IPO underpricing and ex-ante uncertainty

IPO underpricing and ex-ante uncertainty

Empirical evidence from European markets using factor analysis

MSc BA Finance Thesis

Maikel A.J. Munsterhuis

Student at the University of Groningen, Faculty of Economics and Business

April 2011

IPO underpricing and ex-ante uncertainty

Empirical evidence from European markets using factor analysis

Abstract

This thesis examines the influence of ex-ante uncertainty factors on underpricing of IPOs in the three main European stock markets; France, Germany and the UK. The study is based on a sample of 708 IPOs in the period 2001 till 2010. A factor analysis is performed to create ex-ante uncertainty variables of underlying firm-characteristics which could explain the level of underpricing. Besides, investor sentiment is taken into account as an explanation of underpricing. This study finds that significant underpricing exists in the three countries. Regression analysis with the generated factors shows that the factor that captures the size and maturity of the firm has a significant negative influence on the level of underpricing. Furthermore, it is found that investor sentiment has a significant positive influence on underpricing. Finally, cross-country analysis shows no significant differences between explanations of underpricing.

Keywords: IPO, initial public offering, underpricing, first-day return, ex-ante uncertainty

Bill Gates had thought longest about the price. Guided by Goldman Sachs, he felt the market would accord a higher price-earnings multiple to Microsoft than to other personal computer software companies like Lotus or Ashton-Tate, which have narrower product lines. On the other hand, he figured the market would give Microsoft a lower multiple than companies that create software for mainframe computers because they generally have longer track records and more predictable revenues. A price of roughly $15, more than ten times estimated earnings for fiscal 1986, would put Microsoft’s multiple right between those of personal software companies and mainframers.

…By the end of the first day of trading,…Microsoft’s stock stood at $27.75.

TABLE OF CONTENTS

TABLE OF CONTENTS 4

1 INTRODUCTION 5

2 THEORY AND BACKGROUND 7

2.1 Initial public offerings 7

2.2 Underpricing 8

2.3 Asymmetric information 9

2.3.1 Proxies for ex-ante uncertainty 12

2.4 Investor sentiment 15

3 DATA AND METHODOLOGY 16

3.1 Data 16

3.2 Methodology 17

3.2.1 Measuring explanatory variables 17

3.2.2 Factor analysis 21

3.2.3 Regression analysis 22

4 EMPIRICAL RESULTS 23

4.1 Descriptive statistics 23

4.2 Regression analysis with initial variables 25

4.2.1 Country-specific analysis 27

4.3 Principal component analysis 28

4.3.1 Country-specific analysis 29 4.4 Regression analysis with factors 30

4.4.1 Country-specific analysis 32

5 CONCLUSION 33

REFERENCES 37

1 INTRODUCTION

Initial public offerings (IPOs) and anomalies. This combination is heavily researched by academics in the last decades. A vast body of theoretical and empirical literature about IPOs documents that there are several patterns that seem to contradict with the efficient market theory. According to this theory, given a particular information set, it would not be possible for investors to structurally outperform the market. This also means that prices on financial markets cannot be predicted. However, this theory does not seem to hold for some of the so-called anomalies which are associated with IPOs.

This thesis deals with one of these anomalies: underpricing1. It is widely documented that firms that go public sell their shares at an offer price that is lower than the first-day closing market price. This positive first-day offer-to-close return is commonly known as underpricing. This initial return is measured as the percentage difference between the offer price and the market price at which the shares close at the end of the first trading day. Underpricing is thus the initial return investors can make by participating in an IPO of common stock.

If the solution to the puzzling phenomenon of underpricing does not lie in the mechanisms of the financial markets, it has to be searched in other explanations. This search has been the objective of numerous empirical studies. Many underlying theories have been generated which can be classified into a few categories (Ljungqvist, 2005). One of these categories consists of asymmetric information theories where an uneven distribution of information between the issuer and the investor is assumed. This asymmetry creates ex-ante uncertainty about the true value of the shares offered. According to Beatty and Ritter (1986), underpricing is correlated with the level of information asymmetry (uncertainty) between issuers and investors. This correlation is due to the fact that underpricing should serve as a compensation for the ex-ante uncertainty investors are facing.

By analyzing 708 IPOs from 2001 till 2010 on the main European stock exchanges (France, Germany and the United Kindom), this study investigates what ex-ante uncertainty factors influence underpricing in Europe. I am not the first to conduct such an investigation. However,

1

this study is still novel for a few reasons. Because historical market prices do not exist for newly listed firms, it is difficult to calculate ex-ante uncertainty. Therefore this study focuses on observable financial accounting variables as proxies for a firm’s ex-ante uncertainty with the objective to solve the lack of ex-ante uncertainty variables problem. A principal component analysis is conducted with the objective to bundle certain variables which are related to each other into new variables. Moreover other IPO studies typically focus on the U.S. Although several IPO underpricing studies have been conducted in Europe as well, to my knowledge I am the first in investigating in such a way the influence of financial accounting variables on underpricing in Europe in the last ten years.

This study shows that significant underpricing is present in Germany, France and the UK. The generated factors with principal component analysis do have an influence on the level of underpricing. The size and maturity factor has a significant negative sign whereas the investor sentiment factors have a significant positive sign. Besides, it is shown that the offer price of the IPO also plays a role in determining ex-ante uncertainty. Finally, cross-country analysis is performed but no significant differences between determinants of ex-ante uncertainty were found.

2 THEORY AND BACKGROUND

2.1 Initial Public Offerings

An Initial Public Offering (IPO) is known as the first effort of a private firm to raise capital in the public equity market. Ritter (1998) defines an IPO as follows: “An IPO occurs when a security is sold to the general public for the first time, with the expectation that a liquid market will develop”. Going public marks an important phase in the lifecycle of a firm. Ljungqvist (2004) argues that going public provides access to public equity capital and therefore may lower the cost of funding the firm’s operational and investment activities. Once the stock is traded in public, the increased liquidity allows the firm to raise capital on more beneficial terms than in the case it had to compensate investors for the lack of liquidity which is associated with a privately held company (Ritter, 1998). Another reason for firms to issue stock may be to finance future investment projects or eventually acquisitions2.

Various other explanations can be mentioned why firms prefer to go public. One of the first formal theories underlying the going public decision is presented by Pagano et. al (1998). He argues that it is much easier for a potential acquirer to identify a potential takeover target when it is listed on a stock exchange. Furthermore, being listed requires information transparency about the firm towards their investors. As a consequence the competition between lenders may increase. This can result in more bargaining power for the firm which in turn lowers the costs of capital. As a result of the lower cost of capital the firm value will increase. Moreover, a listing can provide good publicity because it might enhance name recognition and could increase a firm’s credibility towards its customers and suppliers. Finally, assuming that publicity for a firm in general is good, recruitment of qualified employees will be easier. According to Pagano et. al (1998) firms appear to go public not with the purpose to finance future investments and possible growth, but to rebalance their accounts after large investments during a period of high growth. He claims that IPOs are also likely to be followed by lower cost of credit and increased turnover in control.

