Influence of family ownership on IPO underpricing

Influence of family ownership on IPO underpricing Abstract The purpose of this thesis is to determine if there is a difference between family ownership and non-family ownership regarding IPO underpricing. Family firms differ in a lot of aspects from non-family firms. Their difference in characteristics and structure has an influence on the determinants of underpricing and underpricing itself. Recent data from multiple databases is used to perform regressions to find answers for the research question. This thesis finds evidence concluding family firms are less underpriced than non-family firms. Family was the only variable in the model that could define the variation in underpricing. Olaf Mollee 25-06-2018 10764577 Supervisor: Evgenia Zhivotova Bsc Economie & Bedrijfskunde Track: Finance & Organization 1

Statement of Originality This document is written by Student Olaf Mollee 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 contents. 2

Table of contents 1 Introduction... 4 2 Literature review... 6 2. 1 IPO... 6 2.2 Theory of underpricing... 6 2. 2 Differences between family and non-family firms regarding IPO underpricing... 7 3 Research method... 13 3. 1 Hypothesis... 13 3.2 Methodology... 13 3.3 Model... 13 3. 4 Control variables... 14 3.5 Data collection... 15 4 Results... 17 4. 1 Descriptive statistics... 17 4. 2 Regression results... 21 4.3 Robustness check... 25 5 Conclusion... 27 6 References... 28 Appendix... 31 3

1 Introduction The process where the offer price, the first price at which investors can buy the share, is beneath the intrinsic value, is called IPO underpricing. The intrinsic value is the real value of the share. Between 1960 and 2011 the share price was on average 17 percent higher at the end of first day trading (Berk and DeMarzo, 2014). For example, the first day return of Tesla motors was 40.53% and Google was 18,4% 1. The first day return of an IPO is often positive and this is an extraordinary occurrence. With as consequence a famous subject in the literature and a starting point for many research. There is sufficient literature available on the causes of IPO underpricing. One of the main causes for underpricing is the information asymmetry (Rock, 1986). After the findings of the causes, the effects of IPO underpricing with different circumstances were examined. European and American firms with different time periods, for example the financial crisis, were examined. A research in 2012 described that family controlled firms from 1996 till 2004 were less underpriced (Daugherty and Jithendranathan, 2012). But besides these research, research papers that investigated IPO underpricing with family ownership are rare. In other words, the relationship between family versus non-family firms regarding IPO underpricing is not clear, but that does not imply that it is not interesting to investigate. When the 20 biggest firms in each country worldwide will be checked, the group that is most presented as owner is the family (Laporte etal., 1999). Shanker and Astrachan (1996) concluded that 61 percent of the corporations in America with an ultimate owner have a controlling family, which made the family the most superior ownership group.this topic is also interesting for the European market. Nearly 40 percent of the European listed firms has a controlling family (Maury, 2015). Well known firms as Novartis, Roche and Volkswagen are family controlled firms. Besides the importance of the topic because of the high percentage of family controlled firms, it is also important because the IPO is the main source of liquidity for family firms (Pukthuanthong and Walker, 2008). The IPO increases the capital of the firm but reduces the influence and authority of the family on the firm (Leitterstrof and Rau, 2014). These are two contradictory interests the firm must deal with when performing an IPO. The paucity of literature on the topic does not imply that there is no connection 1 https://finance.yahoo.com 4

between the two subjects. There is literature that implies a relationship between IPO underpricing and family characteristics. Tsui-Auch and Lee (2003) concluded that it is interesting to focus on family firms because the behavior of those firms is in contrast with those of non-family firms. The research of Daughterly and Jithendranathen (2012) concluded that the difference in underpricing for family and non-family firms is caused by differences in their ownership structure. When firms with family ownership perform an IPO, it causes a change in the ownership structure of the firm. The family want to maintain their ownership of the firm resulting in a higher consciousness of the IPO, because there is more at stake for the family firm. The higher management ownership of family firms cause a reduce in the underpricing (Daughterly and Jithendranathen, 2012). The less separated ownership and control of family firms will reduce the information asymmetry and agency conflicts because there are less conflicts of interest (Habib and Ljungqvist, 2001). Firms with family ownerships tend to have a different style of management due to the strong relationship the family has with the firm. The family firms attach more value to long perspective view. The long-term perspective and better coalignment of interest could increase the willingness for a monitor system and more transparency (Anderson and Reeb, 2003). The differences in management could have an impact on the underwriter's performance to evaluate the firm's value. Thus, there is evidence to assume that family ownership has an influence on IPO underpricing, because of firm characteristics. In the following chapter the existing literature will be discussed. The causes of IPO underpricing and the differences in characteristics between family and non-family ownership will be discussed. Afterwards, the data collection and the descriptive statistics will be explained. Finally, the regression results and the conclusion of the thesis will be shown. 5

