The impact on performance of IPO allocation reform

Description
The purpose of this paper is to investigate the changes in initial public offering (IPO)
underpricing and short-run performance following a regulatory reform (No. 54 [2002] China Securities
Regulatory Commission (CSRC)) of the method of allocating IPO shares in China.

Journal of Financial Economic Policy
The impact on performance of IPO allocation reform: An event study of Shanghai Stock
Exchange A-shares
Fei J iang Lawrence A. Leger
Article information:
To cite this document:
Fei J iang Lawrence A. Leger, (2010),"The impact on performance of IPO allocation reform", J ournal of
Financial Economic Policy, Vol. 2 Iss 3 pp. 251 - 272
Permanent link to this document:http://dx.doi.org/10.1108/17576381011085458
Downloaded on: 24 January 2016, At: 21:38 (PT)
References: this document contains references to 35 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 625 times since 2010*
Users who downloaded this article also downloaded:
Dilip K. Das, (2010),"Financial globalization: a macroeconomic angle", J ournal of Financial Economic
Policy, Vol. 2 Iss 4 pp. 307-325http://dx.doi.org/10.1108/17576381011100847
David Tripe, (2010),"Using DEA to investigate bank safety and soundness – which
approach works best?", J ournal of Financial Economic Policy, Vol. 2 Iss 3 pp. 237-250 http://
dx.doi.org/10.1108/17576381011085449
Irina Barakova, Ajay Palvia, (2010),"Limits to relative performance evaluation: evidence from
bank executive turnover", J ournal of Financial Economic Policy, Vol. 2 Iss 3 pp. 214-236 http://
dx.doi.org/10.1108/17576381011085430
Access to this document was granted through an Emerald subscription provided by emerald-srm:115632 []
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.
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
The impact on performance
of IPO allocation reform
An event study of Shanghai Stock Exchange
A-shares
Fei Jiang and Lawrence A. Leger
Department of Economics, Loughborough University, Loughborough, UK
Abstract
Purpose – The purpose of this paper is to investigate the changes in initial public offering (IPO)
underpricing and short-run performance following a regulatory reform (No. 54 [2002] China Securities
Regulatory Commission (CSRC)) of the method of allocating IPO shares in China.
Design/methodology/approach – On 20 May 2002, the CSRC announced that IPO subscription
and allotment would be based on the market value of investors’ tradable shareholdings. Before the
regulatory change, this was determined by the amount of funds used for subscription. The reform was
intended to increase participation by both smaller and institutional investors. Based on a sample of 209
IPOs in the Shanghai A-share market during the period 2001-2003, the paper employs an event study
methodology to examine the impact of this IPO regulatory reform.
Findings – The paper ?nds that the overall (pre- and post-reform) average abnormal initial return of
116.94 per cent is lower than in earlier studies of Chinese IPOs but higher than in other markets.
Post-reform underpricing decreases by 42.27 per cent compared to pre-reform levels. In the post-listing
aftermarket a pre-reform upward trend of cumulative abnormal returns was reversed to become
downward post-reform. The results suggest that the regulatory change has encouraged well-informed
investors, consistent with Information Cascades and Bandwagon hypotheses. It also appears that the
reform improved market ef?ciency and secondary market liquidity.
Originality/value – The ?ndings shed light on the relationship between IPO costs, IPO pricing,
market liquidity and market microstructure. They also have important implications for issuers,
underwriters and in particular for policy markers.
Keywords Stocks and shares, Regulation, Financial performance, Stock markets, Stock exchanges,
China
Paper type Research paper
1. Introduction
Initial public offering (IPO) refers to the ?rst sale of stocks by an unlisted company to the
public. Stock exchange listing (followed by public trading in open market) allows the
creation of market prices and liquidity. Information asymmetry and agency problems in
the market make the valuation of IPOs more dif?cult than that of listed common stocks
so an essential part of the IPO process is the discovery of an appropriate issue price that
must compensate for both direct costs (such as underwriting and information disclosure
fees) and indirect costs (such as unknown risks speci?c to the offering, as distinct from
systemic risks generally involved in pricing listed common stocks). The complex and
special nature of IPO pricing is re?ected in “IPO underpricing” in which statistically
signi?cant positive abnormal returns are widely observed in the ?rst day of trading.
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1757-6385.htm
JEL classi?cation – G14, G28
IPO allocation
reform
251
Journal of Financial Economic Policy
Vol. 2 No. 3, 2010
pp. 251-272
qEmerald Group Publishing Limited
1757-6385
DOI 10.1108/17576381011085458
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
In this paper, we examine the impact on underpricing of changes in IPO regulations in
the Chinese stock market.
The IPO underpricing anomaly is well documented in the literature and has been
observed almost in every stock market in the world, although the level of abnormal
initial return varies considerably across countries and over time. The degree of
underpricing has also been found to depend on particular market circumstances
(different trading mechanisms, market liquidity, regulation status, etc.). For instance,
Loughran et al. (1994) revealed abnormal initial returns that varied from a low of only
4.2 per cent in France (1983-1992) to a high of 80.3 per cent in Malaysia (1980-1991)
while Ritter (1998) found a higher average level of abnormal initial returns in
13 emerging markets than in 20 developed markets (40.8 and 19.6 per cent, respectively).
In a study of particular interest in the context of this paper, for East Asian markets
Loughran et al. (1994) found a lower average level of underpricing in the 1990s than in
the 1980s that apparently resulted from a reduction in regulatory interference.
With regard to Chinese IPOs in particular, empirical studies have without exception
found extraordinarily high abnormal initial returns, although it has been shown that
the degree of underpricing is sensitive to the period sampled. For example, Su and
Fleisher (1999) documented an abnormal initial return as high as 948.59 per cent for
1987 to 1995; Mok and Hui (1998) found lower (but still substantial) underpricing of
289 per cent for 1990 to 1993 while Chanet al. (2001) reportedunderpricing of 178 per cent
for 1993 to 1998. Gu (2003) offered a year on year comparison of underpricing for 1984
through 2000, ?nding a pattern consistent with that documented above.
We argue that the changing degree of IPO underpricing over time may re?ect the
general (and continuing) regulatory and microstructural development of Chinese
markets. The 1987-1995 sample period of Su and Fleisher (1999) covers the very early
years when the stock exchanges in China were not even established (1987-1990), which
may in part explain the extremely high underpricing. On 1 July 1994, Company Law
was enforced in China and, for the ?rst time, speci?ed formally in law how shares
should be offered and transferred, possibly explaining why underpricing over
1993-1998 (Chan et al., 2001) was lower than for 1990-1993 (Mok and Hui, 1998). These
authors variously argue that IPO size (Mok and Hui, 1998; Su and Fleisher, 1999),
percentage of equity held by the state and legal entities (Mok and Hui, 1998; Chan et al.,
2001) and time lag between offering and listing (Mok and Hui, 1998; Su and Fleisher,
1999; Chan et al., 2001; Gu, 2003) are potentially important determinants
of cross-sectional variation in IPO underpricing and may all be good proxies for the
ex ante uncertainty of an IPO. In the presence of information asymmetry between
issuers, underwriters and investors, and between informed and uninformed investors,
IPOs may be underpriced to compensate for the risks incurred by less well informed
parties (Rock, 1986; Baron, 1982) although this alone may not explain the magnitude of
underpricing in China.
Launched in the 1980s, established in 1990 and developing during a period of
transition from centrally planned economy to market economy, the Chinese stock
market is distinguished for its “Chinese characteristics” (Section 3). The very idea of
private investment in capital markets was new to the vast majority of Chinese
nationals when stock markets ?rst opened in the early 1990s and followed an extensive
period of cultural revolution and extraordinary, unprecedented, social and economic
