RAISING EXTERNAL CAPITAL AND FAMILY CONTROL An Analysis of Large European Firms. Ettore Croci a Università degli Studi di Milano - Bicocca

Transcription

1 RAISING EXTERNAL CAPITAL AND FAMILY CONTROL An Analysis of Large European Firms Ettore Croci a Università degli Studi di Milano - Bicocca Halit Gonenc b University of Groningen January, 2009 Abstract We examine the propensity to raise outside capital, both equity and debt, by family firms and compare it to that of non-family firms. We use a comprehensive sample of 777 large European firms in the period from 1998 to We find that family firms are, indeed, reluctant to make equity issues compared to non-family firms. However, all other things equal, family firms tend to make new equity issues following period in which stock prices substantially increased. Family control does not affect the decision to issue debt, unless the founder is still in the company (positive impact). Finally, family firms are as likely as non-family firms to use rights offerings. Our evidence presents interesting implications for family control firms in emerging countries to understand their long-term aspects of management choices for financing decisions. JEL Classification: G32 Keywords: Family firms, equity issues, debt issues, shareholder s identity. a Università degli Studi di Milano-Bicocca, Facoltà di Economia, P.zza Ateneo Nuovo, 1 ed. U6, Milan, Italy, e- mail: ettore.croci@unimib.it b University of Groningen - Faculty of Economics and Business, P.O. Box 800 Groningen 9700 AV, Netherlands, e- mail: h.gonenc@rug.nl

2 1. Introduction Recently researchers have focused their attention on how families affect performances and strategic decisions in the firms they control, usually defined as family firms. So far, the studies that examine the relationship between family ownership and firm value shows mixed results, with part of the literature documenting a positive overall effect of family control on firm performance (Anderson and Reeb, 2003; Villalonga and Amit, 2006; Andres, 2008; Villalonga and Amit, 2008), while the remaining articles find a negative contribution of family ownership on firm s performance (Claessens et al. (2002), Cronqvist and Nilsson, (2003), Maury (2006), and Bennedsen et al., 2007). It is also well-known that family firms often reduce firm s value using control-enhancing mechanisms like multiple share classes, pyramids, and voting agreements, which substantially reduce their valuation premiums (Villalonga and Amit, 2007, and Barontini and Caprio, 2006) and choosing the new CEO among family members in case of succession (Perez-Gonzales, 2006). In their paper on family firms, Bertrand and Schoar (2006) also suggest that families maximize their utility, not the one of all shareholders. The effect of family firms on acquisition decisions has received some attention as well. The main finding is that families tend to make firms reluctant not only to make acquisitions, but also to accept takeover offers (Sraer and Thesmar, 2007, Ben-Amar and Andre, 2006, Holderness and Sheehan, 1988, Klasa, 2007, and Caprio et al., 2008). Despite this growing interest on the implications of family control on listed corporations, there is a surprising lack of evidence about the differences between family firms and non-family firms in their choices of raise external capital. Cronqvist and Nillson (2005) provide evidence that, in Sweden, control considerations are particularly important in the 2

3 family firm s decision between different types of equity issue methods. In fact, family firms are very unlikely to use private placements, a method that can easily lead to the creation of a large block holder who can threaten their, often unchallenged, control. Bigelli (1998) examine insiders activism around rights offerings in Italy. He observed that large shareholders in Italian companies, often families, participate pro-quota to the underwriting of the new issued shares in rights offerings. However, we still ignore if, compared to non-family firms, family-controlled firms have a lower tendency to raise capital with equity or debt issues. This paper aims to fill this gap. Using a sample of 777 large European firms in the period , we examine and compare family firms and non-family firms in their decisions to issue debt and equity. We first analyze the propensity of raise external capital of these firms. Using data from Thomson One Banker s Equity and Debt databases, we find that 216 sample firms made at least one issue (debt or equity) during the period investigated, with 120 firms issuing new equity. Our main result is that family firms tend to make less issues and to raise less outside capital than nonfamily firms. However, this reluctance of family firms towards outside financing is only limited to equity issues. In fact, family firms are as likely as non-family firms to issue debt. The lack of equity issues by family firms can be explained by the well-documented fear of the controlling family to dilute or relinquish control. Family firms still led by founders, who have been found to positively affect firm s performance in previous studies (Miller et al. 2007, Andres, 2008), are particularly reluctant to issue equity. On the other hand, founder-led firms tend to issue more debt. These results are robust to a series of robustness tests that take into account other 3

