Another value for underwriters is that underwriters are the information manufacturers in the issuing market. According to Allen and Faulhaber (1989), although the issuing company before an IPO can disclose some enterprise's value information through various channels, these disclosure of information can not accurately and directly respond the expected value of the enterprise. The information regarding the enterprise's value still retain in the hands of issuing enterprise. Also, Grinblatt and Hwang (1989) argue that the issuing company obviously knows more than outside investors in terms of financial situation, the operating performance and the future cash flow distribution of the issuing company. Therefore, the issuing enterprise can often hold personal information which the massive investors do not know. As IPOs undervalued arising from information asymmetry between the investors and the issuer, if the issuing enterprise would like to reduce IPOs underpricing level, the issuing company may need to reduce the degree of information asymmetrical. Thus, in this process, the underwriter plays a very important role.
The underwriter's prestige may be an important signal for investors. According to Booth and Smith (1986), it has been argued that the underwriters are witnesses of shares pricing. The issuing companies can obtain more investors to companies of quality approval by employing trustworthy underwriters. It is reported that the higher reputation for underwriters tend to choose more stringent standards to appraise the value of initial shares. Thus, the better the reputation of underwriters who determine the price of IPOs, the more accurate reflect to the intrinsic value of issuing companies. Outside investors also understand that underwriters would take into account their reputation when making underwriting decisions. If the underwriters agree to underwrite business, the investors may obtain enterprise's positive signal. Although investors can not directly judge the quality of new public company is good or bad, it can be indirectly determined by the underwriter's reputation. Therefore, companies tend to employ high reputation underwriters to signal valued information of companies to outside investors when companies carry on IPOs. On the other hand, an underwriter may simultaneously underwrite many stocks and there are many potential customers. Therefore, underwriters may set up the prestige through suitable IPOs underpricing, and relies on the prestige to earn more profits.
However, according to Cheah (2000), there is a positive relation between underwriter's reputation and initial return of IPOs. By using Hong Kong's data, it can be found that high reputation underwriters have higher degree of IPOs underpricing. Nevertheless, it is argued that the underwriters who have better reputation may offer relatively lower rate of IPOs underpricing. Carter and Manaster (1990) have conducted the empirical study by using the US's IPOs data in the 1980s. The research shows that the underwriter's prestige has provided the signal of business risk to the market. As small companies with fewer financing have higher risk, as a result, the high prestige's underwriters generally refuse to accept the young and high risk enterprises. Kirkulak and Davis (2005) study the relationship between the reputation of underwriters and IPOs underpricing using Japan's data from 1998 to 2002. It is discovered that the relationship may depend upon the pricing of IPOs and the demand of sale. When the demand is bigger, underwriters' reputation and the degree of IPOs underpricing are positive correlation, otherwise inverse correlation. Carter and Manaster (1990) point out that the high prestige's underwriters issuing low-risk shares is because they expect to obtain income from enterprise's following financing, such as SEO. Carter (1992) indicates enterprises that have no SEO have the high risk after going on the market.
Table 2.5 Selected studies of Underwriter's reputation
Papers
Findings
Booth and Smith (1986)
The underwriters are witnesses of shares pricing. The issuing companies can obtain more investors to companies of quality approval by employing trustworthy underwriters.
Carter and manaster (1990)
The underwriter's prestige has provided the signal of business risk to the market.
Carter (1992)
Enterprises that have no SEO have the high risk after going on the market.
Cheah (2000)
There is a positive relation between underwriter's reputation and initial return of IPOs by using Hong Kong's data,
Kirkulak and Davis (2005)
Kirkulak and Davis study the relationship between the reputation of underwriters and IPOs underpricing using Japan's data from 1998 to 2002. It is discovered that the relationship may depend upon the pricing of IPOs and the demand of sale. When the demand is bigger, underwriters' reputation and the degree of IPOs underpricing are positive correlation, otherwise inverse correlation.
