Board Independence, Ownership Structure and the Valuation of IPOs in Continental Europe

Authors

  • Fabio Bertoni,

    Corresponding author
    • Address for correspondence: Fabio Bertoni, EMLYON Business School, Research Center on Entrepreneurial Finance (ReCEntFin), Department of Economics, Finance and Control, 23 Avenue Guy de Collongue, 69134 Ecully, France. E-mail: bertoni@em-lyon.com

    Search for more papers by this author
  • Michele Meoli,

  • Silvio Vismara


Abstract

Manuscript Type

Empirical

Research Question/Issue

We combine the value-creation and value-protection views of the board of directors to study the impact of board independence (BI) on the value of the firm at the time of its initial public offering (IPO).

Research Findings/Insights

We conduct our analysis on a sample of 969 firms that went public in France, Germany, and Italy between 1995 and 2011. We show that BI is a critical factor in the valuation of IPO firms. Our results support both the value-creation and value-protection roles of the board of directors. The relative importance of the two roles of the board varies over time, with value-creation (value-protection) dominating in IPOs of young (mature) companies.

Theoretical/Academic Implications

Our theoretical framework combines the agency and resource-dependence theories. The impact of BI on IPO valuation depends on the importance of the value-creation and value-protection roles played by the board. The change in the relative importance of the two roles determines a U-shaped relationship between BI and firm age. Corporate governance is particularly important for young and innovative firms (where the resource-dependence theory applies, and governance acts as a value-creation device), as well as for mature firms and for companies where ownership and control are separated (where the agency theory applies, and governance serves as a value-protection mechanism for minority shareholders).

Practitioner/Policy Implications

We show that corporate governance is a significant factor affecting the valuation of an IPO company. The importance of BI varies substantially with the knowledge intensity of the industry, the separation between ownership and control, and the age of the listing company.

Introduction

The contribution of the board of directors (BoD) to the value of a company may be split along two dimensions. First, an effective BoD protects suppliers of finance from managerial misbehavior, reducing the cost of, and facilitating access to, external capital. Second, the BoD may give the company a competitive advantage by providing reputation, a network of contacts, and strategic advice. Behind these two dimensions lie two different theoretical perspectives and two streams of literature that, only recently, have been converging.

The first theoretical perspective is the agency view of the BoD. According to this view, the BoD is a value-protection device: the role of the BoD is to monitor the behavior of managers by ensuring that they operate in the interests of shareholders. A well-functioning BoD should determine higher firm valuation by inducing managers to exert more effort and to restrain from extracting private benefits (Hermalin & Weisbach, 2001). The second theoretical perspective is the resource-dependence view of the BoD. According to this view, the BoD is a value-creation device: the role of the BoD is to provide valuable resources to the firm, contributing to its competitive advantage. A well-functioning BoD should determine higher firm valuation by giving strategic advice, contributing to the firm's reputation, and expanding the firm's network of business contacts (Coff, 1999).

These two perspectives are not mutually exclusive, and each BoD simultaneously performs both value-protection and value-creation mechanisms. However, the relative importance of the two roles varies across the life-cycle of the firm, with the value-creation function being dominant when the firm is young and the value-protection function taking the lead as the company matures (Filatotchev, Toms, & Wright, 2006). The evolving impact of the BoD on firm value across the firm's life-cycle has relevant implications for the empirical analysis of the relationship between value and governance. However, this aspect has not received sufficient attention in the academic literature, and the present work aims to fill this gap.

In this paper, we argue that the firm characteristics moderating the impact of board independence (BI) on firm value depend on the age of the company, reflecting which function of the BoD (i.e., value-creation or value-protection) dominates. According to the resource-dependence view (which dominates for young companies), BI is particularly important in high-tech and innovative companies. According to the agency view (which dominates for older companies), BI is particularly important when minority shareholders are exposed to the risk of expropriation. Age itself moderates the impact of BI on firm value, and it will have a different marginal effect depending on which function of the BoD dominates. When the value-creation (or value-protection) function is dominant, BI will increase in importance as the age of the firm decreases (or increases). As a result of the board's gradual switch from a value-creation to a value-protection role, the impact of BI on the firm's valuation has a U-shaped relationship with company age.

We study how BI affects firm value across the life-cycle of the firm in a sample of 969 IPOs between 1995 and 2011 in the three largest economies in continental Europe: France, Germany, and Italy. IPOs are a particularly interesting unit of analysis to study the effect of the dual nature of BoDs. At the time of the IPO, the corporate governance of the firm is clearer than at any other point in the firm's history (Bruton, Filatotchev, Chahine, & Wright, 2010). However, companies may reach the IPO at different stages in their life-cycle, with corresponding different dominant functions of the BoD. This is particularly true in continental Europe, where the nature of the firms going public and the mechanisms to separate ownership from control are more diverse than they are in the United States (Ritter, 2003). These characteristics allow us to study the role of BI across a broad spectrum of stages of maturity and ownership structures of the listing companies.

Our work contributes to the literature in three ways. First, we add to the IPO literature by showing that the contribution of BI to the IPO valuation is sizeable across the entire life-cycle of the firm. Second, we show that BI is not equally important for all IPO firms, but that some firm characteristics moderate its impact on valuation. The relevant moderating factors, however, depend on which function of the BoD dominates, which is the third contribution of our paper. In young companies, the impact of BI on firm valuation is higher when the firm is younger and more innovative. In mature companies, the impact of BI increases with age, and it is particularly pronounced when ownership and control are separated.

The rest of the paper is organized as follows. In the next section, we briefly review the literature and illustrate our research hypotheses. In the following section, we illustrate the methodology and sample used in this study. Next, we report the results of our empirical analyses and discuss their robustness. In the final section of the paper, we provide some concluding remarks.

Literature Review and Hypothesis Development

IPO Valuation and Corporate Governance

The valuation of an IPO company is determined by many factors. IPO valuation is influenced by numerous country-specific institutional characteristics, including listing standards (Johan, 2010) and the quality and enforcement of securities laws (Jackson & Roe, 2009; La Porta, Lopez-de-Silanes, & Shleifer, 2006). Firm-specific characteristics influencing IPO valuation include the listing firm's fundamentals (Aggarwal, Bhagat, & Rangan, 2009; Kim & Ritter, 1999), ownership structure (Meoli, Paleari, & Vismara, 2009; Yeh, Shu, & Guo, 2008), prestige of the top management team (Cohen & Dean, 2005; Pollock & Gulati, 2007), and academic affiliation (Bonardo, Paleari, & Vismara, 2011). Finally, a very important factor influencing the valuation of companies at the IPO stage is the quality of their corporate governance (Bell, Moore, & Filatotchev, 2012; Certo, 2003). Overall, the empirical evidence is supportive of the existence of a positive link between corporate governance quality and the valuation of IPO firms.1 Sanders and Boivie (2004) study a sample of 183 publicly traded US internet firms that went public between 1993 and 1999. The valuation of these firms was strongly associated with several measures of corporate governance, including BI. Chahine and Filatotchev (2008) analyze 140 French IPOs from 1996 to 2000. They show that BI was positively correlated with a higher valuation of the company at the IPO. Using a sample of 251 IPOs in the United Kingdom, Filatotchev and Bishop (2002) show that by improving their corporate governance (including BI), firms can go public with less underpricing.

The evidence of a positive relationship between corporate governance and IPO valuation is consistent with both the value-creation and value-protection roles played by the BoD. The actual mechanism driving the influence of corporate governance on IPO valuation, however, differs depending on which of the two roles of the board dominates. According to the resource-dependence view, corporate governance is linked to higher valuation because firms with better corporate governance have better access to valuable resources (Coff, 1999). In contrast, the agency view argues that better corporate governance is linked to higher valuation because more of the potential value of the company is captured by its shareholders, rather than extracted by managers or controlling shareholders (Dyck & Zingales, 2004). Interestingly, although the two theories agree that corporate governance matters, they produce different predictions about when it matters most. In the next two sections, we will develop distinct hypotheses on the importance of BI in IPO valuation, based on the resource-dependence and agency theories.

IPO Valuation and Value-Creation by the Board of Directors

The value-creation role of the BoD is dominant in young companies (Filatotchev et al., 2006). Certo, Covin, Daily, and Dalton (2001) argue that, in young organizations, the agency view is not as prevalent as compared to mature firms. Instead, the resource-dependence view is dominant in young companies, in which the BoD's stock of knowledge, experience, and network of contacts can contribute to the firm's competitive advantage. The BoD may play a strategic role in the decision-making process (Zahra & Pearce, 1989). Outside directors may facilitate the exchange of strategic information between young companies and established players (Geletkanycz & Hambrick, 1997), compensate for the lack of experience and contacts of the firm's executives (Shivdasani & Yermack, 1999), and reduce the tendency of the management to commit excessively to the firm's current strategy (Hambrick, Geletkanycz, & Fredrickson, 1993). As a company matures, it will gradually acquire strategic knowledge and develop its own business contacts, reducing the marginal importance of BI. Accordingly, we formulate the following hypothesis:

  • Hypothesis 1. In young companies, the marginal effect of board independence on the IPO valuation decreases with the firm's age.

The value-creation role of the BoD will be more crucial in high-tech and innovative companies (Bertoni, Colombo, & Croce, 2013). In high-tech and knowledge-intensive sectors, independent directors play a key role in advising managers and supporting the firm's innovation process (Catherine, Corolleur, Carrère, & Mangematin, 2004; Higgins, Stephan, & Thursby, 2008; Lacetera, 2001). Accordingly, we expect, all other things being equal, that when the value-creation role of the BoD is dominant, the marginal effect of BI on IPO valuation will be higher in high-tech and knowledge-intensive industries, which leads to our second hypothesis.

  • Hypothesis 2. In young companies, the marginal effect of board independence on the IPO valuation is higher in high-tech and knowledge-intensive industries.

