What is the effect of hiring an independent financial advisor on IPO underpricing

(1)

WHAT IS THE EFFECT OF HIRING AN INDEPENDENT

FINANCIAL ADVISOR ON IPO UNDERPRICING?

Master Thesis of Jordy Kroon

April 5, 2018

Name: Jordy Kroon

Student Number: 10790012

Supervisor: dhr. dr. J.K. (Jens) Martin

Master Finance

(2)

2

Statement of Originality

This document is written by Jordy Kroon, who declares to take full responsibility

for the contents of this document.

I declare that the text and the work presented in this document is 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 content.

(3)

3

Abstract

On average first-day positive returns in an initial public offering (IPO) are 24% in the United States during the period 1993-2008. Preceding literature studied the reasoning for underpricing of IPOs and classified these into the following theories and models: Asymmetric Information Theory, Behavioural Theory, Institutional explanations and Control Theory. Moral hazard and asymmetric information are the main reasons for underpricing. We will empiricially test the effect of hiring an independent financial advisor on underpricing. In line with previous literature, share overhang ratio and venture-backed capital dummy show significant positive relationships with underpricing. We conclude that financial advisors are not able to increase information transparency.

(4)

4

1. Introduction

On average first-day positive returns in an initial public offering (IPO) are 24% in the United States during the period 1993-2008. However, the market for underwriters is highly competitive. The link between underpricing and the competitive market is contradictive and leads to different research questions (Liu & Ritter, 2011). Preceding literature studied the reasoning for underpricing of IPOs, for example asymmetric information, laddering or spinning. The issue with the IPO market is that positive first-day returns are neither acquired by issuing firms nor the underwriter. Instead, these returns are obtained by the investors in the company.

Initial public offerings can be seen as the first time a company decides to sell its stock to the private investor in order to generate assets for company growth. Therefore, it is most likely that small firms undergo an initial public offering to seek capital. Firms hire an underwriter to obtain advise on the share price, amount of initial offerings, the timing and the issue type. A new trend is rising in the financial industry: firms hire an independent financial advisor for an initial public offering. Obviously, classic finance theories suggest that it is not possible that firms forego substantial amounts of money when going public, because according to the efficient market hypothesis firms are assumed to have all the information available before the issue date. Therefore, potential investors do not need additional compensation for uncertainty in the form of underpricing (Ritter & Welch, 2002).

The most logical and intuitive reason for underpricing is information asymmetry. Investors tend to be risk-averse and demand a risk premium in order to invest into a company (Loughran & Ritter, 2004). This theory suggests that information asymmetry is a result of an information gap between investors and the company itself. Therefore, it is beneficial to have some degree of underpricing in IPOs to keep the investors pleased (Rock, 1986). Furthermore, Akyol et al. (2014) argue that increasing the

standards of corporate governance would reduce the issue of information asymmetry. The corporate governance codes increased transparency for IPOs in member states of the European Union and the United States.

The asymmetric information theory will be used to explain through which channel an independent financial advisor could affect underpricing. According to Loughran and Ritter (2004), potential investors are the uninformed party and use underpricing to gather an extra income. This thesis will attempt to analyse if underpricing will drop by hiring an independent financial advisor. A potential explanation for the decrease in underpricing could be that issuing companies use the valuation report to make a commitment about the potential value of their business. This report could be used by the company itself to compare the company’s fair value estimated by third-parties with the price proposed

(6)

6 by the underwriters. Although many potential theories and models are trying to explain underpricing caused by underwriters or issuing firms, the asymmetric information theory is the most logical and intuitive theory to explain a potential effect of hiring an independent financial advisor on underpricing.

According to Akyol et al. (2012), the Sarbanes-Oxley act was passed by the US congress in 2002. The SOX was an important source of inspiration for the European Commission to make regulatory

changes. The European Commission invented the Report of the High Level Group of Company Law Experts on a Modern Regulatory Framework for Company Law in Europe1, which is similar to the Sarbanes-Oxley act regulations. Although the European countries adopted their own national corporate governance codes, they do line up with the guidelines as proposed by the European Commission.

This thesis seeks to analyse if underpricing could decline significantly by hiring an independent financial advisor when going public. Most of the data was obtained from the Thomson One Database except for family characteristics, underwriter quality, initial returns and returns at lock-up expiration dates. To calculate the returns at those specific dates, we used the option New Request in Excel by using Sedol codes. Another database called Osiris was used to gather data for the Family dummy. A company is controlled by a family when a firm’s shareholders consist of one or more individuals. Underwriter quality was ranked on a scale ranging from one to nine according to Carter and Manaster (1990). Following the activities of Rothschild in Europe as leading independent financial advisor, having advised on 82 IPOs since 2010, we made the choice to select the period 2010 to 2017. A total of 917 IPOs located in the United States, The Netherlands, Belgium, Italy and the United Kingdom were added to the dataset.

The main research question is: Are firms that hire an independent financial adviser during an Initial Public Offering experiencing less underpricing? In addition to the main research question, the paper attempts to identify the different sources of underpricing in IPOs. Existing literature will be reviewed to identify other sources of underpricing and certain control variables will be added to the regression line.

The thesis is structured as follows. Section 2 elaborates on the existing literature review on IPOs and defines the determinants for this phenomena. Section 3 describes the methodology and the hypotheses. Section 4 defines the data used, which are all the IPOs during the period 2010-2017, and the

construction of the sample. Section 5 shows the descriptive statistics and section 6 gives an

interpretation of the test results. Section 7 provides some robustness checks to indicate if the results are still valid after the robustness check is performed. Lastly, section 8 renders the conclusion of this paper.

(7)

7

2. Literature Review

Underpricing is the difference between the price of the initial offering and the closing price on the first day multiplied by the amount of shares sold. The puzzling phenomenon of positive first day returns is studied by several academics. However, the researchers are unable to rationalize this ongoing puzzle. According to Loughran & Ritter (2002) the issuers leave an averaging 9.1 million on the table during the period from 1990 to 1998. This amount of money is twice as large as the underwriter costs and could be seen as an indirect cost for the issuing company. Since the amount left on the table is twice as large as the underwriting costs, the effect of underpricing is not a compensation for risk (Grinblatt & Hwang, 1989). This phenomenon is well documented in the literature and many researchers currently attempt to explain this puzzle with theories and models. Ljungqvist et al. (2004) classified these theories and models into the following: Asymmetric Information Theory, Behavioural Theory, Institutional explanations and Control Theory.

First, the process of going public will be explained, followed by the different theories and models which attempt to fully explain underpricing. The role of hiring an independent financial advisor will be discussed. Lastly, preceding literature describing the effect of lock-up agreements on underpricing in IPOs will be pointed out.

2.1 IPO Process

When a company decides to go public in the form of an IPO, firms have to consider the fact that the process is very intensive and involves large consequences. The first consequence is reporting, because prior to an IPO most of the financial information was only available to private investors. But if a firm wishes to go public it should produce financial reports for both private and public investors (Berk & Demarzo, 2017). A company could have different reasons for going public. For instance, they could boost their reputation and image by getting more market recognition. Furthermore, the firm’s share value can be used for acquisitions and to motivate employees by compensating them in the form of stock bonuses. Moreover, by going public private investors could diversify their portfolio and increase the level of liquidity (Berk & Demarzo, 2017).