4 motivations to issues equity like overvaluation for example. Our tests do not find support to the view that families tend to exploit overvaluation more than other types of firms. Following Cronqvist and Nillson (2005), we turn our attention to the choice between rights offerings and private equity placements. Using Thomson One Data s Equity database data, we document that, consistent with Eckbo (2008), rights offerings are not very common in our sample. In fact, less than 24% of the issues are, at least partially, rights offerings. Private placements are even rarer than rights offerings. There is no evidence that family firms use this equity issue methods more than non-family firms. Finally, we also examine the abnormal returns around the equity issue announcements. We find a relatively small reaction (-0.57%) for the full sample. This paper offers several contributions to the literature. First, it offers needed evidence on whether family firms and non-family firms behave differently when it comes to the decision to raise outside capital. Second, as an ancillary results of the analysis, it contributes to the identification of the main determinant of the decision to issue new equity and new debt. Our results give support to the view that firms issue equity after period of strong stock price performance. Third, this paper presents the first cross-country study at European level of equity and debt issues. The paper is structured as follows. Section 2 presents the hypotheses and the related literature. Section 3 presents the sample and the data. Section 4 analyzes the propensity to make equity and debt issues. Section 5 takes a closer look at equity issues. Section 6 concludes. 4

5 2. Literature Review and Hypothesis Development There is a growing literature on the role of families as controlling shareholders of listed firms. However, the academic literature has not fully investigated the role played by family control in determining how firms are financed. The well known argument for family controlled firms is that families do not take decisions that will put their control at risk. In fact, as originally observed by Demsetz and Lehn (1985) and Holderness and Sheehan (1988), individuals, and thus by extensions family members, value the opportunities to consume perquisite more than corporate majority shareholders, especially because of non-pecuniary and non-transferable private benefits. While there is some evidence that family firms adopt a very conservative strategy when it comes to acquisition decisions, either selling the firm to outsiders (Klasa, 2007 and Holderness and Sheehan, 2007) or making acquisitions (Sraer and Thesmar (2007), Ben-Amar and Andre (2006), Caprio et al. (2008)), not much is known about how differences in the strategies of family firms versus non-family firms when they need outside financing. From a similar perspective, Andres (2008) argues that families will be reluctant to raise new equity since an increase share capital will dilute their equity stake and gradually undermine their controlling positions. Thus the first hypothesis we are going to test is the following: Hypothesis 1: Family firms rely less on outside financing than non-family firms. In particular, family firms will avoid issuing equity. 5

6 Previous literature (Loughran and Ritter (1995) and Spiess and Affleck-Graves (1995), Baker and Wurgler (2002), Henderson, Jegadeesh, Weisbach (2006)) indicates that there is sufficient evidence that firms tend to issue new equity issues when equity is overvalued, that is firms try to time the market. However, the family-controlled firms may be more reluctant than non-family firms to exploit this temporary mispricing because of their well-known attachment to control and the private benefits they enjoy. Moreover, families are usually long-term investors in the firm they run. This can impact the choice of issuing equity when it is overvalued because, exploiting such opportunity may come at a very high cost in terms of loss of reputation. So we test the following hypothesis: Hypothesis 2: Family firms exploit less than non-family firms mispricing opportunities in the stock market. Control by a family firm will reduce the incentive to issue overvalued new equity both because family firms are reluctant to dilute control and because of reputational considerations. Regarding the issue method, as Cronqvist and Nilsson (2005) argue, outside the US, seasoned equity offerings (SEOs) are conducted primarily as rights offerings (ROs) or private equity placements (PPs). Using Swedish data, they provide evidence that family-controlled firms prefer a right offering over a private placement (PP). From the family s point of view, this choice has the advantage of avoiding the creation of a large investor (the PP investor) who would have an incentive to monitor the controlling family and the firm s use of the proceeds. Thus, the 6

7 second goal of our paper is to examine the choice of the equity issue method using a larger European sample. There are of course other factors, most importantly asymmetric information, that can affect the decision to issue equity, debt, or nothing at all. The effect of asymmetric information on the choice between a public offering and a private placement has been examined by Hertzel and Smith (1993). According to them, firms are likely to choose a private placement when the degree of asymmetric information about firm value is high. In fact, private placement investors have the incentives and the ability to learn the true value of the firm at some cost. Following Hertzel and Smith (1993) and Eckbo and Masulis (1992), Cronqvist and Nillson (2005) find that firms with extreme asymmetric information tend to raise new equity funds with private placements. They also find that firms with high asymmetric information tend to ask for an underwriter certification when they do a rights offering. These results are consistent with a more general argument of Eckbo and Masulis (1992) and Eckbo (2008), who argue that the rights offer method is optimal only if current shareholders are expected to buy and hold the new shares. Otherwise, the rights offer fails to mitigate the adverse selection problem and it becomes expensive. Eckbo (2008) also argues that the cost of rights is likely to be high in companies with a fragmented ownership structure. Since family firms are characterized by concentrated ownership structures and families are reluctant to dilute their stakes, we can expect that family firms will use rights offering more than non-family firms when an equity issue takes place. 7