However, innovative companies are found not only in high-tech and knowledge-intensive sectors (Von Tunzelmann & Acha, 2005). An alternative way to identify companies for which the value-creation role of the board is especially important is to look at the segment of the stock exchange where they went public. Several European stock exchanges have created specific segments of the market where small companies with innovative business models may list and raise funds (Vismara, Paleari, & Ritter, 2012). We expect the value-creation role of the BoD to be more relevant in companies listing on these “second” markets.

  • Hypothesis 3. In young companies, the marginal effect of board independence on the IPO valuation is higher for firms listing on second markets.

IPO Valuation and Value-Protection by the Board of Directors

According to the agency view, the main role of the BoD is not to channel valuable resources to the company, but to act as a monitoring device. Outside directors are particularly involved in the monitoring role of the BoD (Hermalin & Weisbach, 1988). Parallel to what is argued in the previous section, the agency view of the role of the BoD becomes dominant in firms in later stages of their life-cycle (Filatotchev et al., 2006). Older companies tend to have stable cash flows in excess of their profitable investment opportunities, which exacerbates the potential for value expropriation (Jensen, 1986). In family firms, misbehavior tends to be more likely when ownership is passed across generations and the social interaction among family members weakens (Lubatkin, Schulze, Ling, & Dino, 2005; Mustakallio, Autio, & Zahra, 2002). The more a company matures, the more the value-protection function of the BoD becomes crucial. Accordingly, we expect that, in mature companies in which the value-protection role of the BoD dominates, the marginal effect of BI will increase with firm age.

  • Hypothesis 4. In mature companies, the marginal effect of board independence on the IPO valuation increases with the firm's age.

In the European context, where widely held companies are relatively uncommon (La Porta, Lopez-de-Silanes, Shleifer, & Vishny, 1998) and the value of control is relatively high (Dyck & Zingales, 2004), the BoD may be especially useful for curbing the extraction of private benefits from controlling shareholders. A dimension of the ownership structure that has a significant impact on firm value is the separation of cash-flow and voting rights (Faccio & Lang, 2002; La Porta, Lopez-de-Silanes, Shleifer, & Vishny, 2002). There are two main reasons why this separation may affect valuation of an IPO company. First, the interest alignment hypothesis predicts that a higher level of cash-flow rights should capture an incentive of large shareholders to maximize a firm's value and minimize agency misconduct. We expect underwriters and aftermarket investors to incorporate these effects into their valuation of the firm's shares. Second, the entrenchment hypothesis claims that the controlling shareholder may be strongly motivated to expropriate minority shareholders opportunistically through opaque transactions, in which profits are transferred to other companies that are controlled by the controlling shareholder (Fan & Wong, 2002). According to the agency view, appropriate corporate governance mechanisms (including BI) can be envisaged to minimize the loss derived from the separation between ownership and control (Yeh & Woidtke, 2005). Thus, we may expect BI to compensate, at least in part, for the negative impact of this separation on firm value (Erickson, Park, Reising, & Shin, 2005). We summarize this discussion in the following hypothesis:

  • Hypothesis 5. In mature companies, the marginal effect of board independence on the IPO valuation is higher when the separation between ownership and control is large.

Sample, Data, and Model

Empirical Design

We perform our analysis on a sample of 969 IPOs that took place in the period from 1995 to 2011 on the stock markets of the three largest economies in continental Europe: France, Germany, and Italy. For France, we consider the Paris Bourse until 2004 and Euronext afterwards; for Germany, we consider the Deutsche Börse; and for Italy, we consider the Borsa Italiana. The list of IPOs is selected from the EURIPO database, which provides IPO prospectuses and detailed information on all companies that have recently gone public in Europe.2 We only include “real” IPOs of newly issues shares; we exclude introductions (i.e., admissions with no initial offer), offers of existing shares by selling shareholders, re-admissions, and cross-listings of companies already listed on other stock markets. The IPOs of investment entities and financial companies are also excluded because they display different characteristics compared to other IPO firms.

To test our hypotheses, we regress a variable measuring IPO valuation against a variable measuring BI, while controlling for other firm characteristics and country, industry, and time dummies. To capture how the impact of BI on IPO valuation varies with firm age, we follow two alternative approaches. First, by combining Hypothesis 1 (which predicts that the impact of BI declines with age in young companies) and Hypothesis 4 (which predicts that the impact of BI increases with age in mature companies), we expect the impact of BI on IPO valuation to be U-shaped with age. We test this hypothesis by interacting BI with age and age squared. Hypotheses 1 and 4 will be supported if the quadratic term is positive and the linear term is negative, such that the effect of BI declines with age for young firms (i.e., age less than the minimum of the quadratic function) and increases with age for mature firms (i.e., age greater than the minimum of the quadratic function).

Second, we estimate a regression by interacting BI with age on two subsamples, split based on the age of the listing company. We expect the interaction term to be negative in the subsample of young firms (i.e., age less than the median age) and positive in the subsample of old firms (i.e., age greater than the median age). We also use the sample-split estimation, which is based on a simpler specification, to test Hypotheses 2, 3, and 5.

Variables

Dependent Variable

To investigate the determinants of the initial market valuation, we rely upon Tobin's Q. This measure is defined as the ratio of the market value of the outstanding financial claims on the firm to the current replacement cost of the firm's assets. Tobin's Q is a widely recognized indicator of the firm's future growth opportunities as assessed by the investors' market. It has been extensively adopted in the literature on IPO valuation (e.g., Bonardo et al., 2011; Daily, Certo, & Dalton, 2005). Tobin's Q is usually computed as the ratio of the market value of assets to the book value of assets. The market value is calculated as the sum of the book value of assets and the market value of common stock, minus the book value of common stock. In our main analysis, we compute the market value of common stock based on the IPO offer price. We verify the robustness of our results by calculating the value of common stock based on the preliminary market price and the first-day market price. We also perform our regressions using alternative valuation measures, such as the market-to-book ratio and the EV/Sales ratio.

Board Independence

In line with prior research, we compute the Board Independence variable as the proportion of non-executive members on the board, as reported in the IPO prospectus (Chancharat, Krishnamurti, & Tian, 2012; Kang, Cheng, & Gray, 2007).3

Separation of Ownership and Control

We employ two variables to test for the effect of the separation of ownership and control on the valuation of IPO companies. First, we compute cash-flow rights (C) by measuring the controlling shareholder's percentage ownership of the profits and dividends of the firm. If multiple chains of ownership exist, then the cash-flow rights along each chain are the products of all ownership rights in the intermediate companies along that chain. The total cash-flow rights are equal to the sum of all cash-flow rights from all ownership chains (Faccio & Lang, 2002).

We measure the controlling shareholder voting rights (V) by following the procedure used by La Porta et al. (1998). Following Faccio and Lang (2002), when multiple control chains exist, the voting rights are the sum of the voting rights along each chain with the weakest link among all holding layers. The ratio of the controlling shareholder's voting to cash-flow rights (V/C) approximates the divergence from the one share/one vote ownership structure. This ratio is used to proxy for the controlling owner's motive to extract wealth from the firm.

Age

Firm Age is the age of the IPO firm, in years, since foundation. In the regression, we use Log (1 + Age).

High-tech and Knowledge-Intensive Industries (HTKIS)

To identify high-tech industries in our sample, we use the Eurostat (2009) sectoral classification. Manufacturing industries are classified as high-tech, medium-tech, or low-tech, according to their technological intensity (R&D expenditure/value added). Services are aggregated into knowledge-intensive services and less knowledge-intensive services, based on their share of tertiary educated persons. The HTKIS is equal to one for listing companies operating in manufacturing sectors classified as high- or medium-tech, or in service sectors classified as knowledge-intensive.

Second Markets

Stock exchanges in Europe are organized in segments, with a main market and one or more second-tier markets dedicated mostly to small companies with innovative business models (Vismara et al., 2012). The Second Markets dummy is equal to one for firms listed in these markets.

Control Variables

Several control variables, chosen on the basis of prior IPO literature, are also included in the model. First, we consider three board-related variables, namely Board Ownership, Board Size, and CEO Duality. Board Ownership is measured as the ratio of total beneficial shares held by executive directors at year-end to the total number of shares outstanding at year-end. We measure Board Size as the number of directors on the board, including the chairperson (Chancharat et al., 2012). In IPO firms, small boards have the advantages of being able to monitor management better and to enact decisions more quickly (Fischer & Pollock, 2004). CEO Duality is a binary variable equal to one when the CEO and the chairman of the board of directors are the same person (Lin & Chuang, 2011). CEO duality may reduce the principal-agent conflicts, but may also result in managerial entrenchment (Young, Tsai, & Hsieh, 2008). On average, the literature shows that CEO duality negatively affects the probability of success of IPOs (Certo et al., 2001).

In terms of firm characteristics, we define size as the logarithm of net sales for the year prior to the IPO date, adjusted for inflation (Vismara et al., 2012). Smaller firms typically show higher Tobin's Q because of their better future growth opportunities. We control for profitability (return on assets) and leverage (debt over total assets), as more profitable and less indebted IPO firms are expected to be worth more (Bonardo et al., 2011). We also control for the size and structure of the offer, to account for the effects of the nature of shares offered to the public. The size of the offer relative to the size of the firm is expected to signal market confidence and, therefore, to be positively related to the firm's valuation. The structure of the offer is measured in terms of secondary shares sold by existing shareholders relative to the total number of shares offered. An offer with a larger fraction of newly issued shares signals a higher commitment by existing shareholders.4 Related to the characteristics of the offer, we control for the reputation of the underwriter5 and for the market momentum, defined as the FTSE Euromid percentage index return calculated over the 6 months before the offer date (Cogliati, Paleari, & Vismara, 2011).