Jenkinson (2001) defined the process of going public in five steps. The first step for the issuing company is to select the market in which it wants to conduct an IPO. Companies can decide to sell their shares abroad in order to boost their reputation and image or to increase the level of liquidity. Following Jenkinson (2001), the next step is choosing an underwriter among all of the investment banks. According to Brau and Fawcett (2006), an issuing company selects their lead underwriter based on overall quality and status of the underwriter, quality and reputation of the research analyst and the underwriter’s level of knowledge about the industry. Multiple underwriters will be hired when the IPO

(8)

8 is too complex for a single underwriter, forming a so-called syndicate. According to Jenkinson (2001), issuing firms hire an underwriter to obtain advise on the share price, amount of initial offerings, the timing and the issue type.

As reported by Jenkinson (2001) the third step in the process is to develop a prospectus, which is a legal document required by the stock exchange. The prospectus contains information about the management team, history of the company, future expectations, risks and share valuations. Firms can attract potential investors by releasing the final prospectus, which is a form of advertisement for the company. The fourth step is to gather information as mentioned by Jenkinson (2001). The underwriter will visit trusted investors in order to gain information on the price range in the initial offering. This phase is called “pilot fishing” and results in a certain price range.

When the final prospectus and price range are known, “the road show” starts. During this phase the underwriter introduces the issuer’s management to the potential investors and takes non-binding offers. This process is also called “book building” and once the offers are finalized the offer price will be determined (Pwc, 2010). The final step is the allocation of shares granted to underwriters according to Jenkinson (2001).

2.2 Asymmetric Information Theory

2.2.1 The Winner’s Curse

Akerlof (1970) published an article about the Lemons problem in which the author explained the issue between uninformed and informed investors. Rock’s model is an extension of the Lemons problem. In Rock’s study (1986) the sample consists of an uninformed and informed group, causing the groups to have different expectations about the future share price of the issuing firm. Informed investors are only buying if they expect positive returns in the future. On the contrary, uninformed investors are not entirely sure whether they earn positive returns by investing into an issuing firm. Uninformed investors will not invest into initial public offerings in the long-term. This eventually leads to what is called ‘The Winner’s Curse’.

2.2.2 Signalling Theory

Another theory suggested that underpricing is used as a signalling mechanism to inform potential investors. Ibbotson and Jaffe (1975) were one of the first who argued that underpricing may be used to signal the firm’s prospects, yet it does not provide a solution for the underpricing puzzle. The idea of the signalling theory is to attract potential investors to a secondary offering by leaving the investors a “good taste in investors’ mouths” (Ljungqvist et al., 2004). Allen and Faulhaber (1988) contributed to the paper of Ibbotson and Jaffe (1975) by creating ‘good’ and ‘bad’ issuing firms in the sample. Their article states that the ‘good’ firms are willing to incur the signalling costs of the first tranche and

(9)

9 regain this in the second tranche. Although ‘good’ firms could create a trustworthy signal, ‘bad’ firms could lose the signalling costs of the first tranche if potential investors reveal their type of firm (Allen & Faulhaber, 1988).

2.2.3 Information Revelation

In general, informed investors reject to reveal information about the share price of the issuing company according to Ljungqvist et al. (2004), while issuing firms would like to know all of the information in advance. Therefore, the issuer must provide incentives to informed investors in order to obtain all of the information. Benveniste and Spindt (1989) introduced underpricing in a repeat game setting where informed investors share information in exchange for special rights in the allocations of shares, though a small margin of information is always required to be shared by informed

investors.(Benveniste & Spindt, 1989).

2.2.4 Principal-Agent Model

Most prominent in mentioning the potential of a principal-agent problem must be Loughran and Ritter (2004). Their model assumes that the underwriter has superior information over issuing firms. As mentioned before, the underwriter could have one of the following motives for underpricing: either potential investors may offer the underwriter an extra payment, or underwriters could give firms priority in the allocations of shares in order to obtain their future projects (Loughran & Ritter, 2004). The allocation of shares to different investors is also known as “spinning” (Loughran & Ritter, 2004). Loughran and Ritter (2004) also studied the “laddering hypothesis”, laddering is simply the agreement between investors and underwriters in which investors agree to buy shares in the aftermarket in exchange for access in the IPO. Underwriters use underpricing to create excess demand for their stocks to induce “laddering”.

Baron (1982) constructed a model in which the issuer (uninformed party) delegated the pricing task to the underwriter (investment banker). The investment banker was allowed to choose between

combinations of share prices and spreads. In this way, the issuing party optimized the unobservable utility of the underwriter (Baron, 1982).

2.3 Behavioural Theory

In spite of the asymmetric information theory, control theory and institutional explanations, many researchers argue that underpricing can not be fully explained. Therefore, Ritter and Welch (2002) supported behavioural explanations in future research. Behavioural finance focusses on cascades, investor sentiment and prospect theory.

(10)

10 According to Ljunqvist et al. (2004) a person is in a cascade if their actions are not influenced by their own thoughts. This means that investors’ actions are depending on historical data, which gives

potential investors more bargaining power. However, underwriters have the ability to keep information to themselves and influence the stream of negative information. Ljunqvist et al. (2004) also studied optimistic investor sentiment, discovering that optimistic investor sentiment is temporary and “smart” issuing companies should capture this effect by limiting the supply of shares. In a “hot” market investors tend to be over-optimistic and overvalue the shares. As a result, the issuer does not increase the supply to capture this price effect due to optimistic investor sentiment (Ljunqvist et al., 2004). Following Ljunqvist et al. (2004), IPOs in a “hot” market will experience more underpricing. Ritter (1984) studied the effect of investor sentiment on the aftermarket performance. The author measured aftermarket performance over a three-year period and concluded that long-term returns are negative. Loughran and Ritter (2002) used prospect theory to explain underpricing. According to this theory the level of wealth does not matter, while the change in wealth does concern the issuing company. Since the increase in share price exceeds the costs of under-pricing, the companies do not hesitate to arrange share offerings (Loughran & Ritter, 2002).

In summary, Ljunqvist et al. (2004) evaluated empirical evidence for the theories on cascades, investor sentiment and prospect theory. The author found supportive evidence for cascades and investor

sentiment. However, Ljunqvist et al. (2004) was ambiguous about the prospect theory.

2.4 Institutional explanations

Asymmetric information in the form of missing information in the prospectus or registration form could lead to litigation risk (Lowry & Shu, 2012). Litigation is the act where companies sue the issuing firm because of asymmetric information. Nevertheless, issuing companies could use underpricing as an insurance against litigation risk (Lowry & Shu, 2012). This is also known as the litigation-risk hypothesis. Another institutional explanation concerns tax benefits associated with underpricing. Of course, these tax benefits must outweigh the level of underpricing (Lowry & Shy, 2012).

2.5 Control Theory

2.5.1 Underpricing in order to retain control

According to Brennan and Franks (1997) managers try to avoid allocating shares among large investors. Because non-value maximizing behaviour could lead to a critical observation from large stakeholders. Small outside stakes imply that the performance of managers will not be monitored

(11)

11 frequently. Another reason to allocate shares to multiple investors is to reduce the chance of a hostile takeover. Besides, Brennan and Franks (1997) argue that managers could only discriminate in the allocations of shares if demand exceeds supply. Consequently, managers use underpricing to generate high demand for their shares.

2.5.2 Underpricing in order to reduce agency costs

Outside shareholders require an additional incentive to invest into a company in order to reduce agency costs in the form of lower offer prices. Following Brennan and Franks (1997) the agency costs could be reduced if managers lower their private benefits of control instead of maximizing them. Managers actually seek to maximize their private utility by increasing control benefits. However, managers are also to some extent owners of the company and should avoid underpricing in IPOs. Brennan and Franks (1997) concluded that managers must share all of the confidential information to reduce agency costs if the costs of their non-value maximizing actions outweigh their private benefits.