Next to the benefits of an IPO, costs are unavoidable. Ritter (1998) categorizes the one-time IPO costs in direct and indirect costs. Direct costs include the underwriting, legal and auditing fees incurred with the preparation of the IPO. Indirect costs include the time and effort management

2

spends to finalize the offering and the costs (‘leaving money on the table’) as a consequence of underpricing the IPO. Besides one-time costs also continuous costs are related to the decision of going public, as the firm is required to provide information on a regular basis to their investors.

2.2 Underpricing

The Securities and Exchange Commission (SEC) (1963), Logue (1973) and Ibbotson (1975) were among the first researchers that investigated underpricing. All three studies showed positive initial returns for the U.S. IPO market in that period. Since then many studies followed. They showed that underpricing was present in almost every country and every time period.

Therefore, underpricing is seen as permanent and economically significant, especially so in certain periods. For example, Loughran and Ritter (2004) reported that, in the years 1999 and 2000, issuers in IPOs could see their stock price rise up to 65% on the first trading day. This extreme form of underpicing was due to the so called hot issue market that period. Especially internet-related firms were seen as very attractive by investors. This excessive underpricing represents a shocking loss to the issuing firm. In case of underpricing, the issuing firm leaves a substantial portion of its market value “on the table” when making access to the public capital markets.

Why is underpricing regarded as a puzzling phenomenon in the world of financial research? Why can it not be explained by a simple misvaluation of the market or a kind of risk premium for stocks? The significant level of underpricing found by numerous researchers, which can be seen in appendix I, makes it impossible to see it as a simple misevaluation. Ritter and Welch (2002) also give another motivation why underpricing can not be seen as just a premium for bearing the risk of the stock. According to them it is not rational that this substantial risk premium is required by first-day investors, who supposed to be diversified, but not by the second-day investors who are purchasing from the initial first-day investors. They assume that this fundamental risk is unlikely to be resolved in one day.

in theories based on institutional aspects and based on ownership and control considerations. In general, these theories are not mutually exclusive. Furthermore, a given reason can be more important for some IPOs than for others. Figure 1 shows the classification of Ljungqvist (2005) schematically.

Figure 1: The classification of underpricing theories

This study deals with the problem of asymmetric information between the issuer and the investors. Information asymmetry causes ex-ante uncertainty about the true value of an IPO. As a result, investors face the problem of assessing the value of the newly offered shares. This is especially difficult because no share price history exists for an IPO firm before the IPO. On the other hand, the so-called adverse selection models assume that issuers and their underwriter face a placement problem, because they don’t know what the true value is of the shares on offer. In other words, they don’t know what the market demand is for these shares. Next to the asymmetric information theories, the influence of investor sentiment on the underpricing of IPOs is investigated. These two theories are discussed in detail in sub sections 2.3 and 2.4 respectively.

2.3 Asymmetric information

shares must on average be underpriced in equilibrium. Otherwise, uninformed investors, buying new issues at random, would end up with a high proportion of overpriced equities because underpriced issues would be severely rationed. Presence of such a winner’s curse for uninformed investors can be demonstrated by showing that underpricing tends to zero when adjusted for rationing. Since most of the markets publish not enough allocation data to adjust underpricing for rationing, the evidence for the winner’s curse is only available for a small group of markets. For instance, Keloharju (1993) and Rogiers et. al (1993) show the presence of this winner’s curse in respectively Finland and Belgium where they use fixed-price rather than bookbuilding mechanisms. Both studies showed that an uninformed strategy in the associated market indeed approximately broke even.

Another key implication of the winner’s curse hypothesis is that initial excess returns are inversely related to ex-ante uncertainty about the true value of the newly offered shares. This was already proposed by Ritter (1984), but formalized for the first time in Beatty and Ritter (1986). Beatty and Ritter show in their model that underpricing serves as a compensation for costs made by investors to become informed. The higher the uncertainty about the true value of the shares is, the higher the costs will be to acquire the needed information. And this means, in turn, higher underpricing will be needed to compensate these costs.

Searching for ex-ante uncertainty factors, Beaver et al (1970) reported a strong positive correlation between systematic risk and individual risk components of a firm. This means that firms with large systematic risk are characterized by large variance in their individual risk as well. Clarkson and Thompson (1990) state that in the case of little information regarding a firm, this firm is perceived by investors as risky as a result of the lack of certainty regarding the exact parameters of the variance of expected return. They examined 198 IPOs in the period 1976 to 1985 and found that systematic risk (beta) for their sample decreased in several periods following to the offer date. The arrival of newly available information reduces the uncertainty about the IPO firm and its systematic risk. This relationship presents the correlation between systematic risk and the uncertainty about a firm. Knowing that the individual risk is an appropriate proxy for measuring ex-ante uncertainty of the IPO and that individual risk is correlated with systematic risk, it turns out that systematic risk is a good proxy for ex-ante uncertainty of the IPO. Therefore, a positive relationship with the level of underpricing can be expected. So, the larger the systematic risk of an IPO firm, the higher the ex-ante uncertainty regarding its true market value and consequently the higher the level of underpricing for that particular IPO.

In order to use systematic risk as a proxy for ex-ante uncertainty, it has to be observable. However, newly public traded firms with no history in the stock market, such as IPOs, lack many of the parameters which are needed to calculate systematic risk. As a consequence, we should look for ex-ante observable proxies for systematic risk. The study of Almisher and Kish (2000) documents a significant relation between accounting betas and the level of underpricing of an IPO. This implies a relation between accounting and market betas for the privately held firms in their sample. Therefore, they argue that accounting betas can be considered as an appropriate proxy for systematic risk of IPO firms.

2.3.1 Proxies for ex-ante uncertainty

Age and size

The most popular firm characteristic proxies are the ones used by Beatty and Ritter (1986): age and size of the firm. By examining a dataset of 545 IPOs conducted from the second quarter of 1981 till the end of 1982 in the United States, Beatty and Ritter (1986) find evidence that the underpricing of an issue increases in the ex-ante uncertainty surrounding the issue. As a proxy for ex-ante uncertainty they use the size (measured as the turnover in the year preceding the IPO) and the age of the floating company. They argue that the larger and older a company is, and thus is more “established”, the more information will be available, which reduces the uncertainty surrounding an issue. In line with Beatty and Ritter (1986), Brau, et. al (2004) argue that large firms may be followed by analysts more, are more likely to be in the news, sell products and/or services to more customers and usually have a longer history of performance and thus exhibit lower information asymmetry and lower risk and thus need lower underpricing. Besides sales, other company characteristics can be used to define firm size. Ritter (1984), for example, uses the book value of the equity, while Lowry and Schwert (2004) use the book value of the firm’s total assets.

