DeepDyve requires Javascript to function. Please enable Javascript on your browser to continue.
Investor sentiment, underwriters’ behaviour and IPO pricing: Empirical analysis from off-line institutional investors’ bids Investor sentiment, underwriters’ behaviour and IPO pricing: Empirical analysis from off-line...
Song, Shunlin; Tang, Siyuan
2015-10-02 00:00:00
China Journal of Accounting Studies, 2015 Vol. 3, No. 4, 348–373, http://dx.doi.org/10.1080/21697213.2015.1100087 Investor sentiment, underwriters’ behaviour and IPO pricing: Empirical analysis from off-line institutional investors’ bids a b ,† Shunlin Song and Siyuan Tang * School of Accountancy, Central University of Finance and Economics, Beijing, People’s Republic of China; School of Economics and Management, Tsinghua University, Beijing, People’s Republic of China (Received 25 June 2014; final version received 20 September 2015) Using detailed bid data of institutional investors from 380 Chinese IPO firms, this paper examines the roles of investor sentiment and underwriters in the IPO pricing process. The results show that the higher the investor sentiment, the higher the bids that institutional investors will offer in the inquiry stage. In addition, underwriters will further raise the issue price based on the bid price offered by the institutional investors. Although the increased issue price is not related to investor sentiment, it has a significant negative correlation with the gap between the bids of institutional investors and the intrinsic value of the companies. Finally, we also find that investor sentiment is negatively correlated with IPO underpricing (intrinsic value less the issue price) and positively correlated with IPO first-day returns. These findings add to the IPO pricing literature and have important implications for market-oriented reforms of IPO pricing. Keywords; investor sentiment; underwriters’ behaviour; IPO pricing 1. Introduction Regulation and politics play a critical role in China’s IPO market (Huang, Cheng, Li, & Wei, 2014; Ke, 2014). Prior to June 2009, the IPO book-building mechanism in China was not market-oriented. Instead, regulators had strict control over IPO pricing caps. This nominal book-building mechanism resulted in unique IPO first-day returns in China. According to statistical data, China’s IPO first-day returns were as high as 149% from 2006–2008 (Song, Tan, & Yi, 2014). To further improve the efficiency of capital market pricing, China initiated IPO book-building reforms in June 2009. Through continuous improvement, the reforms achieved initial success: IPO first-day returns declined significantly, and the efficiency of IPO pricing increased (Liu, Zheng, & He, 2011). However, after the book-building reform, the closing price frequently dropped below the issue price on the first issuing day, and the ‘three highs’ (high issue price, high price-to-earnings ratio, high excess funds raised) were apparent in the Growth Enterprises Market board (the GEM board). Faced with a series of contradic- tions and problems after the market-oriented reform, China’s regulators pledged to adhere strictly to market-oriented IPO pricing and emphasised the decisive role of the *Corresponding author. Email: dalin507@163.com Paper accepted by Heng Yue. © 2015 Accounting Society of China China Journal of Accounting Studies 349 market in it. Therefore, the study of the IPO pricing problem from the perspective of market-oriented pricing is of great practical significance. The previous literature mainly focuses on the interpretation of high first-day returns. One stream of the literature generally assumes that the first-day closing price of new shares is reasonable, and the IPO first-day return is equivalent to IPO underpricing. It tries to explain the high first-day returns from a rational perspective that is mainly based on the asymmetric information theory (Beatty & Ritter, 1986; Benveniste, Ljungqvist, Wilhelm, & Yu, 2003; Guo & Zhao, 2006; Habib & Ljungqvist, 2001; Wang, Xin, & Yang, 2009). Another stream of the literature considers that the first-day closing price may include an irrational component and proposes an influence of investor sentiment on IPO first-day returns from the perspective of complete or incomplete rationality (Baker & Wurgler, 2007; Han & Wu, 2007; Miller, 1977; Shiller, 1990; Song, Cao, & Yang, 2011). However, these two streams of literature treat the IPO pricing process in the pri- mary market as a black box, and they do not analyse the formation process of the issue price because of data limitations – institutional investors’ bid prices are not required to be disclosed in most countries. Thus, the mandatory disclosure of institutional investors’ bid prices in China after 2010 provides us with an opportunity to examine the roles of investor sentiment and underwriters in the IPO pricing process. The market-oriented book-building reforms in China have not been in place for a long time. In the literature, the pricing process is still a black box, and relevant studies are scarce (Shao, Xue, Jiang, Zhao, & Zheng, 2013; Yu, Liu, & Li, 2013). Our study can help to open this black box. Specifically, using detailed bid data of institutional investors from 380 Chinese IPO firms, this paper examines the roles of investor sentiment and underwriters in the IPO pricing process. The results show that the higher the investor sentiment, the higher the bids that institutional investors will offer in the inquiry stage. In addition, underwriters will further raise the issue price based on the bid price offered by the institutional investors. Although the increased issue price is not related to investor sentiment, it has a significant negative correlation with the gap between institutional investors’ bids and the intrinsic value of the companies. Finally, we also find that investor sentiment is negatively correlated with IPO underpricing (intrinsic value less the issue price) and positively correlated with IPO first-day returns. These findings indicate that investor sentiment pushes the IPO issue price higher by affecting the bid price offered by insti- tutional investors in the inquiry stage, thus reducing the IPO underpricing. Underwriters further improve the issue price in the final stage on the basis of the bids and exploit investor sentiment in the institutional investors’ bids. However, underwriters do not make full use of investor sentiment; that is, investor sentiment and the IPO first-day return are still positively correlated. In addition, underwriters may also reduce the issue price to play a buffer role when the bid price is relatively high. The study makes an important contribution to the IPO pricing literature. The existing literature on IPO pricing mainly focuses on the influence of underwriter reputation and other factors on the issue price (IPO underpricing in the primary market) (Derrien, 2005; Li, Li, Yu, & Zhu, 2014; Ljungqvist, Nanda, & Singh, 2006; Yu et al., 2013)oronthe influence of investor sentiment and other factors on the first-day trading price (IPO over- valuation in the secondary market) (Ritter & Welch, 2002; Ljungqvist & Wilhelm, 2003; Jiang, 2007; Han & Wu, 2007). In addition, the literature treats the pricing process of new shares in the primary market as a black box. Using detailed bid data of institu- tional investors that are mandatorily disclosed, this paper examines the roles of investor sentiment and underwriters in the IPO pricing process (from institutional investors’ bids to the issue price), which helps to open the black box of the pricing process. Thus, the 350 Song and Tang study not only enriches the existing literature but also provides important implications for regulatory strategies in market-oriented reforms of IPO pricing. The remainder of the paper is organised as follows. Section 2 discusses the institutional background and reviews the relevant literature. Section 3 develops the hypotheses. Section 4 presents the research design. Section 5 reports the empirical results, and Section 6 concludes. 2. Institutional background and literature review 2.1. Book-building mechanism with Chinese characteristics IPO book building is one of the IPO pricing mechanisms used to determine the issue price from the objects of inquiry, the inquiry objects that include fund management companies, securities companies, trust and investment companies and other institutional investors. In the international capital markets, the book-building mechanism has been widely used in IPO pricing. With the introduction of book building in 2004, the system in China has been in place for ten years. According to the ‘Securities Issuance and Underwriting Management Regulation’ implemented in November 2010, the book- building process includes three stages. The first stage is the preliminary inquiry stage. The issuer and its sponsor institution make inquiries with more than 20 organisations to determine the range of the issue price and the price-to-earnings ratio based on the bid prices offered by those organisations. The second stage is the accumulative tender stage. The issuer and its sponsor institution accumulate tender from the objects of the inquiries within the price range and then determine the issue price according to the results of the accumulative tender. The stocks are allocated to investors based on this price for the shares, which all eligible subjects (i.e. investors who are qualified to sub- scribe new shares in the primary market regulated by China Securities Regulatory Commission (CSRC)) can participate in. The third stage is the final pricing stage. Referring to results of the preliminary inquiry or the accumulative tender stage com- bined with the fundamentals