Entrepreneurs may seek to increase their offering proceeds by opportunistically manipulating earnings through the management of accruals, to deceive investors temporarily before going public (Teoh, Welch, & Wong, 1998). Therefore, we control for pre-IPO abnormal accruals, corrected for accounts receivables, by following DuCharme, Malatesta, and Sefcik (2001). The actual accruals are compared to conditionally expected accruals predicted by a regression model, and the difference (abnormal accruals) is attributed to earnings management activities.

To estimate the expected accruals, each offering firm is pooled with other firms in the same two-digit SIC code industry (among all firms listed simultaneously for at least 2 years on Euronext, Frankfurt, and Milan). Then, the coefficients of the following equation are estimated by ordinary least squares (OLS):

display math

where WACijp represents working capital accruals in year p for the ith firm in the industry group matched with offering firm j; ΔREVijp is the change in revenues in year p for the ith firm in the industry group matched with offering firm j; and ϑijp accounts for the regression disturbances, which are assumed to be cross-sectionally uncorrelated and normally distributed with zero means. As we rely on a cross-country sample, we also control for the market of listing (Deutsche Börse for Germany, Borsa Italiana in Italy, and Paris Bourse or Euronext for France). Given that IPOs tend to cluster in time and industry, we control for IPO year and industry.6

Estimation

One major concern with the analysis of the cross-sectional determinants of valuation is the potential endogeneity between a firm's IPO value and its BoD structure. Tobin's Q and BI can be jointly affected by the firm's unobserved characteristics, which may result in spurious correlations (Hermalin & Weisbach, 2001). We address this issue by employing an instrumental variable approach, using a two-stage least squares (2SLS) regression. In the first stage, BI is instrumented by using the Mimicking Behavior variable, which is defined as the ratio of non-executive members in the BoDs of all the firms belonging to the same industry (ICB code, first digit) and listed in the same stock market in the IPO year. Mimicking is a common behavior to achieve social legitimacy (Deephouse, 1996, 2000; Deephouse & Carter, 2005), and it is particularly important for IPOs (Bell et al., 2012).

Empirical Results

Descriptive Statistics

Table 1 categorizes the sample of 969 IPOs by stock exchange, age, and IPO year. There is a wide distribution of firm age at IPO. Most of the sample firms were younger than 5 years old, and 257 (26.5 percent) of them were less than 1 year old, at the time of listing. At the other extreme, 327 (33.7 percent) companies in our sample went public when more than 10 years old. This diversity is crucial for our analysis, because it allows us to study the roles of BoDs corresponding to a wide range of company ages at the time of listing.

Table 1. Sample of IPOs
 SampleFranceGermanyItaly
N%N%N%N%
Age at IPO (Years)        
Age < 125726.510925.512131.62717.1
1 < Age <524725.514333.48722.71710.8
5 < Age <1013814.27016.45013.11811.4
Age > 1032733.710624.812532.69660.8
IPO Year        
1995–199922222.912328.78221.41710.8
2000–200332433.413230.813334.75937.3
2004–200728629.515135.37720.15836.7
2009–201113714.1225.19123.82415.2
Total969100.0428100.0383100.0158100.0

Table 2 reports the means, correlation coefficients, and variance inflation factors (VIFs) for all variables. On average, the companies in our sample are 19 years old at IPO, a much higher age than is observed for IPOs in the United States and the United Kingdom (Ritter, 2003; Vismara et al., 2012). Firms going public in Italy are relatively more mature than those in France and Germany (on average, 28.76 years). The average size of IPO firms is not statistically different across countries. The highest Tobin's Q is found in German companies (average Tobin's Q 6.43 in Germany, 4.37 in France, and 3.70 in Italy), which also have the highest degree of BI (on average, 56.25 percent in Germany, 51.50 percent in France, and 53.16 percent in Italy) and the lowest separation between ownership and control (average V/C is 1.04 in Germany, 1.24 in France, and 1.47 in Italy). Board Ownership is, on average, 20 percent in France and more than 30 percent in Germany and Italy, where boards are larger (on average, Board Size is 5.84 in France, 6.87 in Germany, and 7.24 in Italy). With smaller boards, the CEO is also the chairman of the board in most French firms (60 percent), whereas these two roles are typically split in German firms (average CEO Duality 14 percent).

Table 2. Descriptive Statistics and Correlation Matrix
 VariablesSampleFranceGermanyItaly12345678910111213141516171819VIF
  1. This table presents averages for the variables employed in the regression analyses on the whole sample and on country subsamples, as well as correlations between variables. Tobin's Q is the ratio of the ratio of the market value of assets (the sum of the book value of assets and the market value of common stock, based on the IPO offer price, minus the book value of common stock) to the book value of assets. Board Independence is the percent of non-executive members on the board, as reported in the IPO prospectus. Age is the logarithm of one plus the age of the company in years. V/C is the ratio of the controlling shareholder's voting rights (V) to its cash-flow rights (C). C measures the controlling shareholder's percentage ownership of the profits and dividends of the firm. If multiple chains of ownership exist, then the cash-flow rights along each chain are the products of all ownership rights in the intermediate companies along that chain. The total cash-flow rights are equal to the sum of all cash-flow rights from all ownership chains. V is computed as in La Porta et al. (1998). When multiple control chains exist, the voting rights are the sum of the voting rights along each chain with the weakest link among all holding layers. Board Ownership is measured as the ratio of total beneficial shares held by executive directors at year-end to the total number of shares outstanding at year-end. Board Size is the number of directors on the board, including the chairperson. CEO Duality is a dummy indicating when the CEO and the chairman of the board of directors are the same person. Firm Size is the logarithm of net sales for the year prior to the IPO date, adjusted for inflation. Profitability is the return on assets in the year before the IPO. Leverage is debt over total assets in the year before the IPO. Offer Size is the ratio (in percent) between the size of the firm and the size of the offer. Offer Structure is the fraction (in percent) of secondary shares sold by existing shareholders relative to the total number of shares offered. Reputation of the underwriter is equal to 100 when the underwriter is in the Carter-Manaster ranking (list taken from Jay Ritter's website), and otherwise as equal to the underwriter's percentage market share in terms of number of IPOs underwritten in Europe during 1995 to 2011. Market Momentum is defined as the FTSE Euromid percentage index return calculated over the 6 months before the offer date. VC backing is a dummy variable indicating that the firm is VC-backed at the IPO. VC Reputation is measured as the cumulative market capitalization of IPOs backed by the VC in continental European markets. Accruals is a measure of pre-IPO abnormal accruals computed following DuCharme et al. (2001). Mimicking Behavior is defined as the ratio of non-executive members in the BoDs of all the firms belonging to the same industry (ICB code, first digit) and listed in the same stock market in the IPO year. For the sake of readability all dummy variables in this Table are multiplied by 100. Stars for averages refer to tests for differences between a single country and the rest of the sample. ***, **, and * indicate significance levels below 1%, 5%, and 10%, respectively. For correlation coefficients, ° indicates a significance level below 1%. Variance Inflation Factors (VIFs) are obtained after estimating an OLS regression of Tobin's Q against all variables, except the instrumental variable (Mimicking Behavior).
 1Tobin's Q5.074.37***6.43***3.70***1.000                  
 2Board Independence53.6551.50***56.25***53.16.0171.000                 1.07
 4Age18.7816.51**17.1928.76***−.238°.0171.000                1.14
 3V/C1.21.241.04***1.47***−.132°.004.0321.000               1.11
 5C46.1947.2248.97***36.65***−.030−.003−.279°.0661.000              1.20
 6Board Ownership27.2419.86***31.59**36.71***−.018.028−.008−.004.0771.000             1.17
 7Board Size6.485.84***6.87***7.24***−.078.243°.063.064.013.203°1.000            1.26
 8CEO Duality38.9159.58***14.36***42.41−.166°−.074.060.010−.038−.027−.293°1.000           1.14
 9Firm Size280.84280.95257.54337.03−.130°.023.000.120°.115°.611°.203°−.0071.000          1.36
10Profitability17.0111.1624.9413.66.020.011−.001.019−.022−.003−.008.038−.0051.000         4.06
11Leverage66.231.86121.5325.08.048.022−.009.009−.024−.001−.005.034−.002.947°1.000        4.08
12Offer Size41.0628.57**41.0674.86***.051.024−.001.044−.056−.002.048−.034−.007.011.0141.000       1.05
13Offer Structure24.1850.45***78.06***66.31.178°.033−.055−.070−.116°.002−.003−.123°−.050−.016−.010.0481.000      1.07
14Underwriter reputation2314.25***27.94***35.44***.034.032.020.105°.003.070.187°−.076.185°.037.066−.005−.0191.000     1.13
15Market Momentum.38.18.47.69.024.002.024−.034.003.022.008−.003.036.016.021.134°.006.0121.000    1.04
16Accruals.00.01.01−.03.008−.035−.032.073.017.050−.013.014.069.034.004.002−.018.001.0011.000   1.02
17VC Backing38.4935.0549.61***20.89***.162°.032−.049−.093°−.177°−.100°.025−.010−.105°.009.040−.014.072−.006.046−.0651.000  1.10
18VC Reputation1.13.57**1.70*1.26.066.004−.016−.025−.133°−.111°.033−.005−.052−.022−.005.006.027.036.062−.015.172°1.000 1.06
19Mimicking Behavior54.4151.99***57.74***52.87**.165°.449°−.062−.048−.005−.024.143°−.146°−.030.059.058.044.081−.014.010−.026.083.0161.000

Several other variables do not differ significantly between countries. On average, Profitability (return on assets) is 17 percent and Leverage is 66 percent. The relative size of offer (shares offered over shares outstanding) is larger in Italy, whereas the percentage of secondary shares (Offer Structure) is higher in Germany. German firms are more often VC-backed (50 percent), as compared to French (35 percent) and Italian IPOs (21 percent). Finally, there appears to be no major concern for multicollinearity among the independent variables in the regression models, because most correlations are low and no VIF exceeds five.