Chapter Three: Features of the Chinese stock market and IPOs
3.1 The type of China's stock
After China's economic reform in 1978, the stock market was established in the 1990s. There are two securities exchanges in China at present, Shanghai Securities Exchange (SHSE) and Shenzhen Stock Exchange (SZSE). Because of the special circumstance, the listed companies in China may be divided into A-share, B-share, H-share, N-share, S-share, etc. A-share namely the Renminbi common share is issued by Chinese domestic companies. Chinese citizens that do not include in Hong Kong, Macao and Taiwan can purchase A-share both on the SHSE and SZSE. B-share is special Renminbi share which is provided for foreign investors to subscribe on SHSE and SZSE. H-share is listed in Hong Kong, N-share means it is listed in New York and S-share in Singapore.
3.2 The pricing of IPOs mechanism
From 2005 to the present, the China Securities Regulatory Commission (CSRC) stipulates that the price of IPOs can be determined by the book-building system. After implementing this new system at the beginning of 2005, it was obvious to reduce the degree of IPOs underpricing on the first trading day. The average IPOs underpricing was about 52% in the first half of 2005. Thus, it can be explained that the book-building system is advantageous in the decrease of information asymmetry among issuers, underwriters and investors (Jiang, 2008)
3.3 The method of share allocation
There are mainly two kinds of share allocation. One is based on the rate of success of winning the lottery; the other is the auction mechanism. The investors who subscribe IPOs are too many and the quota of issuing is limited. As a result, the lottery mechanism was introduced in 1992. At the beginning, there was fixed number of lottery forms to be provided. Investors were limited to buy the number of lottery forms. In 1993, the lottery mechanism was modified that investors can purchase many lottery forms if necessary and need to deposit a certain funds until finishing the process of the lottery. In 1994, an auction mechanism was introduced. That is, the price of IPOs was fixed and investors can bid for the price and the number of shares they want (Chan et al., 2004; Su and Fleisher, 1999). Also, it is argued that the issue method can affect the level of IPOs underpricing. Under lottery system, the level of IPOs underpricing is greater than that under auction mechanism (Basu and Li, 1998).
3.4 The institutional characteristics of the Chinese IPOs
It is common that the abnormal high return of IPOs exists on the first trading day in various countries' stock markets, especially in China. China's IPOs are not only the highest degree of underpricing, but also have the largest capital raise. The Agricultural Bank of China was the world's largest IPOs in 2010 that the amount of financing was $22.1 billion (Anderlini, 2010). Su and Fleisher (1999) research 101 IPO samples and find that there is huge underpricing from 1987 to 1995. The average initial return of IPOs reached to 949%. According to Kim, Rui and Xu (1998), the average initial return of new issues was 594% in 1993. Chi and Padgett (2002) collect 340 Chinese IPOs data during the period of 1996 and 1997 finding that the underpricing of IPOs was 127%. Tian and Megginson (2007) argue that Chinese IPOs underpricing is the highest in the world. The average initial return was 267% from 1991 to 2000.
Some researchers think that special institutional characteristics of IPOs in China may cause IPOs underpricing. The Chinese government controls an annual quota of IPOs and supervises China's Stock market. The supply and demand for IPOs is not balanced because of Chinese government's control (Chau et. al., 1999; Gu, 2000). The number of new shares issuing is less than market demand. As a result, there are fewer new issues of closed price below their opening price and almost no issue failure of new shares. Also, as for whether to issue the new stock, it is decided completely by the government. According to Yau and Steele (2000)，the Chinese government rests on certain standards to filter the request of the companies which ready to list on the stock market every year. These companies that meet those standards set by Chinese government would be allowed to list on the stock market. At the same time, the government can decide the fixed price of new shares. Because of long time between the new stock offering and the going on the market, it has increased the risk and the cost of subscription for investors. Tian and Megginson（2007）argue that the Chinese IPOs underpricing have some determinants such as listing time lags, offering sizes, offering proceed. However, what more encounter is the economic problem. They believe that China's system characteristics have brought huge IPOs underpricing, particularly listing time lags. This argument has been also supported by Chowdhry and Sherman (1996) and Chen et al. (2004). Mok and Hui (1998) discover that long elapsed time between offering and listing and ex an uncertainty are main determinants in China's IPOs underpricing. They adopt 87 Shanghai companies to take the research object from 1990 to 1993 and confirm their viewpoint.