of the issuer, information about its industry, the market conditions and other factors, the issuer and the sponsor (lead underwriter) determine the final issue price. Because companies on the Small and Medium-sized Enterprise board (the SME board) and the GEM board may not experience accumulative tender, these companies usually go through the preliminary inquiry stage and then the issuers and underwriters determine the final issue price. Under the current IPO book-building mechanism in China, underwriters have no rights to allocate new shares (IPO alloca- tions are decided by lottery). The market-oriented pricing reform gives underwriters independent pricing authority, which they greatly cherish (Shao et al., 2013). To improve the efficiency of IPO pricing, in June 2009, China launched a market- oriented reform of the book-building mechanism. So far, the book-building reform has gone through four stages. The main changes are shown in Table 1. To summarise, the most important change is regulators’ attitude toward controls over the pricing process. Before the first stage of the reform, the CSRC ruled that IPOs’ price-to-earnings ratios must not exceed 30; after the first stage of the reform, the Commission liberalised issu- ing price controls and implemented a real market-based pricing process. However, in the third stage of the reform, fearing the ‘three highs’ and ‘break’ problems, the Com- mission re-launched the issuing price controls, thus requiring that, in principle, the price-to-earnings ratio of newly issued shares could not be more than 25% of the aver- age industry level. A more recent reform (fourth stage of the reform) that indicates the direction of the market-oriented pricing reforms requires that issuers and underwriters China Journal of Accounting Studies 351 Table 1. Four stages of the book-building reform in China. Time period Specific Measures Price Control Stage One 1. Improve the restraining mechanisms of None (10 June 2009–1 book building and inquiry systems; further November 2010) marketise pricing mechanism. 2. Optimise the online offering mechanism; separate subscription objects to on-line and off-line. 3. Set limits for the account of on-line single subscription objects. 4. Strengthen IPO risk cautions. Stage Two (1 1. Further improve the subscription and None November 2010–27 rationing constraint mechanisms. April 2012) 2. Expand the scope of inquiry objects. 3. Enhance transparency of pricing information. 4. Improve the clawback mechanism and the suspension of release mechanisms. Stage Three(27 April 1. Strengthen information disclosure. In principle, the 2012–30 November 2. Expand the scope of inquiry objects; company’s PE ratio is 2013) increase the proportion of shares placed with no higher than 25% of off-line investors. the industry average 3. Strengthen the supervision on the issue PE ratio price; indicate the responsibilities of those who are involved. 4. Increase the number of outstanding new shares; effectively alleviate the supply shortage of shares. 5. Strengthen the supervision of new share speculation; maintain the normal order of trading shares. 6. Strengthen supervision and punishment for corruption. Stage Four (30 1. Promote market-oriented mechanisms for Exclude the highest November 2013) issuing new shares. price in advance 2. Strengthen the main responsibilities of the fiduciaries, such as the issuer and its controlling shareholders. 3. Further improve the degree of market- oriented IPO pricing.4. Reform the new share placement process. 5. Strengthen supervision of law enforcement; earnestly implement the ‘three fairness’ principles. Sources: regulations and laws of the CSRC. remove the highest bid price in the inquiry stage in advance, and the removed price should be no less than 10% of the total inquiry amount. This provision may cause an issuer to deliberately exclude a relatively high bid price to avoid triggering the regulations. Thus, in essence, price controls still exist. Regulators dare not open the issuing price controls entirely because of concerns about the series of problems caused by market-oriented pricing. Understanding these issues requires in-depth knowledge of the market-oriented IPO pricing process, which makes this study of great practical significance. 352 Song and Tang 2.2. Literature review Two streams of literature are most relevant to our research. One is the cause of first- day returns, and the other is the influence of investor sentiment on IPO pricing. We will review each of these streams of literature. 2.2.1. The causes of first-day returns The prior literature mainly tries to explain the phenomenon of IPO first-day returns based on two types of assumptions – complete rationality and incomplete rationality. The asymmetric information theory accounts for the rational assumption. The asymmet- ric information theory assumes the existence of asymmetric information between partici- pants. When there is information asymmetry between issuers and investors, issuers know the true value of the company but investors do not. Therefore, investors face an uncertainty risk regarding the value of the shares (Rock, 1986). To compensate the investors for the uncertainty risk and to attract investors to buy new shares, issuers need to underprice the issue, which means that the issue price should be lower than the intrin- sic value of the stock (Beatty & Ritter, 1986). According to the information asymmetry theory, the higher the uncertainty of new shares is in the eyes of investors, the higher will be the IPO underpricing. Researchers have used different methods to identify valua- tion uncertainty, including age (Ritter, 1984), the company’s industry (Benveniste et al., 2003), the number of risk factors listed in the prospectus (Beatty & Welch, 1996) and trading volume (Miller & Reilly, 1987). Most of these methods find a positive correla- tion between valuation uncertainty and IPO underpricing. Habib and Ljungqvist (2001) show that issuers have incentives to reduce information asymmetry to cut the costs of IPO underpricing. To deliver a signal of high quality, the issuer may increase the reten- tion ratio (Brealey, Leland, & Pyle, 1977) and employ underwriters with a strong reputa- tion (Beatty & Ritter, 1986) or prestigious auditors (Titman & Trueman, 1986). Some studies refer to the information asymmetry theory in the Chinese setting. They find that underwriter reputation does not affect IPO underpricing (Guo & Zhao, 2006;Song et al., 2011), but the reputations of the auditor and the venture capital company do influ- ence IPO underpricing (Wang et al., 2009; Zhang & Liao, 2011). In addition, Rock (1986) proposes the ‘winner’s curse’ hypothesis based on information asymmetry among investors. According to this hypothesis, less-informed investors have the opportunity to win the subscription of new shares only when the issue price is overbid and investors who have prior information abandon the purchase. In the long term, according to the ‘winner’s curse’ hypothesis, rational investors who have inferior information would not dare to purchase the new shares. Therefore, the issuer will attract less-informed investors to purchase the shares through underpricing. In China, before the first stage of the book- building reform, the issue price had been strictly controlled, and IPO underpricing was high. Investors were ensured of making money as long as they obtained a ballot, and thus the ‘winner’s curse’ did not exist (Yu et al., 2013). However, after the liberalisation of the issue price controls, IPO underpricing has been reduced, and the ‘winner’s curse’ attached to small investors has begun to appear (Liu et al., 2011). IPO first-day returns are so high that it is difficult to explain the phenomenon solely from the perspective of complete rationality (Ljungqvist et al., 2006; Ritter & Welch, 2002). Thus, Ritter and Welch (2002) recommend that scholars study first-day returns from the perspective of behavioural finance. The influence of investors’ behaviour on secondary market stock prices has been confirmed by a number of studies (Baker & China Journal of Accounting Studies 353 Wurgler, 2007; Shiller, 1990). The reason that investor sentiment can affect a com- pany’s share price is that given the condition of short sale restrictions and heteroge- neous expectations, stock prices only reflect the most optimistic investors’ expectations, so the optimistic investors’ sentiment will lead to overvalued issue prices (Miller, 1977). In the short history of IPO information disclosures, coupled with the absence of secondary market prices as a reference, the valuation of new shares has great uncer- tainty, and stock prices are susceptible to influence by investor sentiment. Currently, many studies have confirmed that the investor sentiment theory can explain IPO first- day returns, which means that investor sentiment affects IPO overvaluation (Cornelli, Goldreich, & Ljungqvist, 2006; Han & Wu, 2007; Jang, 2007; Ljungqvist & Wilhelm, 2003; Ritter & Welch, 2002). Recent studies have found that the limited attention theory of investors can also be used to explain IPO overvaluation (Song et al., 2011). In addition, a few studies have begun to use the detailed off-line bid data of institu- tional investors to examine the formation process of IPO underpricing. Since October 2010, such data are required to be disclosed in China. Yu et al. (2013) find that exces- sive competition among the inquiry investors is one of the causes of IPO overvaluation. Shao et al. (2013) find that underwriters use their pricing power in the inquiry stage to further raise the issue price. Li et al. (2014) find that the more variation there is in the inquiry agencies’ quotes, the greater will be the IPO overvaluation. The available literature has two main deficiencies. First, the previous studies tend to focus on IPO underpricing but pay little attention to the formation of the IPO issue price, which leaves the formation process as a black box. The main reason for the black box is that before 2010, detailed data on off-line bids were not required to be dis- closed. Second, in most of the studies, IPO underpricing is measured based on an IPO’s first-day returns. However, the precondition for this measure is that the first-day closing price properly reflects the intrinsic value of the company. In the developing and transi- tional areas such as China, this precondition may not apply. Recent studies show that IPO first-day returns in the Chinese capital market include not only IPO underpricing in the primary market but also IPO overvaluation in the secondary market, and IPO overvaluation may account for the major portion of the first-day closing price (Liu & Shen, 2011; Song et al., 2014). Underpricing is mainly related to pricing problems in the primary market, and overvaluation is relevant to issuing problems in the secondary market. Thus, dividing an IPO first-day returns into underpricing and overvaluation based on the estimated intrinsic value is more in line with China’s special background and the research needs of this study. 