change. We argue that this revolutionary economic transition has led to a substantial
JFEP
2,3
252
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
period of learning by a rapidly growing number of new and relatively unsophisticated
stock market investors, accompanied by signi?cant uncertainty and severe asymmetry
of information in the process of privatisation in the immature Chinese market,
particularly with respect to IPO. Mok and Hui (1998, p. 457), sourcing an example from
the South China Morning Post, write:
One clear example of information asymmetry in China’s IPO market is demonstrated in
August 1992, when over half a million people poured into Shenzhen from all over China and
queued for days to scramble for the application forms for a lottery to buy shares from 14 new
issues. Few in the queues even knew which companies they would be investing in, let alone
having any knowledge about the earning history of the companies.
We argue that such “herding behaviour” by Chinese investors may be a particularly
important factor in underpricing, so it is of signi?cant interest to discover whether
regulatory reform and development of market microstructure can induce better
dissemination of information and hence a reduction in information asymmetry. Such
asymmetry implies the existence of noise traders (De Long et al., 1991) and raises
the possibility of “informational cascades” (Welch, 1992) and “bandwagon effects”
(Ritter, 1998) in which poorly informed investors may copy the behaviour of other
investors thought to be better informed. In consequence an IPO ?rm has an incentive to
reduce their offer price in order to avoid IPO failure. Improved information ?ows in the
market should lead to a reduction in the proportion of noise traders and hence a
reduction in underpricing.
The regulatory and microstructural development of Chinese markets makes for
interesting natural experiments on the impact of policy reform, particularly with
respect to IPO regulation. In this paper, we exploit a particular natural experiment
created by the IPO allocation reform of May 2002 which saw a change in the lottery
allocation mechanism.
We investigate the impact of the reform by comparing underpricing and cumulative
abnormal returns (CAR) before and after the change, using event study methodology
for data from 2001 to 2003. Any change in IPO policy (and the consequent change in
market microstructure) is likely to alter IPO costs and be re?ected in IPO performance.
This research therefore sheds light on the relationships between IPO costs, IPO pricing,
market liquidity and market microstructure, with implications for issuers,
underwriters and policy markers.
The remainder of this paper is structured as follows. The theoretical background of
the study is discussed in Section 2 while the “Chinese characteristics” of the Chinese
stock market and their implications are explained in Section 3 The methodology is
presented in Section 4 and empirical results are discussed in Section 5. Section 6
provides concluding remarks.
2. Chinese IPO regulatory reform and theories of IPO performance
From 1996 to May 2002, the odds of being allocated IPO shares were determined by the
amount of money in the subscription. For large investors (generally institutional or
other better informed investors) with a large proportion of funds already invested in
the secondary market, the costs of subscribing to IPOs therefore would very likely
have included an opportunity cost of cashing in stocks from the secondary market
(a trade-off between winning the lottery against giving up future gains on existing
shareholdings). This is necessarily a speculation, as we have no data to support it and
IPO allocation
reform
253
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
see no likelihood of such data becoming available, but we argue that large investors
would have been unlikely to enter an IPO unless convinced that such opportunity cost
would be covered. After May 2002, the odds and amount of IPO share allocation were
determined by subscribers’ existing holdings of tradable shares. This greatly reduced
the costs of subscribing to IPOs for large investors, since they were no longer
required to risk their existing shareholdings in order to increase the chance of winning
a “good” IPO. On the other hand, speculators who before the reform could have used
all their funds to chase IPOs in the primary market were “punished” after the reform.
We should therefore ?nd an impact on market demand for IPOs of changing the
incentives for different types of investors. In addition, the new regulation may have
also encouraged continuing investment in the secondary market and thus had a
positive in?uence on its liquidity. There has been some debate over this regulation
change with respect to its impact on investor demand for IPOs. In particular, it has
been argued that the reform neither bene?ted smaller investors nor motivated
institutional investors and hence failed to promote the stability of the market
(Hong, 2006).
Changes in investor demand for IPOs should in the ?rst instance affect IPO pricing,
with consequent effects on IPO performance. In moving from the primary to the
secondary market risky IPO stocks undergo a price discovery process and it is well
established that this process produces anomalies in IPO performance on the initial day
of open trading. Stoll and Curley (1970), Reilly (1973), Logue (1973) and Ibbotson (1975)
have all documented signi?cant and systematic increases from offering to ?rst-day
closing prices in the US market. This underpricing phenomenon appears to violate the
ef?cient markets hypothesis, since investors are apparently willing to pay higher
prices in the secondary market than in the primary market very shortly beforehand,
while issuers are apparently willing to “leave money on the table”.
Since the 1970s, numerous empirical studies have provided international evidence
on the underpricing anomaly in almost all of the world’s ?nancial markets. Theories of
IPO underpricing are generally divided into four groups: information based models,
institutional explanations, ownership and control models and behavioural
explanations ( Jenkinson and Ljungqvist, 2001; Ritter and Welch, 2002; Ljungqvist,
2007) and no single theory completely resolves the anomaly – different theories ?nd
supporting evidence in different markets and different periods, and under different
microstructural and regulatory conditions.
Among the various underpricing theories, Welch (1992) and Ritter (1998) have,
respectively, proposed “informational cascades” and “bandwagon effects” in stock
markets. In this approach, less well informed investors make decisions by judging the
interest of other investors. They subscribe only to IPOs they believe to be popular and
refuse offerings that they believe other investors do not want (even when other
favourable information may be available). Welch (1992) demonstrated that demand
curves can be very elastic in the presence of informational cascades, leading to the
rapid failure of offerings that are priced too high, even when issuers have favourable
private information. To avoid failure therefore, issuers underprice their IPOs to attract
better-informed investors and to induce positive cascades or positive bandwagon
effects. Any change in IPO regulation could alter the elasticity of market demand for
IPOs by changing the population of subscribers. In addition, if the market for IPOs is
JFEP
2,3
254
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
subject to competitive pressure then, ceteris paribus, we should expect any simple
increase in the number of IPO subscribers to lead to a reduction in underpricing.
It is arguable that Chinese culture and social behaviour since the revolution have been
more strongly in?uenced by the perceived desirability of “collective action” than is the
case elsewhere, and we may conjecture that Chinese markets are therefore particularly
vulnerable to informational cascades and bandwagon effects. If we accept this argument,
it follows that if the IPO regulatory reform of May 2002 gave better-informed investors
increased opportunities to subscribe to IPOs at the expense of less well-informed investors
then this should have led to a decrease in the elasticity of IPOdemand, reducing the risk to
issuers of adverse bandwagon effects and IPO failure, inducing higher IPO prices and a
decrease in underpricing. Of course, the same outcome should emerge if the effect of the
reform was either to increase the number of subscribers or the total amount subscribed
while holding informational asymmetry constant – increased competition between
subscribers and/or higher total IPO demand should lead to a higher IPO price (a decrease
in underpricing). Unfortunately, we have no data on the types or numbers of IPO