Though research by Chen et. al. (2004), the proportion of state and quasi-state ownership is quite high in China stock market and the proxy cost is quite high. However, the liquidity of stock is weak. Therefore, greater underpricing possibly needs to make up the increased investment risk for investors. As for the explanation of IPOs underpricing in China, Kim, Rui and Xu (1998) point out that majority of the proportion of IPOs bought by individual investors has more underpricing. If the companies expect high income, then the degree of underpricing is less. Chi and Padgett (2002) argue that the limitation of listing quota is the main explanation of IPOs underpricing, further confirm by Tian (2003).
According to Aharony, Lee and Wong (2000), another feature for IPOs is that China's stock market was established in a short time and stock market regulatory may be incomplete and the level of development is relatively low. Also, public information of companies is less than the developed country that leads to asymmetric information and more ex ante uncertainty of issuing companies. It may lead to the difficulty for investors to appraise IPOs whether to be worth investing. Basu and Li (1998) argue that an explanation for IPOs underpricing may be corruption. Bureaucrats can accept the undervalued shares as bribes, which can make companies have a shortcut to list on the stock market.
Table 3.1
Selected studies of institutional characteristics of the Chinese IPOs
Papers
Findings
Chau et. al., (1999); Gu, (2000)
The Chinese government controls an annual quota of IPOs and supervises China Stock market. The supply and demand for IPOs is not balanced because of Chinese government's control.
Yau and Steele (2000)
The Chinese government rests on certain standards to filter the request of the companies which ready to list on the stock market every year.
Tian and Megginson（2007）
The Chinese IPOs underpricing have some determinants such as listing time lags, offering sizes, offering proceed.
Mok and Hui (1998)
The time lags between offering and listing and ex an uncertainty are main determinants in China's IPOs underpricing.
Chen et. al. (2004)
The proportion of state and quasi-state ownership is quite high in China stock market and the proxy cost is quite high.
Chi and Padgett (2002); Tian (2003)
The listing quota is the main explanation of IPOs underpricing
Basu and Li (1998)
An explanation for IPOs underpricing may be corruption. Bureaucrats can accept the undervalued shares as bribes, which can make companies have a shortcut to list on the stock market.
Chapter Four: Data and methodology
4.1 Data
The Chinese IPOs market has experienced a significant transformation after 2006. The initial shares in China have not been divided into tradable and non-tradable shares as before. Also, there are few seminal papers to study the Chinese IPOs market of recent five years. The majority of attention is the late 90s' IPOs data. Thus, this paper selects Chinese IPOs data as samples which issued A-shares on SHSE or SZSE from 1st January 2006 to 31st December 2008. In addition to three companies were suspended from listing on the stock market because of some reasons. The total of samples in this paper is 267. The source of samples is GAT Database which is well-known China's financial database.
4.2 Variables selection
Table 4.1 Definition of variables
Variable
Definition
Dependent variable
MARi1
Market adjusted return of an IPO
Independent variables
LotteryRate
The rate of winning the lottery
Percent
The percent of state-owned equity
IssueP
The price of new shares
Lnum
The number of IPOs being issued by logarithm
P/E
The P/E ratio of new shares
Lage
The logarithm of the age of company before public
Lday
The time lags between offering and listing shares by logarithm
Turnover
The turnover of IPOs
Exch
The stock exchange dummy which equal to 1 in the Shanghai Stock Exchange and equal to 0 in the Shenzhen Stock Exchange
Variable 1: This paper uses five variables to explain the IPOs underpricing phenomenon in China. Base on the discussion in the features of the Chinese IPOs market above. This article employs the rate of winning lottery as a proxy to test the demand for IPOs in China stock market. The lower is the rate of winning the lottery; the larger is the demand of new shares (Chi and Padgett, 2005). Thus, it can be expected that there is a negative relation between the rate of winning the lottery and the initial return of IPOs in the test of this paper.