2.2.2. Investor sentiment and IPO pricing Previous studies use the investor sentiment theory to explain both IPO overvaluation and IPO underpricing. The literature on IPO overvaluation explained by investor senti- ment as the cause of first-day returns is reviewed above, while this section focuses on the IPO underpricing literature based on the investor sentiment theory. The investor sentiment theory assumes that investors are not completely rational and that their trans- actional behaviour is easily influenced by emotional factors. In Derrien’s(2005) model, IPO pricing selected by underwriters depends on two components: (1) intrinsic value (revealed by institutional investors), and (2) noise traders – the more optimistic the noise traders, the higher the price. Derrien (2005) explains the reason why underwriters do not exploit investor sentiment to set the bid price as high as that willing to be paid by optimistic investors so as not to leave any underpricing, since they know the 354 Song and Tang existence of noise traders. Derrien (2005) believes that underwriters’ knowledge about adjustment information from noise traders is incomplete. Underwriters are also con- cerned about post-IPO market performance (so a conservative price is needed). Using an independent sample from France, Derrien (2005) finds that individual investors’ demand is associated with the market situation. Additionally, individual investors’ demand leads to higher IPO pricing, higher first-day returns and worse long-term mar- ket performance. Ljungqvist et al. (2006) also show that underwriters use investor sen- timent to improve IPO prices. Finally, Dorn (2009) utilises a unique sample from Germany and finds that individual investors always pay higher prices for IPO compa- nies in the WI market (a market before listing), which indicates that individual inves- tors tend to buy IPO shares based on emotional behaviour. Although a small number of studies have examined the influence of investor sentiment on IPO pricing in the pri- mary market, there is a lack of literature based on developing and transitional countries, such as China, that considers the role of investor sentiment in the primary market (its effects on IPO underpricing), which makes this research especially significant in the literature. 3. Theoretical analysis and hypotheses development Before the theoretical analysis, we first review the determinative process of the issue price. First, underwriters obtain bid prices from institutional investors through an inquiry stage. Then, based on the results of the inquiry stage coupled with comprehen- sive consideration of the issuer’s fundamentals, its industry, market conditions and other factors, the issuer and the sponsors (lead underwriter) negotiate the issue price. For the following reasons, we believe that investor sentiment can affect the bid prices of institutional investors in the inquiry stage. There is asymmetric information between issuers and investors, and investors do not know the company’s true value (Allen & Faulhaber, 1989; Beatty & Ritter, 1986). When investor sentiment is opti- mistic, such that some institutional investors may value the company highly, they may offer a high price in the inquiry stage. Some institutional investors are professional investors and their valuations are less susceptible to influence by investor sentiment. However, when they expect investor sen- timent to be high, they expect the high sentiment to increase the bid prices of other institutional investors, as well as the price in the secondary market. In those circum- stances, offering a relatively high price is a rational speculation to obtain the off-line placement. In particular, when there is excessive competition among institutional inves- tors (Yu et al., 2013), it should be an optimal strategy to bid high, given optimistic investor sentiment. Based on the foregoing discussion, we state our first hypothesis as follows: H1: The higher the investor sentiment, the higher the bids that institutional inves- tors will offer in the inquiry stage. Under the current pricing system, underwriters have great pricing autonomy in the final pricing stage. They can raise or lower the issue price on the basis of the bids offered by institutions. Theoretically, the underwriters have incentives to both increase and lower the issue price. There is no exact conclusion regarding whether underwriters will exploit investor sentiment to further increase the issue price on the basis of bid prices in the inquiry stage. On the one hand, the incomes of underwriters are linked to China Journal of Accounting Studies 355 the amount of funds they raise. The amount of raised funds depends on the issue price of new shares on the condition that the number of funds is fixed. Therefore, the under- writers have a significant incentive to further exploit investors’ optimism to raise the issue price, thus charging higher underwriting fees (Shao et al., 2013). Shao et al. (2013) find that on the stock demand curve formed by institutional investors’ bids, underwriters’ actual choice of pricing is relatively high, and the issue price is signifi- cantly higher than the average price of the institutions. Their findings suggest that underwriters have an incentive to raise the issue price. On the other hand, after the market-oriented IPO pricing reforms, regulators are increasingly concerned about the ‘three highs’ and ‘break’ problems, which are related to excessively high issue prices. Therefore, underwriters also have an incentive to avoid an extremely high issue price, thus avoiding the reputation losses caused by a ‘break’. In particular, when a bid already reflects high investor sentiment, if the underwriters use investor sentiment to further push up the issue price, it may be detected by regulators and investors and dam- age the underwriters’ reputation in both channels (Shao et al., 2013). First, the ‘break’ will affect the underwriters’ reputation in the capital markets, thus affecting their long- term interests (Chemmanur & Fulghieri, 1994). If the underwriter raises the issue price of new shares for short-term interests, it will frequently result in a ‘break’ and could damage its reputation in the eyes of investors. Then, there will be no buyers for its sub- sequent underwritten shares, which will lead to huge losses of long-term benefits. In addition, the government does not want to see a large amount of new shares to ‘break’. If the stocks underwritten by a certain underwriter often ‘break’, the underwriter may be interviewed by the CSRC, its image will be damaged in the eyes of regulators and even its rating among the rest of the underwriters will be downgraded. Because we are uncertain whether underwriters will use investor sentiment to further increase the issue price in the final pricing stage, we have to propose a null hypothesis as our second hypothesis: H2: An increased issue price is not related to investor sentiment in the final pricing stage. Ljungqvist et al. (2006) and Derrien (2005) develop models to explain the influence of investor sentiment on IPO pricing. Their models show that underwriters not only consider a company’s intrinsic value but also investor sentiment in the IPO pricing pro- cess. Rational underwriters could exploit investor sentiment to increase the issue price, thereby reducing the IPO underpricing. Our analysis suggests that institutional investors in the inquiry stage may raise the bid price because of optimistic sentiment, and under- writers in the final pricing stage may or may not take advantage of investor sentiment to further raise the issue price. However, even though underwriters will not use investor sentiment to further raise the issue price in the final pricing stage, investor sentiment could still affect the final issue price through the bid prices offered by institutional investors and ultimately reduce IPO underpricing. Based on the above discussion, we state our third hypothesis as follows: H3: Investor sentiment is negatively correlated with IPO underpricing. 