2.6 Independent financial advisors

The independent financial advisors are not hired only for the purpose of selling shares at fair prices, they also make an entire IPO readiness assessment (PwC, 2012). The advisors create the IPO plan, keep track of the plan, monitor the tasks listed, choose IPO delivery dates and gradually determine the fair share price (PwC, 2012). Going public is a complex event nowadays, causing increased demand for independent financial advisers. It seems that managers are not able to deal with a complex IPO. For example, the cost of going public exceeded the expectations of managers according to 48% of the issuing firms (PwC, 2012). The most important task for independent financial advisors is to manage the cost associated with a IPO.

NL Financial Investments (2015) suggested several ideas for ABN AMRO bank to successfully coordinate the costs associated with an IPO process. They found that it is common for issuers to accept a lock-up period after the initial public offering. Potential investors would like to know the amount of shares initially offered, which decreases uncertainty (NL Financial Investments, 2015). Another suggestion made by NL Financial Investments (2015) is to make use of the well-known tool

‘greenshoe’. This tool diminishes the short-term price volatility, because the underwriters are able to purchase up to 15% of the shares from selling shareholders in the first tranche. Lastly, NL Financial Investments (2015) advised issuing firms to sell between 15 and 30 percent of the total shares in the first tranche. Issuing firms could benefit from price fluctuations for the remaining shares. The suggestions made by NL Financial Investments (2015) decreases uncertainty around an IPO and this could influence underpricing in IPOs.

(12)

12

2.7 The Role of Lock-ups in IPOs

Most of the corporate insiders are subject to a lock-up agreement in which they ensure the public that the market will not be oversupplied (Courteau, 1995). Following Courteau (1995), a lock-up

agreement is a deal between the underwriter and pre-IPO investors to withdraw themselves from the market for a pre-specified period. After the lock-up expiration date, managers and pre-IPO investors are allowed to sell their shares. Courteau (1995) highlighted two motives for issuing companies to require a lock-up agreement. The first motive is to signal the quality of the firm whether information asymmetry exists, also known as the signaling hypothesis. The second motive is to reduce the unusual risk that managers take, also called moral hazard.

According to Courteau (1995), high-value companies could create a valuable signal to potential investors by entering into a lock-up agreement. This will result in higher proceeds from the IPO. Leland and Pyle (1977) invented a firm valuation model in which investors bought the shares of the insiders after the lock-up expiration date. Courteau (1995) extended this model by arguing that issuing firms could create a trustworthy signal by increasing the length of lock-up period. Moral-hazard problems could also be solved with lock-up agreement, this is called the commitment hypothesis. Courteau (1995) argued that managers could act in their own best interests instead of maximizing stakeholders’ value. Lock-up agreements could serve as a commitment device to attract potential investors.

(13)

13

3. Hypotheses and Methodology

In this chapter hypotheses will be derived from the existing literature. Afterwards, methodology used to answer the research question will be discussed in detail. Multiple regression analysis will be used to empirically test the hypotheses.

3.1 Hypotheses

Existing literature shows significant results after testing whether issuing firms are underpriced. For example, Neuberger and LaChapelle (1983), Beatty and Ritter (1986), Miller and Reilly (1987) estimated positive initial returns in the first week after the IPOs took place. By looking at the results of discussed literature the first hypothesis reads:

Hypothesis 1: IPOs have been fairly priced from 2010-2017

This hypothesis is rejected if we find significant results for over- or underpricing. According to the literature review we expect the offer prices to be underpriced. If hypothesis 1 could be rejected, then we start testing if independent financial advisors could affect underpricing in IPOs.

The asymmetric information theory is used to explain through which channel an independent financial advisor could affect underpricing. As mentioned in the literature review, potential investors could be considered as the uninformed party and require compensation in the form of underpricing (Loughran and Ritter, 2004). By hiring an independent financial advisor the company could ensure the

underwriters that most of the information is available to the public investor. Besides, the company is provided with an additional valuation report. This report could be used during the IPO process to compare the company’s fair value estimated by third-parties with the price proposed by the

underwriters. Next to estimating the fair value of companies, independent financial advisors also make an entire IPO-readiness assessment, provide specialized advice and keep track of the IPO process (PWC, 2012). After evaluating the information mentioned above the second hypothesis states:

Hypothesis 2: Underpriced IPOs are not affected by independent financial advisors

Upon potential rejection, we expect a negative sign for the dummy variable: independent financial advisor.

3.2 Methodology

This analysis attempts to define the relationship between the dependent and independent variables. Ordinary least squares regression will be used to empirically test the hypotheses. Stock and Watson (2012) specified four key assumptions of multiple regression analysis:

(14)

14 1. The error term must have a conditional mean of zero

2. The variables are independently and identically distributed 3. Large outliers are unlikely

4. There is no perfect multicollinearity

This type of regression is also used by Boulton et al. (2009), who studied the effect of international corporate governance on underpricing. Underpricing is the dependent variable, which is defined as the percentage of first day returns. Daily returns are calculated by offer price minus closing price, divided by the initial share price. Most of the papers use the percentage of first day returns as their dependent variable, like Loughran and Ritter (2002).

The variable of interest in this paper is a dummy, indicating 1 if a firm hires an independent financial adviser and 0 if the firm only takes advise of their underwriters. Several control variables will be used in order to clarify the relationship between the variable of interest and underpricing. The first control variable is “market capitalization” of a firm, which is used as a proxy for ex-ante uncertainty (Beatty and Ritter,1986). We take the natural logarithm of market capitalization because this is useful in describing relationships between certain variables, as the natural logarithm function creates a normal distribution. Another control variable is “underwriting fee per share” ,simply representing the costs of hiring an underwriter. The issuing company pays the gross spread to the underwriter, but the spread also contains payment for valuation uncertainty. Thus, the higher the uncertainty, the higher

underwriting fee per share (Loughran & Ritter, 2002).

To continue, the natural logarithm of “IPO proceeds” will be used as a control variable. Following Beneviste and Wilhelm (1990), there exists a negative relationship between “IPO proceeds” and underpricing, because of the fact that larger firms could be considered as value firms that are more stable over time. Since it is quite likely that older firms have more information available, there should exist less information asymmetry. Firm age will be added to the regressions, as it is likely that younger firms experience more underpricing as a result of information asymmetry according to Lowry and Shu (2002). The control variable “share overhang” will be used and is calculated by the total number of shares issued divided by number of shares in first or second tranche. We expect that a low ratio, implying that a high proportion of new shares offered, results in less underpricing (Habib & Ljungqvist, 2001).

Venture capital backed companies seem to have a negative effect on the level of underpricing according to Megginson and Weiss (1991). Venture capitalists monitor and advice the issuing company, raising the expectation that underpricing decreases. However, Bessler and Seim (2012) suggest the opposite, arguing that venture capital backed companies have a positive effect on the level of underpricing. A dummy will be created which is 1 if the company is venture capital backed and 0 otherwise. From the Thomson Reuters database the companies and their corresponding SIC codes

(15)

15 were downloaded. Following Loughran and Ritter (2004), internet and technological firms are more underpriced, creating necessity to control for this effect. Therefore, a technology-dummy will be added to the regression. To control for a potential time-trend in the data we added a dummy variable for each year. As suggested by Cliff and Denis (2004) the dummy variable for 2017 is removed from the data in order to avoid the problem of perfect multicollinearity. Nation dummies will be used because we are interested in analysing the impact of different nations in the regression. Therefore, by creating nation and year dummies, we add fixed effects to the regression.