Variable 2: For testing information asymmetry hypothesis, it can be chosen the percent of initial shares controlled by the Chinese government as the second variable. In China, it is common for the government to own most of initial shares. The government can determine the pricing of the IPOs, while outside investors can not understand the quality of state-owned equity (SOE). Consequently, fewer investors tend to purchase SOE. Thus, it is conjectured that the percent of state-owned shares and the initial return are negative.
Variable 3: The third variable is the price of IPOs which is a proxy to test signaling hypothesis. It is argued that issue price can be viewed as a signal of the quality of a company. Under the signaling hypothesis, companies tend to underprice IPOs to attract investors. That is because good companies have confidence to earn large income from following SEO. In contract, poor companies do not have ability to bear the losses of underpricing (Datar and Mo, 1998). Therefore, the relation between the price of IPOs and initial return is expected to be negative.
Variable 4: The fourth variable is the number of initial shares. Chen et al. (2004) argue that if the number of new shares being issued is larger, investors may be not urgent to purchase on the first trading day. This may result in lower initial return. Also, in general, small issue size is controlled easily by the speculators so that investors follow to buy. The price of new shares increases so that the degree of IPO underpricing is larger. Thus, it is predicted that there is negative correlation between the number of initial shares and IPOs return on the first trading day.
Variable 5: As the test to signaling hypothesis, the P/E ratio of companies can be regarded another variable. Higher P/E ratio reflects that the company has strong profitability and will have a good development in the future. The investors usually use this proportion to estimate investment value of some shares or carry on the comparison between the two different stocks. Therefore, it is expected to be a significantly positive relation between initial return of IPOs and the P/E ratio.
Control variables: In addition to the five variables above, it can be added some variables as the control variables which are the age of company, the days between offering and listing date, turnover of new shares and stock exchange dummy. According to Tian and Megginson (2007), the age of company is predicted to be negative relation to the initial return of IPOs. The time delay is expected to positively relate to the initial return of IPOs. The turnover of new issues and the initial retune of IPOs are positively proportional (Zhao, 2009). The stock exchange dummy which equal to 1 on the Shanghai Stock Exchange and equal to 0 on the Shenzhen Stock Exchange.
4.3 Hypothesis
According to the variables selection above, it can be proposed some hypotheses.
Hypothesis 1 H0: There is no relation between the rate of winning the lottery and initial return of IPOs. H1: There is a negative relation between the rate of winning the lottery and initial return of IPOs.
Hypothesis 2 H0: There is a positive relation between the percent of state-owned equity and initial return of IPOs. H1: There is a negative relation between the percent of state-owned equity and initial return of IPOs.
Hypothesis 3 H0: There is no relation between the price of IPOs and initial return of IPOs. H1: There is a negative relation between the price of IPOs and initial return of IPOs.
Hypothesis 4 H0: There is no relation between the number of shares being issued and initial return of IPOs. H1: There is a negative relation between the number of shares being issued and initial return of IPOs.
Hypothesis 5 H0: There is no relation between the P/E ratio of company and initial return of IPOs. H1: There is a positive relation between the P/E ratio of companies and initial return of IPOs.
Hypothesis 6 H0: There is no relation between the age of company and initial return of IPOs. H1: There is a negative relation between the age of company and initial return of IPOs.
Hypothesis 7 H0: There is no relation between Lday and initial return of IPOs. H1: There is a positive relation between Lday and initial return of IPOs.
Hypothesis 8 H0: There is no relation between turnover of new shares and initial return of IPOs. H1: There is a positive relation between turnover of new shares and initial return of IPOs.
Hypothesis 9 H0: There is no relation between stock exchange dummy and initial return of IPOs. H1: There is a positive relation between stock exchange dummy and initial return of IPOs.
4.4 Methodology
This paper will use Ordinary Least Squares (OLS) method to conduct the regression analysis. The dependent variable is initial return of IPOs. It is worthy explaining that this initial return is after market adjusted. The formula and the model are as follow:
Formula 1: IRn1=（Pn1/Pn0）-1
Where IRn1 is the initial return of the stock and Pn1 reflects the closing price of an IPO and Pn0 is the offering price of an IPO.