356 Song and Tang 4. Research design 4.1. Sample and data We choose IPO companies in the market-oriented IPO pricing phase as our research sample. Since October 2010, after the second stage of the book-building reform, issuers and underwriters must disclose the bid data offered by institutional investors, which provides us with a good opportunity to study the IPO pricing mechanism. Therefore, we choose November 2010 to October 2012 as our sample interval, which represents the beginning of the second stage of the book-building reform to the suspension of IPOs. Because the book-building mechanism in the main board is quite different from that of the SME and the GEM boards, and there are very few companies on the main board, following Yu et al. (2013), we only use observations from the SME and GEM boards. After excluding firms with missing analyst price forecasts and information about investor sentiment and corporate characteristics, the final sample consists of 380 firms. The sample distribution is shown in Table 2. All of the variables are winsorised at 1% and 99% to eliminate the influence of outliers. The detailed bid data of institutional investors are collected manually from the WIND database. Analyst price forecast data are derived from the WIND and CSMAR databases. Three-year financial data before the IPO are from the CCER database. The other data are derived from the CSMAR database, including the variables used to construct the investor sentiment index and the corporate characteristics variables. Because there are omissions for the forecast prices of new shares in the WIND database, we use analysts’ pre-IPO forecast prices as a supplement. 4.2. Models and variables 4.2.1. Main regression models To test hypotheses H1–H3, we design equations (1)–(3): BIDPRICE ¼ a þ b SENT þ CONTROLS þ (1) ADJUST ¼ a þ b SENT þ b BIDPRICE þ CONTROLS þ (2) 1 2 IRUP ¼ a þ b SENT þ CONTROLS þ (3) The dependent variables are BIDPRICE, ADJUST and IRUP in models (1)–(3), respec- tively. BIDPRICE represents the bid price offered based on investor sentiment in the inquiry stage, so BIDPRICE = (the median of institutional investors’ bids – intrinsic value)/issue price; ADJUST represents the adjusted range in the final pricing stage, so ADJUST = (issue price – the median of institutional investors’ bids) /issue price; and IRUP represents IPO underpricing, so IRUP = (intrinsic value – issue price)/issue price. ADJUST and IRUP both require the intrinsic value of the companies. Referring to Song et al. (2014), we use the median of the analyst price forecasts to measure the intrinsic value. The main independent variable is SENT, which is defined as the investor sentiment. There is no standard way in academia to measure investor sentiment, but the most pop- ular method is to use the principle component analysis of several variables to extract a comprehensive factor to represent it. In this paper, by following the method of Baker and Wurgler (2007) and Wu, Pan, Hu, and Jiang (2012), we use six sentiment-related variables to create an overall investor sentiment index, which includes the following: China Journal of Accounting Studies 357 Table 2. Sample distribution. Panel A: Board and year distribution Board/Year 2010 2011 2012 Sum The SME Board 105 52 181 The GEM Board 122 66 199 Sum 227 118 380 Panel B：Industry distribution Industry Obs. Proportion Industry Obs. Proportion Agriculture/Fishing/Forestry 1.58% Telecom/Technology 4 1.05% Mining 1.84% Wholesale/Retail 55 14.47% Manufacturing 71.84% Real Estate Development 7 1.84% Utilities/Energy 0.26% Social Care & Service 14 3.68% Architectural Services 2.11% Media/Arts 5 1.32% (1) market turnover (monthly data); (2) discounts of closed-end funds (monthly data); (3) the number of shareholders that open new stock accounts (monthly data); (4) the overall price-to-earnings ratio of A shares (monthly data); (5) the three-month stock return of the Shanghai Composite Index; and (6) the premium price of A-H shares. Wu et al. (2012) use market turnover, discounts of closed-end funds, the number of share- holders that open new stock accounts and the IPO premium as four sentiment-related variables. We choose to use the first three of those variables. Because we study IPO pricing, we avoid the use of variables that are directly associated with IPO pricing to measure investor sentiment (such as the IPO premium) and instead choose the overall price-to-earnings ratio of A shares, which is related to the IPO premium. In addition, according to findings from the relevant literature and on the basis of Wu et al. (2012), we add two more variables to better measure investor sentiment: the three-month stock return of the Shanghai Composite Index and the premium price of A-H shares (for example, Song et al., 2014). Following the approach of Wu et al. (2012), SENT1 is computed by first choosing six variables among the six sentiment-related variables and their lags that have the most correlated coefficients and then using principle component analysis of the selected six variables. SENT2 is computed using factor analysis of the six sentiment-related variables. Equations (1)–(3) have the same control variables. To alleviate the influence of asymmetric information, we add the divergence in analyst price forecasts (UNCER) as a control variable. Furthermore, referring to the previous literature (Song, Tan, & Yi, 2014; Yu et al., 2013; Zhang & Liao, 2011), we also control for the underwriter reputa- tion (UW), the third stage of the book-building reform (REGU), the shareholding pro- portion of the largest shareholder (TOPONE), firm age (AGE), issuance size (MV), total assets (SIZE), book value per share (BPS), firm leverage (LEV), earnings per share (EPS), a dummy variable for the GEM board (BOARD), the number of 358 Song and Tang Table 3. Definitions of variables. Variable name Definition of variables IRUP IPO underpricing = (intrinsic value – issue price)/issue price; intrinsic value = the median of analyst price forecasts IR IPO first-day return = (first-day closing price – issue price)/issue price IROP IPO overvaluation = (first-day closing price – intrinsic value)/issue price; intrinsic value = the median of analyst price forecasts CAR30 Post-IPO short-run stock performance = the 30-day abnormal return following the IPO CAR90 Post-IPO short-run stock performance = the 90-day abnormal return following the IPO BIDPRICE Bid price = (the median of institutional investors’ bids – intrinsic value)/issue price; intrinsic value = the median of analyst price forecasts ADJUST Adjust price = (issue price – the median of institutional investors’ bids)/issue price SENT1 Index of investor sentiment, computed by first choosing six variables among the variables six sentiment-related variables and their lagged values that have the most correlated coefficients and then using the principle component analysis of the six selected variables SENT2 Index of investor sentiment, computed using factor analysis of the six sentiment- related variables UNCER Value uncertainty for the IPO as measured by the divergence of analyst price forecasts (variance of analyst price forecasts/mean of price forecasts) UW Underwriter reputation, equal to 1 for the top 10 underwriters measured by size and 0 otherwise REGU Dummy variable for the third stage of the book-building reform, equal to 1 for time before 27 April 2012 and 0 otherwise TOPONE Shareholding proportion of the largest shareholder = shares held by the largest shareholder/total shares AGE Firm age = year of IPO – year of firm establishment MV Natural logarithm of issuance size = ln(number of issued shares * issue price) SIZE Natural logarithm of total assets = ln(total assets); the value is the average for three years before the IPO BPS Book value per share = Book value of net assets/total shares; the value is the average for three years before the IPO LEV Firm leverage = debt/assets; the value is the average for three years before the IPO EPS Earnings per share = net income/total equity; the value is the average for three years before the IPO BOARD Dummy variable for the GEM board, equal to 1 for the GEM board and 0 otherwise INSNUM The number of institutional investors that participate in the book building = Natural logarithm of the number of institutional investors MACRO1 Macro variable = the growth rate of value added of industry MACRO2 Macro variable = the growth rate of social retail sales The six sentiment-related variables are: (1) market turnover (monthly data); (2) discounts of closed-end funds (monthly data); (3) the number of shareholders that open new stock accounts (monthly data); (4) the overall price-to-earnings ratio of A shares (monthly data); (5) the three-month stock return of the Shanghai Compos- ite Index; and (6) the premium price of A-H shares. institutional investors that participate in the book building (INSNUM) as well as year (YEAR) and industry (INDU) dummy variables. Meanwhile, to control the macro fac- tors in IPO underpricing, we add the growth rate of the value added of the industry (MACRO1) and the growth rate of social retail sales (MACRO2) to the model. Definitions of the variables are shown in Table 3. China Journal of Accounting Studies 359 Table 4. Descriptive statistics. Variable N Mean Standard deviation Min. Median Max. IRUP 380 0.084 0.152 −0.336 0.070 0.864 BIDPRICE 380 −0.152 0.174 −0.848 −0.140 0.953 ADJUST 380 0.061 0.117 −1.000 0.050 0.840 SENT1 380 −0.464 0.354 −0.976 −0.610 0.300 SENT2 380 0.253 0.880 −1.084 0.060 0.480 UNCER 380 0.124 0.057 0.020 0.120 0.297 UW 380 0.411 0.493 0.000 0.000 1.000 REGU 380 0.829 0.377 0.000 1.000 1.000 TOPONE 380 0.367 0.137 0.115 0.350 0.706 AGE 380 2.185 0.558 0.693 2.300 3.091 MV 380 10.980 0.542 9.852 10.930 12.320 SIZE 380 19.620 0.792 17.920 19.570 21.860 BPS 380 3.475 2.531 1.223 2.800 18.220 LEV 380 0.482 0.155 0.140 0.490 0.819 EPS 380 1.012 0.967 0.209 0.740 6.598 BOARD 380 0.524 0.500 0.000 1.000 1.000 INSNUM 380 4.427 0.477 3.258 4.390 6.138 Refer to Table 3 for variable definitions. 