Leverage

Another popular firm characteristic that serves as a proxy for ex-ante uncertainty is leverage. The more debt a firm has in its capital structure, the more highly leveraged the firm is considered to be. Several studies have examined the relationship between leverage and the level of underpricing. A negative relationship is found many times. According to Myers and Majluf (1984), a large pre-IPO leverage serves as a credible signal of firm’s quality because debt (with the threat of bankruptcy) imposes a hard budget constraint on managers, limits management’s control over firm’s cash flows, and raises the risk of firm’s undiversified stock ownership. Less valuable firms are unwilling to assume much debt because they are more likely to be forced into bankruptcy. This quality signal of a firm will decrease ex-ante uncertainty for investors and therefore they might face less underpricing.

In line with the above stated, Schenone (2004) shows that having a borrowing relationship at the time of the IPO, in the form of a bank loan, a credit commitment, or long-term debt from a source other than shareholders, was enough to lead to less underpricing. When a bank lends to a firm, the bank obtains valuable firm specific information which cannot easily be obtained by investors. This in fact causes a decrease in information asymmetry. As If banking relationships reduce asymmetric information, and if asymmetric information between the IPO players (the firm, the underwriting bank, and the market) is behind IPO underpricing, then firms with a pre-IPO banking relationship would face less of an asymmetric information problem than an otherwise equal firm would face. As a consequence these firms might exhibit lower IPO underpricing On the other hand, Su (2004) argues that a relative high pre-IPO leverage ratio increases ex-ante uncertainty about the financial strength of a firm. He does not consider debt financing for investment projects as a powerful choice for imposing hard budget constraints on top level managers. Therefore he regards a relative small pre-IPO leverage ratio as a positive signal for investors in the market. By analyzing 587 Chinese IPOs in the period 1994 – 1999 he finds a significant positive relationship between a firm’s pre-IPO leverage ratio and the level of underpricing.

Earnings

The Signalling Hypothesis argues that underpricing is an attempt by the issuing firm to inform (less-informed) investors about the quality of the future prospects of the firm. By deliberately underpricing an IPO issuers can signal the firm’s high value and return to the market on a later date on better terms. This intuition was first put forward by Ibbotson (1975), who stated that “underpricing new issues leaves a good taste in investors’ mouths so that future underwritings from the same issuer could be sold at attractive prices.” According to Allen and Faulhaber (1989) and Welch (1989), “good” firms try to distinguish themselves from “bad” firms by incurring a cost that less successful firms cannot profitably sustain. This cost is the underpricing of the new issue. Allen and Faulhaber’s model implies that the better firms with corresponding higher earnings will underprice more.

Industry

should exhibit higher initial returns on an IPO. Loughran and Ritter (2004) and Guidici and Roosenboom (2004) show that internet and high-tech IPOs (also referred to as the new economy IPOs) in the late nineties are significantly more underpriced than IPOs in other sectors due to the risky nature of these high-tech companies.

Share price level

Low offer prices are usually associated with retail investors, whereas high offer prices are said to attract more institutional investors (e.g. Stoughton and Zechner (1998) and Brennan and Franks (1997)). By making their shares more attractive to retail investors, i.e. choosing a low offer price, issuers increase the winner’s curse problem and consequently the required underpricing. Another reason why lower priced IPOs might be more underpriced is that lower prices are often used by smaller firms, which are regarded as more risky compared to larger firms (Jenkinson, et. al (2006)).

Brennan and Hughes (1991), on the other hand, argue that IPOs with low price levels should be less underpriced. They build a model that implies that lower issue price levels encourage analysts to produce more information, due to the decreasing fractional trading commissions in the share price level. So, lower prices mean less ex-ante price uncertainty and therefore lower underpricing. Next to that, Fernando, et. al (2004) find evidence for nonlinear relationship since they show that low priced IPOs as well as high priced IPOs show a high level of underpricing. Thus, medium priced IPOs are less underpriced. They attribute their finding for the low priced IPOs to the winner’s curse hypothesis, because of the large amount of retail investors subscribing in these IPOs. Their finding for the high priced IPOs is credited to the market feedback hypothesis. High priced IPOs are aimed to institutional investors who require a compensation for their information revelation.

2.4 Investor sentiment

As shown in section 2.3 there is a large body of evidence that the underpricing of an IPO is mainly a result of ex-ante uncertainty regarding the true value of the shares on offering. Underpricing is explained as a deliberate act required by rational investors to compensate the faced uncertainty. But the substantial increase in the initial returns in the late 1990s cast doubt whether information asymmetry theories could possibly be severe enough to explain underpricing on this scale (Ljungqvist (2005)).

Some researchers argue that behavioral theories, such as investor sentiment, are better suited to explain the IPO underpricing in the late 1990s. Behavioral theories claim that high and fluctuating initial returns are caused by irrational or sentiment retail investors who show strong interest in IPOs. Ritter and Welch (2002) posit that when small investors are overoptimistic, they are willing to pay a price exceeding fundamental value. Because the over-optimism of the retail investors can become obscured prematurely, carrying IPO shares in inventory is risky. This risk can be compensated by underpricing the shares.

Dorn (2003) also investigates the hypothesis which concerns the relation between retail interest in an IPO and the market conditions at the time of the IPO. He finds that following a period of high market returns or high returns in related stocks, retail investors are willing to place bolder bets, either because their portfolios have done well causing them to become more confident, or because they form extrapolative expectations, or both. A monthly market return one-standard deviation above its average is associated with 50% greater retail demand. Finding a relationship between market returns and underpricing suggests that underpricing is, at least for a part, driven by investor sentiment. This relationship has been widely studied in the recent literature and is proven to exist in many countries (Guidici and Roosenboom (2004) for several European countries and Lowry and Schwert (2004) for the U.S.). Dorn (2003) also shows that past returns in recent new IPOs are the single best predictor of retail investor interest in IPOs, which, in turn, is a good predictor of underpricing. So, underpricing is positively auto correlated in time.

more risky environment for issuance and behavioral models explain it by arguing that volatility can be seen as a proxy for the level of divergence of opinion among investors.

3 DATA AND METHODOLOGY

3.1 Data

In this study 1297 firms are considered that obtained a listing in the period from January 2001 till November 2010 on the three stock exchanges with the highest market capitalization in Europe. These stock exchanges are situated in France, Germany and the UK. This list of relevant IPOs is collected from Zephyr. From this initial number 589 firms are excluded for several reasons. First, IPOs without an offer price notation on Zephyr are excluded (428). Next to that, financial institutions (105) (i.e. banks, insurance companies, real estate investment trusts, holdings and financial services companies) are excluded, due to the different and more stringent accounting standards for these firms. Also excluded are firms already listed on another stock market (56). After these exclusions a total sample of 708 IPOs remains for analysis. A summary for the number of listings and exclusions is given in Table 3.1.