Econometric Results

We begin our analysis by studying how the relationship between BI and IPO valuation is affected by the firm's age, as illustrated in Hypotheses 1 and 4. The results of the regressions are reported in Table 3. First, we report the results of a baseline regression estimated using OLS (Model 1) and 2SLS (Model 2). Both models report a positive effect of BI on IPO valuation, consistent with the extant literature (e.g., Chahine & Filatotchev, 2008). The sign and significance of BI are consistent across both models. The first stage of the regression shows that the instruments have the expected signs, and that Mimicking Behavior is highly statistically significant.

Table 3. Effect of Board Independence on IPO Valuation: Moderating Effect of Age and Age Squared
 (1)(2)(3)(4)(5) (6)
OLS2SLSOLS2SLS2SLS – split samples
 1st stage2nd stage 2nd stage2nd stage2nd stage
     YoungOld
  1. This table reports the results of regressions in which Tobin's Q is the dependent variable. Regressions are performed on the full sample of 969 European IPOs, or on subsamples split according to age (below and above the median value; Models (5–6)). Model (1) is an OLS regression of the baseline specification. Model (2) is a 2SLS as in Model (1), in which the endogenous variable is Board Independence and the instrument is Mimicking Behavior. Model (3) is an OLS regression, in which the interactions between Board Independence, Age, and Age2 are included. Model (4) is a 2SLS regression as in Model (3). Board Independence and its interactions are instrumented by Mimicking Behavior and its interaction with Age and Age2. Models (5–6) are 2SLS regressions as in Model (4), in which the non-linearity of Age as a moderating factor of the impact of BI is only captured by splitting the sample. In Models (4–6), the first stage is not reported for the sake of readability. Tobin's Q is the ratio of the ratio of the market value of assets (the sum of the book value of assets and the market value of common stock, based on the IPO offer price, minus the book value of common stock) to the book value of assets. Board Independence is the percent of non-executive members on the board, as reported in the IPO prospectus. Age is the logarithm of one plus the age of the company in years. V/C is the ratio of the controlling shareholder's voting rights (V) to its cash-flow rights (C). C measures the controlling shareholder's percentage ownership of the profits and dividends of the firm. If multiple chains of ownership exist, then the cash-flow rights along each chain are the products of all ownership rights in the intermediate companies along that chain. The total cash-flow rights are equal to the sum of all cash-flow rights from all ownership chains. V is computed as in La Porta et al. (1998). When multiple control chains exist, the voting rights are the sum of the voting rights along each chain with the weakest link among all holding layers. Board Ownership (in percent) is measured as the ratio of total beneficial shares held by executive directors at year-end to the total number of shares outstanding at year-end. Board Size is the number of directors on the board, including the chairperson. CEO Duality is a dummy indicating that the CEO and the chairman of the board of directors are the same person. Firm Size is the logarithm of net sales for the year prior to the IPO date, adjusted for inflation. Profitability is the return on assets (in percent) in the year before the IPO. Leverage is debt over total assets (in percent) in the year before the IPO. Offer Size is the ratio (in percent) between the size of the firm and the size of the offer. Offer Structure is the fraction (in percent) of secondary shares sold by existing shareholders relative to the total number of shares offered. Reputation of the underwriter is equal to 100 when the underwriter is in the Carter-Manaster ranking (list taken from Jay Ritter's website), and otherwise as equal to the underwriter's percentage market share in terms of number of IPOs underwritten in Europe during 1995 to 2011. Market Momentum is defined as the FTSE Euromid percentage index return calculated over the 6 months before the offer date. Accruals is a measure of pre-IPO abnormal accruals computed following DuCharme et al. (2001). Mimicking Behavior is defined as the ratio of non-executive members in the BoDs of all the firms belonging to the same industry (ICB code, first digit) and listed in the same stock market in the IPO year. All regressions control for time, industry, and country effects (Germany is the omitted category). Coefficients of time and industry effects are omitted for readability. Heteroskedasticity-robust standard errors are reported in brackets. ***, **, and * indicate significance levels below 1%, 5%, and 10%, respectively.
Board Independence1.567*** 3.309***2.577***5.683***3.675**8.661***
(.437) (.924)(.582)(1.593)(1.497)(1.894)
Board Independence × Age   −.771***−1.540***−.444**1.316***
   (.219)(.478)(.197)(.318)
Board Independence × Age2   .126***.216***  
   (.035)(.072)  
Age−.154.006−.125−.153−.125−.223−1.364
(.226)(.015)(.220)(.224)(.218)(.628)(1.324)
Age2−.027−.001−.034−.028−.033−.046.124
(.043)(.003)(.042)(.043)(.042)(.266)(.179)
V/C−.237***.005−.231***−.253***−.264***−.232*−.233***
(.075)(.007)(.076)(.075)(.075)(.133)(.081)
C.003.000.003.002.000.006−.001
(.004)(.000)(.004)(.004)(.004)(.006)(.005)
Board Ownership−.545**.004−.566**−.529**−.549**−.123−1.028***
(.244)(.015)(.242)(.243)(.241)(.371)(.340)
Board Size−.034.007***−.051**−.033−.052**−.102**−.039
(.024)(.002)(.025)(.024)(.025)(.044)(.037)
CEO Duality−.833***.010−.816***−.839***−.831***−1.011***−1.009***
(.173)(.012)(.171)(.172)(.171)(.269)(.250)
Firm Size−.738***−.003−.730***−.734***−.719***−.850***−.640***
(.047)(.003)(.047)(.047)(.047)(.077)(.072)
Profitability.048−.009**.055.043.025−.0741.778**
(.112)(.004)(.113)(.113)(.111)(.068)(.769)
Leverage.006.001**.004.007.009.006−.255**
(.017)(.001)(.017)(.017)(.016)(.019)(.115)
Offer Size.027−.000.020.033*.031.786.016
(.018)(.001)(.019)(.019)(.021)(.758)(.016)
Offer Structure.313.011.281.298.233.2561.076***
(.361)(.007)(.356)(.360)(.351)(.370)(.363)
Underwriter Reputation.904***−.008.920***.909***.926***1.869***.669**
(.211)(.013)(.208)(.211)(.209)(.319)(.290)
Market Momentum1.594−.0161.7691.4991.579.5783.149
(2.047)(.131)(2.026)(2.056)(2.068)(3.330)(2.613)
Accruals.172−.006.185*.175*.199*.191*1.473***
(.105)(.006)(.101)(.106)(.102)(.111)(.452)
VC Backing.212.002.203.175.127.521**.418
(.171)(.011)−.125(.171)(.170)(.257)(.256)
France−1.464***−.002−.883***−1.435***−.587−1.121***−.107
(.218)(.017)(.322)(.218)(.485)(.392)(.939)
Italy−1.071***.001−.643**−1.069***−.499−.192−.062
(.252)(.020)(.359)(.250)(.395)(.517)(.958)
Instrumental Variable       
Mimicking Behavior .973***     
 (.048)     
Constant15.048***−.03116.786***14.877***16.378***18.679***15.645***
(1.320)(.079)(1.235)(1.343)(1.334)(1.477)(2.867)
Observations 969969969969458511
Statistics for OLS models:      
Adj. R-squared.414 .418   
Statistics for 2SLS models:      
Centered R-squared (2nd stage) .430 .426.436.445
Cragg-Donald Wald F statistic 257.136 138.80040.34560.696
[Stock-Yogo weak ID test critical values: 5% maximal IV relative bias] [13.91] [9.53][9.53][9.53]
Hansen (endogeneity test) 4.600 4.762.57514.910
[Chi-sq(k) p-value] [.032] [.029][.448][.000]

We add the linear and quadratic interactions between BI and age in Model 3 (OLS) and Model 4 (2SLS). Both models confirm that age affects the impact of BI on IPO valuation. We find a negative and significant coefficient for the linear interaction, and a positive and significant coefficient for the squared interaction. Thus, as expected, the marginal effect of BI is U-shaped: the impact of BI decreases with age for young companies (consistent with Hypothesis 1) and increases with age for mature companies (consistent with Hypothesis 4). The turning point is estimated to correspond to an age of 34 years (Model 4), well within the range of variation of age observed for listing companies in our sample. For a company at the turning point, a one standard deviation increase in BI results in an increase in Tobin's Q by 12.5 percent (Model 4). Due to the U-shape of the marginal effect of BI, the effect is larger for companies aged below or above this turning point. If we consider companies one standard deviation above (below) the turning point, we observe (Model 4) a marginal effect of BI larger by 1.5 percent (5.4 percent).

We also study the effect of age on the impact of BI on IPO valuation by splitting the sample between young and mature firms (above or below the median age). The results are consistent with those of Models 3 and 4. In young companies (Model 5), the marginal effect of BI on the IPO valuation decreases with the firm's age, whereas in mature companies (Model 6), the marginal effect increases with the firm's age. Thus, Hypotheses 1 and 4 are validated. The impact of age on the marginal effect of BI is not only statistically significant, but also economically sizeable. The marginal effect of BI in a company whose age corresponds to the first quartile is 10.3 percent larger than that in a median aged company (Model 5). The marginal effect of BI in a company whose age corresponds to the third quartile is 6.1 percent larger than that in a median aged company (Model 6).

Table 4 reports the results obtained by interacting BI with HTKIS, Second Markets, and separation between ownership and control. Each regression is conducted on the full sample and by splitting the sample between young and old companies. Model 1 shows that the interaction between BI and the HTKIS dummy is positive and significant. This result indicates that the marginal effect of BI on the IPO valuation is higher in high-tech and knowledge-intensive industries. Interestingly, when the sample is split on the basis of age, the result holds only for young companies (Model 2) and disappears for mature companies (Model 3). This evidence supports Hypothesis 2. Similarly, the interaction between BI and the Second Market dummy, which is positive and significant for the full sample (Model 4), is significant only in the subsample of young companies (Model 5) and not significant in the subsample of mature companies (Model 6). This result is in line with Hypothesis 3.