2 Literature review 2. 1 IPO The process of a firm selling stock for the first time is called an Initial Public Offer, IPO. Firms got multiple reasons to sell their stock and go public. The two main reasons for a firm to perform an IPO is increasing liquidity and better access to capital (Ritter and Welch, 2002). By selling the stocks, the firm will increase their capital, which can for example be used for investment opportunities. There are disadvantages of performing an IPO for firms. Going public will result in a loss of ownership and control of the firm. Also, the firm will face more rules about financial disclosure (Allen and Faulhaber, 1988). To know whether family ownership has an influence on IPO underpricing, it is necessary to understand the causes of the IPO underpricing. By knowing the causes of underpricing, it can be investigated whether the causes of underpricing are connected to firm characteristics. Two philosophies are the point of focus to explain the IPO underpricing. The first philosophy gives the cause of underpricing to information asymmetry. Information asymmetry arises when one party has more information available about a subject than the other party and therefore achieve an information advantage (Beatty and Ritter, 1986). Information asymmetry will affect the underpricing due to uncertainty about the real price. It is difficult for investors to determine the value of the firm. The investors are dealing with a high uncertainty and risk when buying the stock. To diminish this uncertainty, investors need to gather information about the firm and the intrinsic value of the share. The gathering of information is costly and therefore they expect a lower issue price to compromise for the costs. By lowering the issue price, they diminish the information asymmetry, and increase the underpricing (Beatty and Ritter, 1986). 2.2 Theory of underpricing There are three main theories about the effect of information asymmetry for underpricing. One of the most influential literature research that examines the causes of IPO underpricing, is the research of Rock. Rock (1986) stated that there are two different groups in the market. The investors that are informed and the group of investors that are uninformed. Logically, the informed group has a better idea of the real value of the share and will only invest in interesting, underpriced IPO s. The uninformed group lacks that kind of information and randomly select the shares. Therefore, the bad, overpriced shares will often be chosen. If this is always the case, only the informed investors will gain profit and the uniformed will leave 6

the IPO market. Without any underpricing, uninformed investors would incur a loss if they participate in IPO's. The purpose of the IPO underpricing is to attract uninformed investors and can be seen as a compensation to the uninformed group. The second theory is the signaling theory. This theory declares that the firm has, in contrast with Rocks theory, more advanced information about the value than the market. Because the investors do not know the value of the share, the underpricing is used to signal the true value of the firm to the market. Firms want to prove with the underpricing that they are wealthy enough to effort the loss of the underpricing (Allen and Faulhaber, 1988). The third theory is related to the agency problems. Agency problems arise when the preferences of the principal and the agent are not aligned. A bought deal is a commitment of the underwriter to the issue firm to sell all the shares. Namely in a bought deal, the shares are bought by the underwriters. If the shares of the firm are undersubscribed, the underwriter will make a loss. According to Baron (1982), the firm cannot monitor the behavior of the underwriter and the underwriter has the preferable information. This information asymmetry gives the underwriter the opportunity to sell the shares for a price beneath the firm value, to be sure the shares are sold (Baron, 1982). This is not aligned with the goal of the firm, which is to maximize the issue proceeds. The underpricing is related to the information asymmetry of the underwriter and the issue firm about the real value of the shares. In the second philosophy, the cause for underpricing is the maintaining of control by the family. When a firm performs an IPO, a separation between ownership and control arise. Firms with family ownership desire to maintain control after the IPO. The underpricing can be used as a tool to maintain control after IPO. According to Brennan and Franks (1997), the underpricing will increase the interest of investors in the share and eventually increase the number of actual buyers. Which as consequence, less concentrated shareholders because small shareholders are less interested in the control part of the firm. In addition, the less concentrated shareholders do not have the capability to monitor the firm, which is in favor of the management that seeks to control the firm. 2. 2 Differences between family and non-family firms regarding IPO underpricing To investigate the difference for family firms regarding IPO underpricing, it is important to understand the difference for family and non-family firms. This difference in structure influences the determinants of underpricing. Family ownership influences the determinants of information asymmetry, agency costs and control discussed in the prior section. Ding and Pukthuanthong-Le (2009) concluded that differences between family and non-family firms 7

make it more difficult to check the quality of family firm IPOs. This could have influences on the underpricing of the shares. The separation of ownership and control distinguish non-family controlled firms from firms which are family controlled. Where firms which are not family controlled have a separation between ownership and control, firms which are family controlled often have not. Members of the family are concentrated in the owner part as in the control part of the family controlled firm (Chambers, 2012). The percentage of managers that also possess shares of the firm is higher for family controlled firms (Anderson and Reeb, 2003). This difference in ownership and control is the main difference between family and non-family ownership and has an influence on the information asymmetry of the firm. In family firms, the management and ownership part are more connected because those parts are being influenced by the same group, the family. Because the ownership and control is less separated in family firms, the transparency of information is better. The two parts are better informed resulting in better alignment of the goals and strategies of the firm. And, in family firms due to the close tie of family members and the strong relation to family wealth, the management and shareholders are more willing to share information. A better environment to exchange information will cause a decrease in information asymmetry and agency problems (Fama and Jensen, 1983). The interconnection between ownership and control of the family firm has an influence on monitoring the firm. Monitoring a firm requires firm specific knowledge and the family of the firm has this specific knowledge. Therefore, in family firms the monitor costs are lower resulting in a reduced agency problem without loss of efficiency (Anderson and Reeb, 2003). The concentrated ownership of the family in the firms aligns the interest of managers with the shareholders and increases the willingness to monitor the manager. This will reduce the chance of the free rider problem. So, in family firms, there are less agency problems because there are more incentives to monitor and it is easier to monitor the managers (Demsetz and Lehn, 1985). Managers who possess shares of the firm have a different risk perception towards an IPO. A failure of an IPO will have more influence on the manager because it will damage their wealth directly. Because of this shared risk perception of managers and owners, the objectives regarding the IPO are more aligned. For example, their aim for the targeted price level is more aligned which will diminish the agency conflict between the firm and the underwriter (Anderson and Reeb, 2003). The interests of managers of family firms are more aligned with the shareholders because those managers don t have incentives to underprice their IPO. In firms without family 8