Germany France UK Total

Nr. of IPOs 316 303 678 1297 No offer price 129 25 274 428 Financial institutions 18 21 66 105 Dual listings 21 6 29 56 Total excluded 168 52 369 589 Used IPOs 148 251 309 708 Percentage of total 21% 35% 44% 100% Table 3.1 Number of listings and exclusions

3.2 Methodology

In well-developed financial markets with no share price fluctuating restrictions, the level of underpricing is calculated quickly (Ljungqvist, 2004). Most studies use the first-day closing price to determine the initial underpricing of an IPO. In formula form this gives:

price Offer price Offer price closing day First = ng Underprici (1)

As already indicated, various factors might influence the level of underpricing. Most of the authors claim that the level of underpricing depends on the level of uncertainty surrounding a going public company, so higher uncertainty results in higher underpricing. Based on the insights of the literature reviewed in the literature section several explanatory variables for underpricing are selected. Next to the company specific characteristics, which are mostly represented by financial accounting variables, market segment variables are included to control for investor sentiment. Below the explanatory variables that influence the level of underpricing are presented briefly.

3.2.1 Measuring explanatory variables

Age

Younger companies are seen as more risky. Investors need to be compensated for the greater ex-ante uncertainty. So, the expected sign for age is negative. The age of a company is calculated as the difference between the IPO year and the year of incorporation. If the company results from a merger, the incorporation year of the oldest company is used to determine the age. This method is the standard method used in the empirical literature. To control for the skewness of the variable age the natural log of age is used.

Size

Growth

To investigate whether the level of growth prior to an IPO influences the level of underpricing, the change in total assets from two years to one year prior to the IPO is also taken as an explanatory variable.

Offering Size

Offering size is defined as the total amount of shares offered at the IPO multiplied by the offering price of the IPO. This is also referred to as the gross proceeds of the offering. Higher proceeds are associated with a lower expected level of underpricing (Aggarwal, et. al, 2001).

Leverage

As discussed in the literature section, leverage can both increase and decrease the ex-ante uncertainty surrounding an IPO company. Therefore the expected sign is unknown. To measure leverage the ratio of total long term debt divided by total assets is used. This is in line with previous empirical studies on this topic.

Earnings

The Signalling Hypothesis argues that underpricing is an attempt by the issuing firm to inform (less-informed) investors about the quality of the future prospects of the firm. According to Allen and Faulhaber (1989) and Welch (1989), “good” firms try to distinguish themselves from “bad” firms by incurring a cost that less successful firms cannot profitably sustain. This cost is the underpricing of the new issue. Allen and Faulhaber’s model implies that the better firms with corresponding higher earnings will underprice more. As a consequence the expected sign is positive. Earnings are measured by a firm’s EBITDA divided by the total assets.

Industry

and the according FTSE Global Classification codes of the stocks in the technology and internet industry.

Subsectors _{FTSE Global Classification codes}

Technology

Computer hardware 932; 936

Communications equipment 673; 678; 938

Electronics 215; 216; 252; 253; 343

Navigation equipment 977

Measuring and controlling devices Included in other FTSE codes

Medical instruments 446; 482; 486 Telephone equipment 938 Communications services 542; 543; 545; 546; 547 Software 581; 974; 977 Internet Retailers e-Commerce 525 Internet 974

Table 3.2 Sub sectors and corresponding FTSE Global Classification Codes

IPO sentiment

Underpricing is found to be auto correlated in time (Lowry and Schwert (2004) and Dorn (2003). This can be explained by investor sentiment. Investors do react on market information. So, it is expected that underpricing is positively correlated to the preceding average level of underpricing in the market. IPO segment can also be defined as the number of IPOs in a preceding period. Here the expected sign is positive since months of high issue activity are generally associated with higher underpricing (Lowry (2003) and Loughran and Ritter (2004)).

So, two different measures for the IPO cycle are used in which an IPO is conducted. First, the average underpricing in the sample is calculated for the month prior to the IPO and second, the number of IPOs is calculated, also for the last month before the IPO. To calculate the number of IPOs, the same method as in Helwege and Liang (2004) is applied. The number of IPOs per month is calculated as a three-months moving average. For each IPO the number for the last month is taken. In formula form, the used number for each IPO is:

By using a moving average Helwege and Liang argue to avoid classifying seasonally low months (like January and August) as cold while they are in the middle of a neutral period. Using the number of IPOs that have to be conducted later in time is not in contrast with the condition that all the data used in the analysis has to be observable for potential investors. Since IPOs have to be filed with the financial market regulators well in advance, the market can form an opinion about the number of upcoming IPOs.

Share price level

Fernando, et. al (2004) find a relationship that is not linear. They show that low priced IPOs as well as high priced IPOs display a high level of underpricing. To capture this nonlinear pattern, both share price and its square are used as explanatory variables.

Market index return

Two different variables are used to test the influence of market returns on underpricing. First, the one-month buy-and-hold (prior to the IPO) return of the country’s corresponding MSCI index is calculated. Second, the 1-day return of the corresponding MSCI index is used as a variable to control for investor sentiment. According to Guidici and Roosenboom (2004), the Morgan Stanley Capital International (MSCI) index for the different countries is the most appropriate to use. This is a free float-adjusted market capitalization index that is able to measure the equity market performance in the different countries. There is a MSCI index available for all the three countries separately. Furthermore, this index describes the whole market in a country, and therefore this index makes the market adjustment more representative since it describes an alternative investment in the market (Hunger, 2003). For both of the variables, the expected sign is positive.

Market index volatility

The market index volatility is calculated as the standard deviation of the daily return of the MSCI index for the according country for the one-month period prior to the IPO. As found in previous literature, the expected sign is positive.

Variable Abbreviation Description

Expected sign Dependent variable

Underpricing return Difference between a stock's closing price and offer price on first day of trading Explanatory

variable

Age age Log of age of company on IPO date in years -

Size size Total assets of the company in IPO year-1 -

Growth growth (TA t-1 minus TA t-2) divided by TA t-2 ?

Offersize offersize Number of shares sold times the offer price -

Leverage leverage Long term debt divided by total assets in IPO year-1 ?

Earnings earnings Net operating profit in IPO year +

Technology dummy tech Equals 1 if the company belongs to the technology + sector, 0 otherwise

Internet dummy int Equals 1 if the company belongs to the internet + sector, 0 otherwise

IPO segment avUP Average underpricing 1 month before IPO +

IPO segment nrIPO Average number of IPOs in preceeding month +

Share price price Offer price of the shares on IPO date ?

Market return mrday 1-day return of MSCI index on IPO date +

mrmonth 1-month return of MSCI index prior to IPO date + Market volatility mrvol Standard deviation of daily return of 1 month MSCI +

index to IPO date

Table 3.3 The proposed variables and their expected signs 3.2.2 Regression analysis

To study the influence of the ex-ante uncertainty variables on underpricing, an Ordinary Least Square (OLS) regression analysis is conducted. A regression analysis studies the relation between a dependent and one or more independent variables and has the following form:

t ji j N j i x Y =α+

_∑

Underpricing = α +

β

β

β

β

β

β

+

β

β

β

β

β

`` 12

β

β

β

+

β

ε

3.2.3 Principal component analysis

In order to reduce the set of explanatory variables a factor analysis is conducted. Factor analysis is known as a multivariate statistical technique that is concerned with the identification of a possible structure within a set of observed variables. It involves the study of interrelationships among variables in an effort to define a new set of variables which is fewer in number than the original set of variables. So, factor analysis establishes dimensions within the data and might serve as a data reduction technique (Jackson, 1980). More specifically, the goal of factor analysis is to reduce “the dimensionality of the original space and to give an interpretation to the new space, spanned by a reduced number of new dimensions which are supposed to underlie the old ones or to explain the variance in the observed variables in terms of underlying latent factors” Thus, factor analysis offers not only the possibility of gaining a clear view of the data, but also the possibility of using the output in subsequent analyses (Rietveld and Van Hout, 1993).