Formula 2: MARi1= (1+IRn1)/ (1+Rm1)-1
Where MARi1 is market adjusted return of an IPO and IRn1 is the initial return of the stock and Rm1 is the market return on the first trading day.
Model 1: MARi1=α + β1 LotteryRatei + β2 Percenti+ β3IssuePi + β4 L(num) i + β5 P/Ei+ ui
Model 2: MARi1=α + β1 LotteryRatei + β2 Percenti + β3IssuePi + β4 L(num) i + β5 P/Ei + β6L(age) + β7L(day) + β8Turnover + β9Exch+ ui
First，though regression analysis, it can be analyzed whether the five variables (LotterRate, Percent, IssueP, Lnum, P/E) above are significant. Second, in addition to five variables above, the rest of variables are as control variables, such as Lage, Lday, Turnover and Exch. It will be tested whether the five variables (LotterRate, Percent, IssueP, Lnum, P/E) are still significant when adding the control variables. If these five variables express significant relation all the times, it may prove that these five variables can explain the Chinese IPOs underpricing during the period between 2006 and 2008.
Chapter Five: Empirical findings and discussion
5.1 Test for IPOs underpricing phenomenon in China
Table 5.1 Underpricing of the IPOs in China from 2006 to 2008
Year
Initial Return
Standard
Error
t-value
N
2006
0.8358***
0.0734
11.3801
66
2007
1.9307***
0.0996
19.3810
126
2008
1.1612***
0.1041
11.1505
75
***Significant 1% level.
Though the regression analysis, it is clear that abnormal high initial return truly exists in China from 2006 to 2008 in the Table 5.1. The initial returns of these three years were all significant at 1% level. The rate of IPOs underpricing was 83.58%, 193.07% and 116.12% in 2006, 2007 and 2008, respectively. The initial return in 2007 was the highest among three years. That is because China's A-share IPOs market had a good performance and the issue size was almost more than 4469 billion Yuan which was 2.7 times than that in 2006 (Xu, 2007). The number of new shares issued is twice than that of last years.
5.2 Summary statistics for variables
Table 5.2 Characteristics of variables
Year
Some Variables
LotteryRate
Percent
IssueP
Lnum
P/E
Lage
Lday
2006 Mean
Min
Max
0.4903
0.2638
8.0870
8.8698
15.157
7.5259
2.6151
0.0000
0.0000
2.4000
7.3132
0.0000
5.8972
1.9459
5.3275
0.8444
26.000
14.218
26.430
8.5487
3.1355
2007 Mean
Min
Max
0.3376
0.2249
11.471
8.6176
22.147
7.4525
2.5790
0.0000
0.0000
2.8800
7.1309
0.0000
4.4067
1.7918
2.4677
0.8629
36.990
13.710
92.600
8.9026
5.3799
2008 Mean
Min
Max
0.1046
0.1467
11.850
8.3789
20.486
6.9379
2.4657
0.0000
0.0000
2.1800
7.1701
4.9700
4.8363
1.9459
0.7265
0.7999
26.080
12.612
38.680
8.8786
3.0910
Table 5.2 presents the characteristics of some variables from 2006 to 2008. The rate of winning the lottery (LotterRate) was 49.03% in 2006 and declined to 10.46% in 2008. The percent of state-owned shares (Percent) decreased from 26.38% in 2006 to 14.67% in 2008. There was a slight increase on the price of issues (IssueP) from 2006 to 2008. As for the number of IPOs issuing per company (Lnum), it decreased slowly from 2006 to 2008. In term of the P/E ratio, it was the highest in 2007. It may be because there were more A-shares to be issued in 2007. Also, institutional transforms and the new accounting standard implementation in 2007 may improve greatly the performance of public companies so that it shows a higher P/E ratio. The age of companies (Lage) that ready to list on the stock market was getting younger along with the year. (Lday) was about 2.6 in 2006 and decreased to 2.47 in 2008 indicating that the time lags between offering and listing was more and more shorter from 2006 to 2008. In terms of the stock exchanges dummy, it is worthy noting that there were 222 A-share samples to list on the Shenzhen Stock Exchange, while there were 45 A-share samples to list on the Shanghai Stock Exchange.