Table 4. Effect of Board Independence on IPO Valuation: Value-Creation, Value-Protection, and Role of Venture Capitalists
 Value creation in high-tech firmsValue creation in second marketsValue protection and V/CVenture capital
(1) Full sample(2) Young firms(3) Old firms(4) Full sample(5) Young firms(6) Old firms(7) Full sample(8) Young firms(9) Old firms(10) VC- backed(11) Non-VC- backed
  1. This table reports the results of 2SLS regressions in which Tobin's Q is the dependent variable. Models (1–3) test for the Value Creation effect of Board Independence, interacted with HTKIS, a dummy variable equal to 1 for IPOs belonging to high- and medium-tech manufacturing sectors and knowledge-intensive service industries (Eurostat, 2009). Model (1) refers to the full sample, Model (2) to the sample of young firms (age below the median value), and Model (3) to the sample of old firms (age above the median value). Models (4–6) test for the value-creation effect of BI, interacted with Second Markets, a dummy variable equal to 1 for IPOs that took place on second markets. Models (4), (5), and (6) refer to the full sample, subsample of young firms, and subsample of old firms, respectively. Models (7–9) test for the value-protection effect of BI, interacted with the separation between ownership and control (V/C). Models (7), (8), and (9) refer to the full sample, subsample of young firms, and subsample of old firms, respectively. HTKIS and Second Market dummies are included in the models in which they are interacted. In the first stage (omitted), Board Independence and its interactions are instrumented by Mimicking Behavior, and by the interaction of Mimicking Behavior and the HTKIS (Model 1–3), Mimicking Behavior and Second Markets (Model 4–6), or Mimicking Behavior and V/C (Model 7–9). Models (10–11) are 2SLS regressions in which the sample is split between VC-backed and non-VC–backed firms, respectively. VC-backing is dropped in both models, and VC reputation is included in Model (10). In Models (10–11), the endogenous variable is Board Independence and its interactions. The instrument is Mimicking Behavior and its interactions with Age and Age2. Tobin's Q is the ratio of the ratio of the market value of assets (the sum of the book value of assets and the market value of common stock, based on the IPO offer price, minus the book value of common stock) to the book value of assets. Board Independence is the percent of non-executive members on the board, as reported in the IPO prospectus. Age is the logarithm of one plus the age of the company in years. V/C is the ratio of the controlling shareholder's voting rights (V) to its cash-flow rights (C). C measures the controlling shareholder's percentage ownership of the profits and dividends of the firm. If multiple chains of ownership exist, then the cash-flow rights along each chain are the products of all ownership rights in the intermediate companies along that chain. The total cash-flow rights are equal to the sum of all cash-flow rights from all ownership chains. V is computed as in La Porta et al. (1998). When multiple control chains exist, the voting rights are the sum of the voting rights along each chain with the weakest link among all holding layers. HTKIS is a dummy variable identifying firms in a high-tech and knowledge-intensive industry, according to the sectoral classification by Eurostat (2009). Second Markets is a dummy identifying firms listed in second-tier stock markets. Board Ownership is measured as the ratio (in percent) of total beneficial shares held by executive directors at year-end to the total number of shares outstanding at year-end. Board Size is the number of directors on the board, including the chairperson. CEO Duality is a dummy indicating that the CEO and the chairman of the board of directors are the same person. Firm Size is the logarithm of net sales for the year prior to the IPO date, adjusted for inflation. Profitability is the return on assets (in percent) in the year before the IPO. Leverage is debt over total assets (in percent) in the year before the IPO. Offer Size is the ratio between the size of the firm and the size of the offer. Offer Structure is the fraction of secondary shares sold by existing shareholders relative to the total number of shares offered. Reputation of the underwriter is equal to 100 when the underwriter is in the Carter-Manaster ranking (list taken from Jay Ritter's website), and otherwise as equal to the underwriter's percentage market share in terms of number of IPOs underwritten in Europe during 1995 to 2011. Market Momentum is defined as the FTSE Euromid percentage index return calculated over the 6 months before the offer date. Accruals is a measure of pre-IPO abnormal accruals computed following DuCharme et al. (2001). Mimicking Behavior is defined as the ratio of non-executive members in the BoDs of all the firms belonging to the same industry (ICB code, first digit) and listed in the same stock market in the IPO year. All regressions control for time, industry, and country effects (Germany is the omitted category). Coefficients of time and industry effects are omitted for readability. Heteroskedasticity-robust standard errors are reported in brackets. ***, **, and * indicate significance levels below 1%, 5%, and 10%, respectively.
Board Independence5.517***4.985**3.860**5.040***5.221**4.174**3.446***3.417*3.087**5.970***6.843**
(1.896)(1.998)(1.704)(1.336)(2.139)(1.695)(1.179)(1.876)(1.431)(1.909)(2.765)
Board Independence × HTKIS3.411*2.999**.188        
(2.012)(1.273)(1.833)        
Board Independence × Second Markets   1.572*2.898**−.190     
   (.942)(1.321)(1.363)     
Board Independence × V/C      .593*.1541.042**  
      (.352)(.574)(.466)  
HTKIS2.771**2.153***1.008        
(1.083)(.651)(.985)        
Second Markets   .840.661.485     
   (.534)(.831)(.701)     
Board Independence ×Age         −1.156**−2.436**
         (.504)(1.117)
Board Independence ×Age2         .175**.349*
         (.073)(.186)
Age−.297.027−1.309−.189.250−1.620−.300−.231−1.194.210−.387
(.227)(.635)(1.232)(.229)(.655)(1.220)(.220)(.636)(1.197)(.374)(.300)
Age2−.023−.099.122−.061−.207.162−.039−.051.099−.098.031
(.043)(.264)(.166)(.044)(.270)(.165)(.042)(.270)(.161)(.076)(.055)
V/C−.230***−.229*−.147−.204***−.157−.176**−.264***−.290*−.182**−.286*−.214**
(.069)(.127)(.092)(.067)(.147)(.077)(.078)(.150)(.084)(.149)(.091)
C.005.006.005.003.005.003.005.006.005.007−.001
(.004)(.006)(.005)(.004)(.006)(.005)(.004)(.006)(.005)(.008)(.005)
Board Ownership−.513**−.121−.922***−.639**.057−1.051***−.569**−.095−.932***−.432−.385
(.257)(.383)(.330)(.267)(.400)(.341)(.253)(.374)(.323)(.428)(.305)
Board Size−.051*−.094**−.019−.071**−.110**−.025−.067**−.100**−.024−.061−.039
(.026)(.043)(.033)(.028)(.044)(.035)(.027)(.042)(.034)(.048)(.032)
CEO Duality−.867***−.943***−.795***−.972***−1.039***−.869***−.930***−.943***−.809***−1.172***−.626***
(.184)(.274)(.233)(.186)(.277)(.234)(.182)(.274)(.232)(.288)(.212)
Firm Size−.716***−.775***−.641***−.746***−.868***−.614***−.750***−.847***−.663***−.846***−.623***
(.050)(.079)(.066)(.054)(.079)(.076)(.049)(.076)(.066)(.083)(.061)
Profitability−.021−.0851.669**.009−.0421.783**−.002−.0881.880***−.0391.410*
(.095)(.065)(.742)(.127)(.072)(.746)(.118)(.066)(.723)(.079)(.727)
Leverage.015.003−.236**.012−.009−.255**.011.001−.269**.014.024
(.014)(.021)(.111)(.019)(.021)(.111)(.017)(.019)(.108)(.012)(.016)
Offer Size.006.435.013.010.601.019.010.657.0201.687**.046**
(.013)(.770)(.015)(.014)(.771)(.016)(.014)(.766)(.016)(.695)(.022)
Offer Structure.507.2621.288***.546.3081.524***.550.2831.430***−.1661.058***
(.394)(.343)(.321)(.422)(.377)(.319)(.423)(.374)(.314)(.202)(.291)
Underwriter Reputation1.116***1.883***.619**1.154***1.823***.720**1.179***1.868***.631**1.525***.433
(.217)(.320)(.274)(.223)(.312)(.284)(.218)(.322)(.273)(.330)(.269)
Market Momentum2.954.7323.5823.512*1.5673.8202.981.8733.408−.345.936
(2.104)(3.354)(2.462)(2.121)(3.250)(2.529)(2.055)(3.275)(2.468)(3.938)(2.439)
Accruals.230*.1571.480***.270**.1671.547***.314***.203*1.444***.1111.021***
(.118)(.105)(.459)(.114)(.102)(.466)(.112)(.109)(.433)(.088)(.322)
VC Backing.507***.405.555**.527***.495*.609***.513***.471*.625***  
(.177)(.260)(.239)(.177)(.260)(.235)(.174)(.259)(.230)  
VC Reputation         1.123 
         (1.392) 
France−1.305***−1.264***−1.420**−1.217***−1.216***−1.289*−1.097***−.390−1.437***−1.135**−.013
(.271)(.351)(.601)(.277)(.345)(.665)(.250)(.504)(.279)(.557)(.521)
Italy−.736**−.758**−.357−.884***−.954***−.497−1.335***−1.213***−1.383***−.572−.436
(.358)(.352).853(.340)(.321)(.617)(.225)(.348)(.285)(.803)(.431)
Constant14.462***15.701***15.026***16.075***18.052***15.168***16.666***18.233***15.551***18.132***13.369***
(1.419)(1.849)(2.809)(1.312)(2.177)(2.732)(1.087)(1.495)(2.590)(1.762)(1.540)
Observations969458511969458511969458511373596
Centered R-squared (2nd stage).431.540.603.422.464.449.425.541.578.485.417
Cragg-Donald Wald F statistic188.55496.12829.629211.75185.811113.479222.13790.87769.99674.32754.785
[Stock-Yogo weak ID test critical values: 5% maximal IV relative bias][11.04][11.04][11.04][11.04][11.04][11.04][11.04][11.04][11.04][9.53][9.53]
Hansen (endogeneity test)7.426.45211.7558.415.4119.8692.584.0584.6069.7773.125
[Chi-sq(k) p-value][.006][.501][.000][.004][.521][.002][.108][.810][.032][.007][.077]

In Models 7–9, we analyze the moderating effect of the separation between ownership and control on the impact of BI on IPO valuation. Consistent with our expectations, the separation between ownership and control is negatively associated with IPO valuation. The interaction term between BI and V/C is positive (Model 7), indicating that corporate governance may reduce the risk of value expropriation for minority investors. When the sample is split by age, we find that the result holds only for mature companies (Model 9), whereas the coefficient is not significant among young companies (Model 8). This evidence is in line with Hypothesis 5.