subscribers, so these possibilities cannot be disentangled.
An alternative view arises from the argument given earlier (Hong, 2006), that the
reformmay not onlyhave beenof no bene?t to smaller investors but mayhave also failed
to motivate institutional investors. If Hong is correct, we should expect to see a net
decrease in IPO demand, leading to a net reduction in the IPO price (an increase in
underpricing), other things being equal.
Changes in investor demand for IPOs may also have an impact on market
microstructure through changes in the number and types of traders in the post-issue
market, consequently re?ected in after-market performance. The theory that links
market microstructure, IPO pricing and after-market performance is not well developed
but trader behaviour models of market microstructure may inform our predictions
of post-IPO price change. For example, De Long et al. (1991) showed that noise traders
(less well-informed investors) can survive in the market and even earn high returns at
the expense of informed traders if they can exert pressure on prices collectively. Small
numbers of noise traders have little power to move prices away fromfundamental value
but an increase in the number of noise traders may eventually distort prices. Following
De Long et al. (1991), any impact of the 2002 IPO reform on the type of traders should be
re?ected in the price discovery process in the secondary market for the IPO stocks
(post-IPOperformance). Unfortunately, as before, we have no way of estimating changes
in the classi?cation of traders by type. However, if the reformled to a relative increase in
general market participation by noise traders we might expect, ceteris paribus, a greater
degree of both price volatility and prolonged abnormal returns in the after-market,
while the reverse would be true if the reform encouraged greater participation by
informed traders.
3. “Chinese characteristics” of stock markets and data sampling
3.1 Chinese characteristics
In addition to the powerful in?uences on investor behaviour arising from China’s
unique recent social and cultural history, there are three other uniquely “Chinese
characteristics” of Chinese stock markets.
(i) High equity retention by the state, government control and restricted IPO supply.
The Chinese Government has a large equityholdinginstate-ownedenterprises, consisting
IPO allocation
reform
255
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
of shares not tradable on the stock market but constituting the major part of all
outstanding stock. A reduction in government ownership would promote stock market
growth but this has still to be implemented. Until this problem is resolved the supply of
IPOs to the market may be seriously limited. For example, although investment banks
were introduced to the IPO approval process to certify IPO quality after the 2001 policy
change, the aggregate IPO supply remains largely in the control of government.
Furthermore, because the Chinese stock market is still relatively undeveloped and has
limited investment instruments, IPO supply is barely able to meet investment demand.
Basu and Li (2000) have argued that bureaucrats possess inside information about
companies most likely to succeed, so that underpricing is used to compensate outsiders
and to signal a trustworthy future.
(ii) Segmented A-share and B-share markets. There are two types of tradable shares in
the Chinese stock market: A-shares and B-shares. A-shares are RMB denominated
(RenMinBi, the domestic currency) ordinary shares and sold to Chinese citizens and
qualifying foreign institutions; B-shares are US dollar denominated and sold to foreign
investors. The A-share and B-share markets are segmented. According to Chen (1997),
the two types of share carry equal voting rights and obligations for any given company
but the offeringandtradingprices of A-shares are about twice as highas those of B-shares.
Poon et al. (1998) attributed this to a lack of investment opportunities for domestic
investors. A-shares are also more underpriced than B-shares. Chen et al. (2000) linked
A-share underpricing to ?rm risk and high government shareholdings and B-share
underpricing to seasoned equity offerings and government ownership. There also seemto
be informational differences: Mok and Hui (1998) argued that B-share investors are better
informedthanA-share investors, due todifferent disclosure requirements andunderwriter
reputation; Su and Fleisher (1999) asserted that B-share investors are largely professional
investors who rely more on the intrinsic value of IPO companies, whereas a large
proportion of A-share investors are inexperienced investors with little market knowledge.
Finally, the time lag between offering and listing is longer for A-shares than for B-shares.
Chen et al. (2000) report an average time lag of 10.71 months for A-shares against
1.46 months for B-shares.
(iii) Allocation methods and changes in government regulation. The share allocation
method has been changed several times, but broadly in two phases. Before 1996,
the allocation of IPOshares to investors was made by application through a computerised
system. After 1996, the allocation was carried out through a lottery mechanism. There
was an important regulatory reform implemented in May 2002, during the second phase,
when the mechanism was changed in favour of a lottery allocation based on the market
value of investors’ tradable shareholdings. After May 2002, investors were able to
subscribe to new issues only if they already owned tradable shares with the amount of
their subscription being determined by the quantity of their tradable shareholdings – the
more the existing shareholding, the higher the probability of winning the IPO “lottery”.
In contrast, before the policy reform of May 2002, the odds of winning the “lottery”
depended on the money spent on the subscription. There was considerable debate about
whether this regulation change would (or did) help to promote the stability of the market.
3.2 The sample
The Chinese IPO market is characterised by immaturity and frequent regulatory reform.
To avoid any confounding effects from other regulation changes and taking into
JFEP
2,3
256
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
consideration various other factors (discussed below), the sample period was chosen as
2001 to 2003:
.
In mid-1999, the ?xed pricing method was modi?ed. Book-building became
predominant after 2001. It is well documented that book-building reduces the
degree of underpricing relative to a ?xed pricing method (Benveniste and
Busaba, 1997; Biais and Faugeron-Crouzet, 2002; Sherman and Titman, 2002).
In a sample period starting before 2001 it would be dif?cult to disentangle the
impact of the change in IPO pricing method and the impact of the change in IPO
allocation method.
.
After October 2000, the Shenzhen Stock Exchange (SZSE) was closed to IPOs,
leaving the Shanghai Stock Exchange (SSE) as the only IPO market from 2001 to
2003. Starting the sample period after October 2000 avoids any effects arising
from the suspension of IPO activities in the SZSE.
.
Since 2001, the Offering Censorate has taken on the role of the China Securities
Regulatory Commission in regulating IPO supply. Recommendations on IPOs
have since been made by investment banks, whereas they were previously made
by local regulatory authorities. Thus, IPO activities have become less controlled
by government while investment bankers have taken on a more important role.
As reputation is so important for the continuous operation of investment banks,
they have an incentive to spend resources on accurate determination of offer
prices. This certi?cation function of investment bankers in the IPO process may
have helped in setting fairer offer prices in comparison to the period before 2001
(Booth and Smith, 1986; Beatty, 1989; Gale and Stiglitz, 1989; Carter and
Manaster, 1990). Choosing a sample start date in 2001 avoids any contamination
of the data by the changing role of investment banks.
.
From 2001, the Chinese stock market turned from bull to bear, maintaining a
downward trend during 2001-2003, as shown in Figures 1 and 2 (SSE and SZSE,
respectively). In a bull market, optimistic investor sentiment may result in
overvaluation and thus observed higher underpricing; and vice versa. Choosing a
sample period with consistent market conditions can avoid noise caused by
changing market trends, thereby highlighting the impact of the targeted allocation
reform.
These considerations suggest that the period 2001-2003 is relatively stable in terms of
both policy environment (only one regulation change was made, which is the focus of
this research) and market performance (a consistent trend).
The data include all A-share IPOs ?oated on SSE during 2001-2003 – a total of