ownership, managers seek to underprice the IPO to maximize their wealth in the long term. Managers are limited in the sale of the shares before the IPO and underpricing the IPO draws a lot of attention to the investors in the market. As a consequence, there is a raise in the share price after the IPO. Afterwards they can sell the shares for a higher price to preserve their own benefits. In family firms, the members are represented in the management and the ownership part, but do not preserve the goal to sell their shares. Those managers are trying to keep the shares to ensure the family has the control of the firm. The managers lack the incentive to underprice the share for own private benefits (Daughterly and Jithendranathan, 2012). Another advantage of family controlled firms related to underpricing, rises when the CEO of the firm is a family member. Anderson and Reeb (2003) concluded that when a family member participates as the CEO of the firm, the performance of the firm will rise and the agency conflicts will diminish. Villalonga and Amit (2006) draw the same conclusion: higher performance of family firms is directly related to the family member and the CEO of the firm being the same person. But Villalonga and Amit did not found any evidence that supporting their theory. The difference in firm performance of family firms could lead to a difference in underpricing. Firm performance has an influence on the underpricing because it is related to the firm value. When the firm value is easier to evaluate, it will be less hard for the underwriter to choose the share price that is nearby the real value of the firm. Firms with family ownership are less efficient regarding their goals because the firms lack incentives to maximize the profits of the firm. Because of the concentrated ownership of family firms, the family will attach more importance to their own private benefits (Schleifer and Vishny, 1997). A good management distinguishes from a bad management when it can separate their own private benefits and focus on the maximization of the firm value. Concentrated shareholders are unable to separate financial preferences with as consequence the diminishing of the firm s profits. Although, the benefits of concentrated shareholders, the reducing agency conflicts, outweigh the disadvantage of the goal to maximize their private benefits (Habib and Ljungvist, 2001). Family managers base their strategies less rational than non-family managers. Palmer and Barber (2001) found that managers who also possess shares of the firm involve their own social status in making their strategies. Further, by family firms, the management is more focused to please the interest of the family (DeAngelo and DeAngelo, 2000). Family firms can lack the capability of good leadership. In non-family firms, the management is chosen through a selection, where only the best and most capable leaders are 9

appointed. In family firms, members of the family are participating in the management position and therefore family firms do not have the kind of selection non-family firms have. There is less space for qualified workers from outside the firm and a higher chance of unqualified management (Anderson and Reeb, 2003). Also, family firms are less likely to pay managers on stock based options because it will cause a dilution of the control of the family (Schulze et al., 2001). Therefore, family firms are less likely to attract professional managers to the firm. Family firms have an increase in management entrenchment because of the less competitive and qualified management (Carney, 2005). Where Anderson and Reeb (2003) concluded that the management of family firms lacks capability, James (1999) stated that family members have firm specific qualities outsiders do not have with as consequence that family firms perform better than non-family. The long-term perspective of the family for the firm is one of the reasons for better performance of the family firm (James 1999). Where the value of the family depends on the value of the firm, the firm is more managed to perform well in the long term and to provide the existence for many generations (Chambers, 2012). The family has stronger bonds with the future because it has an incentive for the next generation to succeed. The bigger importance for reputation creates longer economic perspective. The broader and longer perspective of the firm will lead to better investments opportunities, and makes it less likely that a good investment will be abstained (Stein, 1988). Also, the family of the firm deals more thoughtful with their resources because of the incentives to monitor and minimize the costs. Family firms deal with more efficiency in the allocation of resources (Anderson and Reeb, 2003). Further, Astrachan and Mcconaughy (2001) concluded that the family are running their firms with more efficiency than non-family firms resulting in greater performance. Another determinant which influences the underpricing through information asymmetry is the disclosure of firm performances. When a firm has high disclosure of firm performances, it is easier to evaluate the value of the firm which will reduce the information asymmetry between investors and the firm. The wage of managers of family firms depends less on performance based measurements. Therefore, there is less managerial opportunism which reduces the chance of manipulating the reports. Also, the family has more knowledge about the firm, whereby it will be easier to check for this kind of manipulation. The information asymmetry between ownership and control is less for family firms but the information asymmetry and agency problems between the controlling and non-controlling shareholders is bigger (Gilson and 10