There are several ways to conduct a factor analysis. In this study the principal component analysis is used. Strictly spoken this is not a factor analysis. The main difference between factor analysis and principal component analysis lies in the way communalities are used. Since factor analysis assumes error variance, the communalities have to be estimated. Therefore, factor analysis is more complicated. Despite of the different views on error variance, the two procedures often yield the same results (Field, 2005). For that reason principal component analysis is chosen in this study.

4 EMPIRICAL RESULTS

This section will first describe the data collected to study the underpricing phenomenon in Europe. In section 4.1 the descriptive statistics of the dataset are presented and the data characteristics will by analyzed. Secondly, in section 4.2 the regression analysis with the initial variables is shown. The principal component analysis and corresponding tests are presented in section 4.3 and finally in section 4.4 the regression analysis with the generated factors is discussed.

4.1 Descriptive statistics

In table 4.1 the dispersion of IPOs over the years in the sample is presented. It can be seen that there is no equal dispersion of the IPOs over the 10 years. In the first years of this decade a relative low number of IPOs took place. The shock on the (IPO) market at the end of the internet-bubble and the aftermath of 9-11 attacks could be an explanation for this. In the period from 2005 till 2007 about two-third of the IPOs of the sample took place. The climate on financial markets was comfortable and the shock of the dotcom-crisis and 9-11 was moved away. In the last three years there was a substantial downturn in the number IPOs. The uncertainty around the credit crisis and the corresponding high volatility made financial markets unattractive for IPOs.

Table 4.2 shows that the average underpricing in the total sample equals 11.34%. The corresponding t-statistic indicates that this percentage is highly significant (at the 5% significance-level). This finding confirms that underpricing is present on the main financial markets in Europe. With an average underpricing of 13.88%, France leads the other two countries. Germany and the UK show also highly significant underpricing levels (6.43% and 11.48% respectively).

Year Nr. of IPOs Percentage

2001 65 9.2% 2002 31 4.4% 2003 23 3.2% 2004 77 10.9% 2005 119 16.8% 2006 189 26.7% 2007 148 20.9% 2008 39 5.5% 2009 14 2.0% 2010 3 0.4%

Germany France UK Total N 148 251 309 708 Underpricing Mean 6.43% 13.88% 11.48% 11.34% Standard deviation 15.36% 73.15% 20.28% 46.36% T-statistic 5.001 2.995 9.958 6.541 P-value *0.000 *0.005 *0.000 *0.000

*, significant at the 1% level (2-tailed) Table 4.2 Underpricing in Europe and t-test results (test-value = 0)

The found percentages are in line with previous studies. Ritter (2009) found an average underpricing of 10.6% in France for the period from 1983 till 2009. Concerning the UK, Dimson (2008) found an average underpricing of 16.3% in the period 1959-2008. However, the average underpricing in Germany is lower compared to the study of Loughran et. al (2004). They found for Germany in the period 1978-1999 an average underpricing of 25.2%. This is not surprising since the sample considered in this thesis consists for the largest part out of IPOs conducted in a bull market during the internet-hype in which the average level of underpricing was extremely high3.

Variable Mean Standard Median Minimum Maximum

Deviation Age (years) 12.84 14.61 9.00 2.00 125.00 Size (million €) 506.82 3,752.31 20.72 1.35 14,837.5 Growth (%) 1.96 19.91 0.07 -26.28 383.06 Offersize (million €) 299.58 1,384.81 62.74 1.00 2,513.61 Leverage (ratio) 0.36 0.55 0.33 0.00 7.12 Earnings (ratio) 0.05 0.44 0.04 -1.10 1.43 avUP (%) 11.34 20.74 7.39 -12.36 198.66 nrIPO 12.25 6.48 12.00 1.00 23.00 Price (€) 10.30 40.39 4.98 0.81 189.12 Mrday (%) 0.11 1.03 0.12 -4.22 6.43 Mrmonth (%) 0.05 0.20 0.07 -0.63 0.80 Mrvol (%) 0.95 0.45 0.80 0.43 4.48

Table 4.3 Descriptive statistics of IPOs in the sample

From table 4.3 it can be seen that the average age of a going public firm is nearly 13 years. The oldest firm is 125 years and the youngest firm exists 2 years at the moment of the IPO. Also a great dispersion between the sizes of the firms is shown. The average size, measured in total

3

assets, equals 506.82 million Euros. Furthermore, the average offer size is highly skewed with a minimum of € 1 million and a maximum of approximately € 2.5 billion. The highest level of underpricing found in this sample is 198.66%. With respect to the offer price of the IPO shares, a great range is shown with a minimum of € 0.81 and a maximum of € 189.12.

4.2 Regression analysis with initial variables

This section will deal with the relation between the initial variables and underpricing. Besides the variables which are used in the factor analysis, two industry dummies are included. A technology (tech) dummy which equals one in the firm belongs to the technology sector and zero otherwise. For internet (int) firms the same procedure is followed. To determine which variables to exclude in order to reduce multicollinearity issues, a correlation matrix is constructed (appendix II). Except the logical high correlation between price and price² (ρ = 0.963), no extremely high correlations (ρ > 0.8) were found. A constant term is attached to the regression in order to assure that the mean of the residuals will be zero. To detect for heteroskedasticity, White’s (1980) heteroskedasticity-consistent covariance estimators are used. The regression results are shown in table 4.4.

Variable Coefficient Prob

C 0.291 0.681 Age -0.348 0.049 ** Size -0.182 0.082 * Growth 0.029 0.715 Offer size -0.018 0.860 Leverage 0.084 0.414 Earnings 0.218 0.098 * Tech 0.363 0.059 * Int 0.041 0.891 avUP 0.454 0.000 *** nrIPO -0.008 0.919 Price 0.024 0.877 Price² 0.086 0.088 * Mrday 0.25 0.000 *** Mrmonth 0.014 0.485 MRvol 0.126 0.653 Observations 708 F-statistic 10.112 R² adjusted 0.226 Prob 0.000

***, **,*, significant at respectively the 1%, 5% and 10% level (2-tailed)

The negative coefficient of age with a significant p-value of 0.049 is in line with the expected sign based on previous literature. Beatty and Ritter (1986) use both age and size as proxies for ex-ante uncertainty surrounding an IPO and find evidence for a significant negative relationship on underpricing. They argue that the older and larger the firm is, the smaller the level of underpricing of the IPO. In that perspective the negative coefficient of the size-variable is also consistent with their study. With a p-value of 0.082 the relationship of this determinant is significant at the 10% confidence level. Although the coefficient of growth is in line with the expected positive sign, this relationship is not significant. The same applies for offer size and leverage. Both relationships on the level of underpricing are in line with the existing literature, but are not significant. The ‘earnings-part’ of the signaling hypothesis, which was first put forward by Ibbotson (1975), is confirmed in this study. As indicated earlier the signaling hypothesis reads as follows: by underpricing an IPO the issuing firm can signal its high quality by incurring the cost of underpricing. The positive relationship of earnings one year prior to the IPO on the level of underpricing is significant at a 10% confidence level with a p-value of 0.098.