5.3 Test for determinants of IPOs underprcing in China
Table 5.3 Empirical results for IPOs underpricing in China (2006-2008)
Variable
Coefficient
t-value
Expected sign
Intercept
3.1723***
5.8706
LotteryRate
-0.2659**
-2.0573
-
Percent
0.4035*
1.8421
-
IssueP
-0.0210*
-1.8755
-
Lnum
-0.2324***
-4.0000
-
P/E
0.0229***
2.9190
+
R2
0.1641
***Significant 1% level.
** Significant 5% level.
* Significant 10% level.
Model 1: MARi1=α + β1 LotteryRatei + β2 Percenti+β3IssuePi + β4 L(num) i +β5 P/Ei+ ui
Table 5.4 Empirical results of IPOs underpricing with control variables (2006-2008)
Variable
Coefficient
t-value
Expected sign
Intercept
5.5109***
4.6529
LotteryRate
-0.2514**
-2.3517
-
Percent
0.4050*
1.7785
-
IssueP
-0.0225*
-1.8852
-
Lnum
-0.3352***
-3.5850
-
P/E
0.0229***
2.9873
+
Lage
-0.1009
-1.1244
-
Lday
-0.2303
-0.9480
+
Turnover
-0.2435
-0.3982
+
Exch
0.3150
0.9021
+
R2
0.1780
***Significant 1% level.
** Significant 5% level.
* Significant 10% level.
Model 2: MARi1=α + β1 LotteryRatei + β2 Percenti + β3IssuePi + β4 L(num) i + β5 P/Ei + β6L(age) + β7L(day) + β8Turnover + β9Exch+ ui
The regression results above have been modified for the autocorrelation, multicollinearity and heteroscedasticity. Through Table 5.3 and Table 5.4, they shows that these five significant variables, the rate of winning the lottery (LotteryRate), the percent of state-owned shares (Percent), the price of issues (IssueP), the number of IPOs per company (Lnum) and P/E ratio may basically explain Chinese A-shares IPOs underpricing from 2006 to 2008. However, those control variables added in the model are all insignificant.
5.3.1 Discussion for the test result of variable 1 (LotteryRate)
In the results of multivariate regression, the rate of winning lottery is significant at 5% level. As it is expected before, there is an inverse correlation between the rate of winning lottery and the market adjusted return. In line with Chi and Padgett (2005), fewer opportunities of winning the lottery can be explained that investors' investment cost is relatively high. As a result, investors need to be paid for high returns. Therefore, the degree of new shares underpricing may be possibly large. On the other hand, the rate of winning the lottery often reflects the investors' recognition to the value of new shares.If there is lower rate of winning the lottery, investors may think that the issuing price of initial shares is relatively low and there will be an outstanding performance in the future so that the demand of new stock for investors is also large. It leads to the high initial return. In contrast, a higher rate of winning the lottery may be thought that the investors' enthusiasm for purchasing new shares is not strong. Also, the low rate of winning lottery increases the risk of investment for investors. Investors may need IPOs underpricing as the higher income compensation. This is easier to create a huge IPOs underpricng phenomenon. As Chi and Padgett's view, Chinese people like to follow the crowd. If other people tend to purchase the new shares, they may also buy a certain shares whatever there is asymmetrical information. Therefore, it is allowed to reject the null hypothesis, thinking that the rate of winning the lottery is inversely proportional to the market adjusted return of new shares.
