Underpricing at venture-capital-backed IPOs : looking at the differences in IPO underpricing between venture-capital-backed IPOs and non-venture-capital-backed IPOs in Europe from 2011 until 2017

(1)

Underpricing at venture-capital-backed IPOs

Looking at the differences in IPO underpricing between venture-capital-backed

IPOs and non-venture-capital-backed IPOs in Europe from 2011 until 2017

University of Amsterdam

Bachelor’s Thesis Finance & Organisation Topic: Finance

Author: Lennert Flens

Student e-mail: lennert.flens@student.uva.nl Student number: 10796665

Thesis Supervisor: dhr. dr. J.E. Ligterink Date of completion: 25-06-2018

(2)

Abstract

The goal of this thesis is to examine whether IPO underpricing at venture-capital-backed IPOs differs from IPO underpricing at non-venture-capital-backed IPOs. A sample of 2285 European IPOs was retrieved. 40 of the IPOs in the sample were venture-capital-backed. The regression output of a multivariate OLS regression showed that venture-capital-backed IPOs were significantly less underpriced than non-venture-capital-backed IPOs. Secondly, this research adds further empirical evidence to theories that attribute higher degrees of underpricing to IPOs in hot-issue markets.

(3)

Statement of Originality

This document is written by Student Lennert Flens who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(4)

Introduction

On June 13th_{Dutch electronic payments company Adyen brought 13.4% of their shares to the} market at an initial public offering (IPO). The shares were initially priced and offered at €240, but already after one hour of trading in the secondary market, the share price reached a value of €480. The market value of Adyen therefore doubled within an hour and it is therefore needless to say that the investors’ demand for shares in the young Dutch fintech company was extraordinary high (Sterling, 2018).

IPOs are called underpriced when the share price jumps substantially on the first day of trading (Ljungqvist, 2004). An underpriced IPO is costly for the issuing firm, because the existing shareholder leave money at the table at an underpriced IPO. The higher share price in the secondary market shows that the firm has raised less money than they could have raised at the IPO. Plenty of research has been done on the topic of IPO underpricing and research shows that even though underpricing is costly, firms can in fact have good reasons to underprice. Allen and Faulhaber (1989) state that underpricing can serve as a signaling device to investors of the positive expected returns of a firm. The willingness of a firm to underprice their stocks could show that the issuing firm is confident about earning back the loss of leaving money at the table in the future. The underpricing of the IPO of Adyen could therefore help reinforce a statement of the representatives of the firm, who boldly stated that they expect annual growth sales for Adyen between 25% and 35% in the coming years (Sterling, 2018). Some of the existing shareholders of Adyen were big venture capital firms such as General Atlantic, Index Ventures and Iconic Capital (Sterling, 2018). Venture capital firms handle large amounts of capital from private individuals, financial institutions and corporations, for whom it is costly as well when money is left at the table at an underpriced IPO (Silver, 1985). It can therefore be interesting to find out whether IPOs of firms that are backed by venture capitalists are more often associated with high degrees of underpricing such as at the IPO of Adyen.

Research Question

in the existing literature on venture-capital-backed IPOs there is no consensus on whether the presence of venture capital at an IPO leads to lower or higher degrees of underpricing. Krishnan, Ivanov, Masulis and Singh (2011) for instance found higher degrees of underpricing

(7)

at venture-capital-backed IPOs, whereas the sample of Belghitar and Dixon (2012) showed lower degrees of underpricing at venture-capital-backed IPOs. This research will therefore add to the existing literature by looking at the degrees of IPO underpricing of venture-capital-backed IPOs in a new research sample. The research question for this essay therefore is: Are venture-capital-backed IPOs subject to different degrees of IPO underpricing than non-venture-capital-backed IPOs? A sample of 2285 European IPOs between 2011 and 2017 was collected for this research. 40 of these IPOs were venture-capital-backed. The other 2245 IPOs were capital-backed. Average underpricing was higher for the non-venture-capital-backed IPOs. A multivariate OLS regression will be used to test whether the observed degrees of underpricing are significantly different.

The structure of the remainder of this paper is as follows. Chapter 1 consists of an overview of literature that is related to IPO underpricing and to IPO underpricing at venture-capital-backed IPOs in particular. There is a vast amount of literature on these subjects. Therefore, there was chosen to only include the most relevant theories and empirical findings. After the literature review the hypotheses that will be tested in this research are presented. This will be followed by a chapter on the methodology of this research. In this chapter the origins of and the alterations to the data that was used for the regressions will be explained, and the relevant characteristics of the data will be covered. In chapter 4 the regression results will be explained, and the results of the hypothesis tests will be presented and interpreted. Chapter 5 concludes with an overview of the thesis.

1. Related Literature

This chapter covers a review of the literature that is related to the subjects that are being discussed in this essay. First, an introduction to initial public offerings is given. This will be followed by a coverage of theories about the phenomenon of IPO underpricing. Secondly, the workings of venture capital firms will be explained followed by related theories on the effect of venture capital on IPO underpricing.

1.1. Initial Public Offerings

If a firm is in need of funds to finance an investment or to pay back debts, it can do so by selling firm shares. When these shares are sold to the public for the first time, it is called an

(8)

initial public offering (IPO) (Berk & DeMarzo, 2014). Three main parties are involved with an IPO. These are the issuing firm, the outside investors and the IPO underwriter (Berk & DeMarzo, 2014). The underwriter is often an investment banking firm that manages the IPO and takes care of the details of the offering. It for instance deals with choosing the number of shares that will be offered and with what the offer price will be. The offer price is the price each share is initially sold for in the primary market. (Berk & DeMarzo, 2014).

Often the underwriter and the issuing firm agree to a so called firm commitment IPO. In this case the underwriter buys all the shares for a price that is slightly lower than the offer price. The underwriter will subsequently try to sell the shares to the outside investors at the IPO at the offer price (Berk & DeMarzo, 2014). The underwriter can make a profit on the spread there is between the price they paid the firm for the shares and the offer price. However, the underwriter can also make a loss if it does not manage to sell all the shares to the public at the chosen offer price (Chen & Mohan, (2002).

An IPO can be an important milestone for young companies. Not only does the firm get access to capital from outside investors, also the existing shareholders get the opportunity to cash in by selling their shares to the public. One of the indirect benefits of going public is the positive public attention that an IPO might bring to a company. This could for instance allow the firm to attract the curiosity of high quality managers (Ljungqvist,2004). There is also one main disadvantage of going public. After an IPO a firm’s shares might be dispersed among a lot of different investors. The following lack of ownership concentration can make it more challenging for the company’s management to receive effective monitoring and screening by the shareholders (Berk & DeMarzo, 2014). This group of relatively anonymous shareholders might be less willing to actively assist the company’s decision-makers. Instead, they will most likely show their discontent by just selling their shares when they disagree with the management’s performance (Ljungqvist, 2004).