209 IPOs. B-share IPOs were excluded both because the Aand Bmarkets are segmented
and because A-shares can be thought to better evince “Chinese characteristics”.
Offer prices for each of the 209 IPOs, were collected from: www.cnlist.com while
closing prices for 220 post-IPO trading days were collected from Datastream (Thomson
Financial).
The 209 IPOs were divided into two groups, by allocation method. Group I
(88 companies) consists of pre-reform IPOs ?oated from January 2001 to May 2002
(allocation odds determined by subscription amount). Group II (121 companies) consists of
post-reform IPOs ?oated from June 2002 to December 2003 (allocation odds determined
IPO allocation
reform
257
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
by the market value of tradable shareholdings). The main characteristics of the sample are
summarised in Table I.
As shown in Table I, the average IPO size in the post-reform period appears to be
smaller thaninthe pre-reformperiod, althoughthe difference is not statisticallysigni?cant
Figure 2.
Shenzhen Stock Exchange
index, 1996 to 2004
Market trendline of SZSE composite index
January 1996 - January 2004 (monthly data)
700
600
500
400
300
200
100
0
I
n
d
e
x
1996 1997 1998 1999 2000
Year
2001 2002 2003 2004
Source: www.cnlist.com
Figure 1.
Shanghai Stock Exchange
index, 1996 to 2004
Market trendline of SSE composite index
January 1996 - January 2004 (monthly data)
2,000
1,500
I
n
d
e
x
1,000
500
0
2,500
1996 1997 1998 1999 2000
Year
2001 2002 2003 2004
Source: www.cnlist.com
JFEP
2,3
258
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
(t-statistic 1.4468). A “thin trading problem” can be associated with a large proportion of
zeroreturns, causingbias inestimationof betaandreducingthe power of testingabnormal
returns (Berglund et al., 1989; Cowan and Sergeant, 1996). The proportion of zero daily
stockreturns is verysmall for bothperiods (onaverage less than4 per cent), indicatingthat
the sample does not suffer from thin trading problems.
4. Methodology
4.1 Event and estimation windows
Standard event study methodology was used to test for the impact of the regulatory
change on IPO performance. In an event study, the event window is conventionally
de?ned as the period over which new information arrives at the market and changes
the expected value of the stocks in question. For an IPO event, uncertainties exist in the
price discovery process when the stock is moved from the primary market to the
secondary market and new information is conveyed to the market via after-market
trading. In an ef?cient market this is expected to be done in a very short period, say a
day or a week. However, with asymmetry of information (a likely characteristic of the
Chinese market) dissemination of new information may take longer for IPO stocks.
The choice of event window is therefore problematic, implying the need for sensitivity
analysis.
Event and estimation windows are interrelated. In this study, because there are no
observable trading prices before listing, a post-event window was used as an estimation
The whole
sample Group I (pre-reform) Group II (post-reform)
Sample period January 2001-
December 2003
January 2001-May 2002 June 2002-December 2003
IPO allocation
method
Lottery
mechanism
Allocation odds determined
by subscription amount
Allocation odds determined by
tradable shareholdings
Number of IPOs 209 88 121
Money raised (million RMB)
Total amount 137,567 70,179 67,388
Mean 658 797 557
SD 1,185 1,461 929
Min. 100 121 100
Max. 13,130 13,130 6,325
Total number of
daily stock returns 45,980 19,360 26,620
Average daily stock
return 0.0044 0.0055 0.0037
Proportion of zero daily stock returns
Mean (%) 3.50 3.44 3.54
SD 0.0166 0.0161 0.0170
Min. (%) 0.45 0.91 0.45
Max. (%) 12.73 7.73 12.73
Notes: Money raised for each IPO is calculated as the issue price times the total number of IPO shares;
daily stock returns for each of the 209 IPOs include 220 raw daily stock returns during the 220 trading
days after listing
Table I.
Sample descriptive
statistics
IPO allocation
reform
259
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
window for computing alpha and beta parameters. If the event/post-event windows are
not properly de?ned the estimation of beta risks may be affected, with implications for the
calculation of abnormal returns. If the event window used is too short then the post-event
window will inevitably include some event days and actual stock returns over these days
maydeviate fromtheir expectedvalue (abnormal performance instockreturns). Beta risks
estimated froma combination of abnormal and normal performance in stock returns over
the inaccurate post-event windowmay therefore be biased. On the other hand, if the event
window used is too long the post-event window is still post-event, imposing fewer
problems for beta estimations.
To ensure that the estimation window was truly post-event, we tested ?ve
event-estimation pairs, using 100 trading days for the estimation window in each pair.
The start of the estimation window clearly depends on the speci?ed event window.
We tested 20, 40, 60, 90 and 120 trading days (from the listing day) for the event
window. The estimation windowfor the 20-day event windowwas therefore fromthe 21st
until the 120th trading day, the estimation windowfor the 40-day event windowwas from
the 41st until the 140th trading day, and so on. If the betas estimated from the ?ve
estimation windows were not different from each other, the true event window could be
either shorter than 20 days (implying that post-event betas are estimated) or longer than
220 days (implying that event betas are estimated – although a 220-day event window
does not seem sensible for an IPO). In such a case different event-estimation windows
would need to be tested. On the other hand, differences in the beta estimations would
suggest that the ending of the event windowfell between the 20th and 220th trading days.
In this case at least one of the tested event-estimation pairs should be appropriate for
calculating abnormal returns.
4.2 Daily abnormal returns
Daily abnormal returns over the event window were calculated for the sampled IPOs
using the Market Model (1) to compute the expected returns. This allows abnormal
returns to be estimated relative to the beta risk of each stock:
R
it
¼ a
i
þb
i
R
mt
þ1
it
ð1Þ
The actual event window daily return, R
it
, is de?ned as R
it
¼ P
it
=P
i;ðt21Þ
21, where P
it
is the closing price for IPO stock i (i ¼ 1; . . . ; 209) on day t.
Expected daily return (i.e. normal return) over the event window is de?ned as:
EðR
it
X
t
j Þ ¼ a
i
þb
i
R
mt
ð2Þ
Here, X
t
is conditioning information supplied by the market model on day t, a
i
is a
constant term for IPO stock i (measuring the part of returns that is independent of the
market index), b
i
is the systemic risk for stock i and R
mt
is the market return on day t,
calculated from the Shanghai Stock Exchange Composite Index. a
i
and b
i
are
estimated by the market model for the post-event window.
The daily abnormal return over the event window is de?ned as:
AR
it
¼ R
it
2EðR
it
X
t j Þ ð3Þ
4.3 Aggregation of daily abnormal returns
Abnormal performance was aggregated for each IPO stock i fromday 1 until day T. Both
CAR and buy-and-hold abnormal returns (BHAR) were calculated. The CAR method
JFEP
2,3
260
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
assumes that a portfolio is re-balanced in each period (every day in this case) whereas
the BHAR method gives the abnormal return from the initial day until the target day
(no rebalancing at any point). Lyonet al. (1999) have argued that the CARapproachis to be
preferred if the aim is to measure whether or not the sample persistently earns abnormal
returns over time. On the other hand, BHARcan precisely measure investor experience. In
addition, Gompers and Lerner (2003) have suggested that the choice between CAR and
BHAR largely depends on the trading strategy. Since both methods have their own
advantages, we calculated both. We then calculated the group average of these abnormal
returns (CAR and BHAR) from day 1 until day T, using both equal- and value-weighting
methods.
The CAR for IPO stock i over period (1, T) is de?ned as:
CAR
i
ð1; TÞ ¼ AR
i1
þ · · · þ AR
iT
¼
X
T
t¼1
AR
it
ð4Þ
The equally weighted cross-sectional average CARs for the two groups of the sample
are:
CAR
I
ð1; TÞ ¼
1
88
X
88
i¼1
CAR
i
ð1; TÞ ðGroup I; pre-reformÞ ð5Þ
CAR
II
ð1; TÞ ¼
1
121
X
209
i¼89
CAR
i
ð1; TÞ ðGroup II; post-reformÞ ð6Þ
The value-weighted cross-sectional average CARs for the two groups of the sample are:
CAR
I
ð1; TÞ ¼
X
88
i¼1
w
i
ðTÞCAR
i
ð1; TÞ ðGroup I; pre-reformÞ ð7Þ
CAR
II
ð1; TÞ ¼
X
209
i¼89
w
i
ðTÞCAR
i
ð1; TÞ ðGroup II; post-reformÞ ð8Þ
Here, w
i
ðTÞ is the weight (by IPO market value) of each stock in the relevant group on