Looking at the industry dummies it is remarkable that no significant relationship is found for internet firms. This could be a consequence of the dotcom-bubble, because investors may be aware of the riskiness of these firms. On the other hand, for the technology dummy a significant positive relationship on the level of underpricing is found. This finding is in line with the study of Guidici and Roosenboom (2004) and Loughran and Ritter (2004). In both studies evidence is found that IPOs of technology firms are more underpriced than others. Their risky nature and corresponding higher ex-ante uncertainty are the underlying factors of this result.

and Ritter (2004). They claim that periods of high issue activity in so called ‘hot-markets’ are associated by a higher level of underpricing.

At last, the price² variable has a significant positive influence on underpricing with a p-value of 0.088. This means that the argument of Fernando et al. (2004) is confirmed, since they claim that high offer prices attract mainly institutional investors. This leads in turn to more media coverage an ultimately to higher demand and underpricing.

4.2.1 Country-specific regression analysis

In table 4.5 the country-specific regression analysis is shown.

_{Germany France UK} Variable Coefficent Prob Coefficent Prob Coefficent Prob

C 0.408 0.544 0.249 0.608 0.217 0.502 Age -0.283 0.041 ** -0.310 0.037 ** -0.133 0.224 Size -0.252 0.037 ** -0.222 0.063 * -0.872 0.016 ** Growth 0.119 0.372 0.089 0.627 0.142 0.205 Offer size -0.089 0.225 -0.302 0.065 * -0.044 0.315 Leverage 0.072 0.417 0.112 0.458 -0.089 0.274 Earnings 0.167 0.220 0.110 0.186 0.212 0.191 Tech 0.308 0.015 ** 0.473 0.007 *** 0.290 0.033 ** Int 0.156 0.187 0.241 0.071 * 0.108 0.184 Av_up 0.242 0.063 * 0.291 0.091 * 0.276 0.000 *** Nr_ipo 0.208 0.067 * 0.012 0.669 0.211 0.083 * Price 0.047 0.563 0.034 0.819 0.100 0.286 Price² 0.102 0.174 0.126 0.095 * 0.266 0.047 ** Mr_day 0.311 0.000 *** 0.254 0.000 *** 0.487 0.000 *** Mr_month 0.115 0.081 * 0.018 0.551 0.204 0.066 * mr_vol 0.129 0.425 0.097 0.516 0.099 0.321 Observations 148 _{251 309} R² adjusted 0.204 _{0.193 0.174} F-statistic 11.967 _{8.286 7.007} Prob 0.000 _{0.000 0.000}

***, **,*, significant at respectively the 1%, 5% and 10% level (2-tailed)

Table 4.5 Country-specific regression analysis

with a p-value of 0.047. Therefore, the non-linear relationship between price and underpricing of Fernando et. al (2004) applies the most in the UK. Furthermore, except the market volatility, all the market sentiment variables have a significant influence on underpricing in the UK. Therefore, for these data the influence of market sentiment is the highest in the UK.

With respect to the industry variables it can be seen that in France they play the largest role. Both the technology – as the internet dummy show a significant positive relationship. In Germany and the UK the internet dummy has no significant influence. Finally, it might be worth noting that only in France the offer size has a significant negative relationship on the level of underpricing.

4.3 Principal component analysis

As indicated earlier, with the aim to reduce the set of explanatory variables a principal component analysis is performed. The starting point of a principal component analysis is a correlation matrix, in which the intercorrelations between the explanatory variables are shown. With respect to this matrix, the following is important: the variables have to be intercorrelated, but they should not correlate too highly. Too highly correlated variables cause extreme multicollinearity and singularity (Brooks, 2002). The intercorrelations can be checked by using the Bartlett’s test of spherity, which tests the null hypothesis that the original correlation matrix is an identity matrix. When this is the case there are no correlations between the variables. Therefore, this test has to be significant (i.e. have a p-value of less than 0.05). As can be seen in appendix II, for these data the Bartlett’s test is highly significant with a p-value of 0.000 and as a consequence principal component analysis is appropriate. Multicollinearity, then, can be detected via the determinant of the correlation matrix. If this determinant is greater than 0.00001, then there is no multicollinearity (Brooks, 2002). For these data the determinant is 0.182. The correlation matrix with its corresponding determinant can also be found in appendix II

The variables that score high on the first factor all seem to be related to size and experience. However, the existence of ‘growth’ might be unexpected since growth firms are more often associated with younger firms. From now on the first factor will be called ‘size and maturity’. The second factor consists of the earnings ratio and the leverage ratio of a firm. Firms which experience higher earnings have a more levered capital structure. This factor will be named by ‘leverage and profit’. The third factor is called the ‘price’ factor. The last two factors are dealing with investor sentiment. The fourth factor is labeled by ‘medium-term sentiment’ and consists of three underlying variables. The average level of underpricing one month prior to an IPO is positively related to the market return in the month before the IPO whereas the market volatility is negatively related to the other two variables. The fifth factor will be called ‘short-term sentiment’, since the market return on the IPO date and the number of IPOs one month before are underlying.

4.3.1 Country-specific analysis

The same procedure as discussed is the section above is followed for the three different countries in the sample. To check whether the subsamples are large enough to conduct factor analysis, the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy is calculated. Values between 0.5 and 1.0 imply that factor analysis is appropriate, while values below 0.5 indicate that it may not be appropriate. For France and UK the KMO measure lies between 0.5 and 1. However, for Germany a KMO measure of 0.470 is given due to the smaller sample size (148) compared to

Factor Variable 1 2 3 4 5 Age 0.982 Growth 0.931 Size 0.911 Offersize 0.774 Earnings 0.989 Leverage 0.877 Price 0.910 Price² 0.910 Mrvol -0.740 Mrmonth 0.722 avUP 0.684 Mrday 0.692 nrIPO 0.660

France (251) and UK (309). As a consequence the results for Germany have to be interpreted carefully. In appendix III, IV and V the KMO results and the factor loadings for the different countries are presented. Except the differences in factor loadings, the factors consist of the same underlying variables for each of the countries. For that reason, the factor names used for the factors of the whole sample are also used in the regression analysis for the sub samples.