1.2. IPO Underpricing

IPO underpricing occurs when the share price jumps substantially on the issuance day of the IPO (Ljungqvist, 2004). Underpricing is therefore often measured as the percentage change from the offer price to the closing price at the end of the first trading day of an IPO. (Loughran & Ritter, 2004). Underpricing is costly for firms as they leave money at the table when the actual value of the shares at the end of the first trading day, turns out to be higher than the

(9)

offer price (Ljungqvist, 2004). Chen and Mohan’s (2002) regression results on a sample of 608 IPOs in the United States from 1990 until 1992 showed that IPO underpricing was negatively correlated with the reputation of underwriters. This could support the idea that firms going public indeed look for underwriters with a reputation of keeping underpricing low.

1.3. Theories of IPO Underpricing

When underpricing does occur, the informed investors who expect the share price to jump up after the offering would seem to be the only party that is able to benefit from IPO

underpricing (Ljungqvist, 2004). The following section shows however why underpricing can occur in the first place and how other parties than the outside investors can have reason to encourage underpricing as well.

One reason why IPOs are underpriced can be found by looking at the incentives of underwriters. Beaty and Ritter (1986) argue that individual firms are dependent on the underwriters to be able to bring their shares to the market in the first place. Individual firms cannot make a credible commitment to the market that their IPO is fairly priced, because their only interest would be their own profit. Underwriters are therefore hired because they are repeat players in the market and therefore they repeatedly have to deal with the interests of both the issuing firms as well as the interests of the outside investors (Beaty & Ritter, 1986). Because underwriters are repeat players, they have more information about market demand conditions of the outside investors compared to the issuing firm (Baron, 1982). The outside investors can demand a low offer price however, to cover their risk of overpaying for the shares. The offer price that best fulfills the demands of the outside investors can therefore be a different price than the offer price that best matches the actual value of the issuing firm. Because underwriters deal with the outside investors repeatedly but with the issuing firm only once, Ljungqvist (2004) states that underwriters attach more value to the interests of the outside investors than of the issuing firm that has hired them. Underwriters therefore have an incentive to choose the offer price that best matches investor demand, even though money will be left at the table by the issuing firm.

The offer price that is chosen by the underwriters, can therefore defer from the actual value of the shares. Once the shares are offered, the market will recognize the real value of the shares, therefore leading to underpricing when the actual value of the shares turns out to be higher than the offer price (Loughran and Ritter, 2004). This occurrence of underpricing

(10)

might even be reinforced by the fact that investors look for IPOs that are managed by underwriters that have a reputation of having underpriced IPOs in the past. By investing in an IPO that is managed by an underwriter with a reputation of underpricing IPOs in the past, investors hope to make a profit by buying shares that they expect to be underpriced as well (Beaty and Ritter, 1986).

These theories show that both the underwriters and the outside investors have reason to encourage IPO underpricing. What is quite surprising however, is that firms themselves can have good reasons to underprice their IPO as well. Outside investors often do not possess all the information there is about the quality of a firm going public. Firms can therefore deliberately underprice an IPO to use it as a quality signaling device. Outside investors know that only good firms can recoup from the losses that are made by leaving money at the table at an IPO. IPO underpricing can therefore show the outside investors that the firm going public is of high quality, which can be beneficial for the expected future performances of the firm after the IPO (Allen & Faulhaber, 1989).

Another reason for a firm to underprice an IPO is to be sure that there is sufficient investor demand for the IPO. Rock (1986) assumes that underpricing is unavoidable when trying to attract uninformed investors. In his so-called Winner’s Curse model, the assumption is made that some investors are better informed than others. Informed investors have the ability to only bid for IPOs that are attractively priced. The uninformed investors can make a loss however, when due to their lack of information, they invest in an IPO that turns out to be overpriced. When uninformed investors do manage to bid for an underpriced IPO, they will have to share their profits with the informed investors (Ljungqvist, 2004).

As the expected profits of the uninformed group can be negative due to the risk of overpricing, their total expected returns can turn negative (Ljungqvist, 2004). The uninformed investors have an incentive to leave the market when their expected returns are negative (Rock, 1986). This can be a problem for the issuing firm, as there might not be enough demand for an IPO when the uninformed investors have left the market. This problem can be solved by deliberately underpricing an IPO so that the expected returns of the uninformed investors go up until their expected returns are non-negative. This way the uninformed investors will stay in the market allowing the issuing firms to sell all the shares they offer at an IPO (Ljungqvist, 2004).

(11)

However, because firms only go public once, they will most probably only care about their own profits at the day of the IPO. Firms would have little reason to be bothered with whether overall IPO demand is high enough after their own IPO has taken place, as they have already gone public by then. It can therefore for be tempting for individual firms to free ride on the underpricing of others by eventually underpricing too little themselves (Ljungqvist, 2004). Thus, even though firms can have good reasons to let underpricing occur, the question still remains to what extent each firm will allow underpricing to happen. This research will therefore also add to the existing literature on IPO underpricing, by measuring the overall observed degrees of IPO underpricing on a new sample.

1.4. Venture Capital

Next to the issuing firm, the underwriters and the outside investors, sometimes other parties are involved with IPO underpricing as well. Studies show that venture capitalists are one of those players that can have an important effect on the observed degrees of IPO underpricing in the IPO market (Francis & Hasan 2001). In this section the activities of venture capitalists are briefly covered, followed by related literature on the effect of venture capital on IPO underpricing.

Venture capital firms have the ability to raise large amounts of capital from private individuals, financial institutions and corporations (Silver, 1985). They use this capital to specifically invest in the private equity of high potential growth firms (Berk & DeMarzo, 2014). Some examples of currently well-known and successful firms that obtained investments from venture capital firms when they were still considered as start-up companies are Apple, Microsoft, Intel, FedEx and Cisco (Cumming, 2010). Venture capital firms are set up as limited partnerships and the limited partners of venture capital firms are the investors. The managers of the venture capital firms are the general partners and they can also be called venture capitalists (Berk & DeMarzo, 2014).

Venture capitalists are often people with a lot of direct experience in the venture capital market (Silver, 1985). Therefore, in exchange for the limited partners’ capital,

venture capitalists can offer the limited partners their expertise. The venture capitalists goal is to offer the limited partners high returns and access to portfolio diversification opportunities through the investments they make into multiple high potential firms. Venture capitalists can make substantial earnings themselves as well by charging the limited partners

(12)

management fees between 1.5% and 2.5%. In addition, venture capitalists often also earn carried interest, which can be seen as a percentage of the positive returns of the venture capital fund (Berk & DeMarzo, 2014).

