100, Forbes China Rich
List, and New Fortune Richest 500. We find that sample stocks have significant negative
cumulative abnormal returns within the event period. Using the propensity
score-matching method, we construct a control group that is the same size as the
reported sample and test the returns of the two portfolios. One week after the event date,
 CMS the reported portfolio shows significant negative returns that remain significant after
7,4 they are adjusted by market return and Fama-French three-factors. The returns for the
matched portfolio are not significant. This indicates that the low-attention portfolio does
not cause the media-effect. Rather it comes directly from the low return of the
high-attention portfolio. The trading volume reaction and test on return of calendar-time
portfolio show that the trading volume of sample stocks is magnified significantly
600 within the event period, and the significant positive stock return before the event date
turns to significant negative return afterward, which is basically consistent with the
over-attention underperformance hypothesis. The time when trading volume is
magnified does not accord with the time of positive return, showing that
attention-driven buying is not significant as limited-attention theory predicts.
Researchers who have studied how investor attention affects stock price, trading
Downloaded by University of Southern Queensland At 22:38 01 April 2016 (PT)

volume, and investors’ trading behavior have paid too little attention to price movement
after attention-driven buying. As Barber and Odean (2008) commented, attention-based
purchases by many investors could temporarily inflate a stock’s price, but it may lead
to disappointing subsequent returns. For stocks that rise merely because of attention
and not because of changes in fundamental value, it is intuitive that fundamental
prices will return after a short “bubble” period. Another explanation lies in investors’
bounded rationality. Rational investors are likely to postpone taxes by selling their past
losers, while behaviorally motivated investors are likely to sell past winners, thereby
postponing regrets associated with losses, aka, the “disposition effect” (Shefrin and
Statman, 1985). No matter whether it is the price return or disposition effect, for investors
holding “attention-grabbing” stocks, selling them in time is the most rational decision.
In addition to its theoretical contribution, our sample also provides value in that the
fortune lists document and witness the evolution of China’s private entrepreneurs as a
special and important part of the market economy, as well as the development of the
market economy in China. They impact the social culture by shaping people’s
behavior, not only economically but also through broad news reports. Through the
wide report of wealthy entrepreneurs, it encourages the entrepreneurship and
innovation in the whole society. Moreover, the old Chinese wisdom “happiness lies in
contentment” is gradually turning into aggressive pursuit of personal success,
especially for the young people. This change of people’s thinking is natural while the
market economy deepens, but is greatly accelerated and promoted by the media. Thus,
our sample has value beyond testing the media-effect following publications of rich
lists, but also helps us understand the change in China’s cultural background.
Our findings also have some implications for practitioners. First, they confirm that
media disseminate information to the market even though it does not contain “genuine”
news. The attention effect caused by media report is big enough to affect trading behavior
in the market. It further means that instead of seeking analyst coverage, companies’ media
relations activities may have a similar effect and may reduce the capital cost of the
company. Second, for individual investors, the attention-grabbing stocks may not be as
valuable an investment as it seems. The suddenly pouring attention may cause temporary
upward price reaction, but is very likely to have disappointing subsequent earnings. Not
to mention that when individual investors get the “news”, it is not at all new to the market
and all that is left to them is pure attention effect.
This research provides some understanding of the media-effect and the economic
consequences of attention-driven buying. As with all research, this paper has some
 deficiencies that may serve as a springboard for further research. Although the sample Impact of
is valuable, the data have potential country-specific properties that may affect media on
generalization. Also, the sample firms in this research are private enterprises, and thus
fail to represent the overall market. Our results are sufficient to confirm that the stock returns
Chinese media are powerful enough to impact the private sector, but caution must be
taken in generalizing findings to the whole market. Questions about how media
coverage affects asset pricing are growing, so further work will enrich our knowledge 601
about this interesting and anomalous capital market phenomenon.

Note
1. Along with the rapid development of China’s stock market, “corporate class” appears in
Downloaded by University of Southern Queensland At 22:38 01 April 2016 (PT)

recent years. Usually several listed firms are related, and the frequent capital operation and
complicated cross-shareholding relationship makes it quite difficult to identify each firm
controlled by the same entrepreneur. However, most entrepreneurs develop their corporate
class or capital class from an earlier-listed firm. They use the one firm to enter the capital
market and then reach to other listed firms to complete capital operation, so the firm is
usually directly and the strongly connected with the controller, as is the entrepreneur on the
rich list in our research. Researchers on corporate class or capital class in China can see Shao
and Liu (2007) and Ma and Wang (2009), both in Chinese.