4.4 Regression analysis with factors

This section will deal with the regression analysis in which the relation between the produced factors and the level of underpricing is investigated. First the results of the regression on the whole sample are discussed and thereafter the regression results of the three countries are discussed in section 4.4.1.

The results of the regression analysis with the generated factors as explanatory variables and the level of underpricing as response variable are shown in table 4.7 below. The technology and internet dummies are included.

First of all, table 4.7 shows that the adjusted R-squared is substantial higher with a value of 0.422 than for the regression with the initial variables (0.226). Therefore, the generated factors have much more power in explaining underpricing compared to the initial individual variables. This higher value indicates that using principal component analysis is an appropriate tool for this data.

Factor Coefficent Prob

Constant 0.154 0.653

Size and maturity -0.228 0.019 **

Leverage and profit 0.097 0.454

Price 0.313 0.094 * Medium-term sentiment 0.321 0.000 *** Short-term sentiment 0.326 0.000 *** Tech 0.214 0.087 * Int 0.805 0.329 Observations 708 R² adjusted 0.422 F-statistic 3.994 (Prob) 0.000 ***

***, **,*, significant at respectively the 1%, 5% and 10% level (2-tailed)

Regarding the generated factors, it can be seen that the ‘size & maturity’ factor is negatively related to underpricing. This result is significant at the 1% confidence level with a p-value of 0.014. The factor includes the following underlying variables: age, growth, size and offer size. All these variables are positively related to each other. The negative sign of the factor is in line with previous empirical studies. Beatty and Ritter (1986) find that the older and larger the firm is, the less underpricing is found. They use size and age as proxies for ex-ante uncertainty. So, older and larger firms are surrounding with less ex-ante uncertainty and therefore the IPO is less underpriced.

The ‘leverage and profit’ factor is positively related to underpricing. However, this relation is insignificant with a p-value of 0.513. In previous literature the influence of leverage on underpricing is ambiguous, since both positive and negative relations are found. Su (2003) argues that a relative high pre-IPO leverage ratio increases ex-ante uncertainty, which in turn leads to more underpricing. In that perspective the positive sign of this factor can be explained. With respect to a firm’s earnings as the other underlying variable, the positive sign is in line with the signaling hypothesis of Allen and Faulhaber (1989) in which “good” firms try to distinguish themselves from “bad” firms by incurring a cost that less successful firms cannot profitably sustain. This cost is the underpricing of the new issue.

For the ‘price’ factor, which consists of the offer price and its quadrate, the sign is positive and significant at a 10% confidence level with a p-value of 0.091. This is in contrast with the findings of Jenkinson et. al (2006). They argue that lower prices are more often used by smaller firms which are regarded as more risky than larger firms. On the other hand, the positive relationship is confirmed by several other studies. Fernando et al. (2004), for example, find a positive relation due to behavioral factors.

sentiment’ factor may be the fact that influence of both the market return and the average underpricing offset the effect of market volatility.

Also for the ‘short-term sentiment’ factor a highly significant (p-value of 0.000) positive relationship with underpricing is found. The market return on the IPO date and the weighted number of IPOs one month before are underlying this factor. Both relationships are confirmed in previous studies. According to Guidici and Roosenboom (2004) underpricing is driven by investor sentiment. They also use the market return on the IPO date as a proxy for investor sentiment and find a significant positive relationship with underpricing. With respect to the second underlying variable the result is in line with the findings of Dorn (2003). He claims that underpricing is positively auto correlated in time.

Finally, it can be seen that for the industry dummies only technology firms face more underpricing. The positive relationship is significant with a p-value of 0.087. The fact that these firms are associated with a more risky environment might be an underlying reason for this finding.

4.4.1 Country-specific analysis

The results of the country-specific regression analysis with the factors are shown in table 4.8.

_{Germany France UK}

Factor Coefficent Probability Coefficent Probability Coefficent Probability

Constant 0.214 0.542 0.232 0.487 0.131 0.766

Size and maturity -0.188 *0.054 -0.293 **0.038 -0.142 0.294

Leverage and profit 0.111 0.651 0.068 0.799 -0.133 0.671

Price 0.104 0.369 0.335 *0.076 0.365 **0.037 Medium-term sentiment 0.267 ***0.000 0.194 0.109 0.408 ***0.000 Short-term sentiment 0.396 ***0.000 0.388 ** 0.035 0.237 ***0.008 Tech 0.215 *0.077 0.343 ***0.004 0.308 *0.059 Int 0.138 0.224 0.219 *0.095 0.098 0.312 Observations 148 251 309 R² adjusted 0.391 0.463 0.367 F-statistic 4.207 7.415 2.942 (Prob) ***0.000 ***0.000 ***0.000

As mentioned earlier the same factors are used as for the regression analysis of the whole sample. The technology and internet dummies are included. Due to the smaller sample size of Germany the generated results should be taken with care. As can be seen from table 4.8 there is similarity between the results of the different countries. For Germany the signs of the regression coefficients are equal to the signs for the whole sample. Size and maturity have a significant negative relationship with underpricing which is in line with other empirical evidence. Also in Germany the short- and medium-term investor sentiment have a highly significant positive influence.

Regarding the regression results of France it is remarkable that investor sentiment plays a smaller role in predicting underpricing. Although the coefficient of the medium-term sentiment factor is positive, the relationship is not significant. On the other hand, the positive influence of short-term sentiment, including the market return on the day of the IPO, is significant at the 5% confidence level with a p-value of 0.028. In accordance with the results of the main regression analysis, the negative influence of size and maturity on underpricing is also in France significant.

In line with the overall results, investor sentiment plays an important role in predicting underpricing in the UK. Both the coefficients of the short- and the medium-term sentiment factors are highly significant at the 1% confidence level. However, in the UK size and maturity do not have a significant influence on the level of underpricing. This result is not in line with previous literature where a negative relationship is claimed. Furthermore, the UK is the only country where the price factor does play an important role. The positive sign is significant at the 5% confidence level with a p-value of 0.032. Behavioral explanations could be underlying the results in the UK where investor sentiment and offer price are of great importance.

With respect to the industry dummies, the results are almost the same as in the regression analysis with the initial variables. Again it is shown that in France technology and internet firms are associated more with underpricing than in Germany and the UK.

5 CONCLUSION

R-squared and reduce the set of variables, principal component analysis is performed which generated five ex-ante uncertainty factors. These factors were used as explanatory variables in a subsequent regression analysis.

Most of the results of the initial regression analysis were in line with the expected signs based on previous literature. Age and size both have a significant negative influence on the level of underpricing since older and larger firms are expected to face less ex-ante uncertainty (Beatty and Ritter, 1986). When two industry dummies for technology and internet firms are added, it is shown that these firms are associated with more underpricing. This finding confirms the study of Guidici and Roosenboom (2004), who explain this by the more risky nature of these firms. With respect to the individual market segment variables, two highly significant relationships were found. Both the average underpricing in the market one month before the IPO and the market return on the IPO day have a positive influence on the level of underpricing. These results are in line with previous literature in which evidence is found for the presence of auto correlation of underpricing in time (Dorn, 2003).