5.3.2 Discussion for the test result of variable 2 (Percent)
In terms of testing information asymmetry hypothesis, at 10% significant levels, there is a positive between the percent of state-owned shares and the market adjusted return of IPOs though the empirical analysis. This finding is opposite to the expected result. According to Jones et al. (1999), research discover that if the proportion of the new shares controlled by the government is small, there will be more stocks to be sold to the outside investors. This may cause high market adjusted return and high A-share IPOs underpricing. However, as for the test result of this paper, it can be explained that the financing ability of Chinese state-owned shares may be stronger than that of non-state-owned shares (Liu, 2010). The financial crisis erupted after the period of 2007. This possibly causes the downturn of Chinese stock market. The government offers IPOs underpricing to attract investors to purchase state-owned shares. Also, in the period of financial crisis, the performance of stock market is unstable. Under China's special institutional system, the percentage of new shares which are owned by the government may have smaller risk. Investors tend to purchase the new shares controlled by the government. That is because, when necessity the government will sacrifice own benefit to save the controlled companies. Thus, it is allowed to accept the null hypothesis that the percent of government-owned new issues are positively related.
5.3.3 Discussion for the test result of variable 3 (IssueP)
As it can be seen from Table 5.4, at 10% significant level, the price of new stock issuing and the market adjusted return are the reversely related. Generally speaking, if the offering price of new issues is higher at first, the growth level for the price of new issues may be smaller. In addition, the test result for the price of issues is the same as Chang et al. (2008). They use 891 China's A- share IPOs from 1996 to 2004 to conduct regression analysis and find that the offering price of IPOs have a significantly inverse impact on the initial return. That is because the higher is the offering price of IPOs, the fewer are the investors with less funds. Datar and Mo (1998) use 226 Chinese IPOs data during 1990 and 1996, proving that the offer price is inversely related to the initial return. Also, Lui and Li (2001) conclude the same empirical evidence using 781 IPOs samples in China from 1991 to 1999. Thus, the null hypothesis can not be acceptable that there is a negative relation between the price of IPOs and initial return of IPOs.
5.3.4 Discussion for the test result of variable 4 (Lnum)
Table 5.4 shows that the number of IPOs issuing on the first day have a negative impact on the initial return of new issues. This test result is significant at 1% level. As it can be known, the greater is the number of IPOs issuing, the more difficulty to control the IPOs. Thus, there are fewer opportunities of speculation. Because of this, the market-adjusted initial return may be low and the rate of IPOs underpricing is low. In contrast, fewer IPOs issuing can be easy to speculation so that the degree of IPOs underpricing is greater. Moreover, the large capitalization stocks tend to employ good reputation underwriters to scientifically pricing IPOs. This may make the price of issuing close to the real price of IPOs and the level of IPOs underpricing may be low. According to Shleifer (1986), the demand for new shares may be not elastic. When issuing companies offer more shares, investors may be not hurry to buy the new shares. This may be a reason of lower initial return on the first day (Chan et al. 2004). Therefore, it needs to reject the null of assumption and think that there is a negative relation between the number of shares being issued and initial return of IPOs.
5.3.5 Discussion for the test result of variable 5 (P/E)
As for the P/E ratio of the companies who are ready to list on the stock market, it is positively related to IPOs underpricing at 1% significant level. P/E ratio can be viewed as the proxy of signaling hypothesis. A higher P/E ratio is generally believed that the price of stock is overvalued in the primary market. Investors can require IPOs underpricing as the compensation. A high P/E ratio also indicates good prospects for the development of companies. Although Zhao (2009) have tested this proxy using regression analysis, the result was not significant. However, in this paper, the result of P/E ratio is positively significant. Thus, the null of hypothesis may be refused that can receive the hypothesis of H1 which proposes that there is a positive impact between the P/E ratio and the initial return of new issues in China from 2006 to 2008.
5.3.6 Discussion for the test results of control variables
The following control variables (Lage, Lday, Turnover and Exch) are all insignificant. The longer is the time between company establishment and listing, the more the investors can get companies information. The information exposure of the company may cause a lower underpricing of IPOs. Otherwise, the shorter is the age of a company, the less information the investors can get. In order to compensate the lack of information for investors, companies tend to underpricing to attract investors. As for the USA IPOs market, the age of companies may be inversely related to IPOs underpricing (Ritter, 1984; Megginson and Weiss (1991). In contrast with the Chinese IPOs market, Yu and Tse (2006) select 343 A-share from 1995 to 1998 finding that the age of firm has a negative impact on the initial return of IPOs. However, this variable can not explain the underpricing phenomenon from 2006 to 2008 in China's stock market. Thus, the null of hypothesis can not be refused that there is no relation between the age of company and initial return of IPOs.