Successful venture capital investments can thus be beneficial to both the venture capitalists as well as the limited partners. Besides that, venture capital can also be very beneficial to the firms that receive the venture capital investments (Cumming, 2010). First of all, the capital that start-up firms receive, can help the company grow and reach their goals. What is especially characteristic to venture capital however, is the monitoring and advising role that venture capitalists play (Cumming, 2010). Venture capitalists help young firms reach success by assisting for instance in the development of high quality management teams or by improving the overall efficiency of the firm (Chemmanur, 2010).

1.5. Venture Capital Exits

Cumming (2010) states that venture capitalists are no long-term investors. This would mean that venture capitalists want to have an exit-option available within a couple of years following the initial investment in a start-up firm. When the venture capitalists exit from an investment it allows the limited partners and the venture capitalists to cash in on the return of the investment. Exiting a venture capitalist investment also allows evaluation of the results. A positive evaluation can play an important role in the follow-up fund-raising venture capital firms receive after an exit (Cumming, 2010)

There are multiple ways in which venture capitalists can exit an investment. The type of exit that is viewed as most profitable and that has been most studied however, is exiting through an IPO (Cumming, 2010). Empirical evidence from Gompers (1995) in the United States from 1961 until 1992 showed that the average annual rate of return for venture capital investments was 60% when the shares of the firm went public through an IPO. The annual rate on return over the same period in the same region of venture capital investments that exited through a trade sale, which is when the firm is sold to another company instead of to the public, was significantly lower at 15%. (Gompers, 1995).

1.6. Theories on the effect of venture Capital on IPO Underpricing

When a venture-capital-backed firm goes public, the venture capitalists sell all or part of the shares that the venture capital firm has in a company to outside investors in the primary

(13)

market. The position of the venture capitalist is therefore quite similar to that of the issuing firm. However, related literature shows that when a venture capitalist is involved with an IPO, additional factors might affect the observed degrees of underpricing. Therefore, this section covers several theories on the effect of venture capital on IPO underpricing.

As stated before, when investors are well informed, they feel that they run less risk of overpaying at an IPO. When the risk of overpaying is low, investors do not have to demand a low offer price to cover their risks and thus underpricing may be reduced (Fitza & Dean, 2016). Venture capitalists can reduce investors’ risk of overpaying, by informing the investors about the value of the firm that goes public in a couple of ways. First of all, venture capitalists undertake rigorous screening and evaluation procedures before they decide to invest in a firm in the first place (Cumming, 2010). Therefore, investors know that venture capitalists only back promising companies that have high expected future returns, so that the risk of overpaying for the stock is probably low (Belghitar & Dixon, 2012).

Secondly, the IPO of a venture-capital-backed firm is not always the end of the venture capitalists’ investment. Many venture capitalists retain equity in a firm after it has gone public. Loughran and Ritter’s (2004) empirical findings which showed that overall retained earnings ware higher at venture-capital-backed IPOs than at non-venture-capital-backed IPOs, support this statement. Typically, venture capitalists retain shares until between 90 and 180 days after the IPO. This period is called the lockup period. Venture capitalists often keep their positions on the board of directors after the IPO as well. This way they can continue to provide the firm with advice and access to capital, which might be very beneficial for the future prospects of the firm (Belghitor & Dixon). Thus, by retaining shares and continuing support to a firm after an IPO, venture capitalists can signal the possible positive future prospects of the firm to the outside investors (Fitza & Dean, 2016).

Additionally, contrary to issuing firms, venture capitalists are repeat players in the IPO market. They invest in start-up firms repeatedly and they bring multiple firms to the market every year. Venture capitalists therefore want to establish a reputation for accurately signaling the value of the firms they take public to the investors, as they probably will have to deal with the same outside investors in the future (Neus & Walz, 2005). This is a big difference with the issuing firm who is a one-time player in the IPO market. The outside investors in turn know that the venture capitalists want to maintain their reputation of accurately pricing IPOs to prevent outside investors from overpaying. They might therefore regard

(14)

venture-capital-backed IPOs as less risky than IPOs which are non-venture-capital-venture-capital-backed, leading to lower degrees of underpricing (Belghitar & Dixon, 2012).

Belghitar and Dixon (2012) performed a regression analysis on a sample of IPOs on the London Stock Exchange between 1992 and 1996. 217 of these IPOs were venture-capital-backed. The fact that in their regression analysis IPO underpricing turned out to be negatively correlated with the presence of venture capitalists could support the previous theories that state that venture capital investments can lead to lower degrees of underpricing. Lee and Wahal (2004) however, found results that state the opposite. In their sample consisting of data of 6,413 different worldwide IPOs between 1980 and 2000, the average first day returns of the venture-capital-backed IPOs was 26.8% versus 19.4% for the non-venture-capital-backed IPOs (Lee & Wahal, 2004). This shows that IPO underpricing was in fact higher at venture-capital-backed IPOs.

It is therefore interesting to also look at studies that explain why venture-capital-backed IPOs can be more underpriced than non-venture-capital-venture-capital-backed IPOs. First of all, contrary to Belghitar and Dixon (2012), Fitza and Dean (2016) state that high degrees of retained equity by venture capitalists after an IPO can in fact lead to higher degrees of IPO underpricing. The more shares venture capitalists retain, the less relevant IPO underpricing becomes for the venture capitalists’ return. When venture capitalists retain equity, it is the longer-term value after the lockup period that becomes more relevant. Venture capitalists would therefore be less bothered to accurately inform investors to prevent underpricing (Fitza & Dean, 2016).

Another theory on why venture capitalists have little incentive to prevent IPO underpricing is also based on venture capitalist reputation, but in a different way than as mentioned before. Lee and Wahal (2004) state that venture capitalist reputation depends on how many firms venture capitalists bring to the market every year. The future fundraising of venture capital firms would subsequently depend on this reputation. Venture capitalists would therefore have an incentive to spend less time on accurately pricing and signaling the value of each IPO to outside investors, so that they can take more firms public in a year. This way venture capitalists thus bear the costs of higher underpricing to be able to raise more capital in the future (Lee and Wahal, 2004). This thus contradicts with Neus and Walz (2005) who stated that venture capitalist reputation is based on how accurately venture capitalists signal firm value to the outside investors.

(15)

Before elaborating on the research of this paper, one more theory will be covered. Rosetto (2008) states that when in a certain period many IPOs take place, the average degree of IPO underpricing increases. These periods of economic expansion combined with many firms going public are called hot-issue markets. An extreme example of hot-issue markets can be found by looking at the US IPO markets from 1999 and 2000. In this period the average US IPO is said to have been underpriced by an average of 71% and 57% respectively. (Ljungqvist, 2004).

What is especially interesting, is that Rosetto also (2008) states that in hot-issue markets, venture-capital-backed IPOs experience higher degrees of underpricing compared to non-venture-capital-backed IPOs. During these hot-issue periods, many attractive investment opportunities arise for venture capitalists. Venture capitalists would therefore be willing to exit their current investments early to raise capital to invest in attractive new ventures (Rosetto, 2008). In Rosetto’s (2008) model it is therefore assumed that when the expected proceeds from a new venture are higher than those of an existing investment, venture capitalists might be willing to accept a low offer price for the shares they own in a firm. This way venture capitalists in hot-issue markets can quickly free up funds for a new investment opportunity by underpricing at an IPO.