The next control variable which is added to the regression is offer price. Miller and Reilly (1987) studied the effect of issue size on underpricing. They defined issue size in terms of offer price and argued that there must be a negative relationship between offer price and underpricing. According to Miller and Reilly (1987), uncertainty is related to the offer price: the higher the offer price the more prestigious underwriters are linked to the IPOs. Therefore, the uncertainty among investors will decrease and underpricing is expected to decline.

Lins et al (2013) constructed a family dummy to indicate whether a firm is controlled by a family. They used data concerning the firm’s ultimate owner to make a distinction between family and non-family controlled firms. Families are larger stakeholders and could influence underpricing. By adding this control variable to the regression we could study the potential effect of separating control and ownership in non-family controlled firms. Daugherty and Jithendranathen (2012) studied the effect of family-controlled firms on underpricing and found that the effectiveness of family-managers leads to less-underpricing.

Finally, Carter and Manaster (1990) created a ranking scale for each underwriter based on tombstone announcement in order to classify them as prestigious or non-prestigious underwriters. The

announcements are ranked from 0 to 9. An underwriter ranked 0 has no prestige and an underwriter ranked 9 has the most prestige. The classification for each underwriter was available at the website of Jay Ritter2_{. If an issuing company hired a syndicate for the IPO, then the underwriter with the highest} rank is considered in this analysis.

Additional information on the construction of the variables can be found in appendix A.

The following models will be used to evaluate the main research question. The first model:

Regression (1): Underpricingi = β0 + β1Independent Financial Advisor

Regression (2): Underpricingi = β0 + β1Independent Financial Advisor + β2Underwriting fee per

sharei + β3Share overhangi + β4(Firm age) +β5ln(Proceeds of IPO)i + β6ln(Market Capitalization)i +

(16)

16 β7(VC-backed dummy)i + β7(VC-backed dummy)I + β8(Technology dummy)I + β9(Offer Price)i + β10(Family-dummy)I + β11(Underwriter Quality)i + ui

Regression (3): Underpricingi = β0 + β1Independent Financial Advisor + β2Underwriting fee per

sharei + β3Share overhangi + β4(Firm age) +β5ln(Proceeds of IPO)i + β6ln(Market Capitalization)i + β7(VC-backed dummy)i + + β8(Technology dummy)I + β8(PE-backed dummy)I + β9(Offer Price)i + β10(Family-dummy)i + β11(Underwriter Quality)i + Σdt Nation Dummyt + Σdt Year Dummyt + ui To empirically test the effect of hiring an independent financial advisor on the returns at lock-up expiration dates, we calculated the returns for each company after the lock-up agreement expired. Positive returns at lock-up expiration dates could indicate that issuing firms set their offer price too low and thus leave money on the table. The option ‘New request in DataStream is used to gather all the returns at different expiration dates. SEDOL codes were obtained in order to find the returns for each company at each date. The new dependent variable are the returns at lock-up expiration dates. The regressions below have the same independent variable and control variables.

Regression (4): Returns(Lock-up expiration date)i = β0 + β1Independent Financial Advisor

Regression (5): Returns (Lock-up expiration date)i = β0 + β1Independent Financial Advisor +

β2Underwriting fee per sharei + β3Share overhangi + β4(Firm age) +β5ln(Proceeds of IPO)i + β6ln(Market Capitalization)i + β7(VC-backed dummy)i + + β8(Technology dummy)I + β8(PE-backed dummy)I + β9(Offer Price)i + β10(Family-dummy)I + β11(Underwriter Quality)i + ui

Regression (6): Returns (Lock-up expiration date)i = β0 + β1Independent Financial Advisor +

β2Underwriting fee per sharei + β3Share overhangi + β4(Firm age) +β5ln(Proceeds of IPO)i + β6ln(Market Capitalization)i + β7(VC-backed dummy)i + β8(Technology dummy)I + β9(PE-backed dummy)I + β9(Offer Price)i + β10(Family-dummy)i + β11(Underwriter Quality)i + Σdt Nation Dummyt + Σdt Year Dummyt + ui

Dependent variable:

-Underpricing: First-day returns Independent variable:

-Dummy: Indication for independent Financial Advisor Control variables:

-Share overhang ratio: Total shares outstanding after the issue divided by primary shares offered -Firm age: Issue date minus founding date

-Proceeds of IPO: natural logarithm of offer price multiplied by amount issued at IPO

-Market Capitalization: natural logarithm of total shares outstanding after issue multiplied by offer price

(17)

17 -Venture Capital Backed dummy: Indication whether the firm is financed by venture capital

-Technology dummy: indicator for technology firms

-Private Equity Backed dummy: Indication whether the firm is financed by private equity -Offer Price: Share price as indicated in the prospectus

-Family dummy: Dummy indicating whether a firm is controlled by a family -Underwriter Quality: Quality of underwriter ranging from 1-9

Robustness is a crucial requirement for these regression to make valid interpretations. The sample will be divided into subsamples based on issue date. Regression 1-3 will be re-evaluated for both groups. Finally, we use the option New Request in Excel to find the stock prices 7 days after the IPO took place. Underpricing will be calculated as the initial returns in the first week of the IPO. By using another definition of underpricing we tested if the results from regression 1-3 are robust. The last test to evaluate the robustness of regression 1-3 is by splitting the sample into a group with prestigious underwriters and un-prestigious underwriters. Prestigious underwriters are ranked eight or nine according to Carter and Manaster (1990), whereas un-prestigious underwriters ranked lower than eight.

(18)

18

4. Data

Most of the data was obtained from the Thomson One Database. Data was downloaded over the period 2010 to 2017. Following the activities of Rothschild in Europe as leading independent financial advisor, they have advised on 82 IPOs since 2010. Considering these activities, we made the choice to select the period 2010 to 2017. In total 917 IPOs located in the United States, The Netherlands, Belgium, Italy and the United Kingdom were selected. IPOs with an offer price of less than 5 euro were deleted from the dataset. Moreover, the dataset is also reduced by excluding firms with SIC codes 6000-6799, which are firms from the financial industry. The dataset from Thomson One contains financial information about the company and the IPO, including issue date, offer price, lock- up expiration date, primary shares offered, market value before and after IPO, number of underwriters, total underwriting fee, issuer’s name, nation, SEDOL codes, venture capital backed flag, private-equity backed capital flag, outstanding shares before and after IPO from prospectus.

The dummy variable for an independent financial advisor was not available in one of the databases at the UvA. Some of the information could be found in the IFRS comments, but this was far from sufficient to study the effect of hiring an independent financial advisor on IPO underpricing. We searched in the prospectus of each firm, which is a legal document offered by the issuing firm to inform potential investors, for companies who served as independent financial advisors during the IPO. Companies that took financial advice of independent third-parties or hired a company solely for the purpose of independent financial advice will get 1 for the dummy variable ‘independent financial advisor’. In practice, some of the companies hired underwriters which also provided financial advice. This was not classified as an independent financial advisor in this research.

We used the Osiris database to indicate whether a firm is owned by a family. The information in this database could indicate whether the firm’s shareholders exist of one or more individuals, corporate companies, bank and financial companies or other firms. As mentioned in the literature review, family companies are indicated by one or more individuals. Otherwise the company is not controlled by a family and therefore the dummy is 0. Cusip codes were used to match the data from Thomson one and Datastream.