In a country-specific regression analysis with the five factors, no significant differences are found. The signs of the coefficients were, except for the leverage and profit factor in the UK, the same in the three countries. In all three countries the size and maturity factor and the investor sentiment factors were found to be significant. As a conclusion, the regression model of the explanation of underpricing using factors of firm characteristics and market segment variables fits for the three largest stock markets in Europe.

Returning to the framework of Ljungqvist (2005) in figure 1, this study shows that both rational as irrational theories might explain underpricing. Parts of the rational asymmetric information theory between issuer and investor are confirmed, since some of the ex-ante uncertainty variables which are associated with information asymmetry have a significant influence on underpricing. Also, parts the irrational theory which is represented by investor sentiment variables are confirmed in this study. The two generated investor sentiment factors show in all the cases significant positive relationships with the level of underpricing.

Evaluating this study, some limitations have to be considered. First of all the German results should be interpreted with care, since the number of German IPOs in this sample might be too low. However, no significant differences were found, a few outliers on some explanatory variables could easily distort the results in a smaller sample. A second limitation could the fact that the different introduction methods of the IPOs are not taken into account in this study. Although, Frederikslust and Van der Geest (2000) find no evidence for a possible influence of the introduction method on underpricing, Eijgenhuijsen (1989) shows that it has an influence on underpricing. Another limitation might be the use of only one method of calculation of underpricing. Although there are several definitions of underpricing known in the literature on IPOs, most of the authors use the method used in this study. Nevertheless, other calculations of underpricing might lead to different results.

REFERENCES

Aggerwal, R., Bagath, S. and Rangan, S. (2009), “The impact of fundamentals on IPO valuation”, Journal of Financial Management 38, p. 253-284.

Aggarwal, R., Krigman, L. and Womack, K. (2002), "Strategic IPO Underpricing, Information Momentum and Lockup Expiration Selling”, Journal of Financial Economics 66 (1), p. 105-137. Beatty, R.P. and Ritter, J.R. (1986), “Investment Banking, Reputation, and the Underpricing of Initial Public Offerings”, Journal of Financial Economics 15, p. 213-232.

Booth, J. and Chua, L. (1996), “Ownership Dispersion, Costly Information, and IPO Underpricing”, Journal of Financial Economics 41, p. 291-310.

Brau, J.C., Brown, R.A. and Osteryoung, J.S. (2004), “Do Venture Capitalists Add Value to Small Manufacturing Firms?”, Journal of Small Business Management 42 (1), p. 78–92.

Brav, A. and Gompers, P.A. (1997), “Myth or Reality? The Long-Run Underperformance of Initial Public Offerings: Evidence from Venture and Nonventure Capital-Backed Companies”, Journal of Finance 52 (5), p. 1791–1822.

Brennan, M. and Franks, J. (1997), “Underpricing, Ownership and Control in Initial Public Offerings of Equity Securities in the UK”, Journal of Financial Economics 45, p. 391-413. Brennan, M.J. and Hughes, P.J. (1991), “Stock Prices and the Supply of Information”, Journal of Finance 46, p. 1665- 1692.

Brooks, C., (2008), “Introductory Econometrics for Finance”, University Press, Cambridge. ISBN: 0521873061

Chambers, D. (2009), “IPO underpricing over the very long run”, Journal of Finance 64, p. 1407-1438.

Clarkson, P. and Thompson, R. (1990), “Empirical Estimates of Beta When Investors Face Estimation Risk”, Journal of Finance 45, p. 431-453.

Derrien, F. and Womack, K. (2003), “Auctions vs. Bookbuilding and the Control of Underpricing in Hot IPO Markets”, Review of Financial Studies 16, p. 31-61.

Dorn, D., (2003), “Does Sentiment Drive the Retail Demand for IPOs?”, Working Paper, Drexel University, Philadelphia.

Eijgenhuijsen, H.G. (1989), “Aandelenintroducties op de Amsterdamse Effectenbeurs en het Verschijnsel Underpricing”, Maandblad voor Accountancy & Bedrijfseconomie 63, p. 119-127. Fernando, C., Krishnamurthy, S. and Spindt, K. (2004), “Are Share Price Levels Informative? Evidence from the Ownership, Pricing, Turnover and Performance of IPO Firms”, Journal of Financial Markets 7, p. 377-403.

Frederikslust, van, R., Van der Geest, R. (2000), “Rendementsontwikkeling van Private Equity Ondersteunde Beursintroducties”, Maandblad voor Accountancy & Bedrijfseconomie 74, p. 403-414.

Guidici, G. and Roosenboom, P. (2004), “Pricing Initial Public Offerings on Europe’s New Stock Markets”, Advances in Financial Economics 10, p. 25-59.

Ibbotson, R. (1975), “Price Performance of Common Stock New Issues”, Journal of Financial Economics 2, p.235-272.

Jenkinson, T. (1990), “Initial Public Offerings in the United Kingdom, the United States, and Japan”, Journal of the Japanese and International Economies 4, p. 428-449.

Jenkinson, T., Morrison, A. and Wilhelm, W. (2006), “Why Are European IPOs So Rarely Priced Outside the Indicative Price Range?”, Journal of Financial Economics 80, p. 185-209.

Keloharju, M., (1993), “The Winner’s Curse, Legal Liability, and the Long-Run Price Performance of Initial Public Offerings in Finland”, Journal of Financial Economics 34, p. 251-277.

Ljungqvist, A. (1999), “IPO Underpricing, Wealth Losses and the Curious Role of Venture Capitalists in the Creation of Public Companies”, Working Paper, Oxford University.

Ljungqvist, A., Nanda, V. and Singh, R. (2001), “Hot Markets, Investor Sentiment and IPO Pricing”, Journal of Business 79, p. 1667-1698.

Ljungqvist, A. and Wilhelm, W. (2003), “IPO Pricing in the Dot-Com Bubble”, Journal of Finance 58, p. 723-752.

Logue, D., (1973), “On the Pricing of Unseasoned Equity Offerings”, Journal of Financial and Quantitative Analysis 8, p. 91-103.

Loughran, T. and Ritter, J. (2002), “Why don't issuers get upset about leaving money on the table in IPOs?” Review of Financial Studies 15, p. 413-443.

Loughran, T. and Ritter, J. (2004), “Why has IPO Underpricing Changed Over Time?”, Journal of Financial Management Association 33, p. 5-37.

Loughran, T., Ritter, J. and Rydqvist, K. (1994), “Initial Public Offerings: International Insights,” Pacific-Basin Finance Journal 2, p. 165-199.

Lowry, M. and Schwert, G. (2004), “Is the IPO Pricing Process Efficient?”, Journal of Financial economics 71, p. 3-26.

Miller, R. and Reilly, K. (1987), “An Examination of Mispricing, Returns, and Uncertainty for Initial Public Offerings”, Financial Management 16, p. 33-38.