2. Hypotheses

The literature that is related to IPO underpricing shows that there are different theories on why underpricing occurs. There also seems to be no consensus about whether venture capital leads to higher or lower degrees of IPO underpricing. This research will therefore add to the existing literature by empirically testing the effect of venture capital on IPO underpricing on a relatively new sample of both venture-capital-backed as well as non-venture-capital-backed IPOs. Lee and Wahal (2004), Loughran and Ritter (2004) and Krishnan, Ivanov, Masulis and Singh (2011) all found higher degrees of underpricing at venture-capital-backed IPOs compared to the IPOs that were not venture-capital-backed in their samples. Therefore, in the sample of this research also a higher degree of IPO underpricing is expected at te venture-capital-backed IPOs. The main alternative hypothesis that is tested in this paper therefore is:

(16)

H1: Venture-capital-backed IPOs experience higher degrees of IPO underpricing than non-venture-capital-backed IPOs.

Within Rosetto’s (2008) model two important assumptions are made. First of all, that IPO underpricing is higher in hot-issue markets in general. Therefore, a distinction is made within the regression of this research between hot-issue markets and non-hot-issue markets. This will make it possible to see whether there is a significant effect of hot-issue markets on IPO underpricing. The second hypothesis that therefore is tested is:

H2: IPOs in hot-issue markets experience higher degrees of IPO underpricing than IPOs in non-hot-issue markets.

Second of all, Rosetto‘s (2008) assumes that the presence of hot-issue markets affects venture-capital-backed IPOs in particular, due to the possible incentive for venture capitalists to early-exit investments in hot-issue markets when many new investment opportunities arise. Therefore, a separate regression is performed on solely the IPOs in the sample that were venture-capital-backed. This will allow to test whether venture-capital-backed IPOs in hot-issue markets are indeed more underpriced. The third alternative hypothesis that therefore is tested is:

H3: Venture-capital-backed IPOs in hot-issue markets experience higher degrees of IPO underpricing than venture-capital-backed IPOs in non-hot-issue markets.

3. Methodology

In this part of the essay the methodology of the empirical regressions for this research is covered. The hypotheses are tested by performing two OLS multivariate regressions that are comparable to the regression performed by Loughran and Ritter (2004). First the origins and relevance of the chosen data are explained. This is followed by a description of how the independent, dependent and control variables were defined. Thirdly, the regression model, the kind of estimation method and the performed statistical tests are covered. Lastly, there is a review of the descriptive statistics of the regression samples.

(17)

3.1 Sample Data

As of today, there is little existing literature on the degrees of venture-capital-backed IPO underpricing in the years after the financial crisis. Bessler and Seim (2012) looked at data from European IPOs from 1996 until 2010. This research will build on their findings by looking at European IPOs as well, but in the period from 2011 until 2017. The data used for this research is retrieved from ThomsonONE, containing all European IPOs from January 1st_{, 2011 until the} 31st _{of December 2017. The sample used for the regression consists of data of 2285 different} IPOs, as can be seen in Table I in the appendix. 40 of the IPOs were venture-capital-backed and the other 2245 were non-venture-capital-backed.

3.2 Dependent variables

The one variable that is key in all the alternative hypotheses that are tested in this research, is the observed degree of IPO underpricing for each IPO. IPO underpricing therefore serves as the dependent variable in the regression models of this empirical research. IPO underpricing is defined as the percentage difference between an IPO its offer price and the closing share price at the end of the first trading day (Heeley, Matusik and Jain, 2007). The IPO underpricing 𝑈𝑃_# of each firm i was therefore calculated as:

𝑈𝑃_# = 𝑃#,'(− 𝑃#,*( 𝑃_#,*(

To be able to assume an asymptotically normal distribution of variables, which is recommended for an OLS multivariate regression, the skewness of the variables has to be between -2 and 2 (West, Finch & Curra, 1995). However, 𝑈𝑃# showed a high level of skewness

due to the fact that there were some great outliers in the sample. Therefore, to reduce skewness, IPO underpricing was Winsorized. This was done by removing the biggest 3.11% of the observed values and replacing it with a maximum upper value of 158.82%. The lower values were kept intact as no extreme negative outliers of less than -100% were observed in the sample. An overview of the effect of Winsorizing on 𝑈𝑃# can be found in Figure I and Figure

II in the appendix. For the venture-capital-backed only sample Winsorizing was not necessary so the data was kept intact, as can be seen in Figure III in the appendix.

(18)

3.3 Independent variables

To test hypothesis H1: Venture-capital-backed IPOs experience higher degrees of IPO underpricing than non-venture-capital-backed IPOs, a distinction was made between the venture-capital-backed and non-venture-capital-backed IPOs. This was done by creating a dummy variable. The dummy variable had 1 for venture capital-backed IPOs and 0 for non-venture-capital-backed IPOs. The independent dummy variable that was created is:

𝑉𝐶𝐵 = .1,_0, 1 = 𝑉𝑒𝑛𝑡𝑢𝑟𝑒 − 𝐶𝑎𝑝𝑖𝑡𝑎𝑙 − 𝐵𝑎𝑐𝑘𝑒𝑑_{0 = 𝑁𝑜𝑛 − 𝑉𝑒𝑛𝑡𝑢𝑟𝑒 − 𝐶𝑎𝑝𝑖𝑡𝑎𝑙 − 𝐵𝑎𝑐𝑘𝑒𝑑}

Alternative hypothesis H2: IPOs in hot-issue markets experience higher degrees of IPO underpricing than IPOs in non-issue markets, and H3: Venture-capital-backed IPOs in hot-issue markets experience higher degrees of IPO underpricing than venture-capital-backed IPOs in non-hot-issue markets, were tested by looking at IPO underpricing in hot-issue-markets. Therefore, first it had to be clear when a market could be considered hot. Helwege and Liang (2004) defined hot-markets by looking at the underpricing of the IPOs itself. When the underpricing of a particular IPO in their sample was part of the highest 25% of underpricing of the whole sample, the market was to be considered hot. For this research, the same approach was used to decide whether a firm went public in a hot-issue market. The dummy variable had 1 for IPOs in hot-issue markets and 0 for IPOs in non-hot-issue markets. The independent dummy variable for hot-issue markets that was created therefore is:

𝐻𝐼 = A1,_0, 𝑥 = 𝐻𝑜𝑡 − 𝐼𝑠𝑠𝑢𝑒_{𝑥 = 𝑁𝑜 𝐻𝑜𝑡 𝐼𝑠𝑠𝑢𝑒}

3.4 Control variables

Fitza and Dean (2016) state that to measure the actual effects of venture capitalists on IPO underpricing, one should control for the fundamentals of the firm going public itself. A firm with strong and successful fundamentals might be able to signal its actual value and quality to the outside investors, independent of whether venture capitalists are involved or not (Fitza & Dean, 2016). Therefore, three control variables that relate to the characteristics of the firms going public were added to the regression as well.