According to Chen and Mohan (2002) underpricing and underwriter spread, which is the difference between the amount paid by the underwriters and the proceeds of the sale, are both determined in a simultaneous game setting. Subsequently, underwriter spread is endogenous and therefore influences the level of underpricing. Underwriter spread is the floating fee for the underwriters (syndicate). Following Chen and Mohan (2002), both high and low level reputation underwriters show a positive effect of initial underpricing on underwriter spread, whereas mid-level reputation underwriters show a negative effect of initial underpricing on underwriter spread. This is one of the many reasons why

(19)

19 companies hired as underwriters which also provided financial advice, will not be classified as an independent financial advisor in this research.

Furthermore, we used DataStream in Excel to gather the returns for each company in a specific period. First the initial returns were calculated by:

First − day Return =Closing price 1 day after IPO − Offer Price Offer Price

The option New Request in DataStream gives the opportunity to gather returns for companies at different periods in time. To avoid a potential error in DataStream the term “UK” was added to all SEDOL codes. With the new SEDOL codes and specific issue dates the closing prices were obtained from the database. Afterwards the option New Request was used once more, enabling lock-up

expiration dates insertion into the request table. To calculate the returns at lock-up expiration dates the following formula was used:

Return (Lock − up experation date)

=Share price 1 day before experation date − Share price 1 day after experation date Share price 1 day after experation date

Finally, we downloaded the returns for the first week after the IPO from Datastream. Again we used the option New Request and applied the SEDOL codes. The following formula was used to calculate the first week returns.

First − week Return =Closing price 7 days after IPO − Offer Price Offer Price

(20)

20

5. Descriptive Statistics

Table 1 describes the summary statistics and the distributions for the entire sample for the period 2010-2017. First-day returns show on average a positive return of 22.5% and a standard deviation of 34.3%. The results are within a range of -15.1% and 117.4%. The summary statistics results show that the IPOs are on average heavily underpriced. According to the statistics for the dummy variable for the independent financial advisor, 42.8% of all issuing firms obtained advice from a third party. Since 42.8% of all issuing firms hired an independent financial advisor, we conclude that IPOs are quite complex nowadays. Since almost half of the issuing firms hired an independent financial advisor, we may distinguish in our analysis between two groups. Underwriting fee per share is on average $0.407 and has a standard deviation of approximately 29.7%, which is an indication for high variance between firms. Share overhang averages on 5.56, meaning that the total amount of shares outstanding is 5.56 greater than the amount of shares offered in IPOs. We also noticed that underwriting fee per share ranges from one to thirteen, indicating a wide variation in share overhang ratio between the issuing firms. The maximum age of a firm is 35 years, so a couple of firms existed decades before they decided to go public. On average, the age of an issuing firm is 8.4 years and the minimum age is 1 year. Natural logarithm is used to make a normal distribution for proceeds, requiring these results to be interpreted differently. Furthermore, the sample shows that on average 49.1% of the issuing firms are financed with venture capital. Since almost half of the issuing firms are backed by venture-capital, this control variable needs some additional insights in the results section. The summary statistics for the technology dummy and private equity-backed dummy are approximately the same. On average, 27.5% of the issuing firms are internet/technology firms with a standard deviation of approximately 44.7%. Private equity backed firms represent 27.9% of all firms in the complete sample and like the

technology dummy the standard deviation is approximately 44.9%. Seeing that these results are in line with each other, we will address this comparison in the correlation section. The offer price is on average $14.5 with a standard deviation of approximately 5.302. An averaged 15% of all issuing firms are controlled by families, which is quite low when compared to other articles which used 2000-2015 as time-frame. Finally,underwriter quality varies between three and nine with a mean of 7.843. Interestingly, Carter and Manaster (1990) qualified an 8 or higher as prestigious underwriters.

(21)

21 Table 1: Summary Statistics for complete sample

The data was obtained from 2010 to 2017 for 916 observations. We selected a minimum offer price of $5 and focused on common shares. Data for first-day returns was obtained from Datastream and information on Underwriter’s Quality was downloaded from Jay Ritter’s website. The Family-Dummy was constructed by downloading stakeholder’s characteristics from the Osiris Database. The independent variable: Underpricing is calculated by first-day closing price minus offer price divided by offer price. The dummy-for financial advisor equals one if the issuing firms gathered independent valuations. Underwriting fee equals total underwriting fee divided by number of shares issued. Share overhang is the number of shares offered divided by total amount of shares outstanding. Age is defined as issue date minus founding date. The natural logarithm of proceeds is calculated by taking ln of offer price multiplied by primary shares offered. The natural logarithm of Market Capitalization is the share price multiplied by total amount of shares outstanding. Venture-backed capital is a dummy variable which equals one if the issuing firm is backed by venture capital. The technology dummy equals one if the issuing firm is a technology/internet firm and zero otherwise. The next dummy equals one if the firm is backed by private equity and zero otherwise. The offer price is the price at the IPO. The Family-dummy equals one if the firm’s stakeholders are one or more individuals and zero otherwise. Underwriter’s Quality is ranked by using the Carter and Manaster (1990) method, with one as the lowest possible value and nine the highest.

Variable:

Mean

Median

Std. Dev.

1%

99%

Obs.

First-day Returns

0.225 .121 0.343 -0.151 1.174 745

Dummy: Financial advisor

0.428 0 0.495 0 1 916

Underwriting fee

0.407 0.26 0.297 0.072 1 364

Share overhang

5.56 4.782 3.005 1.008 13.085 799

Age

8.425 6 9.051 1 35 916

Ln proceeds

13.109 13.821 2.623 6.745 16.366 916

Ln Market Capitalization

20.038 20 1.208 17.027 22.198 841

Venture-Backed capital

0.491 0 0.5 0 1 916

Dummy: Technology firm

0.275 0 0.447 0 1 916

Private Equity Backed

0.279 0 0.449 0 1 916

Offer Price

14.509 14 5.302 6 26 916

Family Dummy

0.15 0 0.357 0 1 916

(22)

22 Table 2: Summary Statistics for sub-sample by year

The data was obtained from 2010 to 2017 for 916 observations. We selected a minimum offer price of $5 and focused on common shares. Data for the independent variable was obtained from DataStream and was calculated by using the following formula.: first-day closing price minus offer price divided by offer price.

Table 2 represents the summary statistics for the dependent variable underpricing grouped by issue year. Underpricing fluctuates between 13.1% and 30.5% with the highest peak in 2013 and 2015. The investor sentiment theory could explain the relation between the amount of IPOs from 2013 to 2015 and underpricing. According to this theory, issuing firms must capture the positive investor sentiment by issuing their shares in the “hot” period. This over-sentiment will cause more demand for the shares and increases the first-day returns. However, the most important observation from this table is that there exists underpricing for each year in our sample.

Graph 1a shows the distribution of underpricing for the complete sample. Obviously, the distribution is skewed to the right and has approximately 50 observations with returns larger than 120%. The option Winsorize in Stata is used to eliminate potential outliers to create a histogram which is approximately normally distributed. Graph 1b represents the distributions of underpricing for the complete sample by issue date. The horizontal axis defines underpricing and the y-axis represents the frequencies for each issue date.