4.2.2. Descriptive statistics and correlation analysis Table 4 reports the descriptive statistics of the main variables. The results show that dur- ing the sample period, the average IPO underpricing is 8.4% when using the analyst price forecasts to estimate companies’ intrinsic value. This result indicates that the issue price is relatively high and the IPO underpricing is low in the market-oriented pricing stage. In the inquiry stage, the average bid price of the institutional investors is below the intrinsic value by 15%, which indicates that the bid price is not high overall. In the final pricing stage, underwriters increase the issue price on the basis of the inquiry results by 6% on average, which suggests that the underwriters pushed up the issue price in the pricing stage. The descriptive statistics of the other variables are reasonable. Table 5 presents the correlation analysis of the main variables. The results in Table 5 show that (1) SENT1 and SENT2 are highly correlated, which indicates that the two methods mentioned above to estimate investor sentiment have a degree of certain con- sistency. (2) IRUP is significantly and negatively correlated with SENT1 and SENT2. The higher the investor sentiment, the lower the IPO underpricing. (3) BIDPRICE is positively and significantly related to SENT1 and SENT2, which indicates that the higher investor sentiment is in the inquiry stage, the higher the bid price that the insti- tutional investors will offer. (4) ADJUST is significantly and negatively associated with SENT1 and SENT2. When the underwriters believe that investor sentiment is too high, they consciously depress the issue price to prevent damage to their reputation owing to an excessively high issue price. However, the correlation analysis only provides a pre- liminary conclusion. Further regression analyses based on theory are needed to test the specific causality among each variable. 5. Empirical results 5.1. Investor sentiment and IPO pricing To test hypothesis H1, based on the sample of the SME and GEM boards from 2010 to 2012, this paper regresses equation (1) to examine the influence of investor 360 Song and Tang Table 5. Pearson correlation matrix. IRUP BIDPRICE ADJUST SENT1 SENT2 UNCER UW REGU TOPONE AGE IRUP 1.000 BIDPRICE −0.711*** 1.000 ADJUST −0.180*** −0.530*** 1.000 SENT1 −0.155*** 0.265*** −0.194*** 1.000 SENT2 −0.111** 0.218*** −0.186*** 0.959*** 1.000 UNCER 0.020 −0.004 −0.103** 0.162*** 0.144*** 1.000 UW 0.071 −0.067 0.007 0.037 0.029 −0.048 1.000 REGU 0.135*** −0.075 −0.063 0.510*** 0.548*** 0.154*** 0.010 1.000 TOPONE −0.020 0.039 −0.050 0.085* 0.075 0.116** 0.036 0.015 1.000 AGE −0.019 0.012 0.006 −0.104** −0.121** 0.098* 0.048 −0.109** 0.005 1.000 MV −0.235*** 0.175*** 0.044 0.406*** 0.383*** 0.038 0.117** 0.187*** 0.086* −0.044 SIZE −0.074 −0.033 0.135*** −0.007 0.003 −0.035 0.037 −0.034 0.014 0.074 BPS −0.002 −0.054 0.075 −0.125** −0.119** −0.008 0.039 0.025 −0.009 0.057 LEV −0.007 −0.046 0.084 0.145*** 0.126** −0.016 −0.035 0.160*** 0.008 0.023 EPS 0.004 −0.043 0.042 −0.079 −0.076 −0.016 0.063 0.005 0.005 0.001 BOARD 0.162*** −0.019 −0.169*** −0.053 −0.057 0.073 −0.040 −0.041 −0.131** −0.057 INSNUM −0.291*** 0.338*** −0.153*** 0.197*** 0.182*** 0.097* −0.043 −0.133*** 0.102** 0.000 MV SIZE BPS LEV EPS BOARD INSNUM MV 1.000 SIZE 0.525*** 1.000 BPS 0.162*** 0.095* 1.000 LEV 0.230*** 0.580*** 0.003 1.000 EPS 0.186*** −0.030 0.829*** −0.024 1.000 BOARD −0.299*** −0.573*** −0.049 −0.303*** 0.029 1.000 INSNUM 0.157*** −0.015 −0.030 −0.145*** −0.070 −0.092* 1.000 *, **, and *** indicate statistical significance at the 10, 5 and 1% levels (two-tailed). Refer to Table 3 for variable definitions. China Journal of Accounting Studies 361 Table 6. Investor sentiment, underwriters and IPO pricing. BIDPRICE ADJUST IRUP (1) (2) (3) (4) (5) (6) SENT1 0.157*** −0.009 −0.095*** (4.11) (–0.26) (–2.99) SENT2 0.050*** 0.000 −0.036*** (3.24) (0.04) (–2.97) BIDPRICE −0.392*** −0.394*** (–3.95) (–4.07) UNCER −0.145 −0.133 −0.094 −0.094 0.012 0.002 (–0.96) (–0.86) (–1.03) (–1.02) (0.08) (0.02) UW −0.024 −0.023 −0.014 −0.014 0.027* 0.027* (–1.57) (–1.54) (–1.36) (–1.37) (1.75) (1.74) TOPONE −0.112** −0.114** 0.008 0.007 0.070** 0.074** (–2.28) (–2.36) (0.29) (0.24) (2.09) (2.21) EPS −0.011 −0.005 −0.051 −0.051 0.050 0.047 (–0.20) (–0.10) (–1.18) (–1.19) (0.83) (0.79) AGE 0.010 0.011 −0.001 −0.001 −0.002 −0.003 (0.64) (0.71) (–0.11) (–0.11) (–0.14) (–0.20) MV 0.051** 0.062*** 0.053** 0.051** −0.076** −0.080*** (2.27) (2.81) (2.39) (2.42) (–2.46) (–2.72) SIZE −0.038** −0.045** −0.037** −0.036** 0.052** 0.056** (–2.10) (–2.46) (–2.08) (–2.07) (2.11) (2.31) BPS 0.002 0.002 0.005 0.005 −0.004 −0.004 (0.41) (0.45) (1.25) (1.25) (–0.61) (–0.64) LEV 0.008 0.022 0.086 0.084 −0.077 −0.085 (0.10) (0.29) (1.59) (1.57) (–1.07) (–1.19) EPS −0.013 −0.014 −0.015 −0.015 0.015 0.016 (–0.89) (–1.01) (–1.51) (–1.50) (0.88) (0.93) BOARD −0.005 −0.004 −0.048*** −0.048*** 0.054** 0.054** (–0.21) (–0.17) (–3.36) (–3.37) (2.48) (2.48) (Continued) 362 Song and Tang Table 6. (Continued). BIDPRICE ADJUST IRUP (1) (2) (3) (4) (5) (6) INSNUM 0.078*** 0.071*** 0.002 0.003 −0.056* −0.051* (3.83) (3.48) (0.10) (0.11) (–1.84) (–1.70) MACRO1 0.001 −0.002 −0.011* −0.011* 0.010** 0.012** (0.06) (–0.14) (–1.85) (–1.77) (2.11) (2.46) MACRO2 0.013*** 0.014*** 0.001 0.002 −0.012*** −0.013*** (3.86) (3.88) (0.62) (0.73) (–3.66) (–3.81) CONSTANT −0.376 −0.382 0.248 0.251 0.136 0.134 (–0.98) (–0.98) (0.90) (0.90) (0.36) (0.36) INDU/YEAR Yes Yes Yes Yes Yes Yes N 380 380 380 380 380 380 R-squared 0.245 0.236 0.395 0.395 0.242 0.241 *, **, and *** indicate statistical significance at the 10, 5 and 1% levels, respectively. T-statistics in parentheses are computed by the standard errors adjusted for heteroscedasticity using the White test. Refer to Table 3 for variable definitions. China Journal of Accounting Studies 363 sentiment on the issue price in book building. Column (1) in Table 6 shows that after controlling for the other variables, SENT1 (investor sentiment) and BIDPRICE (institutional investors’ bids) are positively correlated at the significance level of 1%. This result suggests when investor sentiment is optimistic, institutional investors will bid high in the inquiry stage, which supports hypothesis H1. Column (2) shows that SENT2 is also significantly and positively correlated with BIDPRICE, which further supports the results of column (1) and indicates that our results are not sen- sitive to the different measurements of investor sentiment. Overall, the results from columns (1) and (2) in Table 6 support hypothesis H1, whereby the more optimistic the investor sentiment, the higher the bid price that institutional investors will offer in the inquiry stage. To test hypothesis H2, we regress equation (2) to examine the effect of investor sentiment on the adjustment range of the bid price in the final pricing stage. The results of column (3) in Table 6 show that, after controlling for the other variables, SENT1 (investor sentiment) and ADJUST (adjustment) are negatively correlated, but the correlation is not significant, which indicates that there is no clear evidence that underwriters further exploit investor sentiment to raise the issue price in the final pricing stage. The results shown in column (4) support the results in column (3). The correlation between ADJUST and SENT is still not significant. In addition, the results in columns (3) and (4) consistently show that BIDPRICE and ADJUST are signifi- cantly and negatively related, which suggests that when the bid price is rather high in the inquiry stage, the adjustment range of the bid price decided by the underwrit- ers will be less. The overall results reveal that although underwriters raise the offer price on the basis of the bid prices offered by institutional investors, they have an incentive to reduce the issue price when the bid prices are high and thus play a buf- fer role. Underwriters make no use of investor sentiment in the final pricing stage. On the contrary, they reduce the upward adjustment of the issue price when the bid price is high. We suppose that underwriters mainly tend to play a buffer role out of consideration for their reputation. They worry that a ‘break’ will cause the loss of their reputation and lead to government regulation. Of course, this buffer behaviour may bring certain financing losses to the companies. We assume that the main rea- sons why companies can accept such losses are as follows: first, the issue price is much higher in the market-oriented pricing stage, so entrepreneurs are willing to ‘leave some money on the table’ (Loughran & Ritter, 2002). In addition, companies do not want a ‘break’ because a ‘break’ will affect their reputation and refinancing plans (Allen & Faulhaber, 1989) and may also lead to government regulation. Sec- ond, our sample only includes companies on the SME and GEM boards, the size of which are relatively small. Thus, they may sit in a relatively inferior position in the game of final pricing with underwriters. After the detailed study of the channel of investor sentiment on the issue price in the IPO pricing process, we can finally settle on the influence of investor sentiment on IPO underpricing from a holistic perspective. The last two columns in Table 6 report the regression analysis, which consistently shows that after control- ling for other factors, SENT (investor sentiment) and IRUP (IPO underpricing) are negatively and significantly (p = 0.01) correlated. This result indicates that IPO underpricing becomes lower when investor sentiment is higher, which supports our H3. The main results in Table 6 do not