day T:
w
i
ðTÞ ¼
P
i
ðTÞn
i
P
88
i¼1
P
i
ðTÞn
i
ð9Þ
P
i
(T) is the closing price for IPO i on day T and n
i
is the number of IPO shares offered.
The BHAR for IPO stock i over the period (1, T) is:
BHAR
i
ð1; TÞ ¼
Y
T
t¼1
ð1 þ R
it
Þ 2
Y
T
t¼1
ð1 þ EðR
it
jX
t
ÞÞ ð10Þ
The equally weighted cross-sectional average BHARs for the two groups of sample are:
BHAR
I
ð1; TÞ ¼
1
88
X
88
i¼1
BHAR
i
ð1; TÞ ðGroup I; pre-reformÞ ð11Þ
IPO allocation
reform
261
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
BHAR
II
ð1; TÞ ¼
1
121
X
209
i¼89
BHAR
i
ð1; TÞ ðGroup II; post-reformÞ ð12Þ
The value-weighted cross-sectional average BHARs for the two groups of sample are:
BHAR
I
ð1; TÞ ¼
X
88
i¼1
w
i
ðTÞBHAR
i
ð1; TÞ ðGroup I; pre-reformÞ ð13Þ
BHAR
II
ð1; TÞ ¼
X
209
i¼89
w
i
ðTÞBHAR
i
ð1; TÞ ðGroup II; post-reformÞ ð14Þ
Here, w
i
(T) is calculated as:
w
i
ðTÞ ¼
P
i
ðTÞn
i
P
121
i¼1
P
i
ðTÞn
i
ð15Þ
As before, P
i
(T) is the closing price for IPO i on day T and n
i
is the number of IPO
shares offered.
4.4 Hypotheses
The null hypotheses are that the IPO reform had no impact on either IPO underpricing
or post-listing performance. From our earlier discussion, we propose alternative
underpricing hypotheses:
.
success in stimulating IPO participation, particularly by better-informed
investors, should lead to a reduction in underpricing; and
.
reduction in IPO participation should lead to an increase in underpricing.
After-market performance is necessarily related to the degree of underpricing since a
reduction (increase) in underpricing should be followed by a reduction (increase) in
abnormal performance. However, the time paths of after-market performance are not as
predictable. Alternative after-market hypotheses are:
.
a relative increase in market participation by better-informed investors should
induce quicker and smoother adjustment to normal pricing; and
.
an increase in participation by noise traders should lead to prolonged and more
volatile abnormal performance.
Both parametric and non-parametric tests were performed (two-sample
mean-comparison t-test and two-sample Wilcoxon signed rank test):
(1) Testing for CAR:
.
H0. CAR
I
ð1; TÞ is not signi?cantly different from CAR
II
ð1; TÞ.
.
H1. CAR
I
ð1; TÞ is signi?cantly different from CAR
II
ð1; TÞ.
(2) Testing for BHAR:
.
H0. BHAR
I
ð1; TÞ is not signi?cantly different from BHAR
II
ð1; TÞ.
.
H1. BHAR
I
ð1; TÞ is signi?cantly different from BHAR
II
ð1; TÞ.
JFEP
2,3
262
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
5. Empirical results and analysis
5.1 Event and estimation windows
Table II reports descriptive statistics of betas for the ?ve estimation windows while
Table III presents t-statistics of the beta equality tests. The results suggest that beta
estimations indeed change as the estimation window moves. The betas estimated for
trading days 21 to 120 and 41 to 120 are not signi?cantly different from each other but
are both signi?cantly different from the remaining estimated betas. However, the three
remaining estimated betas are not signi?cantly different from each other.
The observed changes in beta during the post-IPO period suggest that stock
performance during the ?rst 60 trading days post-IPO is likely to differ from that in
following days. On the other hand, stock performance appears to stabilise after the
initial 60 trading days in the sense that stock returns from the relevant periods give
consistent beta estimations (the relevant periods are the estimation windows from days
61 to 160, 91 to 190 and 121 to 220). Choosing 20-120 or 40-140 as the event-estimation
pairs could therefore result in biased beta estimation. We chose 90-190 as our
event-estimation window, although the 60-160 and the 120-220 event-estimation
windows could be expected to give similar results.
5.2 Pre- and post-reform IPO underpricing
As shown in Table IV, average underpricing (measured by the percentage change from
the offer price to the closing price of the ?rst trading day) is 116.94 per cent for the
whole sample period 2001-2003 – a high level, although less than reported for earlier
Chinese IPOs.
It is evident that the average ARfor the ?rst trading day is very similar to the average
return for the initial day (the degree of underpricing). In addition, both underpricing and
AR were substantially reduced after the May 2002 IPO allocation reform, with an
Obs. Mean SD Minimum Maximum
Beta 20-20 209 1.1163 0.2826 20.1144 1.7920
Beta 40-140 209 1.1038 0.2867 20.1266 1.8854
Beta 60-160 209 1.0856 0.2824 20.1131 1.9248
Beta 90-190 209 1.0732 0.2941 20.2184 1.8385
Beta 120-220 209 1.0703 0.2914 20.1826 1.7069
Notes: Five event windows are used: 20, 40, 60, 90 and 120 trading days from the listing day;
100 trading days are used for each estimation window; thus beta 20-120 is estimated for the 21st to
120th trading day after listing, with other estimations being analogous
Table II.
Beta estimation using
different event-estimation
windows
Beta 40-140 Beta 60-160 Beta 90-190 Beta 120-220
Beta 20-120 1.4676 2.7501
* *
2.8798
* *
2.4005
*
Beta 40-140 2.4919
*
2.3917
*
1.8981
Beta 60-160 1.3903 1.0210
Beta 90-190 0.2475
Notes: Signi?cance at:
*
p , 0.05,
* *
p , 0.01 and
* * *
p , 0.001; t-statistics for paired t-tests on the
equality of the means of the different beta estimations are presented
Table III.
Equality tests on betas
estimated using different
estimation windows
IPO allocation
reform
263
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
average decrease of 42.27 and 42.19 per cent, respectively. As a check, we also tested the
difference between offer to ?rst-day opening price and offer to ?rst-day closing price for
both the whole sample and the sub-periods. The differences are not signi?cant,
con?rming that there is no pro?t opportunity for day traders on the initial day and that
the only way to bene?t from initial day returns is by subscribing to the IPO.
We explainthe decrease inunderpricingbyconjecturingthat the post-reformallocation
method increased the attractiveness of IPO subscription, in particular to better-informed
investors, thereby increasing overall IPO demand and reducing information cascade and
bandwagon effects. This would have simultaneously decreased the need for issuers with
little information about demand to underprice their IPOs to ensure success. With no data
on investor type or trading accounts this remains a conjecture, but it would be a matter of
signi?cant interest for further research should the data become available. It appears that
the IPO reform may have increased the odds of winning an IPO “lottery” for large
investors, thereby encouraging subscription, increasing overall IPO demand and leading
to a decrease in underpricing.
One mayargue that the decrease inunderpricingmayhave beencausedbythe decrease
in market returns (the continuing bear market conditions). However, post-reform
decreases in market returns were actually smaller than pre-reform decreases. Therefore,
decreasingmarket returns ingeneral do not appear to explainthe decrease inunderpricing
in our case.
5.3 Pre- and post-reform CARs and BHARs
The CAR and BHAR paths are shown graphically in Figures 3-6. Table V summarises
t-test and Wilcoxon signed-rank test statistics for the impact of the reform on average
aggregated abnormal returns (CAR and BHAR). Table VI compares the pre- and
Pre-reform
January 2001-May 2002
(88 IPOs)
Post-reform
June 2002-December 2003
(121 IPOs)
Overall period
January 2001-December 2003
(209 IPOs)
Underpricing (raw return)
Mean 1.4141 0.9914 1.1694
SD 0.8970 0.6771 0.8030
Min. 20.0044 0.1073 20.0045
Max. 4.1379 4.2825 4.2825
Underpricing 90-190
Mean 1.4148 0.9929 1.1706
SD 0.8951 0.6784 0.8026
Min. 20.0066 0.0962 20.0066
Max. 4.1432 4.2870 4.2870
Underpricing 120-220
Mean 1.4153 0.9931 1.1708
SD 0.8948 0.6788 0.8027
Min. 20.0062 0.0963 20.0062
Max. 4.1437 4.2867 4.2867
Notes: Underpricing (raw return) is the actual ?rst day return calculated as the difference between the
?rst day closing price and offer price scaled by the offer price; underpricing 90-190 and 120-220 are
?rst day abnormal returns calculated using the 90-190 and 120-220 event-estimation windows,
respectively
Table IV.
IPOunderpricing pre- and
post-reform
JFEP
2,3
264
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
post-reform variability of the paths, ?rst by ?nding the mean absolute daily change for
each CARand BHARpath and then byusing a t-test of the difference between the pre- and
post-reform means for each of the four aggregation methods.
There are several noticeable features of the aggregated abnormal return paths (CAR
and BHAR) over the event window:
Figure 3.
Equally weighted average
CAR (CARð1; TÞ),
post-IPO trading days
1 to 90
1.5
1.4
1.3
1.2
1.1
1
0.9
1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86
Trading day
E
W