Issue Year Underpricing Median SD Skewness Kurtosis N

2010 0.175 0.081 0.315 2.17 7.277 61 2011 0.189 0.089 0.309 1.875 6.417 65 2012 0.235 0.138 0.317 1.478 4.75 80 2013 0.305 0.194 0.377 1.043 3.096 121 2014 0.234 0.115 0.357 1.483 4.367 147 2015 0.305 0.135 0.432 1.147 2.963 98 2016 0.161 0.077 0.268 1.697 6.038 61 2017 0.134 0.107 0.214 .814 3.339 82 2018 0.131 0.016 0.321 1.326 3.938 30

(23)

23

Graph 1a. First-day Returns by frequency. Graph 1b. First-day returns by Issue date and frequency. Data for first-day returns is obtained from Datastream from 2010 to 2017. The independent variable:

Underpricing (first-day returns) is calculated by first-week closing price minus offer price divided by offer price. Graph 1a is a histogram with first-day returns on the x-axis and frequencies on the y-axis. Graph 1b is the same as graph 1a, only the first-day returns are now represented by frequency as well as corresponding issue year.

Table 3a represents the summary statistics for the sample in which the dummy for the independent financial advisor equals 1. On average, underpricing is increased when compared to the results from table 1. If hypothesis 2 was rejected, we expected a negative sign for the dummy variable of

independent financial advisors. Yet these summary statistics indicate that by hiring an independent financial advisor underpricing will increase. The dummy variable venture-backed capital is increased on average by 18.5%. It seems like small emerging firms hire independent financial advisors during the IPO, which are labeled as high-growth firms. A possible explanation for the increase in

underpricing could be that venture-backed capital firms are more risky compared to value firms and therefore investors require additional compensation in the form of underpricing. This will be further examined in the results section of this analysis. The remaining variables in table 3 do not show significant changes compared to table 1.

0 2 0 4 0 6 0 8 0 1 0 0 F r e q u e n c y -.5 0 .5 1 1.5

First day Returns

0 5 1 0 1 5 2 0 0 5 1 0 1 5 2 0 0 5 1 0 1 5 2 0 0 .5 1 1.5 0 .5 1 1.5 0 .5 1 1.5 2010 2011 2012 2013 2014 2015 2016 2017 2018 F re q u e n c y

First day Returns

(24)

24 Table 3a: Summary statistics for independent financial advisor sample

The data was obtained from 2010 to 2017 for 916 observations. We selected a minimum offer price of $5 and focused on common shares. Data for first-day returns was obtained from Datastream and information on Underwriter’s Quality was downloaded from Jay Ritter’s website. The Family-Dummy was constructed by downloading stakeholder’s characteristics from the Osiris Database. The independent variable: Underpricing is calculated by first-day closing price minus offer price divided by offer price. The dummy-for financial advisor equals one if the issuing firms gathered independent valuations. Underwriting fee equals total underwriting fee divided by number of shares issued. Share overhang is the number of shares offered divided by total amount of shares outstanding. Age is defined as issue date minus founding date. The natural logarithm of proceeds is calculated by taking ln of offer price multiplied by primary shares offered. The natural logarithm of Market Capitalization is the share price multiplied by total amount of shares outstanding. Venture-backed capital is a dummy variable which equals one if the issuing firm is backed by venture capital. The technology dummy equals one if the issuing firm is a technology/internet firm and zero otherwise. The next dummy equals one if the firm is backed by private equity and zero otherwise. The offer price is the price at the IPO. The Family-dummy equals one if the firm’s stakeholders are one or more individuals and zero otherwise. Underwriter’s Quality is ranked by using the Carter and Manaster (1990) method, with one as the lowest possible value and nine the highest

Table 3b highlights the differences in the subsample based on the dummy variable: independent financial advisor. The summary statistics suggest that there exists a significant difference between the means in column 1. The difference between hiring an independent financial advisor and not hiring an independent financial advisor will be further examined with the use of a t-test and a t-test with Welch approximation.

Mean

Median

Std.Dev

1%

99%

Obs.

First-day Returns

0.251 0.165 0.331 -0.151 1.174 332

Dummy: Financial advisor

1 1 0 1 1 392

Underwriting fee

0.419 0.252 .298 0.098 1 157

Share overhang

5.514 4.742 2.923 1.2 13.085 360

Age

8.105 7 7.756 1 35 392

Ln proceeds

13.066 13.816 2.553 6.745 16.366 392

Ln Market Capitalization

19.99 20.001 1.036 17.538 22.198 372

Venture-Backed capital

0.676 1 0.469 0 1 392

Dummy: Technology firm

0.344 0 0.476 0 1 392

Private Equity Backed

0.179 0 0.383 0 1 392

Offer Price

14.532 14 4.81 6 26 392

Family Dummy

0.168 0 0.375 0 1 392

(25)

25 Table 3b: Summary statistics by independent financial advisor

The results from table 3c and 3d show that the difference in mean for the independent variable is not significant at a 5% level. However, it is significant if we apply a 10% significance level. The outcome of the two-sample t-test without Welch approximation is described in table 3c. The t-test value -1.8177 is smaller than -1.96 and is therefore not significant at a 5% level. However, we also need to apply Welch approximation because our results in table 3c could be invalid. The degrees of freedom decline from 743 to 727.262, but the results are still similar.

Table 3c: Summary statistics by independent financial advisor and two sample t-test

Diff = mean(0) - mean(1) t-value = -1.8177

Ho: diff = 0 Degrees of freedom = 743

Table 3d: Summary statistics by independent financial advisor and two sample t-test with Welch approximation

Dummy for Financial Advisor

Mean Median SD 95% Confidence Interval Obs.

0 0.205 0.093 0.352 0.171 0.239 413

1 0.251 0.165 0.331 0.215 0.286 332

Combined 0.225 0.343 0.200 0.250 745

Difference | -0.046 0.025 -0.095 0.003

Diff = mean(0) - mean(1) t-value= -1.8301

Ho: diff = 0 Welch's degrees of freedom = 727.262 Dummy:

Financial Advisor

Mean Median SD Skewness Kurtosis Obs.

0 0.205 0.093 0.352 1.716 5.132 524 1 0.251 0.165 0.331 1.254 4.089 392 Dummy for Financial Advisor

Mean Median SD Skewness Kurtosis Obs.

0 0.205 0.093 0.352 1.716 5.132 413

(26)

Tabel 4: Correlation matrix of dependent and independent variables

The data was obtained from 2010 to 2017 for 916 observations. We selected a minimum offer price of $5 and focused on common shares. Data for first-day returns was obtained from Datastream and information on Underwriter’s Quality was downloaded from Jay Ritter’s website. The Family-Dummy was constructed by downloading stakeholder’s characteristics from the Osiris Database. The independent variable: Underpricing is calculated by first-day closing price minus offer price divided by offer price. The dummy-for financial advisor equals one if the issuing firms gathered independent valuations. Underwriting fee equals total underwriting fee divided by number of shares issued. Share overhang is the number of shares offered divided by total amount of shares outstanding. Age is defined as issue date minus founding date. The natural logarithm of proceeds is calculated by taking ln of offer price multiplied by primary shares offered. The natural logarithm of Market Capitalization is the share price multiplied by total amount of shares outstanding. Venture-backed capital is a dummy variable which equals one if the issuing firm is backed by venture capital. The technology dummy equals one if the issuing firm is a technology/internet firm and zero otherwise. The next dummy equals one if the firm is backed by private equity and zero otherwise. The offer price is the price at the IPO. The Family-dummy equals one if the firm’s stakeholders are one or more individuals and zero otherwise. Underwriter’s Quality is ranked by using the Carter and Manaster (1990) method, with one as the lowest possible value and nine the highest. *, **, and *** indicate significance at 10%, 5%, and 1%, respectively.