change when the regressions are clustered by year. 364 Song and Tang 5.2. Supplementary analyses The higher the investor sentiment, the lower will be the IPO underpricing. However, this result does not necessarily mean that IPO first-day returns are lower when the sentiment is optimistic because investor sentiment may not only increase the issue price but could also improve the first-day closing price, thus resulting in a premium (overvaluation on the first day of the IPO). If the effect of investor sentiment on the closing price of the new shares is greater than its effect on the issue price, then investor sentiment will be positively related to the IPO first-day returns. The models of Ljungqvist et al. (2006) suggest that underwriters do not take full advantage of investor sentiment to raise the issue price, so investor sentiment should be significantly and positively correlated with first-day returns. To test this hypothesis, we further consider the correlation of investor sentiment with IPO overvaluation (IROP) and IPO first-day returns (IR) where IPO overvaluation = (first-day closing price – intrinsic value)/issue price, and intrinsic value is the median of the analyst price forecasts. The rate of IPO first-day returns = (first-day closing price – issue price)/issue price. The results in Table 7 show that investor sentiment is significantly and positively associated with IPO overvaluation and IPO first-day returns, which is consistent with the existing studies (Han & Wu, 2007; Song et al., 2014). The results also support the conclusion of Ljungqvis et al. (2006) that underwriters do not entirely exploit investor sentiment to raise the issue price. The findings above reveal that, although institutional investors will provide a high issue price when investor sentiment is high, they can still earn a relatively high first- day return because investor sentiment not only influences the issue price in the primary market but also the pricing effect in the secondary market, and it may have a greater impact on the latter. However, in the long run, when investor sentiment is cooled, the secondary market stock price should reverse, which means that when investors are opti- mistic, market performance after the IPO will be poor. To test this hypothesis, we examine the relationship between investor sentiment and short-term stock performance after an IPO. Short-term stock performance is measured by the cumulative abnormal rate of returns at 30 days (CAR30) and 90 days (CAR90). The results in Table 8 show that short-term stock performance has a significantly negative correlation with investor sentiment, which suggests that a high issue price driven by investor sentiment may yield a high first-day return in the short term, but it may bring relatively poor market returns in the long run after the IPO. Our main results reveal that investor sentiment increases the issue price by influenc- ing the bid price offered by institutional investors. However, this result can only be established on the condition that the issue price is not regulated. Due to the absence of price controls, investor sentiment has the opportunity to influence the issue price through the IPO pricing process, thereby reducing IPO underpricing. Thus, a negative correlation between investor sentiment and IPO underpricing can only be established in the market-oriented pricing stage. To test this inference, we extend the sample to the years 2006–2012 to verify the influence of the release of price controls (that is, market- oriented pricing reform beginning on 10 June 2009) on the relation between investor sentiment and IPO underpricing. Because there are no detailed bid data before 2010, we only examine the relationship between investor sentiment and IPO underpricing. REGU2009 represents a dummy variable for market-oriented pricing reform (equal to 1 for the period before 10 June 2009 and 0 otherwise). The results are shown in column (1) and column (2) in Table 9. SENT and IRUP are significantly and positively corre- lated in the price control stage, which means that investor sentiment has a significantly China Journal of Accounting Studies 365 Table 7. Investor sentiment, IPO overvaluation and IPO first-day return. IROP：IPO Overvaluation IR：IPO first-day return (1) (2) (3) (4) SENT1 0.340*** 0.226* (2.98) (1.79) SENT2 0.348*** 0.222** (3.53) (2.18) UNCER 0.248 0.225 0.405 0.387 (0.72) (0.62) (1.07) (0.97) UW 0.073 0.073 0.100* 0.099* (1.49) (1.47) (1.81) (1.79) REGU −0.054 −0.086 0.003 −0.016 (–0.73) (–1.05) (0.04) (–0.19) TOPONE 0.035 0.023 0.089 0.081 (0.16) (0.11) (0.35) (0.32) AGE 0.016 0.014 0.010 0.008 (0.50) (0.43) (0.30) (0.25) MV −0.411*** −0.419*** −0.505*** −0.510*** (–3.47) (–3.48) (–3.69) (–3.65) SIZE 0.126 0.129 0.194* 0.196* (1.34) (1.41) (1.79) (1.84) BPS −0.001 −0.001 −0.007 −0.007 (–0.09) (–0.08) (–0.49) (–0.50) LEV −0.351 −0.345 −0.461 −0.458 (–1.38) (–1.35) (–1.62) (–1.60) EPS 0.037 0.039 0.061 0.062 (1.05) (1.09) (1.53) (1.55) BOARD −0.069* −0.072* −0.020 −0.021 (–1.87) (–1.92) (–0.54) (–0.58) INSNUM 0.379*** 0.394*** 0.327*** 0.337*** (4.20) (4.84) (3.17) (3.63) MACRO1 0.027*** 0.015 0.034*** 0.026** (3.05) (1.53) (3.69) (2.48) MACRO2 0.007 0.000 −0.005 −0.010 (0.97) (0.02) (–0.72) (–1.27) CONSTANT 0.256 0.510 0.312 0.471 (0.23) (0.47) (0.25) (0.38) INDU/YEAR Yes Yes Yes Yes N 380 380 380 380 R-squared 0.371 0.367 0.329 0.327 *, **, and *** indicate statistical significance at the 10, 5 and 1% levels, respectively. T-statistics in parenthe- ses are computed by the standard errors adjusted for heteroscedasticity using the White test. Refer to Table 3 for variable definitions. positive association with IPO underpricing. However, the interaction term SENT * REGU2009 and IRUP are negatively correlated, which indicates that the relationship between investor sentiment and IPO underpricing reversed after the market-oriented IPO pricing reforms. To analyse the regression coefficients of SENT and SENT * REGU2009 comprehensively, investor sentiment and IPO underpricing are negatively (the sum of the coefficients of SENT and SENT * REGU2009 is negative) and signifi- cantly (the F-values of SENT and SENT * REGU2009 are 9.42 and 16.76, respectively, in a joint significance test) related during the market-oriented pricing process. Next, with reference to Table 1, we investigate the influences of the first three stages of the 366 Song and Tang Table 8. Investor sentiment and stock performance after IPO. CAR30 CAR90 (1) (2) (3) (4) SENT1 −0.218*** −0.173** (–4.29) (–2.34) SENT2 −0.224*** −0.108 (–3.80) (–1.25) UNCER −0.220 −0.205 −0.168 −0.137 (–1.43) (–1.31) (–0.77) (–0.62) UW −0.005 −0.005 0.002 0.003 (–0.29) (–0.27) (0.08) (0.10) REGU −0.021 −0.000 0.022 0.023 (–0.48) (–0.00) (0.41) (0.42) TOPONE −0.053 −0.045 −0.017 −0.013 (–0.89) (–0.76) (–0.19) (–0.14) AGE −0.008 −0.007 −0.012 −0.010 (–0.64) (–0.52) (–0.54) (–0.46) MV −0.008 −0.003 0.004 0.003 (–0.35) (–0.12) (0.15) (0.11) SIZE −0.003 −0.005 0.002 0.000 (–0.14) (–0.26) (0.08) (0.00) BPS 0.007 0.007 0.008 0.008 (1.43) (1.42) (0.85) (0.91) LEV 0.005 0.001 −0.033 −0.036 (0.07) (0.02) (–0.31) (–0.33) EPS −0.024* −0.025* −0.048** −0.048** (–1.68) (–1.79) (–2.11) (–2.18) BOARD 0.025 0.027 0.031 0.031 (1.14) (1.21) (1.06) (1.01) INSNUM 0.034 0.024 0.017 0.004 (1.43) (1.06) (0.51) (0.11) MACRO1 0.008* 0.016*** 0.003 0.008 (1.66) (3.23) (0.46) (1.19) MACRO2 −0.014*** −0.010** −0.005 0.000 (–3.31) (–2.59) (–0.72) (0.05) CONSTANT 0.051 −0.113 −0.348 −0.405 (0.14) (–0.30) (–0.77) (–0.85) INDU/YEAR Yes Yes Yes Yes N 380 380 380 380 R-squared 0.126 0.114 0.161 0.151 *, **, and *** indicate statistical significance at the 10%, 5% and 1% levels, respectively. T-statistics in parentheses are computed by the standard errors adjusted for heteroscedasticity using the White test. Refer to Table 3 for variable definitions. market-oriented pricing reform (REGU1-REGU3) on the relation between investor sen- timent and IPO underpricing. The results in columns (3) and (4) of Table 9 show that the regression coefficients of SENT * REGU1 and SENT * REGU2 are significantly negative, and the coefficient of SENT * REGU2 is greater than that of SENT * REGU1, while the regression coefficient of SENT * REGU3 is not significant. These results indicate that the first reforming stage of book building strengthened the negative correlation between investor sentiment and IPO underpricing, and the influence was more significant in the second stage of the reform. However, the third stage of the reform regained control of the issue price, so there is no significant difference in the China Journal of Accounting Studies 367 Table 9. Investor sentiment and IPO underpricing: Regulatory effects. IRUP：IPO underpricing (1) (2) (3) (4) SENT1 0.203*** 0.170*** (7.24) (6.70) SENT2 0.114** 0.112** (2.16) (2.10) SENT1*REGU2009 −0.256*** (–8.33) SENT2*REGU2009 −0.213*** (–3.98) REGU2009 −0.387*** −0.555*** (–9.84) (–12.83) SENT1*REGU1 −0.171*** (–4.33) SENT1*REGU2 −0.397*** (–8.53) SENT1*REGU3 −0.186 (–0.75) SENT2*REGU1 −0.148** (–2.55) SENT2*REGU2 −0.294*** (–4.87) SENT2*REGU3 −0.112 (–0.44) REGU1 −0.447*** −0.576*** (–11.35) (–12.90) REGU2 −0.443*** −0.575*** (–9.69) (–11.62) REGU3 −0.434*** −0.599*** (–3.78) (–5.19) UNCER 0.527*** 0.542*** 0.454*** 0.551*** (4.62) (4.58) (3.83) (4.42) UW 0.051*** 0.048*** 0.049*** 0.047*** (3.62) (3.25) (3.45) (3.18) TOPONE 0.058 0.050 0.053 0.044 (1.06) (0.88) (0.96) (0.77) AGE −0.001 0.000 −0.007 −0.002 (–0.09) (0.01) (–0.67) (–0.21) MV −0.085*** −0.085*** −0.102*** −0.082*** (–4.23) (–4.16) (–4.80) (–3.90) SIZE 0.020 0.017 0.041** 0.019 (1.26) (1.03) (2.46) (1.16) BPS −0.007 −0.007 −0.009* −0.009* (–1.24) (–1.27) (–1.65) (–1.67) LEV −0.027 −0.028 −0.097* −0.046 (–0.48) (–0.48) (–1.73) (–0.80) EPS 0.027* 0.026* 0.032** 0.028* (1.93) (1.75) (2.28) (1.95) BOARD 0.001 −0.005 0.007 −0.007 (0.06) (–0.29) (0.48) (–0.45) MACRO1 0.011*** 0.015*** 0.006* 0.012*** (4.02) (5.90) (1.90) (3.85) (Continued) 368 Song and Tang Table 9. (Continued). IRUP：IPO underpricing (1) (2) (3) (4) MACRO2 −0.000 −0.011*** −0.008** −0.014*** (–0.08) (–3.38) (–2.05) (–3.53) CONSTANT 0.842*** 1.191*** 0.962*** 1.266*** (3.48) (4.38) (3.76) (4.48) INDU Yes Yes Yes Yes YEAR No No No No N 902 902 902 902 R-squared 0.686 0.652 0.682 0.656 *, **, and *** indicate statistical significance at the 10%, 5% and 1% levels, respectively. T-statistics in parentheses are computed by the standard errors adjusted for heteroscedasticity using the White test. Refer to Table 3 for variable definitions. coefficients between the time before the market-oriented reform and the third stage of the reform. Overall, the results in Table 9 reveal that the negative relation between investor sentiment and IPO underpricing is only established in the market-oriented pricing stage. 