(19)

First of all, IPOs of bigger firms could be regarded as less risky by investors. Bigger firms can be more transparent in providing information which could reduce the information asymmetries surrounding an IPO. Firm size was therefore included into the regressions as a bigger firm size might lead to lower degrees of IPO underpricing (Fitza & Dean, 2016). The control variable for firm size was defined as the value of the total assets of the firm before going public (Fitza & Dean, 2016).

The observed values of total assets were highly skewed. Therefore, to decrease the skewness of firm size, the observed values of firm size were transformed into the natural logarithms of the variables in both regressions (Feng, Wang, Lu , Chen, He, Lu , et al., 2014). By taking the natural logarithms of variables, no data is replaced with a certain minimum or maximum value such as is the case with Winsorizing. Therefore, if possible transforming variables into natural logarithms was preferred above Winsorizing in this research.

Secondly, Carter, Dark and Singh (1998) state that investors could associate bigger offers with financially strong and established issuers. This could reduce the risk investors might think they take by investing in a particular IPO, leading to possible lower degrees of underpricing. Krishnan, Ivanov, Masulis, and Singh (2011) therefore state that in many well-known IPO studies, IPO offer size is included into the regression as a control variable. Offer Size therefore is included in this research as a control variable as well. Offer Size is defined as the gross proceeds from each offering, exclusive of overallotment options (Krishnan, Ivanov, Masulis, and Singh, 2011). Once again, to reduce skewness the variables were transformed into natural logarithms in both regressions.

Lastly, the age of the issuing firms was included as a control variable in the regression as well. When a firm has existed for a longer period of time, more information about a firm is often available which can reduce the risk investors attribute to a certain IPO, also leading to lower degrees of underpricing (Fitza & Dean, 2016). Therefore, issuer age in days was incorporated as a control variable as well. Following related literature, the age of each issuing firm was calculated as (Krishnan, Ivanov, Masulis & Singh, 2011):

𝐹𝑖𝑟𝑚𝐴𝑔𝑒_# = 𝐼𝑃𝑂 𝐼𝑠𝑠𝑢𝑒 𝐷𝑎𝑡𝑒 − 𝐼𝑠𝑠𝑢𝑖𝑛𝑔 𝐹𝑖𝑟𝑚 𝐹𝑜𝑢𝑛𝑑𝑖𝑛𝑔 𝐷𝑎𝑡𝑒

Krishnan, Ivanov, Masulis, and Singh (2011) also state that they chose to use the natural logarithm of 1 + issuer age. For the full sample therefore, the natural logarithm of 1 + issuer

(20)

age was used. For the sample that only consisted of venture-capital-backed IPOs however, the skewness of Issuer age was higher after transforming the variables into natural logarithms of 1 + issuer age in days. Therefore, for the venture-capital-backed-only regression the regular values of issuing firm age in days were chosen instead of natural logarithms.

3.5 Regression Model

Now follow the estimation models that were used to estimate the effects of the chosen independent and control variables on IPO underpricing. The following equation is used for the OLS multivariate regression on the full sample:

𝑈𝑃# = 𝛽K+ 𝛽M𝑉𝐶𝐵#+ 𝛽N𝐻𝐼#+ 𝛽O𝐹𝑖𝑟𝑚𝑆𝑖𝑧𝑒#+𝛽R𝑂𝑓𝑓𝑒𝑟𝑆𝑖𝑧𝑒#+ 𝐹𝑖𝑟𝑚𝐴𝑔𝑒 + 𝜀#

To test hypothesis H3: Venture-capital-backed IPOs in hot-issue markets experience higher degrees of IPO underpricing than venture-capital-backed IPOs in non-hot-issue markets, a separate regression was performed. This second regression was performed on only the venture-capital-backed IPOs of the sample. Therefore, the dummy variable 𝑉𝐶𝐵# was

removed as all IPOs were venture-capital-backed, resulting in the following equation for the venture-capital-backed-only sample:

𝑈𝑃# = 𝛽K+ 𝛽M𝐻𝐼#+ 𝛽N𝐹𝑖𝑟𝑚𝑆𝑖𝑧𝑒# + 𝛽O𝑂𝑓𝑓𝑒𝑟𝑆𝑖𝑧𝑒# + 𝛽R𝐹𝑖𝑟𝑚𝐴𝑔𝑒# + 𝜀#

Currit (2002) states that the variables used in an OLS-regression do not all have to be perfectly normally distributed, but an OLS regression might work better if the variables are indeed asymptotically normally distributed. Therefore, if necessary the levels of skewness of the independent and control variables were reduced to values between -2 and 2 as well. This made it possible to assume that the variables used were acceptable for an OLS-regression (West, Finch and Curran, 1995). Table II and III in the appendix show the final observed levels of skewness of all the regression variables.

Secondly, for OLS regressions the assumption is made that no perfect multicollinearity exists among variables. Multicollinearity shows to what extent the independent variables and control variables are dependent on each other. When the degrees of multicollinearity are too high, the ordinary least squared method can predict false estimates of the parameters (Zhang

(21)

and Ibrahim (2005). Table IV shows the correlation degrees that were found in the full regression sample. The correlation levels of the variables of the venture-capital-backed-only sample are found in table V in the appendix.

Table IV

Full Sample Correlations

Correlations IPOUnderpricing VCBdummy HIdummy lnFirmSize lnOfferSize lnFirmAge

IPOUnderpricing_w 1 VCBdummy -0.0435 1 HIdummy 0.7611 -0.0276 1 LnFirmSize -0.0259 -0.0678 -0.0326 1 LnOfferSize -0.0954 0.051 -0.1322 0.7560 1 LnFirmAge 0.0584 -0.0833 0.0405 0.2417 0.115 1

Both correlations tables show that there is no perfect correlation between any set of two different variables, ensuring that no perfect multicollinearity exists within the full regression sample. The correlation between firm size and offer size is quite high however in both models at respectively 0.7560 and 0.7176. This does make sense however, as bigger firms are naturally able to go public with a bigger offer as well. What also stands out in both correlations tables, is that the dummy variable for hot-issue markets is highly correlated with the degree of IPO underpricing in both tables at 0.7611 and 0.6873 respectively. It will therefore be interesting to see what the regression results can show about the effect of the presence of Hot-Issue Markets on IPO underpricing.