Base on the literature review before, the time lags between offering and listing and initial return of IPOs are positively related. This has been proved by many researchers (Chan et al., 2004; Mok and Hui, 1998; Su and Fleisher 1999; Lui and Li, 2001). The longer time delay between offering and listing of IPOs, the higher is the degree of IPOs underpricing. Long period delay for listing may take up investors' funds. This can increase investors' risk. Because of some institutional reasons in China, the listing time of some IPOs delay more than three years. The long time lags between offering and listing of IPOs greatly increase the uncertainty of the pricing of IPOs. However, in the period of 2006 and 2008, as the reform of China's issuing system, the time between offering and listing gets shorter. This may be the reason why the variable is not significant in the empirical analysis of this paper. Thus, this variable can not explain the phenomenon of China's IPOs underpricing from 2006 to 2008. There is no relation between listing time lag and initial return of IPOs.
Generally speaking, turnover of new shares has a significantly positive correlation with the initial return of IPOs on the first trading day. Though the research of China's A-share IPOs by Chang et al. (2008), some underwriters may purchase the shares themselves in order to higher initial turnover and higher price of initial shares. This may lead to higher initial return. Also, turnover is a useful indicator to judge whether the stock market is excessive speculation or not. The higher is turnover on the first day, the greater attraction to investors. This may cause more serious market speculation. However, the variable for turnover is not significant in this paper. It needs to receive the null of hypothesis indicating that there is no relation between turnover of new shares and initial return of IPOs from 2006 to 2008. In terms of stock exchange dummy, as the same with Chan et al. (2004)'s test result, the relation between initial return of IPOs and stock exchange dummy is also not significant. It is possible that the initial return of IPOs can not be affected whatever the stock exchange is the Shanghai Stock Market or Shenzhen Stock Market.
Chapter Six: Conclusion
In conclusion, this paper has demonstrated the underpricing phenomenon of 267 A-share IPOs issued in China from 1st January 2006 to 31st December 2008. From the empirical results above, it is clear that the IPOs underpricing phenomenon exists in China from 2006 to 2008 and the average initial return of China's A-share IPOs arrives to 130%.
It can be seen that the underpricing of A-share IPOs in China can be explained by some determinants. First, there is a negative relation between the rate of winning the lottery and the underpricing of IPOs. It is suggested that a low rate of winning the lottery may lead to high demand of new shares for investors, resulting in higher IPOs underpricing. Second, IPOs underpricing is positively related to the percentage of government-owned shares. Although this result is not consistent with previous hypothesis, investors in China may be confident in the companies which are controlled by Chinese government in recent years. Third, it can be shown that the price of IPOs and the underpricing of initial shares are inversely related. It may be argued that low price of new shares may make companies leave money on the table. It may signal the quality of companies to investors. Fourth, IPOs underpricing has an inversely proportional to the number of new shares, suggesting that the fewer supply for new issues can lead to a huge underpricing of IPOs. Furthermore, the relationship between P/E ratio and underpricing of new issues is significantly positive. It is deduced that high P/E ratio predict that listed companies will have a good expectation in the future so that cause high IPOs underpricing. Finally, some control variables such as time lags between offering and listing of IPOs, the age of company, the turnover of IPOs on the first trading day and stock exchange dummy are insignificant though empirical analysis, suggesting that these variables can not explain China's A-share IPOs underpricing in the period of 2006 to 2008.
As for huge underpricing of IPOs in China, this paper can provide some suggestions. First, China needs to develop the capital market and insist on the innovation of financial products. More financial products may reduce the stock market bubble and make the price of shares decline to a reasonable position. At the same time, it can alleviate the supply and demand imbalance in the primary market. Also, the rate of winning the lottery should be improved that may ensure every investor to have one lottery form at least.