a
g
g
r
e
g
a
t
e
d

C
A
R
Pre-reform
Post-reform
Figure 4.
Equally weighted average
BHAR (BHARð1; TÞ),
post-IPO trading days
1 to 90
1.5
1.4
1.3
1.2
1.1
1
0.9
0.8
1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86
Pre-reform
Post-reform
Trading day
E
W

a
g
g
r
e
g
a
t
e
d

C
A
R
IPO allocation
reform
265
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
.
in all cases aggregated abnormal returns are persistently greater pre-reform than
post-reform;
.
paths are uniformly more variable pre-reform than post-reform;
.
there is evidence of a slightly greater 90-day reduction in value-weighted paths
post-reform than pre-reform, and of much greater reduction (or smaller increase)
over the ?rst 60 to 64 days in all post-reform paths;
Figure 5.
Value-weighted average
CAR (CARð1; TÞ),
post-IPO trading days
1 to 90
0.8
0.75
0.7
0.65
0.6
0.55
0.5
0.45
0.4
0.35
0.3
1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86
Trading day
V
W

A
g
g
r
e
g
a
t
e
d

C
A
R
Pre-reform
Post-reform
Figure 6.
Value-weighted average
BHAR (BHARð1; TÞ),
post-IPO trading days
1 to 90
0.8
0.75
0.7
0.65
0.6
0.55
0.5
0.45
0.4
0.35
0.3
1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86
Trading day
V
W