Gordon, 2003). In family firms, the board of the family is more dominated by family members whereby the chance of manipulation of accounting earnings is higher (Anderson and Reeb, 2003). Also, Chen (2005) stated that the transparency for corporate governance is worse for family firms, because family firms seek to have more family in the boards. Less disclosure about corporate governance is a way to preserve the goal of more family members. Therefore, firms with family ownership tend to have more problems in terms of this transparency. Overall, family firms have better disclosure of financial performances. The reducing of managerial opportunism and the better ability to monitor come before less disclosure about corporate governance (Ali and Cheng, 2007). The differences in disclosure of performances have an impact on the underwriter's performance to evaluate the firm's value and therefore underpricing. Another determinant that can cause a difference in underpricing for family firms is their attachment to their socioemotional wealth (SEW). The SEW is a non-economic value the family acquires from participating in the business. Chrisman and Patel (2012) found that families attach more importance to their SEW than their economic wealth. When normally, firms are trying to sell their shares at the highest possible price, family firms because of their goal to maintain their SEW, are not. There are multiple reasons for family firms to underprice their IPO, to maximize their SEW. An IPO will affect the family SEW through a loss of ownership and control. To minimize the loss of authority and to limit the decrease of the SEW, family firms will underprice their shares. This is known as the ownership dispersion hypothesis. Investors can identify underpricing prior to the IPO with as consequence more investors are interested to buy the shares and eventually an oversubscription of the shares. The oversubscription will diminish the ownership concentration of non-family shareholders and the family keeps their influence and authority. So, a higher underprice will minimize the danger of a concentrated non-family ownership. Keeping their SEW is more important for the families than their loss of money as a consequence of the underpricing (Leitterstrof and Rau, 2014). Underpricing as a method to secure the SEW also plays a part with the reputation of the firm. Because of the connection between family wealth and firm value, the reputation of the firm is more important for family firms. With family firms, the reputation of the firm is linked to the reputation of the family and to their SEW. Families are longer present in the firm and a bad reputation will also affect the firm value in the long run (Leitterstrof and Rau, 2014). An unsuccessful IPO will not only harm the capital of the firm: it will also harm the 11

reputation of the firm (Dyer and Whetten, 2006). Therefore, it is important to have a successful IPO. IPO underpricing will reduce the chance of a failed IPO because it will increase the interest of the investors (Welch, 1992). IPO underpricing also decrease the possibility of lawsuits. IPO underpricing minimizes the chance of an unsuccessful IPO and secures the firm against reputation damage and lawsuits. All to secure the family s SEW. Family firms differ in a lot of aspects from non-family firms. This uniqueness creates an uncertainty for investor about the firm value. Family firms have a long-term perspective because of their long-term interest in the firm. The managers are not as high skilled as managers in non-family firms due to the lack of selection. Because the members of the family are present in the management and control part of the firm, the interests are more aligned. The better alignment and the more willingness to monitor the managers decrease the agency problems and the information asymmetry. Their goal to secure their SEW and the fear of damaging their reputation could also influence the IPO. The differences in organizational structures will lead to a difference in underpricing for family firms (Villalonga and Amit, 2006). 12

3 Research method 3. 1 Hypothesis This thesis examines the influence of family ownership on IPO underpricing for European firms between 2000 and 2017. The expected outcome of the thesis is that firms with family ownership undergo less underpricing than firms with non-family ownership. This outcome is consistent with the conclusion of Daugherty and Jithendranathan (2012). The fact that Family firms have less information asymmetry and agency problems because of their characteristics will outweigh the goal of the family to maintain and secure their SEW. H0= There is no difference in underpricing for firms with family ownership 3.2 Methodology To test the hypothesis, data is gathered from multiple databases to form a multiple regression model. Thereafter a two-sided t-test will be conducted and multiple regressions will be performed regarding OLS. The t-test will test the differences in the mean of the underpricing for the two different groups. When the hypothesis is rejected with an indicated significance level, firms with family ownership have a difference in underpricing. 3.3 Model We conduct several regressions to control for robustness of the outcome. The first regression contains only the dependent and main explanatory variables. Underpricing = α + β1family + β2manager For the second regression, the control variables are included. Underpricing = α + β1family + β2manager + β3lnage + β4lnsize + β5lnproceeds + β6technology + β7underwriterranking +є And for the last regression year fixed effects are included. Underpricing = α + β1family + β2manager + β3lnage + β4lnsize + β5lnproceeds + β6technology + β7underwriterranking + β8year +є The dependent variable in the model is Underpricing. The underpricing is calculated by P1 closing price minus the offer price at P0, divided by the offer price at P0. This is called the first day return of the share (Beatty and Ritter, 1986). 13