When variables are highly dependent on each other, the standard errors of these variables will increase as well, making the variances of the coefficients in a model inflated. Daoud (2018) therefore states that the degree of multicollinearity can be measured by looking

(22)

at the inflation of the standard errors of the parameters. This can be tested by performing a Variance Inflation Factors (VIF) test. Therefore, within STATA a VIF test was performed on the standard errors of the sample variables. Daoud (2018) states that VIF-values above 5 can be regarded as highly correlated. All VIF-values and the mean VIF-value of the full sample and the venture-capital-backed-only sample of this research were between 1 and 5. This corresponds to moderately and therefore acceptable levels of multicollinearity. The exact results of the VIF tests on both samples can be found in Table VI in the appendix.

Lastly, before the OLS regression results are presented, Table VII follows which shows a summary of the values that were estimated for the full sample OLS regression.

Table VII

Summary Statistics Full Sample

Variable Observations Mean Std. Dev. Min Max

IPOUnderpricing_w 2.285 0.1282 0.3858 -0.9998 1.5882 VCBdummy 2.285 0.0175 0.1312 0.0000 1.0000 HIdummy 2.285 0.0411 0.1987 0.0000 1.0000 LnFirmSize 2.285 18.4967 2.6995 11.5129 27.9414 LnOfferSize 2.285 16.6033 2.2645 8.5172 23.3451 lnFirmAge 2.285 8.7140 1.0155 2.9444 10.6491

It for instance shows that underpricing indeed was present in the full sample and that, after Winsorizing, the average estimated degree of underpricing in the full sample was 12.82%. The highest degree of underpricing observed in this sample was 158.82% and the highest degree of overpricing was 99.98%. The average estimated degree of venture-capital-backed IPO underpricing was 0.24%. The highest degree of underpricing observed in this sample was 42.36 % and the highest degree of overpricing was 23.25%. The other statistics of the venture-capital-backed-only sample are found in Table VIII in the appendix.

4. Results and Analysis

This chapter will focus on whether the regression outputs support the hypotheses that were formulated for this research. The regression results are interpreted in a way that is similar to Brav and Gompers (1997). The first two hypotheses are tested by looking at the regression results of the regression on the full sample. The last hypothesis is tested by looking at the

(23)

results of the regression on the venture-capital-backed-only sample. This will be followed by a discussion of the remarkable and relevant other regression output characteristics. The regression output and characteristics are found in Table IX.

Table IX

Regression Results

This table contains the regression results of the OLS multivariate regressions. In the first column the names of the independent variables are found. The second and third column contain the estimates of the regression coefficients for respectively the full sample and the venture-capital-backed-only sample. Robust standard errors were used to ensure that the OLS-assumption of no homoskedasticity among variables was fulfilled. Following Gompers (2005), the p-values for the regression coefficients are in parentheses underneath the coefficient estimates. Full Sample IPOUnderpricing_w VCB Sample IPOUnderpricing Constant -0.024 0.265 (0.696) (0.204) VCBdummy -0.068*** (0.001) HIdummy 1.479*** 0.439*** (0.000) (0.000) lnFirmSize -0.004 0.004 (0.220) (0.759) lnOfferSize 0.004 0.023 (0.276) (0.095) lnFirmAge 0.011** FirmAge (0.023) 0.00001** (0.021) 𝑅N _0.5808 _0.5831 Adjusted 𝑅N _0.5799 _0.5355 F-Statistic 3753.12 277.04 Prob>F 0.000 0.000 * : p<0.10, ** : p<0.05, ***: p<0.01

(24)

4.1 Hypotheses test results

First of all, the coefficient on VCBdummy of -0.068 in the second column can be seen as an interesting result. The VCBdummy coefficient is highly significant at the 99% confidence level and is negatively correlated with IPO underpricing. This suggests that in this sample, even though the coefficient is only -0.068, the venture-capital-backed IPOs are in fact less underpriced than the non-venture-capital-backed IPOs. Thus, the estimation for this sample is that when an IPO is venture-capital-backed, IPO underpricing is 6.8% lower. This contradicts the empirical findings of Lee and Wahal (2004), Loughran and Ritter (2004) and Krishnan, Ivanov, Masulis and Singh (2011). In their samples higher degrees of underpricing were found at venture-capital-backed IPOs compared to the IPOs that were not venture-capital-backed. The null hypothesis of alternative hypothesis H1: Venture-capital-backed IPOs experience higher degrees of IPO underpricing than non-venture-capital-backed IPOs is therefore not rejected and cannot be replaced with the alternative hypothesis.

Secondly, Table IX shows that the coefficient on HIdummy of 1.479 in the second column is highly significant at the 99% confidence level. This suggests that, as expected following Rosetto’s (2008) assumptions, hot-issue markets are positively correlated with higher degrees of underpricing. The null hypothesis of alternative hypothesis H2: IPOs in hot-issue markets experience higher degrees of IPO underpricing than IPOs in non-hot-hot-issue markets is therefore rejected and can be replaced with the alternative hypothesis.

Lastly, the regression output shows that the coefficient on HIdummy of 0.4394 in the third column is highly significant at the 99% confidence level as well. These results suggest that hot-issue markets are positively correlated with the degrees of underpricing at venture-capital-backed IPOs in particular as well. The null hypothesis of alternative hypothesis H3: Venture-capital-backed IPOs in hot-issue markets experience higher degrees of IPO underpricing than venture-capital-backed IPOs in non-hot-issue markets is therefore rejected as well and can be replaced with the alternative hypothesis.

4.2 Other relevant and remarkable regression output results

Next to the coefficient estimates of the independent variables that were used to test the hypotheses of this research, some other interesting characteristics can be found in the regression output. The adjusted RN_{of both regressions of respectively 0.5799 and 0.5355}

(25)

were both considered sufficiently high enough to infer that the regression models had some explanatory power. Not all estimation coefficients were significant however. Firm size and offer size did not seem to have a significant effect on the observed degrees of underpricing in both samples at any confidence level above 90%. This contradicts with the existing literature on IPO underpricing where often is assumed that bigger firms and bigger offers are associated with lower degrees of underpricing (Krishnan, Ivanov, Masulis, and Singh, 2011).

The coefficients of firm age in days of 0.011 and 0.0001 respectively, were significant at the 95% confidence level in both samples. For the venture-capital-backed-only sample for instance, this would roughly suggest that each additional year a firm exists before it goes public, would in fact lead to a 365*0.00001 = 0.365% increase in the degree of IPO underpricing. This also contradicts with the existing literature, as older firms are often associated with lower instead of higher degrees of underpricing (Fitza & Dean, 2016).

(26)

5.Conclusion

In this thesis the degrees of IPO underpricing of a sample of Europeans IPOs were examined. Within the sample, a distinction was made between venture-capital-backed and non-venture-capital-backed IPOs. The motivation for this research was that the existing literature shows that there is no consensus about whether venture-capital-backed IPOs experience higher degrees of IPO underpricing compared to non-venture-capital-backed IPOs. The research question for this thesis therefore was: Are venture-capital-backed IPOs subject to different degrees of IPO underpricing than non-venture-capital-backed IPOs?