In each regression, we control for several factors that can impact valuation. Among them, prior research has focused particularly on the role of Venture Capital (VC), also comparing different institutional environments (Bruton et al., 2010) and different control rights (Cumming, 2008). We do not find clear evidence regarding the role of VC investors. The VC-backing variable is always not significant in the regressions on the subsample of young firms, whereas it is typically significant in mature firms. When we split the sample into VC-backed (Model 10) and non-VC-backed firms (Model 11), the general results are confirmed in both samples. We also assess the impact of the reputation of VCs, in terms of the market shares of all VCs backing an IPO. This parameter is measured as the cumulative market capitalization of IPOs backed by the VC in continental European markets, similar to Nahata (2008). The coefficient of this variable in the regression on the subsample of VC-backed IPOs (Model 10) is positive, as expected, but not significant.

In our sample, CEO duality leads to lower valuations for firms going public. The coefficient of the CEO Duality variable is persistently negative and significant across all models. From an agency theory perspective, CEO duality limits the BI and the board's ability to fulfill its governance role. This observation is in line with the finding by Chahine and Tohmé (2009), who report that IPO firms with CEO duality report more underpricing, due to their higher uncertainty.

Firms taken public by more reputable underwriters obtain higher valuations. The reputation of underwriters acts as a certification mechanism and positively affects IPO valuations (Carter & Manaster, 1990). Other control variables that are significant in some, but not all, of our models include Board Size, Board Ownership, and Firm Size, which are negatively correlated to the firm's Tobin's Q. Consistent with the value-relevance hypothesis by DuCharme et al. (2001), Accrual Accounting is positively associated with IPO valuation.

Overall, our results support the idea that, as the firm matures, the BoD switches from a value-creation to a value-protection mechanism. Different theoretical perspectives are needed when analyzing young versus mature companies, and different moderating factors affect the impact of BI on IPO firm valuation, depending on the firm's age. Measures related to innovativeness, such as HTKIS and listing on second markets, are relevant moderating factors for young companies but not for mature ones, which instead have separation between ownership and control as a key moderating factor. Age is a relevant moderating factor for both young and mature companies, but acting in opposite directions: BI is more important the younger a young company is and the older a mature company is.

Robustness Tests and Sample Splits

In the previous section, we measure Tobin's Q at offer price. However, the existing literature indicates that underwriters systematically discount their value estimates, in order to augment investor participation and compensate for information disclosure by informed investors (Loughran & Ritter, 2002). Accordingly, we test whether our results persist by calculating Tobin's Q with the preliminary offer price (the midpoint of the initial price range, as disclosed in IPO prospectuses) and the first-day market price. In this way, we control for the price revision and underpricing phenomenon. The former allows us to adjust the price by canvassing investors' demands, whereas the latter incorporates first-day market effects.7

In Model 1 of Table 5, the dependent variable is Tobin's Q calculated by replacing the offer price with the preliminary offer price. In Model 2, Tobin's Q is calculated with the first-day price. We also perform the regression analysis using alternative valuation measures. In Model 3, the dependent variable is the M/B ratio, calculated as market value of equity over the book value of equity. In Model 4, we use the EV/Sales ratio, in which EV is calculated as the sum of the book value of assets and the market value of common stock, minus the book value of common stock. The results do not change significantly across the different models.

Table 5. Effect of board independence on IPO valuation: robustness tests on the dependent variable
 (1) Tobin's Q (preliminary offer price)(2) Tobin's Q (first-day closing price)(3) M/B ratio(4) EV/Sales
  1. This table reports the results of the 2SLS regressions, performed on the full sample of 969 European IPOs. The first stage is not reported for the sake of readability. In all cases, Board Independence and its interactions are instrumented by Mimicking Behavior and its interactions with Age and Age2. Dependent variable are: Tobin's Q calculated by replacing the offer price with the preliminary offer price (Model 1); Tobin's Q calculated by replacing the offer price with the first-day price (Model 2); M/B ratio, defined as the ratio of market value of equity to the book value of equity (Model 3); and EV/Sales ratio, in which EV is calculated as the sum of the book value of assets and the market value of common stock minus the book value of common stock (Model 4). Tobin's Q is the ratio of the ratio of the market value of assets (the sum of the book value of assets and the market value of common stock, based on the IPO offer price, minus the book value of common stock) to the book value of assets. Board Independence is the percent of non-executive members on the board, as reported in the IPO prospectus. Age is the logarithm of one plus the age of the company in years. V/C is the ratio of the controlling shareholder's voting rights (V) to its cash-flow rights (C). C measures the controlling shareholder's percentage ownership of the profits and dividends of the firm. If multiple chains of ownership exist, then the cash-flow rights along each chain are the products of all ownership rights in the intermediate companies along that chain. The total cash-flow rights are equal to the sum of all cash-flow rights from all ownership chains. V is computed as in La Porta et al. (1998). When multiple control chains exist, the voting rights are the sum of the voting rights along each chain with the weakest link among all holding layers. Board Ownership (in percent) is measured as the ratio of total beneficial shares held by executive directors at year-end to the total number of shares outstanding at year-end. Board Size is the number of directors on the board, including the chairperson. CEO Duality is a dummy indicating that the CEO and the chairman of the board of directors are the same person. Firm Size is the logarithm of net sales for the year prior to the IPO date, adjusted for inflation. Profitability is the return on assets (in percent) in the year before the IPO. Leverage is debt over total assets (in percent) in the year before the IPO. Offer Size is the ratio (in percent) between the size of the firm and the size of the offer. Offer Structure is the fraction (in percent) of secondary shares sold by existing shareholders relative to the total number of shares offered. Reputation of the underwriter is equal to 100 when the underwriter is in the Carter-Manaster ranking (list taken from Jay Ritter's website), and otherwise as equal to the underwriter's percentage market share in terms of number of IPOs underwritten in Europe during 1995 to 2011. Market Momentum is defined as the FTSE Euromid percentage index return calculated over the 6 months before the offer date. Accruals is a measure of pre-IPO abnormal accruals computed following DuCharme et al. (2001). Mimicking Behavior is defined as the ratio of non-executive members in the BoDs of all the firms belonging to the same industry (ICB code, first digit) and listed in the same stock market in the IPO year. Each regression controls for time, industry, and country effects (Germany is the omitted category). Coefficients are omitted for the time and industry effects. Heteroskedasticity-robust standard errors are reported in brackets. ***, **, and * indicate significance levels below 1%, 5%, and 10%, respectively.
Board Independence3.366**3.259**13.48***8.628**
(1.604)(1.644)(4.985)(3.393)
Board Independence × Age−.856**−1.031**−3.930***−2.649***
(.433)(.438)(1.361)(1.015)
Board Independence × Age2.143**.170***.586***.406***
(.0580)(.0588)(.193)(.150)
Age−.0854−.304.713−1.547***
(.254)(.249)(.662)(.544)
Age2−.0522−.0212−.243*.200**
(.0473)(.0468)(.133)(.101)
V/C−.174**−.144−.695***−.389*
(.0748)(.0885)(.221)(.226)
C.00571.00331.00453−.0114
(.00409)(.00421)(.0109)(.00810)
Board Ownership−.355−.354−.287−3.664***
(.248)(.258)(.685)(.553)
Board Size−.0267−.0318−.132*.105*
(.0262)(.0275)(.0784)(.0615)
CEO Duality−.690***−.843***−1.568***−.790**
(.179)(.190)(.490)(.377)
Firm Size−.723***−.719***−1.127***−2.340***
(.0473)(.0487)(.144)(.120)
Profitability−.0926.0268−.0129−1.177**
(.0947)(.110)(.467)(.475)
Leverage.0304**.0113−.0174.212***
(.0144)(.0165)(.0701)(.0711)
Offer Size.0478***.0358**.291***.110***
(.0152)(.0161)(.0672)(.0356)
Offer Structure−.116.270.2081.470**
(.239)(.381)(1.068)(.598)
Underwriter reputation1.073***.900***1.531**3.782***
(.227)(.231)(.608)(.479)
Market Momentum−.9331.676−7.5622.578
(2.260)(2.378)(6.109)(4.477)
Accruals.308***.271**.786***−.0376
(.112)(.118)(.220)(.258)
VC backing.358**.279−.667.949**
(.182)(.189)(.493)(.377)
France−.070−.304−2.097**−.339
(.401)(.394)(.879)(.743)
Italy−.135−.462−.9271.033
(.406)(.392).922(.713)
Constant15.852***16.674***25.397***42.201***
(1.102)(1.177)(3.913)(2.557)
Observations969969969969
Centered R-squared (2nd stage).388.371.288.536
Cragg-Donald Wald F statistic138.800138.800138.800138.800
[Stock-Yogo weak ID test critical values: 5% maximal IV relative bias][9.53][9.53][9.53][9.53]
Hansen (endogeneity test).485.6322.4973.325
[Chi-sq(k) p-value][.486][.427][.114][.068]

Our sample of IPOs covers 17 years and three countries. While the span and breadth of the sample is one of the distinctive merits of this paper, there is still a risk that the results may be driven by a single country or special period. To rule out this possibility, we run subsample regressions distinguishing IPOs in the tech bubble years from 1999 to 2000 (Loughran & Ritter, 2004) and for each stock exchange. Table 6 reports the results of these analyses.