The main explanatory variable of the model is Family. For family ownership, a dummy variable is chosen. Firms with family ownership are defined by their ultimate owner. The ultimate owner is classified as an entity that exceeds 50 percent of the ownership. Firms with an ultimate owner classified as one or more named individuals receive the value 1. Firms with another ultimate owner are non-family firms and will receive the value of 0. This distinction allows us to see the difference in underpricing for family ownership. There is discussion about the accurate percentage of ultimate ownership. Some research stated this percentage should be at 50 percent and some at 25 percent. In the research of Böhren (2011) family firms were defined as firms were the family controls more than 50% of the shares. 50 percent ownership is chosen due to financial and accounting rules according to Zephyr. The second explanatory variable is Manager. Firms where the ultimate owner is also the manager of the firm, the value 1 is given. The variable is used to investigate the impact of participating on the ownership and the control parts on the underpricing. Anderson and Reeb (2003) stated that this interconnection will reduce agency costs and could have an influence on the underpricing of the firm. 3. 4 Control variables There are several control variables in the regression model. The first control variable is the age of the firm when performing an IPO. The age of the firm has an influence on the underpricing due to the available information related to the age. Stakeholders receive information of a firm from the market. Firms that are less old have less information and less data available and this makes it more difficult to measure the value or risk. Younger firms are expected to have more underpricing than older firms (Loughran & Ritter, 2004). Age is calculated by the issue date minus the incorporation date. The second control variable is the size of the firm. Beatty and Ritter (1986) stated that larger firms are less risky. This is in line with the explanation of the variable age where larger firms have more information available which could reduce underpricing. For firm size, market capitalization at the offer price was used. Due to high correlation with the Proceeds, for market capitalization the assets of the firm prior to the IPO are used. The third control variable is the Proceeds. The proceeds are the total quantity of money raised when performing an IPO. The proceeds are calculated by the number of shares multiplied by the offer price. In contrast with the size of the firm, the higher the issue proceeds, the more difficult it is to evaluate (Baron, 1982). The fourth control variable is Technology. Consistent with the Research of Loughran 14

and Ritter (2004), a dummy variable is chosen, equal to 1 for technological and internet firms. The classification of those firms is based on their SIC-codes. The fifth control variable is the Underwriter ranking. The underwriter has an influence on the underpricing because it composes the offer and the offer price. The paper of Carter and Manaster (1990) concluded that skilled underwriters experience lower underpricing. High ranked underwriters are better in reducing the information asymmetry that is involved in the IPO. Carter and Manaster ranked the underwriters by their skills from 1 to 9, but this research used mainly American underwriters. The European market contains more small European underwriters. Therefore, the research of Migliorati and Vismara (2014) is used. They investigated which ranking system is the best for European underwriters. European markets are more complicated, for instance it makes a difference in which market the underwriter participates, in contrast with the USA where Nasdaq and Nyse are more correlated (Migliorati and Vismara, 2014). Because of this complicity, the equally weighted metric, based on numbers of IPO managed, which is better with small issuers, is the best to use. The equally weighted method is first used by Megginson and Weiss (1991) who ranked the reputation of the underwriters based on the market share of the underwriter on the IPO. The last control variable is Year. For this variable fixed year effects are used, to control for differences in year of the IPO. 3.5 Data collection To test our hypothesis, the acquired date for family ownership was found in the Zephyr database. To find the acquired data, there are multiple steps to go through. This thesis focuses on the European firms with a timeline from 2000 till 2017. Therefore, the first search criteria reads that the companies should have done their IPO in this time line. The IPO is in the phase of completeness, not just a rumor and the firm who has given the IPO should be the target firm is the second criteria. These steps are the standard criteria to find the data for the IPO underpricing. Because the focus of this thesis is going further than only the underpricing, namely to examine the effect of the family ownership on the IPO underpricing, there are more steps included. First, the firms must have an ultimate ownership of 50 % percent. Thereafter, it is crucial to classify this ownership and make a distinction between firms that have family ownership and firms that have not. Zephyr gives multiple options to characterize the ultimate ownership as a type of shareholder. The shareholders include: bank and financial companies, 15

one or more named individuals, industrial companies and so forth. To know which type of shareholder to apply, the paper of Lins et al. (2013) is used. In this paper shareholders of firms with family ownership are classified as shareholders of one or more named individuals. This makes all the other types of shareholders, shareholders for non-family ownership. The last requirement is that the ultimate owner of the firm should be in Europe, to focus only on Europe. From the zephyr database, the target name, incorporation date, issue date, target shareholders type, offer price, ticker symbol and ISIN number are taken. A sample of 1779 firms are found. The closing prices of the firms are found in the Thomson One database. Because multiple datasets are used, identifiers are necessary to combine the datasets. Firms of which the closing price is not available in Thomson one, are found with DataStream. For this, the ISIN numbers are used. In some cases, Thomson One and DataStream have differences in the issue date of the firm, DataStream occasionally has the issue date a couple months after the issue date of Thomson One. Firms that has differences in the Issue date, DataStream isn t used. Firms that still haven t got an offer price, are excluded. The underwriter of the firms is found in the Thomson one database. From Thomson one, the closing price, underwriter, shares outstanding and the SIC codes are found. Many firms are due to lack of information, excluded from the sample. Firms with family ownership are generally smaller and have a lower offer price. Smaller firms tend to have less information available. Therefore, a lot of firms with family ownership are excluded due to lack on information on one or more required variables. 46 firms with family ownership and 166 firms without family ownership are remaining. 16