References
Abadie, A. and Imbens, G.W. (2007), “Bias corrected matching estimators for average treatment
effects”, working paper, Harvard University, Cambridge, MA.
Bae, S.C. (2004), “Futures trading, spot market volatility, and market efficiency: the case of
the Korean index futures markets”, Journal of Futures Markets, Vol. 24 No. 12,
pp. 1195-1228.
Barber, B.M. and Odean, T. (2008), “All that glitters: the effect of attention and news on the
buying behavior of individual and institutional investors”, Review of Financial Studies,
Vol. 21 No. 2, pp. 785-818.
Bushee, B.J. (2010), “The role of the business press as an information intermediary”, Journal of
Accounting Research, Vol. 48 No. 1, pp. 1-19.
Chan, W.S. (2003), “Stock price reaction to news and no-news: drift and reversal after headlines”,
Journal of Financial Economics, Vol. 70 No. 2, pp. 223-260.
Easley, D., Hvidkjaer, S. and O’Hara, M. (2002), “Is information risk a determinant of asset
returns?”, Journal of Finance, Vol. 57 No. 5, pp. 2185-2221.
Engelberg, J.E. and Parsons, C.A. (2011), “The causal impact of media in financial markets”,
Journal of Finance, Vol. 66 No. 1, pp. 67-97.
Fama, E.F. and French, K.R. (1993), “Common risk factors in the returns on stocks and bonds”,
Journal of Financial Economics, Vol. 33 No. 1, pp. 3-56.
Fang, L. and Peress, J. (2009), “Media coverage and the cross-section of stock returns”, Journal of
Finance, Vol. 64 No. 5, pp. 2023-2052.
Harris, L. (1989), “S&P 500 cash stock price volatilities”, Journal of Finance, Vol. 44 No. 5,
pp. 1155-1175.
Haw, I.M. (2006), “Securities regulation, the timing of annual report release, and market
implications: evidence from China”, Journal of International Financial Management &
Accounting, Vol. 17 No. 2, pp. 110-139.
 CMS Hong, H. and Stein, J.C. (2007), “Disagreement and the stock market”, Journal of Economic
Perspectives, Vol. 21 No. 2, pp. 109-128.
7,4
Huberman, G. and Regev, T. (2001), “Contagious speculation and a cure for cancer:
a non-event that made stock prices soar”, Journal of Finance, Vol. 56 No. 1, pp. 387-396.
Kahneman, D. (1973), Attention and Effort, Prentice-Hall, Upper Saddle River, NJ.
Malmendier, U. and Tate, G. (2009), “Superstar CEOs”, Quarterly Journal of Economics, Vol. 124
602 No. 4, pp. 1593-1638.
Merton, R.C. (1987), “A simple model of capital market equilibrium with incomplete information”,
Journal of Finance, Vol. 42 No. 3, pp. 483-510.
Peterson, P.P. (1989), “Event studies: a review of issues and methodology”, Quarterly Journal of
Business and Economics, Vol. 28 No. 3, pp. 36-66.
Downloaded by University of Southern Queensland At 22:38 01 April 2016 (PT)

Quan, X. and Wu, S. (2010), “Investor attention, earnings announcement effect and
announcement timing of management”, Journal of Financial Research, No. 11, pp. 90-107
(in Chinese).
Rao, Y., Cheng, D. and Peng, D. (2010), “Does media attention cause abnormal return? Evidence
from China’s stock market”, Systems Engineering – Theory & Practice, Vol. 30 No. 2,
pp. 287-297.
Ritter, J.R. (1991), “The long-run performance of initial public offerings”, Journal of Finance,
Vol. 46 No. 1, pp. 3-27.
Rosenbaum, P.R. and Rubin, D.B. (1985), “Constructing a control group using multivariate
matched sampling methods that incorporate the propensity score”, American Statistician,
Vol. 39 No. 1, pp. 33-38.
Schmitz, P. (2007), “Market and individual investors reactions to corporate news in the media”,
working paper, available at: http://ssrn.com/abstract¼1004488
Seasholes, M.S. and Wu, G. (2007), “Predictable behavior, profits, and attention”, Journal of
Empirical Finance, Vol. 14 No. 5, pp. 590-610.
Shefrin, H. and Statman, M. (1985), “The disposition to sell winners too early and ride losers too
long: theory and evidence”, Journal of Finance, Vol. 40 No. 3, pp. 777-790.
Shiller, R.J. (1999), “Human behavior and the efficiency of the financial system”, Handbook of
Macroeconomics, North-Holland, Amsterdam, pp. 1305-1340.
Shiller, R.J. (2000), Irrational Exuberance, Princeton University Press, Princeton, NJ.
Tetlock, P.C. (2007), “Giving content to investor sentiment: the role of media in the stock market”,
Journal of Finance, Vol. 62 No. 3, pp. 1139-1168.
Tetlock, P.C. (2008), “More than words: quantifying language to measure firms’ fundamentals”,
Journal of Finance, Vol. 63 No. 3, pp. 1437-1467.
Tetlock, P.C. (2011), “All the news that’s fit to reprint: do investors react to stale information?”,
Review of Financial Studies, Vol. 24 No. 5, pp. 1481-1512.
Vega, C. (2006), “Stock price reaction to public and private information”, Journal of Financial
Economics, Vol. 82 No. 1, pp. 103-133.
Zhang, X. and Zhu, H. (2003), “The economic analysis of insider trading”, China Economic
Quarterly, No. 4, pp. 71-96 (in Chinese).

Further reading
Chen, C.W. (2009), “Press coverage and stock prices’ deviation from fundamental value”,
working paper, available at: http://ssrn.com/abstract¼1359261
 About the authors Impact of
Yahui Zhang graduates from the School of Management, Xi’an Jiaotong University. She is a
postdoctoral Research Fellow in China Merchants Bank. Her research interests focus on media on
corporate finance and behavioural finance. Yahui Zhang is the corresponding author and can be stock returns
contacted at: [email protected]
Difang Wan is a Professor of organization management at the School of Management,
Xi’an Jiaotong University. He received his PhD degree in management through Xi’an Jiaotong
University/University of Alberta Joint PhD Program under the Canada-China Management 603
Education Project. His research interests include corporate governance, merger and acquisition,
organization innovation, and experimental research.
Leiming Fu is a PhD candidate at the School of Management at Xi’an Jiaotong University
China. His research interests are corporate governance, corporate finance and innovation.
Downloaded by University of Southern Queensland At 22:38 01 April 2016 (PT)

To purchase reprints of this article please e-mail: [email protected]

Or visit our web site for further details: www.emeraldinsight.com/reprints