Table 6. Effect of Board Independence on IPO Valuation: Sample Split by Time Period and Listing Market
 (1)(2)(3)(4)(5)
BubbleOut of bubbleFranceGermanyItaly
  1. This table reports the results of 2SLS regressions in which Tobin's Q is the dependent variable, performed on the full sample of 969 European IPOs. The first stage is not reported for the sake of readability. In all cases, Board Independence and its interactions are instrumented by Mimicking Behavior and its interactions with Age and Age2. In Models (1–2), the sample is split between IPOs carried out during the internet bubble (1999–2000) and all other periods. In Models (3–5), the sample is split according the reference market (Euronext, Frankfurt, or Milan). Board Independence is the percent of non-executive members on the board, as reported in the IPO prospectus. Age is the logarithm of one plus the age of the company in years. V/C is the ratio of the controlling shareholder's voting rights (V) to its cash-flow rights (C). C measures the controlling shareholder's percentage ownership of the profits and dividends of the firm. If multiple chains of ownership exist, then the cash-flow rights along each chain are the products of all ownership rights in the intermediate companies along that chain. The total cash-flow rights are equal to the sum of all cash-flow rights from all ownership chains. V is computed as in La Porta et al. (1998). When multiple control chains exist, the voting rights are the sum of the voting rights along each chain with the weakest link among all holding layers. Board Ownership (in percent) is measured as the ratio of total beneficial shares held by executive directors at year-end to the total number of shares outstanding at year-end. Board Size is the number of directors on the board, including the chairperson. CEO Duality is a dummy indicating that the CEO and the chairman of the board of directors are the same person. Firm Size is the logarithm of net sales for the year prior to the IPO date, adjusted for inflation. Profitability is the return on assets (in percent) in the year before the IPO. Leverage is debt over total assets (in percent) in the year before the IPO. Offer Size is the ratio (in percent) between the size of the firm and the size of the offer. Offer Structure is the fraction (in percent) of secondary shares sold by existing shareholders relative to the total number of shares offered. Reputation of the underwriter is equal to 100 when the underwriter is in the Carter-Manaster ranking (list taken from Jay Ritter's website), and otherwise as equal to the underwriter's percentage market share in terms of number of IPOs underwritten in Europe during 1995 to 2011. Market Momentum is defined as the FTSE Euromid percentage index return calculated over the 6 months before the offer date. Accruals is a measure of pre-IPO abnormal accruals computed following DuCharme et al. (2001). Mimicking Behavior is defined as the ratio of non-executive members in the BoDs of all the firms belonging to the same industry (ICB code, first digit) and listed in the same stock market in the IPO year. Each regression controls for time and industry dummies. Country effects are included in Models (1–2) (Germany is the omitted category). Coefficients of time and industry effects are not reported for the sake of readability. Heteroskedasticity-robust standard errors are reported in brackets. ***, **, and * indicate significance levels below 1%, 5%, and 10%, respectively.
Board Independence6.246*8.958***4.024***.545**.989**
(3.772)(1.913)(1.368)(.265)(.461)
Board Independence × Age−1.879*−2.254***−.603**−.758***−1.107***
(1.043)(.640)(.240)(.235)(.269)
Board Independence × Age2.319***.275***.070.216***.259***
(.131)(.104)(.053)(.0507)(.0671)
Age.156−.176−.880**.438−1.223**
(.479)(.267)(.386)(.322)(.527)
Age2−.160−.0174.102−.152**.128
(.106)(.0489)(.0741)(.0689)(.0890)
V/C−.732**−.286***−.196*−.0955.0357
(.307)(.0849)(.102)(.512)(.168)
C−.00158.000503.00507.00259.00665
(.00810)(.00467)(.00607)(.00536)(.0101)
Board Ownership−.576−.562**−.138.275−1.143***
(.580)(.282)(.325)(.467)(.436)
Board Size.0451−.118***−.0221−.0497−.0287
(.0520)(.0296)(.0381)(.0431)(.0513)
CEO Duality−.261−1.190***.0614−.774**−.298
(.325)(.211)(.232)(.360)(.305)
Firm Size−.902***−.623***−.676***−.786***−.785***
(.0972)(.0561)(.0618)(.0978)(.107)
Profitability.176−.0469−.0722.4701.538
(.711)(.129)(.0971)(.459)(1.180)
Leverage−1.391.0197.0246−.0570−1.518*
(.991)(.0190)(.0191)(.0688)(.848)
Offer Size1.486.0208.985−1.500**.0618**
(1.062)(.0156)(.648)(.626)(.0273)
Offer Structure.182.567*1.092***−.169**−.367
(.474)(.295)(.321)(.0786)(.456)
Underwriter reputation1.471***.735***.938***.971***1.394***
(.387)(.267)(.342)(.324)(.373)
Market Momentum12.85***1.6101.5403.2492.707
(3.990)(2.430)(2.304)(3.796)(4.173)
Accruals.235.236**.294.0111.614*
(.292)(.0989)(.381)(.107)(.371)
VC backing−.0304.383*−.111.630**−.284
(.383)(.208)(.227)(.271)(.368)
France−.054−.803**   
(1.135).338   
Italy−.163−.838**   
(.641).339   
Constant20.093***14.702***15.645***20.340***21.069***
(2.005)(1.244)(1.217)(2.078)(2.300)
Observations289681428383158
Centered R-squared (2nd stage).452.370.513.416.718
Cragg-Donald Wald F statistic26.081120.96680.83690.96378.607
[Stock-Yogo weak ID test critical values: 5% maximal IV relative bias][9.53][9.53][9.53][9.53][9.53]
Hansen (endogeneity test)10.98013.5845.1178.50811.281
[Chi-sq(k) p-value][.004][.002][.023][.003][.001]

We find that the key results are confirmed for IPOs in the tech bubble (Model 1) and outside the bubble (Model 2), although with lower statistical evidence for the former. In young companies, the marginal effect of BI on the IPO valuation decreases with the firm's age; in mature companies, the marginal effect increases with the firm's age. These results are strongly supported at the single-country level for Germany and Italy. However, we do not find statistical evidence of a positive marginal effect of BI increasing with age (coefficient of age squared) for French IPOs, possibly because of the relative scarcity in our sample of mature companies going public in that country (see Table 1).

Finally, we consider the robustness of our results across different quartiles of Tobin's Q using quantile regression. To obtain estimations that account for the potential endogeneity of the relationship between BI and Tobin's Q, we employ the estimator developed by Lee (2007), which combines the control function approach with quantile regressions. Table 7 presents the bootstrapped standard errors obtained after 500 replications. Overall, our results verify the effects of the interaction between BI and age on the firm's value. The impact of BI on IPO valuation is moderated by age and age squared, with the expected signs and acceptable statistical significance.

Table 7. Effect of Board Independence on IPO Valuation: Quantile Regressions
Independent variablesQuantile
.25.50.75
  1. This table reports the results of quantile regressions on Tobin's Q, performed on the full sample of 969 European IPOs. The control-function approach developed by Lee (2007) is used to control for endogeneity. In each regression, Board Independence and its interactions are instrumented by Mimicking Behavior and its interactions with Age and Age2. Board Independence and its interactions are instrumented by Mimicking Behavior and its interactions with Age and Age2. In Models (1–2), the sample is split between IPOs carried out during the internet bubble (1999–2000) and all other periods. In Models (3–5), the sample is split according the reference market (Euronext, Frankfurt, or Milan). Tobin's Q is the ratio of the ratio of the market value of assets (the sum of the book value of assets and the market value of common stock, based on the IPO offer price, minus the book value of common stock) to the book value of assets. Board Independence is the percent of non-executive members on the board, as reported in the IPO prospectus. Age is the logarithm of one plus the age of the company in years. V/C is the ratio of the controlling shareholder's voting rights (V) to its cash-flow rights (C). C measures the controlling shareholder's percentage ownership of the profits and dividends of the firm. If multiple chains of ownership exist, then the cash-flow rights along each chain are the products of all ownership rights in the intermediate companies along that chain. The total cash-flow rights are equal to the sum of all cash-flow rights from all ownership chains. V is computed as in La Porta et al. (1998). When multiple control chains exist, the voting rights are the sum of the voting rights along each chain with the weakest link among all holding layers. Board Ownership (in percent) is measured as the ratio of total beneficial shares held by executive directors at year-end to the total number of shares outstanding at year-end. Board Size is the number of directors on the board, including the chairperson. CEO Duality is a dummy indicating that the CEO and the chairman of the board of directors are the same person. Firm Size is the logarithm of net sales for the year prior to the IPO date, adjusted for inflation. Profitability is the return on assets (in percent) in the year before the IPO. Leverage is debt over total assets (in percent) in the year before the IPO. Offer Size is the ratio (in percent) between the size of the firm and the size of the offer. Offer Structure is the fraction (in percent) of secondary shares sold by existing shareholders relative to the total number of shares offered. Reputation of the underwriter is equal to 100 when the underwriter is in the Carter-Manaster ranking (list taken from Jay Ritter's website), and otherwise as equal to the underwriter's percentage market share in terms of number of IPOs underwritten in Europe during 1995 to 2011. Market Momentum is defined as the FTSE Euromid percentage index return calculated over the 6 months before the offer date. Accruals is a measure of pre-IPO abnormal accruals computed following DuCharme et al. (2001). Mimicking Behavior is defined as the ratio of non-executive members in the BoDs of all the firms belonging to the same industry (ICB code, first digit) and listed in the same stock market in the IPO year. Each regression controls for time, industry, and country effects (Germany is the omitted category). Estimated coefficients of time and industry effects are not reported for the sake of readability. Bootstrapped standard errors obtained after 500 replications are reported in brackets. ***, **, and * indicate significance levels below 1%, 5%, and 10%, respectively.
Board Independence7.062***4.983***11.000***
(1.116)(1.239)(1.692)
Board Independence × Age−1.447***−1.200***−2.943***
(.316)(.366)(.561)
Board Independence × Age2.165***.164***.489***
(.048)(.058)(.091)
Age−.552***−1.141***−.173
(.159)(.176)(.245)
Age2.066**.126***−.093*
(.033)(.037)(.051)
V/C−.059−.174**−.294***
(.056)(.079)(.091)
C−.002−.006*−.010***
(.003)(.003)(.004)
Board Ownership−.373**−.304*.098
(.168)(.177)(.238)
Board Size−.078***−.066***−.098***
(.020)(.022)(.031)
CEO Duality−.616***−1.180***−1.317***
(.116)(.128)(.188)
Firm Size−.551***−.749***−.732***
(.034)(.037)(.054)
Profitability−.268***−.032.361
(.085)(.133)(.384)
Leverage.056***.017−.048
(.013)(.020)(.058)
Offer Size.038***−.007−.039***
(.008)(.009)(.012)
Offer Structure1.021***1.842***1.670***
(.171)(.180)(.254)
Underwriter reputation.594***1.684***1.203***
(.136)(.155)(.219)
Market Momentum5.941***4.597***3.740**
(1.397)(1.377)(1.826)
Accruals.081***−.033.199***
(.031)(.074)(.073)
VC backing.346***.751***.995***
(.122)(.130)(.188)
France−1.001***−1.528***−1.821***
(.106)(.271)(.295)
Italy−.779***−1.258***−1.809***
(.126)(.317)(.347)
Constant10.916***16.951***16.340***
(.775)(.813)(1.162)
Observations969969969