4 Results 4. 1 Descriptive statistics Table 1 shows the descriptive statistics from the variables included in the regression. The mean, median, standard deviation, minimum, maximum and observations of the included variables are shown. Table 1 Descriptive statistics all variables Variable Mean Median Standard deviation Minimum Maximum N Underpricing 0.283 0.282 0.183-0.244 0.986 222 Family 0.217 0 0.413 0 1 222 manager 0.208 0 0.407 0 1 222 Lnage 2.203 2.333 1.482-3.067 5.314 222 LnSize 11.818 11.749 2.652 5.05 18.581 222 LnProceeds 17.372 17.440 2.215 7.444 22.019 222 Technology 0.146 0 0.351 0 1 222 Ranking 0.241 0.164 0.252 0.007 1 222 The variable Family has a mean of 0.217, which means that 21.7% of the sample used is family controlled. Family firms and non-family firms are not equally distributed in the sample. It is not necessary to have an equally distributed sample, because of the dummy variable for the family. As a result, all available firms for the research could be used. The number of observations are for all variables the same, firms that missed one or more variables were excluded from the sample. The mean of the underpricing is 28,3%, and higher than the average underpricing of the research of Berk &DeMarz (2014). But comparable to the average underpricing in the research of Daughterly and Jithendranathan (2012). They found an underpricing of 20.03% for family and 29.61% for non-family firms. The minimum of the underpricing is -0,233, indicating an overpricing of 23,3 percent at the end of the first day trading. The maximum of the first day return is an underpricing of 98,59 percent. 17

Table 2 provides the descriptive statistics for the sample of family firms, the non-family firms are excluded. Table 3 provides the same descriptive statistics for non-family firms. It gives a better insight of the influence of the two groups on the variables. Table 2 Descriptive statistics sample Family firms Variables Mean Median Standard deviation Minimum Maximum N Underpricing 0.208 0.161 0.188-0.097 0.986 46 Family 1 1 0 1 1 46 Manager 0.609 1 0.493 0 1 46 Lnage 2.180 2.44 1.159-0.701 4.167 46 LnSize 10.592 10.588 2.005 5.165 13.995 46 LnProceeds 16.515 16.358 2.264 8.710 21.042 46 Technology 0.152 0 0.363 0 1 46 Ranking 0.203 0.141 0.193 0.016 0.884 46 Table 3 Descriptive statistics sample Non-Family firms Variables Mean Median Standard deviation Minimum Maximum N Underpricing 0.304 0.303 0.176-0.244 0.984 166 Family 0 0 0 0 0 166 Manager 0.010 0 0.296 0 1 166 Lnage 2.210 2.281 1.563-3.07 5.315 166 LnSize 12.157 12.260 2.713 5.050 18.581 166 LnProceeds 17.611 17.696 2.125 7.444 22.020 166 Technology 0.145 0 0.353 0 1 166 Ranking 0.252 0.164 0.266 0.007 1.000 166 The mean of the underpricing for family firms is 20,8 % and lower than the mean of 30,4 % from table 3 where non-family firms are presented in the sample. The mean of the manager in Table 2 is 60,9 %, implying that 60,9 percent of the firms have a family member that is the owner and the manager of the firm. This percentage is higher than the mean of the manager of table 3 for non-family firms. This is in line with the consisting literature about managers and family ownership. Family members often play a part in both the ownership and control position of the firm (Anderson and Reeb, 2003). Comparing these two tables, the mean of LnProceeds and LnSize are lower for family 18

firms than the mean of the non-family firms in table 3, implying that family firms are smaller when performing an IPO. Table 2 and 3 show the mean and standard deviation for the two subsamples. The tables make it easier to compare the results of the two samples for underpricing. The average underpricing for family firms is lower than the average underpricing for non-family firms. To find out whether the difference in underpricing is strong enough to conclude that family firms are less underpriced than non-family firms, a two-sample t-test will be conducted. To know which test to use, Levene s test is conducted to check for equal variances as can be seen in Table 8 (Appendix). Because the H0 could not be rejected, a two-sample t-test for equal variances is used. The t value and the corresponding p value will be used to reject or not reject the hypothesis and conclude if family firms are less underpriced than non-family firms. Table 4 Two-sample t test Group Mean Standard deviation N T statistic df Family 0.208 0.188 46 Non-Family 0.304 0.176 166 3.21* 210 The assumptions regarding the validity of the data are made. These assumptions are required to make use of the t test (Appendix). *P value <0.05 The P value is smaller than the significance level at 5 percent and therefore enough evidence to reject the null hypothesis. From the t-test it can be concluded that family firms are less underpriced. 19