(1) (2) (7) (8) (9) (10) (11) (12) (13) (1) First-day Returns 1.000

(2) Dummy: for Financial advisor 0.067* *

1.000 (0.07)

(3)Underwriting fee per share 0.007 0.035 (0.90) (0.51) (4) Share Overhang Ratio 0.138***

*** -0.014 (0.00) (0.70) (5) Age 0.093** ** -0.031 (0.01) (0.36) (6) Natural Logarithm of 0.014 -0.014 Proceeds (0.71) (0.67) (7) Market Capitalization 0.007 -0.036 1.000 (0.86) (0.30) (8) Venture-Backed Capital firm 0.196*** *** 0.320*** *** -0.225*** *** 1.000 (0.00) (0.00) (0.00) (9)Technology dummy 0.110*** *** 0.134*** *** 0.055 0.236*** *** 1.000 (0.00) (0.00) (0.11) (0.00) (10)Private Equity Backed firm -0.128*** *** -0.194*** *** 0.338*** *** -0.607*** *** -0.166*** *** 1.000 (0.00) (0.00) (0.00) (0.00) (0.00) (11) Offer Price 0.043 0.004 0.697*** *** -0.164*** *** -0.029 0.233*** *** 1.000 (0.24) (0.91) (0.00) (0.00) (0.38) (0.00) (12) Family Dummy 0.028 0.046 0.072** ** -0.026 0.098** ** -0.016 0.060* * 1.000 (0.45) (0.17) (0.04) (0.43) (0.01) (0.64) (0.07) (13) Underwriter Quality 0.034 0.139*** *** 0.507*** *** -0.006 0.093** ** 0.213*** *** 0.370*** *** 0.058 1.000 (0.43) (0.00) (0.00) (0.88) (0.02) (0.00) (0.00) (0.14)

(27)

The matrix highlights linear relationship between all the variables. Underpricing is positively linked to all other variables except for the private equity backed dummy variable. A possible explanation could be that private equity funds are driven by one of the largest pension funds and this party does not require additional independent financial advice. The most interesting result from table 4 is the positive correlation between the independent financial advisor dummy variable and the venture-backed capital dummy variable. The matrix indicates a positive relationship of 0.320 between those variables. As mentioned previously, it seems like small emerging firms (high-growth firms) hire independent financial advisors during the IPO. This relation will be further examined with the use of an interaction variable. In line with previous literature, share overhang ratio, age, venture-backed capital dummy, technology/internet dummy and private equity backed dummy show significant relationship with first-day returns. For share overhang we expected that a high ratio, implying a low proportion new shares offered, results in higher under-pricing (Habib & Ljungqvist, 2001). Age which was defined as issue date minus founding date. Since it is likely that older firms have more information available, there should exist less information asymmetry, thus underpricing should decline. However, in table 4 the correlation between age and underpricing is positive. Venture capital backed companies seem to have a negative effect on the level of underpricing according to Megginson and Weiss (1991). But in this analysis the correlation coefficient is positive and significant at a 1% level which is in line with Bessler and Seim (2012). Following Loughran and Ritter (2004), internet and technological firms are more underpriced which is in line with the results from table 4.

(28)

28

6. Results

Table 5 presents regression estimates for five ordinary least squares regressions. In column 1 the basic regression results are showed. The next column adds several control variables to the regression and column 3 adds the fixed effects. The adjusted R-squared increased from 0.004 in column 1 to 0.220 in column 3. Due to the fact that returns should be unpredictable, an adjusted R-squared value of 0.220 is definitely high. Besides the F-value in column 2 and 3 findings are statistically different from zero and thus the variables explain the variation in underpricing.

Column 1 shows a positive significant effect of the dummy variable for independent financial advisor on underpricing at a 10% significance level. However, these results need to be interpreted with care due to an adjusted R-Squared value of 0.004 and an insignificant F-value. On the contrary, the results in column 2 show an altogether insignificant relationship between the dummy variable for independent financial advisor and underpricing. Column 3 repeats the regression in column 2 and adds the fixed effects. Surprisingly, adding this variable to the regression in column 2 renders the effect of the dummy variable for independent financial advisor on underpricing positive and significant at a 10% level.

Other variables such as share overhang, age, offer price and venture-backed capital dummy are significant at a 5% level in column 2. Moreover, the family-dummy is significant at a 10% level. The results from share overhang, venture-backed capital and offer price are in line with previous literature. Whereas age and family dummy are not in line with existing literature. The effect of age on

underpricing is significant but very small. Therefore, the results of age would not be further examined since the coefficient is 0.006.

As mentioned before, fixed effects were added in column 3 to the regression from column 2 and the effect is surprisingly positive and significant at a 10% level. We would like to further examine if there is an interaction between all the significant variables and the dummy for financial advisor. From table 4 we found that the following variables were significantly correlated with financial advisor: venture-backed capital, private-equity venture-backed capital, technology/internet firm and underwriter’s quality. Yet the results from table 5 only show a significant effect of venture-backed capital on underpricing.

We the interaction term venture-backed capital with financial advisor in column 4 to the regression from column 3. In the last column we added share overhang, age, venture-backed capital and family-dummy with financial advisor to the regression from column 3 to see if there is an interaction between those variables. The regressions from column 4 and 5 show coefficients which are all insignificant, except for the family-dummy. We could now argue that the dummy for financial advisor has no interaction with other significant variables in this analysis.

(29)

29 Column 6 repeats the regression from column 3 and focuses on issue dates ranging from 2012 to 2015. These are the years with higher underpricing and table 5 contains interesting results. The coefficient for financial advisor increased from 0.087 to 0.139. Which means that underpricing increased 5.2% on average when the dummy for financial advisor indicated 1 compared to the results from column 3. Share overhang and age are not significant anymore at a 10% level. Besides, the venture-backed capital dummy is now significant at a 1% level instead of a 5% significance level. On the contrary, offer price is now significant at a 5% level. The adjusted R-squared increased from 0.221 to 0.346.

Table 5: Regression complete sample

Column 1 represents is a regression with dummy Financial advisor as dependent variable. Column 2 represents a regression with dummy Financial advisor as dependent variable and the control variables. Column 3 represents a regression with dummy Financial advisor as dependent variable, control variables and fixed effects. Column 4 and 5 contains also interaction terms. Column 6 repeats the regression from column 3, but focuses on the period from 2012 to 2015. The data was obtained from 2010 to 2017 for 916 observations. We selected a minimum offer price of $5 and focused on common shares. Data for first-day returns was obtained from Datastream and information on Underwriter’s Quality was downloaded from Jay Ritter’s website. The Family-Dummy was constructed by downloading stakeholder’s characteristics from the Osiris Database. The independent variable: Underpricing is calculated by first-day closing price minus offer price divided by offer price. The dummy-for financial advisor equals one if the issuing firms gathered independent valuations. Underwriting fee equals total underwriting fee divided by number of shares issued. Share overhang is the number of shares offered divided by total amount of shares outstanding. Age is defined as issue date minus founding date. The natural logarithm of proceeds is calculated by taking ln of offer price multiplied by primary shares offered. The natural logarithm of Market Capitalization is the share price multiplied by total amount of shares outstanding. Venture-backed capital is a dummy variable which equals one if the issuing firm is backed by venture capital. The technology dummy equals one if the issuing firm is a technology/internet firm and zero otherwise. The next dummy equals one if the firm is backed by private equity and zero otherwise. The offer price is the price at the IPO. The Family-dummy equals one if the firm’s stakeholders are one or more individuals and zero otherwise. Underwriter’s Quality is ranked by using the Carter and Manaster (1990) method, with one as the lowest possible value and nine the highest.. The fixed effects are based on nation and year. *, **, and *** indicate significance at 10%, 5%, and 1%, respectively.