5.3. Robustness tests To ensure the rigour of our empirical results, we conduct the following robustness tests. (1) Following Purnanandam and Swaminathan (2004), we use the method of paired companies to measure companies’ intrinsic value. The calculation processes are as fol- lows: for each IPO company in our sample, we first find a non-IPO industry peer with comparable sales, and then we use the product of the IPO firm’s industry peers’ PEs and the IPO firm’s EPS to measure the firm’s intrinsic value. (2) Considering the opti- mism bias of analysts’ forecasts, we use the average, instead of the median, of the ana- lysts’ price forecasts to measure the intrinsic value of the company in the robustness analysis. (3) We follow the approach of Wu et al. (2012) to recalculate the investor sentiment variable. (4) We regress the six indicators of investor sentiment with the institutional investor bids (BIDPRICE), the adjustment range (ADJUST) and IPO underpricing (IPUP) (results are not reported but are available upon request). The results in Table 10 are consistent with the main results above (Table 6), thus demonstrating the robustness of the main conclusions in this paper. 6. Conclusions and implications While the market-oriented IPO pricing mechanism significantly reduces IPO first-day returns, it has resulted in serious problems such as ‘breaks’ and the ‘three highs’ in the GEM board. In the market-oriented pricing stage, the question of how to prevent the excessive pricing of new shares from harming the interests of investors is an urgent problem for regulators. Using the unique detailed bid data of institutional investors, this paper examines the roles of investor sentiment and underwriters in the IPO pricing pro- cess. The results reveal that the higher investor sentiment, the higher the bids are that institutional investors will offer in the inquiry stage. In addition, underwriters will fur- ther raise the issue price based on the bid price offered by the institutional investors, China Journal of Accounting Studies 369 Table 10. Robustness tests. Panel A：Regression following Purnanandam and Swaminathan (2004) to measure intrinsic value BIDPRICE ADJUST IRUP (1) (2) (3) (4) (5) (6) SENT1 0.265*** −0.049 −0.195** (2.89) (–1.33) (–2.21) SENT2 0.053* −0.015 −0.041* (1.67) (–1.25) (–1.91) BIDPRICE −0.080* −0.082* (–1.68) (–1.78) Variables Yes Yes Yes Yes Yes Yes INDU/YEAR Yes Yes Yes Yes Yes Yes N 380 380 380 380 380 380 R-squared 0.480 0.472 0.190 0.188 0.465 0.459 Panel B：Regression using the average of analysts’ forecasts to measure intrinsic value BIDPRICE ADJUST IRUP (1) (2) (3) (4) (5) (6) SENT1 0.160*** −0.014 −0.087*** (4.36) (–0.42) (–2.85) SENT2 0.051*** −0.001 −0.031*** (3.41) (–0.12) (–2.66) BIDPRICE −0.366*** −0.369*** (–3.36) (–3.48) Variables Yes Yes Yes Yes Yes Yes INDU/YEAR Yes Yes Yes Yes Yes Yes N 380 380 380 380 380 380 R-squared 0.251 0.241 0.371 0.370 0.251 0.249 Panel C： Regression following Wu et al. (2012) to measure investor sentiment BIDPRICE ADJUST IRUP (1) (2) (3) (4) (5) (6) SENT 0.008** 0.036*** 0.002 0.002 −0.007** −0.035*** (2.56) (4.13) (0.87) (0.82) (–2.33) (–3.76) BIDPRICE −0.374*** −0.364*** (–3.89) (–3.52) Variables Yes Yes Yes Yes Yes Yes INDU/YEAR Yes Yes Yes Yes Yes Yes N 380 380 380 380 380 380 R-squared 0.198 0.345 0.371 0.344 0.188 0.308 *, **, and ***indicate statistical significance at the 10%, 5% and 1% levels, respectively. T-statistics in parentheses are computed by the standard errors adjusted for heteroscedasticity using the White test. Intrinsic value is measured following Purnanandam and Swaminathan (2004) in Panel A and is computed as the average of the analyst price forecasts in Panel B. Investor sentiment is calculated following Wu et al. (2012) in Panel 3. The other control variables are still included in the regressions. Refer to Table 3 for variable definitions. 370 Song and Tang but the increased issue price is not related to investor sentiment. We also find that investor sentiment is negatively correlated with IPO underpricing (intrinsic value less the issue price) and positively correlated with IPO first-day returns. These results indicate that investor sentiment may push the offer price up by affecting the bids of institutional investors, but there is no evidence that underwriters will further use inves- tor sentiment to raise the issue price on the basis of the bid price. In fact, the exploita- tion of investor sentiment by underwriters is inadequate (because investor sentiment and IPO first-day returns are positively correlated). In addition, we find that the issue price is higher than the bid prices of the institutions in general, which suggests that underwriters have incentives to further increase the issue price after the inquiry stage. At the same time, in the final pricing stage, the increased issue price has a significant negative correlation with the gap between the bids from institutional investors and the intrinsic value of the companies, which indicates that underwriters tend to avoid exces- sively high issue prices in consideration of their reputation. Finally, we find that inves- tor sentiment is positively correlated with IPO first-day returns and negatively correlated with after-IPO stock performance. This result suggests that offering a high price when investors are optimistic may yield poor market performance after the IPO, but it can lead to higher first-day returns, which demonstrates that the high bidding behaviour of institutions could be a rational short-term speculative strategy. Our findings have important implications for both the literature and policy. With regard to the literature, first, in the market-oriented pricing stage, the interpretation of IPO underpricing needs to consider investor sentiment, and studies examining the influ- ence of investor sentiment on IPO first-day returns need to consider the influence of investor sentiment on both IPO underpricing and IPO overvaluation. Second, studies analysing the IPO pricing process under the book-building mechanism need to consider underwriters’ two motivations to increase or decrease the issue price in the final pricing stage. With regard to policy, first, regulators should pay close attention to the positive influence of investor sentiment on the issue price in the market-oriented pricing stage, particularly because institutional investors may be too optimistic to push up the issue price blindly, thus harming IPO pricing efficiency. At the same time, regulators should notice speculative IPOs where a high valuation of shares on the first day might inspire high bidding behaviour (locking-up system may reduce the incentive to offer a high price because after the first day of trading, the first-day prices will reverse to their intrinsic value). Second, underwriters have an incentive to raise the issue price in their own interest, but they also have an incentive to reduce the issue price for their reputa- tion. In the next stage of reform, regulators may try to further develop the role of underwriters’ reputation to constrain their self-interested behaviour. It should be noted that, although we find that investor sentiment could push up the issue price in market- oriented pricing, this finding does not mean that our study supports price controls. After all, price controls may lead to other negative effects. Finally, it is necessary to remind readers that our main empirical results depend on the accuracy of the intrinsic value estimations. Although we use methods that are based on previous studies to measure the intrinsic value of the companies, because of the unobservability of intrinsic value, it is difficult for us to ensure an accurate estimation of a company’s intrinsic value. Acknowledgements We would like to thank two anonymous referees and the editors for their insightful comments. China Journal of Accounting Studies 371 Disclosure statement No potential conflict of interest was reported by the authors. Funding This work was supported by National Natural Science Foundation of China (71502183) and the Humanities and Social Science Foundation of the Ministry of Education (grant no. 14YJC790101), grants from the Beijing Municipal Commission of Education ‘Pilot Reform of Accounting Discipline Clustering’, and grants from the Beijing Municipal Commission of Educa- tion ‘Joint Construction Project’. Notes 1. On 30 November 2013, CSRC released Announcement [2013] No. 42 ‘The opinions of CSRC on further promoting IPO reform’. The Announcement puts forward the general principles of the reform: to adhere to a market orientation and a legal orientation; to implement the reform strategies comprehensively; to further rationalise the mechanisms of issuing, pricing and rationing; to strengthen the decisive role of market; to enhance market regulation; to maintain market fairness; and to effectively protect the legitimate rights and interests of investors, especially small investors. 