This research adds to the existing literature on venture-capital-backed IPO underpricing, by examining the degrees of IPO underpricing of a sample that has not been examined extensively before. The research sample consisted of 2285 different European IPOs that took place between January 1st_{, 2011 and December 31}st_{, 2017. 40 of these IPOs were} venture-capital-backed. A multivariate OLS regression was performed to test whether the venture-capital-backed IPOs in the sample experienced higher degrees of IPO underpricing compared to the non-venture-capital-backed IPOs. Furthermore, it was tested whether IPO underpricing was indeed higher in hot-issue markets. A second regression was performed on only the venture-capital-backed IPOs in the sample, to test whether the presence of hot-issue markets had a significant effect on the degrees of IPO underpricing of the venture-capital-backed IPOs in particular as well.

The regression output showed that the venture-capital-backed IPOs were significantly less underpriced than the non-venture-capital-backed IPOs. These results add more empirical evidence to the theories that claim that venture capital investments can lead to lower degrees of underpricing. Second of all, IPOs in hot-issue markets experienced significantly higher degrees of IPO underpricing in both the full sample, as well as in the venture-capital-backed-only sample. This adds empirical evidence to the existing literature that relates hot-issue markets to higher degrees of underpricing. Further notable empirical findings that followed from the regression output were that firm size and offer size were not significantly correlated to IPO underpricing in both samples. This is contrary to the existing literature as firm size and offer size are often incorporated as control variables when IPO underpricing is examined. Firm age did have a significant effect on IPO underpricing, but also contrary to existing literature, older firms showed higher instead of lower degrees of underpricing.

(27)

5.1 Limitations and Drawbacks

There were a couple of limitations to this essay. First of all, not all relevant variables that could affect the degree of venture-capital-backed IPO underpricing were included in the sample yet. The only way a distinction was made between venture-capital-backed IPOs and non-venture-capital-backed IPOs was by the creation of a dummy variable. The relevant literature on underpricing at venture-capital-backed IPOs however, shows empirical evidence that variables such as the reputation and the age of venture capitalists might affect the degrees of underpricing among venture-capital-backed IPOs as well (Lee & Wahal, 2004). The implementation of certain data could therefore add to the explanatory power of the regression models.

Secondly, in this research, only the observed degree of IPO underpricing was used to determine whether a market was to be considered hot. As a result, not all the characteristics that are distinctive to hot-issue markets were used. Hot issue markets are not only characterized by high degrees of underpricing, but also by a high concentration of firms that go public in a particular period of time (Helwege & Liang, 2004). More value could therefore be added to this research by incorporating the other characteristics of hot-issue markets as well when determining whether a market is to be considered hot at the time of a certain IPO.

5.2 Suggestions

More research can still be done on venture-capital-backed IPO underpricing. Overcoming the limitations from this research could be a starting point, but one could also search for other ways in which the degrees of venture-capital-backed IPO underpricing can be affected. Because underpricing was significantly lower at the venture-capital-backed IPOs in this sample, there is reason to investigate whether venture capitalists employed different strategies in comparison to periods where IPO underpricing was higher at venture-capital-backed IPOs. Moreover, one could also compare the results of this research to a sample of IPOs in a different market such as the United States. This will allow see whether there are differences in the observed degrees of underpricing among venture-capital-backed IPOs during the same period in different markets. If so, there could be reason to further investigate if and how venture capitalists in different markets employ different strategies that can affect the degrees of IPO underpricing at venture-capital-backed IPOs.

(28)

References

Allen, & Faulhaber. (1989). Signalling by underpricing in the IPO market. Journal of Financial Economics, 23(2), 303-323.

Baron, D.P., 1982, A Model of the Demand for Investment Banking Advising and Distribution Services for New Issues, Journal of Finance 37, 955-976.

Beatty, R. P., & Ritter, J. R. (1986). Investment banking, reputation, and the underpricing of initial public offerings. Journal of Financial Economics, 15(1-2), 213-232.

Belghitar, Y., & Dixon, R. (2012). Do venture capitalists reduce underpricing and underperformance of IPOs? Applied Financial Economics, 22(1), 33-44.

Berk, J., & DeMarzo, Peter M., author. (2014). Corporate finance (Third edition. Global ed.).

Bessler, W., & Seim, M. (2012). The performance of venture-backed IPOs in Europe. Venture Capital, 14(4), 215-239.pdf

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

Chemmanur, T. (2010). Venture capital, private equity, IPOs, and banking: An introduction and agenda for future research. Journal of Economics and Business, 62(6), 471-476. Chen, C., & Mohan, N. (2002). Underwriter Spread, Underwriter Reputation, and IPO Underpricing: A Simultaneous Equation Analysis. Journal of Business Finance & Accounting, 29(3-4), 521-540.

Cumming, D. (2010). Venture capital : Investment strategies, structures, and policies (The Robert W. Kolb series in finance). Hoboken, N.J: Wiley.

Currit, N. (2002). Inductive regression: Overcoming OLS limitations with the general regression neural network. Computers, Environment and Urban Systems, 26(4), 335 -353.

Daoud, J. I. (2018). Multicollinearity and Regression Analysis. Journal of Physics: Conference Series, 949(1).

Feng C, Wang H, Lu N, Chen T, He H, Lu Y, et al. (2014) Log-transformation and its implications for data analysis. Shanghai Arch Psychiatry, 26, 105–109.

(29)

Fitza, M., & Dean, T. (2016). How much do VCS and underwriters matter? A comparative investigation of venture capitalist and underwriter effects on IPO underpricing. Venture Capital, 18(2), 95-114.

Francis, B., & Hasan, B. (2001). The Underpricing of Venture and Nonventure Capital IPOs: An Investigation. Journal of Financial Services Research, 19(2), 99-113.

Gompers, P. (1995). Optimal Investment, Monitoring, and the Staging of Venture Capital. Journal of Finance, 50(5), 1461-1489.

Heeley, M. B., Matusik, S., & Jain, N. (2007). Innovation, appropriability, and the underpricing of initial public offerings. Academy of Management Journal, 50(1), 209-225.

Helwege, J., & Liang, N. (2004). Initial Public Offerings in Hot and Cold Markets. Journal of Financial and Quantitative Analysis, 39(3), 541-569.

Krishnan, C., Ivanov, V., Masulis, R., & Singh, A. (2011). Venture Capital Reputation, Post- Performance, and Corporate Governance. Journal of Financial and Quantitative Analysis, 46(5), 1295-1333.

Loughran, T., & Ritter, J. (2004). Why Has IPO Underpricing Changed over Time? Financial Management, 33(3), 5-37.