Table 5 Correlation matrix Variable 1 2 3 4 5 6 7 8 1. Underpricing 1.000 2. Family -0.216 1.000 (0.002) 3.Manager -0.143 0.520 1.000 (0.038) (0.000) 4.LnAge -0.035-0.009 0.095 1.000 (0.578) (0.902) (0.170) 5.LnSize -0.082-0.244-0.157 0.047 1.000 (0.235) (0.000) (0.022) (0.495) 6.LnProceeds 0.008-0.205-0.101-0.033 0.648 1.000 (0.908) (0.003) (0.144) (0.638) (0.000) 7.Technology 0.019 0.009-0.047-0.115-0.288-0.256 1.000 (0.782) (0.898) (0.494) (0.095) (0.000) (0.000) 8.Ranking -0.009-0.08-0.06 0.095 0.072 0.046 0.078 1.000 (0.902) (0.247) (0.385) (0.167) (0.296) (0.508) (0.257) The correlation matrix contributes to the evidence found in the first 4 tables. In contrast with the literature about the SEW, but in line with the other literature, underpricing and family ownership are negatively correlated. In accordance with the theory, the variable manager is negatively correlated with the underpricing and highly correlated with Family. The variables Proceeds and Size are negatively correlated with the variable Family. This means that family firms are smaller during the IPO and perform smaller IPO s. The same applies for technological firms, were the variable Technology is negatively correlated with Lnproceeds and LnSize. These six variables are all significant. lnsize and lnproceeds are highly correlated which makes sense because the two variables are both related to the size of the firm. But as theory suggests, the two variables have differences in the influence of underpricing. Lnsize is negatively correlated and LnProceeds positive. But both variables are not significant. 20

4. 2 Regression results The hypothesis is tested by an OLS regression. To control for heteroscedasticity, robust standard errors are used in the regression. Multicollinearity is analyzed with the correlation matrix. There are no strong correlations between two variables and therefore no signs of multicollinearity are found. In line with the literature, the variables Size, Age and Proceeds could have nonlinear distribution, to control for the difference in distribution, logarithms are used. Tabel 6 Results of the regression. Year fixed effects are included in the regression but excluded from the table for brevity (1) (2) (3) Variables Underpricing Underpricing Underpricing Family -0.086* -0.098* -0.095* 0.035 0.036 0.037 Manager -0.019-0.022-0.018 0.035 0.036 0.031 LnAge -0.003-0.006 0.008 0.007 LnSize -0.014-0.007 0.006 0.006 LnProceeds 0.006 0.002 0.007 0.007 Technology -0.011-0.011 0.037 0.035 Ranking -0.010-0.046 0.047 0.051 Year fixed effects no no yes Robust standard errors in parentheses. * Denotes significance at 5% The coefficient family does not change in significance in the three different models. Adding control variables and year fixed effects did reduce the p value but not enough to affect the significance at a 5 percent level. Also, the variable manager stays insignificance with a negative coefficient in the three models. As can be seen from table 6 model 3, the coefficient of Family is negative which implies that family firms undergo less underpricing than non-family firms. With a p value of 0.013 the variable is significant at a significance level of 5%. The results from table 6 provide evidence to approve our hypothesis that family firms undergo less underpricing. The results 21

are in line with the research of Daughterly and Jithendranathan (2012) which also concluded that family firms were less underpriced. According to the literature, family firms experience less information asymmetry and agency problems because of the ownership structure of the firm (Demsetz and Lehn, 1985). The managers and shareholders interests are better aligned and have more incentives to monitor the firm. Because of the unification of ownership and control, the risk perception is more aligned which causes less incentives for managers to underprice the share (Anderson and Reeb, 2003). Also, family firms have more disclosure about firm performances which makes it more easy to evaluate. The outcome is in contrast with the literature about the SEW of the family firms. Leitterstorf and Rau (2014) concluded that family firms accept higher underpricing to secure their SEW. Higher underpricing will reduce the loss of ownership and control through oversubscription and minimize the chance of lawsuits and therefore damaging the reputation. Daughterly and Jithendranathan (2012) concluded that because of the higher management ownership in family firms, the managers have no incentive to underprice the IPO, and therefore will maximize the issue proceeds. Anderson and Reeb (2003) also stated that the manager plays a part in the underpricing. When a family member has a function as CEO of the firm, the firm performance will rise and there will be a reduction of agency conflicts. Thus, prior literature describes the manager of the family firm as an important factor to diminish agency conflicts and reduce the underpricing. The negative coefficient in table 6 model 3 implies the same as the literature, except the variable is with a p value of 0.509 not even close to be significant. Because of the insignificance, from this regression it cannot be concluded that the manager has an influence on the underpricing. The difference is to be found in the classification of the dummy variable. Anderson and Reeb examined the influence when a family member is the CEO of the firm and in this research the value 1 is also given when the family member has a function in the management. The regression variables included in the research were based on prior research. As can be seen from table 6, Lnsize has a negative coefficient. The negative relation with underpricing is in line with the research of Beatty and Ritter (1986). Beatty and Ritter declared the higher the size of the firm, the less underpricing. Baron (1982) stated that the higher the issue proceeds, the more difficult it is to evaluate the firm resulting in more underpricing. In table 6, the variable Proceeds has a positive coefficient. But in contrast with the literature, the variables Size and Proceeds are not significant at a 5% significance level. LnSize and Lnproceeds cannot be interpret as variables that explain 22

the underpricing of the firm. The same insignificance of the proceeds was found in the research of Ding and Pukthuanthong-Le (2009). In line with the theory of Carter and Manaster (1991), the ranking of the underwriter has a negative coefficient, implying that the ranking has a negative influence on the underpricing. But the coefficient was not significant and the variable cannot be explained as a variable that influences the underpricing. An explanation could be that Carter and Manaster used a dummy variable based on a ranking scale from 1 to 9 and did not contain many European Underwriters. In contrast with the research of Loughran & Ritter (2002), the variables Age and Technology were not significant. One of the possible reasons that could explain the negative coefficient and the insignificance of Technology is the fact that the research of Loughran and Ritter contained American IPO s from 2001 till 2008. As expected from the literature, the coefficient age has a negative impact on the underpricing. The older the firm, the more data available with as consequence a decline of information asymmetry. But with no significance at a 5% level, in contrast with the literature, it cannot be concluded that Age has an influence on the underpricing. The fixed year effects are included to control for differences across years and are used to control for omitted variable bias. The R2, the percentage of variation of underpricing explained by the model, raised when the year fixed effects were included. But including year fixed effects does not lead to changes in significance for the variables at a 5% level. 23