a
g
g
r
e
g
a
t
e
d

C
A
R
Pre-reform
Post-reform
JFEP
2,3
266
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
.
CAR and BHAR paths appear to be different, with the former tending to show
increasing aggregated abnormal returns; and
.
all pre-reform paths appear to change at about 60 to 64 days.
In theory, trading by investors on the open market after listing should rapidly move
IPO offer prices to their market equilibrium and, under our alternative after-market
hypotheses, this should happen more rapidly and smoothly if the regulatory reform
encouraged greater numbers of better-informed investors to enter the after-market,
Aggregated
abnormal
Pre-reform January
2001-May 2002
Post-reform June
2002-December 2003
Two-sample
mean-comparison
Wilcoxon
signed-ranks
performance Obs. Mean SD Obs. Mean SD t-test (paired) test (paired)
Equally weighted
average CAR
(CARð1; TÞ) 90 1.4346 0.0110 90 1.0021 0.0103 703.4507
* * *
8.2390
* * *
Equally weighted
average BHAR
(BHARð1; TÞ) 90 1.3963 0.0242 90 0.9457 0.0246 167.4939
* * *
8.2390
* * *
Value-weighted
average CAR
(CARð1; TÞ) 90 0.7736 0.0093 90 0.3625 0.0056 350.8672
* * *
8.2390
* * *
Value-weighted
average BHAR
(BHARð1; TÞ) 90 0.7634 0.0124 90 0.3431 0.0156 193.9386
* * *
8.2390
* * *
Note: Signi?cance at:
*
p , 0.05,
* *
p , 0.01 and
* * *
p , 0.001
Table V.
Average aggregated
abnormal performance,
pre- and post-reform
Aggregation
Pre-reform: average
absolute daily
change in aggregate
abnormal return
( January 2001-May
2002)
Post-reform: average
absolute daily
change in aggregate
abnormal return
( June 2002-
December 2003)
Two-sample
mean-comparison
Wilcoxon
signed-ranks
method Obs. Mean SD Obs. Mean SD t-test (paired) test (paired)
Equally weighted
average CAR
(CARð1; TÞ) 89 0.0014 0.0010 89 0.0012 0.0009 1.4118
* * * *
1.2870
Equally weighted
average BHAR
(BHARð1; TÞ) 89 0.0039 0.0032 89 0.0018 0.0012 6.1567
* * *
5.4760
* * *
Value-weighted
average CAR
(CARð1; TÞ) 89 0.0024 0.0022 89 0.0012 0.0012 5.5626
* * *
5.6520
* * *
Value-weighted
average BHAR
(BHARð1; TÞ) 89 0.0033 0.0034 89 0.0014 0.0014 4.8764
* * *
5.3080
* * *
Note: Signi?cance at:
*
p , 0.05,
* *
p , 0.01,
* * *
p , 0.001 and
* * * *
p , 0.1
Table VI.
Variability in pre-
and post-reform CAR
and BHAR paths
IPO allocation
reform
267
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
since this should increase informational ef?ciency. The ?rst three results listed above
suggest support for the conclusion that the after-market became more ef?cient
post-reform. Further weight is therefore added to the conjecture that there was an
increase in the number of better-informed investors in the IPOmarket, reducing the level
of underpricing required for issuers to avoid IPO failure, consistent with the predictions
of informational cascades or bandwagon hypotheses and contrary to the suggestions of
Hong (2006). If traders tend to overestimate the precision of their information (Daniel
et al., 1998) then prices may over- or under-react to new information. Before the reform,
given the low odds of obtaining an allocation, the high initial returns gained by
successful subscribers may have caused traders to over-estimate the value of the IPO
stocks, leading to both accumulating (CAR) and relative accumulating (BHAR)
abnormal performance over the event window, at least over the ?rst 60 days. This
effect could have been magni?ed by the aggregate behaviour of less well-informed
traders or irrational noise traders (De Long et al., 1991). However, after the reform it
seems that offer prices were either closer to market value, or moved more smoothly and
rapidly to market value, and provided weaker signals for over-reaction and irrational
trading.
Another possible explanation is that the reform may have improved secondary
market liquidity, both in general and for newly listed (IPO) stocks in particular. Tying
the odds of a successful IPO subscription to tradable shareholdings may have
encouraged investors to trade in the secondary market and to increase their
shareholdings, so as to become eligible for future IPO subscriptions. New investors
may also have been encouraged to enter the secondary market for the same reason.
Given that successful IPOs are “good” stocks, at least in the short-term, investors may
have been more encouraged post-reform to trade these stocks in particular, increasing
market liquidity and improving market ef?ciency relative to the pre-reform period.
Unfortunately, as remarked above, with no data on the numbers and types of traders
we are unable to disentangle these various possible explanations for the after-market
results, but this remains an interesting possibility for future research.
The results appear to be robust with respect to the method of weighting, although
the equally weighted paths appear to be somewhat smoother (probably re?ecting
greater volatility for larger-capitalisation stocks that would be smoothed overall by the
change from value-weighting to equal weighting). On the other hand, the results
appear to be sensitive to the method of aggregation, with CAR tending to increase and
BHAR tending to decrease. This result does not change any of our conclusions with
respect to the overall impact of the reform on underpricing and market ef?ciency since
pre- and post-reform paths are similarly affected. However, the results con?rm ?ndings
elsewhere (Lyon et al., 1999; Gompers and Lerner, 2003) that the method of aggregation
is not a neutral choice in general.
Here, we see that returns to buy-and-hold strategies (BHAR) would have shown
reasonable bene?ts to pre-reform investors for up to about 64 days but would have
shown decreasing bene?t thereafter. In contrast, post-reform buy-and-hold strategies
would have shown a fairly uniform decrease in bene?t as the holding period
lengthened (presumably because of more ef?cient movement toward normal pricing).
On the other hand, daily realisation of abnormal returns (CAR) would have allowed
each day’s return to be retained and so aggregated over time, albeit with considerably
less bene?t (if any) after the reform.
JFEP
2,3
268
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
5.4 Pre- and post-reform changes in beta risk
The distributions of beta (Table VII) during the pre- and post-reform periods are
slightly different, with the average beta being slightly higher after the reform (1.078
compared to 1.067) although this change is not statistically signi?cant.
There are three negative pre-reform betas whereas the post-reform betas are all
positive. Assets with negative betas (countercyclical returns) are uncommon but may
occur in quite restricted samples where “outlier” returns dominate. Outlier returns are
likely to be greater when information is less well disseminated (the market is likely to
be more volatile), so we take the reduction in the number of negative betas as a sign of
improved information ?ow. Although this beta result is weak, it is consistent with the
view that the reforms may have shortened the average post-listing length of time for
which newly listed stocks were particularly subject to “herd” behaviour and/or
investor over-reaction. In the pre-reform period such behaviour could have led to both
high returns and low sensitivity of the IPO stocks to otherwise relevant market signals.
Increased market ef?ciency in the post-reform period may have encouraged investors
to pay more accurate attention to market signals when trading IPO stocks, leading to
the observed increase in average beta. For the whole sample period the average beta is
a little more than 1, implying that on average the IPO stocks are slightly more risky
than the market index for A-shares in China. This ?nding is consistent with the results
of Balvers et al. (1988) who found that the systematic risk of new issues was greater
than that of the market index.
The above results reported were based on the 90-190 event-estimation windows.
We also conducted sensitivity test to check how the results would respond to the choice
of the event-estimation windows. We used the 60-160 and the 120-220 event-estimation
windows for the sensitivity analysis and found the results (available on request) to
be robust.
6. Conclusions
This paper has reported an event study of the impact of the May 2002 IPO allocation
reform on the short-run performance (90-day post-listing) of 209 Chinese A-share IPOs,
with abnormal initial returns computed using CAR and BHAR methods. The IPO
allocation reform altered the way that odds of winning the IPO “lottery” were
determined. Prior to the reform the odds of success in the lottery allocation depended
Beta Obs. Mean SD Min Max
Pre-reform 88 1.0671 0.2973 20.2184 1.3893
Post-reform 121 1.0777 0.2929 0.4239 1.8385
Whole sample 209 1.0732 0.2941 20.2184 1.8385
Range of beta , 0 0-0.5 0.5-1 1-1.5 . 1.5
Pre-reform
Number 3 0 18 67 0
Percentage 3.41 0 20.45 76.14 0
Post-reform
Number 0 4 37 72 8
Percentage 0 3.31 30.58 59.50 6.61
Whole sample
Number 3 4 55 139 8
Percentage 1.44 1.91 26.31 66.51 3.83