(1) (2) (3) (4) (5) (6)

Underpricing Underpricing Underpricing Underpricing Underpricing Underpricing (2012-2015)

Dummy: Financial advisor 0.046* 0.066 0.087* 0.059 0.076 0.139*

(0.025) (0.049) (0.051) (0.074) (0.124) (0.074) Underwriting fee -0.086 -0.044 -0.048 -0.032 -0.136 (0.077) (0.103) (0.103) (0.104) (0.181) Share Overhang 0.019** 0.017* 0.017* 0.017 0.015 (0.010) (0.010) (0.010) (0.012) (0.018) Age 0.006** 0.005* 0.005* 0.004 0.006 (0.003) (0.003) (0.003) (0.004) (0.004) Ln Proceeds 0.012 0.009 0.009 0.010 0.018 (0.010) (0.010) (0.010) (0.010) (0.015)

(30)

30

Ln Market Capitalization -0.035 -0.017 -0.019 -0.017 -0.047

(0.038) (0.039) (0.040) (0.040) (0.063)

Venture-Backed Capital 0.130** 0.132** 0.103 0.094 0.286***

(0.064) (0.066) (0.086) (0.086) (0.098)

Dummy: Technology Firm 0.064 0.038 0.042 0.044 -0.001

(0.054) (0.055) (0.055) (0.055) (0.080)

Private Equity Backed firm 0.056 0.055 0.052 0.034 0.101

(0.066) (0.068) (0.069) (0.070) (0.100) Offer Price 0.013** 0.010 0.010 0.011 0.019** (0.006) (0.007) (0.007) (0.007) (0.009) Family-dummy 0.107* 0.126** 0.124** 0.199** 0.070 (0.060) (0.062) (0.062) (0.080) (0.095) Underwriter Quality -0.006 -0.007 -0.006 -0.006 -0.014 (0.020) (0.021) (0.021) (0.021) (0.032)

Financial advisor * Age 0.054

(0.102)

0.052 (0.102) Financial advisor * Share

Overhang

-0.001 (0.016) Financial advisor * Venture

backed capital

0.003 (0.006) Financial advisor * Family

dummy

-0.192 (0.128)

Nation Fixed Effects NO YES YES YES YES

Time Fixed Effects NO YES YES YES YES

Intercept 0.205*** 0.375 0.237 0.080 0.001 0.077 (0.017) (0.643) (0.710) (0.688) (0.706) (1.171) Obs. 745 211 210 210 210 114 R-squared F-value 0.004 3.30 0.155 3.02 0.220 2.18 0.221 2.09 0.233 1.96 0.346 2.62

Table 7 shows the results of three regressions with returns at lock-up expiration dates as dependent variable. After the lock-up expiration date, managers and pre-IPO investors are allowed to sell their shares Courteau (1995). We examined the effect of hiring a financial advisor for returns at lock-up expiration dates, to which end we calculated the returns after the lock-up agreement expired for each company. Positive returns at lock-up expiration dates could indicate that issuing firms set their offer price too low and thus leave money on the table. We made a graphical representation and conducted a t-test to examine whether underpricing exists at lock-up expiration dates.

(31)

31

Graph 2: Returns at lock-up expiration dates by frequency

Data for first-day returns is obtained from Datastream from 2010 to 2017. The independent variable: Underpricing (first-day returns) is calculated by first-week closing price minus offer price divided by offer price. Graph 1a is a histogram with first-day returns on the x-axis and frequencies on the y-axis.

The returns at lock-up expiration dates are normally distributed with zero mean. Therefore we expect no positive returns at lock-up expiration dates. We conducted a t-test and the results are shown below in table 6.

Table 7. T-test for returns at lock-up expiration dates Returns at lock-up dates Mean Std. error SD Obs. 0.0046 0.0025 0.0596 579 T-value = 1.8456 Ho: Mean = 0 Degrees of freedom = 578

H1:Mean>0 P-value: 0.0327

By looking at table 7 and graph 2 we conclude that there are no significant positive returns at lock-up expiration dates. However, we still conduct regression 4-6 to examine whether hiring an independent financial advisor affects the returns at lock-up expiration dates. As expected, the sign coefficient for financial advisor changed from positive to negative. The results in column 2 show that the dummy for technology/internet firm is significant at a 5% level. But if we apply the fixed effects to the regression from column 2, this effect disappears. The negative effect of underwriter quality on returns at lock-up expiration dates is significant at a 10% level with and without the fixed effects in the regression. The effect of financial advisor on returns at lock-up expiration dates is far from significant, suggesting that there exists no relation between hiring an independent financial advisor and returns at lock-up

expiration dates 0 2 0 4 0 6 0 8 0 F re q u e n c y -.2 -.1 0 .1 .2

(32)

32 Table 7: Regression complete sample based on returns at lock-up expiration dates

Column 1 represents is a regression with dummy Financial advisor as dependent variable. Column 2 represents a regression with dummy Financial advisor as dependent variable and the control variables. Column 3 represents a regression with dummy Financial advisor as dependent variable, control variables and fixed effects. The data was obtained from 2010 to 2017 for 916 observations. We selected a minimum offer price of $5 and focused on common shares. Data for first-day returns at lock-up expiration dates was obtained from Datastream and information on Underwriter’s Quality was downloaded from Jay Ritter’s website. The Family-Dummy was constructed by downloading stakeholder’s characteristics from the Osiris Database. The independent variable: First-day Returns at lock-up expiration dates is calculated by first-day closing price minus offer price divided by offer price. The dummy-for financial advisor equals one if the issuing firms gathered independent valuations. Underwriting fee equals total underwriting fee divided by number of shares issued. Share overhang is the number of shares offered divided by total amount of shares outstanding. Age is defined as issue date minus founding date. The natural logarithm of proceeds is calculated by taking ln of offer price multiplied by primary shares offered. The natural logarithm of Market Capitalization is the share price multiplied by total amount of shares outstanding. Venture-backed capital is a dummy variable which equals one if the issuing firm is backed by venture capital. The technology dummy equals one if the issuing firm is a technology/internet firm and zero otherwise. The next dummy equals one if the firm is backed by private equity and zero otherwise. The offer price is the price at the IPO. The Family-dummy equals one if the firm’s stakeholders are one or more individuals and zero otherwise. Underwriter’s Quality is ranked by using the Carter and Manaster (1990) method, with one as the lowest possible value and nine the highest.. The fixed effects are based on nation and year. *, **, and *** indicate significance at 10%, 5%, and 1%, respectively.

(1) (2) (3) Returns: Lock-up expiration date Returns: Lock-up expiration date Returns: Lock-up expiration date

Dummy: Financial advisor 0.002 -0.003 -0.003

(0.004) (0.008) (0.009) Underwriting fee 0.014 0.008 (0.013) (0.016) Share Overhang -0.001 -0.001 (0.002) (0.002) Age 0.000 0.000 (0.000) (0.000) Ln Proceeds 0.000 0.000 (0.002) (0.002) Ln Market Capitalization 0.009 0.011 (0.007) (0.007) Venture-Backed Capital 0.017 0.020* (0.011) (0.012)

Dummy: Technology Firm 0.018** 0.015

(0.009) (0.009)

Private Equity Backed firm 0.002 0.004

(0.011) (0.012)

Offer Price -0.000 -0.000

(0.001) (0.001)

What is the effect of hiring an independent financial advisor on IPO underpricing