2. Although Derrien (2005) and Ljungqvist et al. (2006) have already examined the effects of investor sentiment on the issue price, using Chinese data to test investor sentiment’s role in IPO pricing has two advantages: first, the range of issue prices for Chinese IPOs in the preliminary inquiry stage is not limited, so the bid prices of institutions and underwriters have a greater impact on the final issue price. Second, detailed bid data in the inquiry stage are mandatorily disclosed in China, which helps us to examine the role of investor sentiment in the IPO pricing mechanism, that is, to determine whether investor sentiment influences the bids of institutional investors in the preliminary inquiry stage or the beha- viour of underwriters in the final pricing stage. 3. The ‘three highs’ problem indicates the phenomenon where the issue price, price-to-earnings ratio and excess-raised funds of listed companies are rather high in the Chinese capital market. The ‘break’ problem indicates the phenomenon that the market price of listed companies falls below the issue price. 4. The detailed policies can be referred to in ‘The guidance for the further deepening of reform of the IPO system (28 April, 2012)’. 5. The detailed policies can be referred to in ‘The opinions of CSRC on further promoting IPO reform (30 November, 2013)’. 6. In this paper, IPO underpricing refers to the phenomenon where the issue price is less than the intrinsic value of the new shares in the primary market, and IPO overvaluation refers to the phenomenon where the first-day closing price is higher than the intrinsic value of the new shares in the secondary market. 7. Only when the offered bid price is higher than the issue price are the underwriters eligible to participate in the off-line placement. If other institutional investors offer high bids, institutional investors that offer low bids will lose the opportunity to participate in the place- ment. 8. According to reports, the rewards for an investment bank are often linked to the excess raised funds, and the rewards from excess raised funds are much larger than underwriting fees from normally raised funds (available at http://finance.ifeng.com/roll/20120315/ 5750718.shtml). 9. The definitions of IPO underpricing and IPO first-day returns are not the same. IPO underpricing equals the company’s intrinsic value minus the issue price and then divided by the issue price, while the first-day return equals the first-day closing price minus the issue price and then divided by the issue price. 10. The calculation steps for the median of the analyst price forecasts are as follows: first, we take the median of each analyst price forecast range as that analysts’ forecast, and then we choose the median price of all of the analysts’ forecasts. We remove the forecast prices 372 Song and Tang provided by the sponsor analysts and take the last forecast if the same underwriter provides more than one forecast. 11. To verify this assumption, we examine the influence of underwriter reputation and corporate size on the relationship between investor sentiment and the adjustment range of the issue price. We find that when companies are large, underwriters are more likely to increase the final issue price when investor sentiment is optimistic, while underwriters with a strong reputation tend to reduce the final issue price when investor sentiment is optimistic. These findings agree with our analysis. Unfortunately, although the directions of the regression coefficients are consistent with our expectations, the coefficients are not significantly correlated, which leaves this issue for further exploration. References Allen, F., & Faulhaber, G. R. (1989). Signaling by underpricing in the IPO market. Journal of Financial Economics, 23, 303–323. Baker, M., & Wurgler, J. (2007). Investor sentiment in the stock market. Journal of Economic Perspectives, 21, 129–151. Beatty, R. P., & Ritter, J. R. (1986). Investment banking, reputation, and the underpricing of initial public offerings. Journal of Financial Economics, 15, 213–232. Beatty, R. P., & Welch, I. (1996). Issuer expenses and legal liability in initial public offerings. Journal of Law and Economics, 39, 545–602. Benveniste, L. M., Ljungqvist, A., Wilhelm, W. J., Jr, & Yu, X. (2003). Evidence of information spillovers in the production of investment banking services. Journal of Finance, 58, 577–608. Brealey, R., Leland, H. E., & Pyle, D. H. (1977). Informational asymmetries, financial structure, and financial intermediation. Journal of Finance, 32, 371–387. Chemmanur, T. J., & Fulghieri, P. (1994). Investment bank reputation, information production, and financial intermediation. Journal of Finance, 49,57–79. Cornelli, F., Goldreich, D., & Ljungqvist, A. (2006). Investor sentiment and pre-IPO markets. Journal of Finance, 61, 1187–1216. Derrien, F. (2005). IPO pricing in “hot” market conditions: Who leaves money on the table? Journal of Finance, 60, 487–521. Dorn, D. (2009). Does sentiment drive the retail demand for ipos? Journal of Financial and Quantitative Analysis, 44,85–108. Guo, H., & Zhao, Z. Y. (2006). An empirical analysis of the effect of underwriter reputation on IPO pricing, initial return and long run stock performances. Management World (Chinese), (3), 122–128. Habib, M. A., & Ljungqvist, A. (2001). Underpricing and entrepreneurial wealth losses in IPOs: Theory and evidence. Review of Financial Studies, 14, 433–458. Han, L. Y., & Wu, Y. R. (2007). Investor sentiment and the puzzle of IPOs: Underpricing or overpricing. Management World (Chinese), 3,51–61. Huang, Q. Y., Cheng, M. Y., Li, W. J., & Wei, M. H. (2014). Listing approach, political favours and earnings quality: Evidence from Chinese family firms. China Journal of Accounting Studies, 2,13–36. Jiang, H. B. (2007). The behavior of A-share IPO in inefficient market. Journal of Financial Research (Chinese), (8), 90–102. Ke, B. (2014).A personal perspective on protecting investors of publicly listed Chinese firms, China Journal of Accounting Studies, 2, 253–263. Li, D. X., Li, X. D., Yu, H. H., & Zhu, W. Y. (2014). Disagreement of institutional investors’ bids and IPO pricing mechanism. Economic Research Journal (Chinese), (7), 151–164. Liu, Y. H., & Shen, K. T. (2011). Is the underpricing in primary market or the overvaluation in secondary market: A new examination and explanation of high underpricing in China. Journal of Financial Research (Chinese), (11), 183–196. Liu, Z. Y., Zheng, K., & He, Y. N. (2011). Is the first-stage book-building reform efficient? Journal of Financial Research (Chinese), (4), 158–173. Ljungqvist, A., Nanda, V., & Singh, R. (2006). Hot markets, investor sentiment, and IPO pricing. Journal of Business, 79, 1667–1702. China Journal of Accounting Studies 373 Ljungqvist, A., & Wilhelm, W. J. (2003). IPO pricing in the dot-com bubble. Journal of Finance, 58, 723–752. Loughran, T., & Ritter, J. R. (2002). Why don’t issuers get upset about leaving money on the table in IPOs? Review of Financial Studies, 15, 413–444. Miller, E. M. (1977). Risk, uncertainty, and divergence of opinion. Journal of Finance, 32, 1151–1168. Miller, R. E., & Reilly, F. K. (1987). An examination of mispricing, returns, and uncertainty for initial public offerings. Financial Management, 16,33–38. Purnanandam, A. K., & Swaminathan, B. (2004). Are IPOs really underpriced? Review of Financial Studies, 17,811–848. Ritter, J. R. (1984). The “hot issue” market of 1980. Journal of Business, 57, 215–240. Ritter, J. R., & Welch, I. (2002). A review of IPO activity, pricing, and allocations. Journal of Finance, 57, 1795–1828. Rock, K. (1986). Why new issues are underpriced. Journal of Financial Economics, 15, 187–212. Shao, J. X., Xue, K., Jiang, P., Zhao, Y. X., & Zheng, W. C. (2013). Investor sentiment, underwriters pricing and IPO first-day return. Journal of Financial Research (Chinese), (4), 127–141. Shiller, R. J. (1990). Speculative prices and popular models. Journal of Economic Perspectives, 4,55–65. Song, S. J., Cao, H., & Yang, K. (2011). Investor attention and IPO anomalies: Evidence from google trend volume. Economic Research Journal (Chinese), (S1), 145–155. Song, S. L., Tan, J. S., & Yi, Y. (2014). IPO initial returns in China: Underpricing or overvaluation? China Journal of Accounting Research, 7,31–49. Titman, S., & Trueman, B. (1986). Information quality and the valuation of new issues. Journal of Accounting and Economics, 8, 159–172. Wang, B., Xin, Q. Q., & Yang, D. M. (2009). Does auditor’s reputation affect the stock pricing: Evidence from marketization of IPO pricing. Accounting Research (Chinese), (11), 73–81. +96. Wu, Y. R., Pan, K., Hu, S. M., & Jiang, J. (2012). A new explanation of industry analyst earnings forecast bias. Economic Research Journal (Chinese), (4), 149–160. Yu, H. H., Liu, Y., & Li, X. D. (2013). IPO book-building reform and the problem of IPO “three highs” in China stock market: A perspective offline institutional investors’ quotes. Journal of Financial Research (Chinese), (10), 167–180. Zhang, X. Y., & Liao, L. (2011). VCs’ backgrounds, IPO underpricing and post-IPO performance. Economic Research Journal (Chinese), (6), 118–132.
http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png
China Journal of Accounting Studies
Taylor & Francis
http://www.deepdyve.com/lp/taylor-francis/investor-sentiment-underwriters-behaviour-and-ipo-pricing-empirical-9JK2eKXu2o