Table VII.
Comparison of betas for
pre- and post-reform IPOs
IPO allocation
reform
269
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
on the amount of the investor’s subscription bid, while after the reform the odds were
determined by the size of the investor’s existing holding of tradable shares.
It is found that from 2001-2003, the average abnormal initial return (underpricing) of
Chinese A-share IPOs was 116.94 per cent, lower than for earlier Chinese IPOs but still
very high by international standards. Signi?cant pre- and post-reform differences in
underpricing and post-listing cumulative abnormal returns (CAR and BHAR) were
found, suggesting that the reform of the IPO allocation lottery mechanism signi?cantly
lowered the degree of underpricing and probably improved market ef?ciency. We
argue that these results are consistent with the view that the reform encouraged
greater participation by larger better-informed investors, decreasing the elasticity of
IPO demand and increasing total IPO demand. Following the Information Cascades
and Bandwagon hypotheses, which may be particularly relevant to Chinese markets,
these effects would reduce the need for issuers to underprice their IPOs to ensure
success. There is also a suggestion (albeit non-signi?cant) that the average sensitivity
of IPO returns to the market index (beta) increased after the reform, which is consistent
with an increase in market ef?ciency. The observed after-market post-reform changes
may also have arisen from the in?uence of the reform on investors’ post-listing demand
for stocks in general and IPO stocks in particular. Finally, the post-reform reduction in
the degree of IPO underpricing may have increased secondary market liquidity, and
hence reduced the CAR in the after-market. Further research on the market
microstructure of the Chinese stock market is suggested, to clarify the impact of IPO
regulatory reform on IPO uncertainty, market ef?ciency and investor con?dence.
References
Balvers, R., McDonald, B. and Miller, R. (1988), “Underpricing of new issues and the choice of
auditor as a signal of investment banker reputation”, The Accounting Review, Vol. 63,
pp. 605-22.
Baron, D.P. (1982), “A model of the demand for investment banking advising and distribution
services for new issues”, Journal of Finance, Vol. 37, pp. 955-76.
Basu, S. and Li, D.D. (2000), “A theory of the reform of bureaucratic institutions”, available at:http://fmwww.bc.edu/RePEc/es2000/1271.pdf (accessed 16 June 2010).
Beatty, R.P. (1989), “Auditor reputation and the pricing of initial public offerings”, Accounting
Review, Vol. 64, pp. 693-709.
Benveniste, L. and Busaba, W. (1997), “Bookbuilding vs. ?xed price: an analysis of competing
strategies for marketing IPOs”, Journal of Financial and Quantitative Analysis, Vol. 32,
pp. 383-403.
Berglund, T., Liljeblom, E. and Lo?und, A. (1989), “Estimating betas on daily data for a small
stock market”, Journal of Banking & Finance, Vol. 13, pp. 41-64.
Biais, B. and Faugeron-Crouzet, A.M. (2002), “IPO auctions: English, Dutch, . . . French, and
internet”, Journal of Financial Intermediation, Vol. 11, pp. 9-36.
Booth, J. and Smith, R. (1986), “Capital raising: underwriting and the certi?cation hypothesis”,
Journal of Financial Economics, Vol. 15, pp. 261-81.
Carter, R. and Manaster, S. (1990), “Initial public offerings and underwriter reputation”, Journal
of Finance, Vol. 45, pp. 1045-67.
Chan, K., Wang, J. and Wei, K.C. (2001), “Underpricing and long-term performance of IPOs in
China”, Journal of Corporate Finance, Vol. 183, pp. 1-22.
JFEP
2,3
270
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Chen, G.M. (1997), “Harmonization of socialism and capitalism: a case of capital market
development in socialist China”, paper presented at the Eighth World Congress of
Accounting Education and Research, Paris, October.
Chen, G.M., Firth, M. and Kim, J. (2000), “The post-issue market performance of IPOs in China’s
new stock markets”, Review of Quantitative Finance and Accounting, Vol. 14, pp. 319-39.
Cowan, A.R. and Sergeant, A.M.A. (1996), “Trading frequency and event study test
speci?cation”, Journal of Banking & Finance, Vol. 20, pp. 1731-57.
Daniel, K., David, H. and Avanidhar, S. (1998), “Investor psychology and security market under-
and overreactions”, Journal of Finance, Vol. 53, pp. 1839-86.
De Long, J.B., Shleifer, A., Summers, L.H. and Waldmann, R.J. (1991), “The survival of noise
traders in ?nancial markets”, Journal of Business, Vol. 64 No. 1, pp. 1-19.
Gale, I. and Stiglitz, J. (1989), “The information content of initial public offerings”, Journal of
Finance, Vol. 44, pp. 469-77.
Gompers, P.A. and Lerner, J. (2003), “The really long-term performance of initial public offerings:
the pre-Nasdaq evidence”, Journal of Finance, Vol. 58, pp. 1355-92.
Gu, A.Y. (2003), “A trend towards being normal: the ‘A’ share experience on the Shanghai Stock
Exchange”, Applied Financial Economics, Vol. 13 No. 5, pp. 379-85.
Hong, L. (2006), “The new development of the Chinese capital market”, China Securities Journal,
17 May, p. 7.
Ibbotson, R.G. (1975), “Price performance of common stock new issues”, Journal of Financial
Economics, Vol. 2, pp. 235-72.
Jenkinson, T. and Ljungqvist, A. (2001), Going Public: The Theory and Evidence on How
Companies Raise Equity Finance, 2nd ed., Oxford University Press, Oxford.
Ljungqvist, A. (2007), “Chapter 7: IPO underpricing”, in Espen Eckbo, B. (Ed.), Handbook of
Corporate Finance: Empirical Corporate Finance, Vol. 1, North-Holland, Amsterdam,
pp. 375-422.
Logue, D.E. (1973), “On the pricing of unseasoned equity issues: 1965-1969”, Journal of Financial
and Quantitative Analysis, Vol. 8, pp. 91-103.
Loughran, T., Ritter, J.R. and Rydqvist, K. (1994), “Initial public offerings: international insights”,
Paci?c-Basin Finance Journal, Vol. 2, pp. 165-99.
Lyon, J., Barber, B. and Tsai, C. (1999), “Improved methods for tests of long-run abnormal stock
returns”, Journal of Finance, Vol. 54, pp. 165-201.
Mok, H.M.K. and Hui, Y.V. (1998), “Underpricing and after market performance of IPOs in
Shanghai, China”, Paci?c-Basin Finance Journal, Vol. 6, pp. 453-74.
Poon, W.P.H., Firth, M. and Fung, H.G. (1998), “Asset pricing in segmented capital markets:
preliminary evidence from China-domiciled companies”, Paci?c-Basin Finance Journal,
Vol. 6, pp. 307-19.
Reilly, F.K. (1973), “Further evidence on short-run results for new issues investors”, Journal of
Financial and Quantitative Analysis, Vol. 8, pp. 83-90.
Ritter, J.R. (1998), “Initial public offerings”, Contemporary Finance Digest, Vol. 2 No. 1, pp. 5-30.
Ritter, J.R. and Welch, I. (2002), “A review of IPO activity, pricing and allocations”, Journal of
Finance, Vol. 57, pp. 1795-828.
Rock, K. (1986), “Why new issues are underpriced?”, Journal of Financial Economics, Vol. 15,
pp. 187-212.
Sherman, A. and Titman, S. (2002), “Building the IPO order book: underpricing and participation
limits with costly information”, Journal of Financial Economics, Vol. 65, pp. 3-29.
IPO allocation
reform
271
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
Stoll, H.R. and Curley, A.J. (1970), “Small business and the new issues market for equities”,
Journal of Financial and Quantitative Analysis, Vol. 5, pp. 309-22.
Su, D. and Fleisher, B.M. (1999), “An empirical investigation of underpricing in Chinese IPOs”,
Paci?c-Basin Finance Journal, Vol. 7, pp. 173-202.
Welch, I. (1992), “Sequential sales, learning, and cascades”, Journal of Finance, Vol. 47,
pp. 695-732.
About the authors
Fei Jiang received her PhD in 2009 from Loughborough University, UK. She is a Research
Associate in the Department of Economics at Loughborough University. Her research interests
include initial public offerings, corporate ?nance and corporate governance issues in developing
and industrial countries, market microstructure, ?nancial system and legal origin. Fei Jiang is
the corresponding author and can be contacted at: [email protected]
Lawrence A. Leger holds a PhD from the University of Western Ontario and has published
journal articles in both ?nancial economics and international trade. His current research interests
lie in the performance evaluation of professionally managed investment portfolios and more
generally in ?nancial economics. He teaches in the areas of ?nancial economics and ?nancial
derivatives and has special interest in training international students in literature review by
constructive synthesis.
JFEP
2,3
272
To purchase reprints of this article please e-mail: [email protected]
Or visit our web site for further details: www.emeraldinsight.com/reprints
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)
This article has been cited by:
1. Congsheng Wu. 2013. Underpricing of homecoming A-share IPOs by Chinese firms already listed abroad.
Review of Quantitative Finance and Accounting . [CrossRef]
D
o
w
n
l
o
a
d
e
d

b
y

P
O
N
D
I
C
H
E
R
R
Y

U
N
I
V
E
R
S
I
T
Y

A
t

2
1
:
3
8

2
4

J
a
n
u
a
r
y

2
0
1
6

(
P
T
)

doc_141421417.pdf