The same regression as in table 6 is conducted, but non family firms are exlcuded from the sample. Table 7 Results of the regression family sample (1) (2) (3) Variables Underpricing Underpricing Underpricing Manager 0.012 0.057-0.023 0.054 Lnage 0.005 0.017 LnSize -0.024 0.016 LnProceeds 0.002 0.016 Technology -0.005 0.078 Ranking 0.223 0.179-0.046 0.067-0.004 0.020 0.001 0.028-0.005 0.023-0.021 0.082 0.029 0.109 Year fixed effects no no yes The coefficient of the variable manager is in line with table 6, negative, indicating that when a family member is the manager of the firm, the underpricing of the family firm will decrease. But because of the lack of significance it cannot be concluded that manager influences the underpricing of family firms. The positive coefficient of ranking is in contrast with table 6, where ranking has a negative coefficient. Because of the lack of significance, it cannot be concluded that the higher the underwriter, the higher the underprice for family firms. By adding the year fixed effect, the variable Proceeds changed from a positive to a negative coefficient. The negative coefficient of Proceeds is in contrast with the findings of table 6 model 3. But because of the lack of evidence, it cannot be concluded that the size of the issue proceeds has a negative influence on the underpricing of the family firms. The year fixed effect does not lead to changes in significance of the variables, but does increase the R2 of the model. 24

4.3 Robustness check To check the results for robustness, a robustness test is performed. The regression results from table 6 are for some variables in contrast with the existing literature. The robustness test is to check if the results are robust for a change in calculation of the dependent variable. Like the research of Pukthuantong andwalker (2008), the dependent variable is now the underpricing after 14 days. Table 8 Robustness test for change in measurement of underpricing. To test the underpricing, the end of the day price after 14 days is used. Robustness Variables Family -0.053 Underpricing 0.081 Manager -0.101 0.152 LnAge -0.119 0.121 LnSize 0.032 0.043 LnProceeds 0.016 0.038 Technology -0.140 0.154 Ranking -0.081 Year fixed effects 0.153 yes The replacement from the end of the day price after 1 day to the end of the day price after 14 days, has led to a difference in significance of the variable family. Family has a negative coefficient but in comparison with table 6, the family has a p value of 0.098 which leads to insignificance at a 5 percent level. The regression results of table 6 are not robust for a difference in calculation of the underpricing. The conclusion that family firms are less underpriced than non-family firms does not hold when the underpricing is measured by end of the day price after 14 days. 25

The other variable that changed because of the change in dependent variable is LnSize. The robustness test leads to a change of the coefficient LnSize from negative to positive but with no significance nothing can be concluded. 26

5 Conclusion The research tries to answer the question if family firms have a difference in underpricing to non-family firms. The literature section explained the determinants of underpricing and which characteristic of the firm could contribute to the possible difference in underpricing. The variables which could influence the underpricing are included in the regression. Afterwards multiple regressions are conducted regarding OLS. The regression results indicate that there is a difference between family and nonfamily firms regarding underpricing. From the results, it can be concluded that family firms were less underpriced between 2000 and 2017 in Europe. The findings are consistent with the research of Daughterly and Jithendranathan (2012) who found similar results for the American market. The evidence found for less underpriced family firms is not robust for a change in measurement of the dependent variable. Family is the only variable from the analysis that could explain the underpricing, all other variables were insignificant, in contrast with the literature. Due to limited time the research has some limitations. First, the classification of family firms is not clearly defined in the literature. In this research, the paper of Lins et al. (2013) is used to classify family firms as one or more named individuals. Other classifications of family firms could lead to different results because the sample of firms that would be analyzed, would be different. The data sample used for the regression is small and contains 42 family firms. The small amount of family firms could have lead to a less reliable result. Another limitation is the possible problem of endogeneity. Even though year fixed effects control for omitted variable bias, the problem of endogeneity still exists. With as consequence that OlS could be not the appropriate model to use, even though it is mainly used as model in the existing literature. To solve the endogeneity, an 2SLS regression could be conducted. Because of time limits, this could not be done. Family firms have differences in the characteristics that influence the underpricing of the firm. However, it is still not clear what kind of difference in ownership structure is the dominant factor. Thus, an extending of the research would be recommended, with the limitations of this research kept in mind. A bigger data sample, an improved look at the classification of family firms, more variables that are related to family ownership and a 2SLS regression could improve the research. 27