Ljungqvist, A. (2004). Chapter III.4: IPO Underpricing. In B. Eckbo, (Ed.), Handbooks in Finance: Empirical Corporate Finance (pp. 1-76). Amsterdam: Boston: North Holland/Elsevier.

Neus, & Walz. (2005). Exit timing of venture capitalists in the course of an initial public offering. Journal of Financial Intermediation, 14(2), 253-277.

Rock, K. (1986). Why new issues are underpriced. Journal of Financial Economics, 15(1), 187-212.

Shorack, G. (1996). Linear Rank Statistics, Finite Sampling, Permutation Tests and Winsorizing. The Annals of Statistics, 24(3), 1371-1385.

Rossetto, S. (2008). The price of rapid exit in venture capital-backed IPOs. Annals of Finance, 4(1), 29-53.

Silver, A. (1985). Venture capital: The complete guide for investors (Small business management series). New York Chichester: John Wiley.

Sterling, T. (2018, June 13). Payments Company Adyen Evokes Tech Bubble as Share Price Doubles on Market Debut. Retrieved from https://www.reuters.com/article/us

(30)

adyen-ipo/payments-company-adyen-evokes-tech-bubble-as-share-price-doubles on-market-debut-idUSKBN1J90QZ

West, S.G., Finch, J.F. and Curran, P.J. (1995) Structural Equation Models with Non-Normal Variables: Problems and remedies. In: Hoyle, R.H., Ed., Structural Equation Modeling: Concepts, Issues, and Applications, Sage, Thousand Oaks, 56-75.

Zhang, J., & Ibrahim, M. (2005). A simulation study on SPSS ridge regression and ordinary least squares regression procedures for multicollinearity data. Journal of Applied Statistics, 32(6), 571-588.

(31)

Appendix

Table I

Sample Statistics

The sample originally consisted of 19722 IPOs. From this sample duplicates and IPOs with missing information were removed. A sample of 2285 IPOs remained. This table shows which of these IPOs were venture-capital-backed and which were issued in a hot-issue market.

SAMPLE HOT-ISSUE MARKET

NON-HOT-ISSUE-MARKET TOTAL

VCB 2 38 40

NVCB 94 2151 2245

TOTAL 96 2189 2285

Figure I

Full Sample Distribution of IPO Underpricing before Winsorizing

The variable for IPO underpricing in the full regression sample showed a high skewness level. The skewness of 𝑈𝑃# was 38.28885. The IPO of Swedish company Eniro AB for example showed an increase in share price on the

first trading day from $0.009 to $21.31, which is equal to an increase in share price of 2366.77%. Figure 1 shows the distribution of IPO Underpricing before the observations were Winsorized.

0 .0 05 .0 1 .0 15 D en si ty

0.0000 500.0000 1.0e+03 1.5e+03 2.0e+03 2.5e+03 IPO Underpricing %

(32)

Figure II

Full Sample Distribution of IPO underpricing after Winsorizing

To reduce the skewness and to control for the extreme values in the sample, 𝑈𝑃# was Winsorized (Shorack,

1996). In the initial sample, the negative values of IPO underpricing were not smaller than -100%. Therefore, there was decided to keep the lower side of the sample data intact. Due to the extreme positive values in the sample data, the upper 3.11% of IPO underpricing was Winsorized by removing and replacing the maximum positive values with the upper value of 158.82%. The Skewness of IPO underpricing after Winsorizing was 1.896072. The following figure shows the distribution of IPO underpricing of the full sample after Winsorizing.

Figure III

VCB Sample Distribution of IPO Underpricing

The observed degrees of IPO underpricing of the sample that consisted of only venture-capital-backed IPOs, showed no great outliers and a skewness level of 1.1957. Therefore, the data did not have to be Winsorized to decrease the skewness level. Figure 3 shows the distribution of IPO Underpricing for the venture-capital-backed-only sample. 0 1 2 3 D en si ty -1 -.5 0 .5 1 1.5 IPO Underpricing % 0 1 2 3 4 5 Densit y -0.2000 0.0000 0.2000 0.4000 IPO Underpricing %

(33)

Table II

Degrees of Skewness of the Full Sample

After Winsorizing IPO underpricing and transforming Firm Size, Offer Size and Firm Age into natural logarithms, all the levels of skewness of the full regression sample were reduced to values between -2 and 2. This made it possible to assume that the variables were acceptable for an OLS regression (West, Finch and Curran, 1995). As the dummy variables VCB and HI could only take on the values of 1 and 0, their skewness levels were considered to be less relevant. The eventual levels of skewness that were found in the full sample are:

Variableable Skewnessewness

IPO underpricing_w 1.8961 VCB dummy 7.3582 HI dummy 4.6208 Ln Firm Size 0.2880 Ln Offer Size -0.0383 Ln Firm Age -1.0338

Table III

Degrees of Skewness of the VCB Sample

First of all, the VCB sample distribution in Figure III shows that the skewness of IPO underpricing is at an acceptable level. Secondly, the skewness level of Firm Age is also at an acceptable level at 0.3228. Therefore, for the venture-capital-backed-only sample, only the skewness levels of Firm Size and Offer Size had to be reduced by transforming the values into natural logarithms. Table III shows the observed skewness levels of the venture-capital-backed-only sample: Variable Skewness IPO underpricing 1.1957 HI dummy 4.1295 Ln Firm Size 0.3723 Ln Offer Size - 0.0966 Firm Age 0.3228

(34)

Table V

VCB Sample Correlations

Correlations IPOUnderpricing HIdummy lnFirmSize lnOfferSize FirmAge

IPOUnderpricing 1 HIdummy 0.6873 1 LnFirmSize 0.1444 0.3071 1 LnOfferSize -0.0082 0.2847 0.7176 1 FirmAge 0.1747 -0.1374 0.1168 -0.1201 1

Table VI

Results VIF Tests

Full Sample VCB Sample

Variable VIF Variable VIF

ln Issuing Firm Size 2.55 ln Firm Size 2.35

ln Offer Size 2.48 ln Offer Size 2.26

ln Firm Age at Issue 1.08 Hot-Issue dummy 1.14

VCB dummy 1.03 Firm Age 1.14

Hot-Issue dummy 1.03

Mean VIF 1.63 Mean VIF 1.72

Table VIII

Summary Statistics VCB Sample

Variable Observations Mean Std. Dev Min Max

IPOUnderpricing 40 0.0024 0.1254 -0.2325 0.4236

HotIssuedummy 40 0.0500 0.2207 0.0000 1.0000

lnFirmSize 40 17.1253 1.4471 14.3461 20.4480

lnOfferSize 40 17.4686 1.2320 15.0304 20.0291

Underpricing at venture-capital-backed IPOs : looking at the differences in IPO underpricing between venture-capital-backed IPOs and non-venture-capital-backed IPOs in Europe from 2011 until 2017