Conclusions

The innovative aspect of our study is that it explores the relationship between the firm's valuation and board structure across its life-cycle. Although several prior works have studied the impact of corporate governance mechanisms on IPO performance and valuation, we focus our attention on how this relationship evolves as the company matures. We rely on a large sample of firms going public in continental Europe, where the firms going public are generally more diverse than in the United States (Ritter, 2003). In our sample, more than one-fourth (or one-third) of the IPO companies in our sample were younger than 1 year old (or older than 10 years old) at the time of listing. This great variety among our sample allows us to study the role of the BoD across many stages of maturity in the IPO companies.

In our study, the impact of BI on valuation (Tobin's Q) is investigated through the interaction with age, innovativeness, and separation between ownership and control. Control variables, such as the characteristics of the firm and offering, are also incorporated into the model. A key finding is that the effect of BI on firm value is U-shaped with age. In young companies, the value-creation role of the BoD dominates, and the impact of BI on firm valuation decreases with firm age. In mature companies, the value-protection role of the BoD dominates, and the impact of BI on firm valuation increases with firm age. We also find that young and mature companies are characterized by different factors that moderate the impact of BI on the firm's value. Measures of firm innovativeness are crucial in terms of moderating the impact of BI on IPO valuation for young companies (but not mature ones). Meanwhile, the separation between ownership and control is a critical moderating factor for mature companies (but not young ones). This evidence is consistent with the changing function of the BoD across the life-cycle of the company, with a value-creation role being dominant in young companies and a value-protection role being dominant in mature ones.

This paper has relevant implications for empirical analyses on the relationship between value and governance, whose changing nature has not yet received sufficient attention in the academic literature. Our findings have direct implications for entrepreneurs who are contemplating an IPO and would like to shape the process to their advantage. Investors and financial market participants will find information regarding the valuation of an asset class (IPO firms) that could deliver large satisfaction, if we consider that the top 100 IPOs earned over 1,000 percent in their first 3 years of trading (Field & Lowry, 2009). Our research may be of interest to policy makers, who are interested in setting best-practice standards regarding board structure. Finally, regulators and corporate governance advocates should consider the stage in the company's life-cycle when recommending how boards should be filled with non-executive members.

This study introduces the effect of the corporate life-cycle into the study of the impact of corporate governance on the valuation of IPOs, and thereby uncovers several avenues for further research. For example, future work could build on our findings to determine the impact of the corporate life-cycle with reference to different industries and business models. The impact of the firm's age might vary across industries and phases of the economic cycle. In particular, the importance of rapid growth has increased over time in many industries, due to the increasing speed of technological innovation. As a result, profitable growth opportunities may potentially be lost if they are not quickly seized. This situation impacts both the decision to go public and the relationship between age and valuation.

Additionally, whereas our study is based on a sample of IPO companies, an alternative research setting might consider the valuation impact of BI in seasoned firms. IPO companies tend to be relatively older in Europe than elsewhere, which explains our choice of sample. However, IPO companies are, almost by definition, relatively younger than listed companies. It would be interesting to extend our analysis to a sample of listed companies, to explore BI in even older companies than the ones we can observe in our sample. Another potentially interesting aspect is the role played by institutional factors in affecting the pace of evolution in the role of the BoD. To this extent, extending the analysis to countries outside continental Europe (especially market-based economies) would be a valuable addition.

Acknowledgments

We thank the editor Praveen Kumar, the associate editors of the special issue Rajesh Chakrabarti and Douglas Cumming, and two anonymous referees for their valuable and constructive comments. We are grateful to Hisham Farag, William Q. Judge, Erik Lehmann, Stefano Paleari, J. Ari Pandes, Shaker A. Zahra, and all the participants to the workshop “Global Perspectives on Entrepreneurship: Public and Corporate Governance” held at the Schulich School of Business, York University, Toronto (Canada). We also thank Mauro Seghezzi for his precious research assistance.

Notes

  1. 1

    In contrast, empirical evidence about the link between corporate governance and stock performance is much more nuanced. Gompers, Ishii, and Metrick (2003) and Bebchuk, Cohen, and Ferrell (2009) find evidence suggesting that better governance translates into higher stock valuation and returns. After controlling for the endogeneity of corporate governance, Bhagat and Bolton (2008) find no relationship between stock performance and measures of corporate governance quality. Finally, Faleye, Hoitash, and Hoitash (2011) and Fischer and Swan (2013) provide evidence that excessive monitoring causes lower stock returns.

  2. 2

    See Vismara et al. (2012) for a description of the database.

  3. 3

    In unreported regressions, we test the robustness of our results when using a BI dummy that is equal to one for IPOs with BI that is higher than the sample median of 57.1 percent, and zero otherwise (Chahine & Saade, 2011). Overall, the results are consistent with the ones shown here.

  4. 4

    We find similar results when the offer characteristics are measured by the dilution ratio (new shares issued at listing over market capitalization) and the participation ratio (percentage of the IPO composed of existing shares).

  5. 5

    The Carter-Manaster measure ranks underwriters based on their placement in the IPO “tombstone announcements,” which are marketing brochures in which banks deemed more reputable appear above those considered less prestigious (Carter & Manaster, 1990). This measure is not directly applicable to IPOs in Europe, where underwriting syndicates are not as large, and “tombstone announcements” do not exist. Its indirect application overweights US banks and does not include underwriters dealing only outside the United States. Therefore, the Carter-Manaster rankings alone do not correctly measure the reputation of underwriters in most European IPOs. For this reason, we rank the reputation of the underwriter as equal to 100 when the underwriter is in the Carter-Manaster ranking (list taken from Jay Ritter's website), and otherwise as equal to the underwriter's percentage market share in terms of number of IPOs underwritten in Europe during 1995 to 2011.

  6. 6

    Sample firms are categorized into 10 industries, according to the first digit of the Industry Classification Benchmark, the official industry classification adopted by the European stock exchanges. The sample includes IPOs in the 1995–2011 period. Therefore, firms are categorized in 17 IPO-years.

  7. 7

    Ritter (2011) and Levis and Vismara (2013) provide reviews of the literature on IPO pricing. See Jay Ritter's website for current data on the IPO activity across countries.

Biographies

  • Fabio Bertoni (PhD, CFA) is Associate Professor of Corporate Finance at the Department of Economics, Finance and Control of EMLYON Business School (France). His research activity focuses on the relationship between financing and firm performance, new listings, sovereign wealth funds, venture capital and corporate governance. He is author of articles in journals such as: Journal of Banking & Finance, International Finance, Technovation, Small Business Economics, Research Policy, Venture Capital, and European Financial Management. He was visiting professor at the Copenhagen Business School, Universidad Computense de Madrid, Centre for European Economic Research (ZEW) in Mannheim, and the University of Oxford – Saïd Business School.

  • Michele Meoli (PhD) is Assistant Professor of Corporate Finance at the Department of Engineering, University of Bergamo. He is a member of the CISAlpino Institute for Comparative Studies in Europe (CCSE), University of Bergamo and University of Augsburg. He was Marie Curie Research Fellow at the Centre for Econometrics Analysis, Cass Business School (City University London). His research interests include corporate governance, IPO valuation, academic entrepreneurship, and governance in higher education systems.

  • Silvio Vismara (PhD) is Associate Professor of Corporate Finance at the University of Bergamo, Italy. His research activity is mainly on initial public offerings. Silvio is associate editor of Small Business Economics, member of the Editorial Review Board of Entrepreneurship Theory and Practice, and is author of articles in journals such as Entrepreneurship Theory and Practice, Small Business Economics, European Financial Management, and Journal of Technology Transfer. He is co-founder and co-director of the CISAlpino Institute for Comparative Studies in Europe (CCSE). He is scientific consultant for the Italian Stock Exchange and founder of Universoft, a spin-off company from the University of Bergamo.

Ancillary