8 Hypothesis 3: Conditional to an equity issue, family firms are expected to use rights offers more than non-family firms. Moreover, family firms rely less on private equity placements than nonfamily firms. 3. Data and Descriptive Statistics The starting point of our analysis is the sample of 777 firms used by Caprio, Croci, and Del Giudice (2008). This sample includes all publicly traded Continental European non-financial companies available on Worldscope, whose total assets (Worldscope item WC07230) exceed US$ 250 million at the end of Firms are from the following countries: Belgium (24), Denmark (38), Finland (37), France (161), Germany (144), Italy (72), Luxembourg (2), Netherlands (77), Norway (40), Spain (46), Sweden (64), and Switzerland (72). 1 As argued in Caprio et al. (2008), this sample permits to examine firms with a large dispersion of ownership structures both in terms of amount of the largest shareholders cash-flows and voting-rights, and in terms of family/non family control, which cannot be replicated with a UK or a US sample. We have detailed ownership data for the 777 companies in our sample. 2 In fact, cashflow rights and voting rights of the largest shareholders computed according to the now standard methodology developed by La Porta et al. (1999), and used by Claessens et al. (2000) and Faccio and Lang (2002), are available for 3,510 firm-year observations. We consider as family-controlled any company in which a family or individual is the largest shareholder with more than 10% of voting rights. 1 No firm from Austria, Portugal and Greece survives screening criteria. 2 We thank Lorenzo Caprio and Alfonso Del Giudice to make their database available. 8

9 In addition to ownership structure variables, we control for variables that are known to affect with both the propensity to issue debt and equity and the stock market reaction at the announcement. We present the definition of these variables in the Appendix. In all tables, the values of these variables (including ownership ones) refer to the end of the previous calendar year. The variables are the following: - Size: the market value of the company s equity. 3 Franck and Goyal (2007) in their survey of trade-off and pecking order theories of debt document that small firms actively use equity financing, while large companies rely more on corporate bonds. - Sales growth: the growth rate of the company s sales. A strong sales growth likely increases the company s cash flows, thus reducing the need of external financing; - Profitability (ROA): firms that are doing well generate more cash flows, decreasing the need to raise capital from external sources; - Cash holdings: cash-rich firms can use their cash reserves to fund their investment projects without issuing any new security; - Leverage: a high debt ratio can increase the likelihood of issuing new equity since the firm could have almost exhausted its debt capacity and may want to rebalance its capital structure. A high debt level should also reduce the probability of taking on more debt; - Collateral: it is defined as tangible assets over total assets. It increases debt capacity and therefore makes it easier for a firm to raise new debt capital; - Market-to-book: firms have incentives to issue equity when their stock price is overvalued. Moreover, high market-to-book may also signal growth opportunities. If a firm has growth 3 In the regressions, we use the log of the market value of the company s equity as proxy for size. 9

10 opportunities but not enough free cash flows, it may have to issue new debt or equity. Thus, a high market-to-book is expected to positively affect new issues. - Stock Price Performance: this variable controls for the possibility that firms raise new external capital after a period of excellent stock price performance. - Idiosyncratic volatility and Analysts: these variables are proxies for the asymmetric information problem a firm can face when it decides to issue equity or debt. The idiosyncratic volatility is computed as the standard deviation of the firm s daily stock price returns over the previous calendar year. The variable "analysts" is a measure of the analyst coverage for a given firm and is the number of analysts following the firm that issue recommendations on its stock. In Table 1 we summarize descriptive statistics about these variables for the 777 companies in our sample. 4 [Please insert Table 1 about here] We can observe from Panel A that, as common in Continental European listed companies, the controlling shareholder owns on average a remarkably large stake (30%). As noted by Faccio and Lang (2002), an important divergence from the one-share-one-vote principle exists in many firms (the median is 0, but the average wedge is 10.67%). Moreover, the average firm does not have a second large block holder who can monitor and challenge the controlling shareholder. In fact the average voting rights of the second block holder is just 6.55%. 4 To avoid that large outliers affect results, we winsorized all financial variables at 0.01 and

11 Following Caprio and al. (2008), Panel B compares of the same statistics between family firm-years and non-family firms-years. In family firms, the controlling shareholder owns a larger fraction of the cash flows rights in their firms than other types of controlling shareholders (35.57% vs %). They also rely more on control enhancing devices to create a positive wedge. There are also differences in the remaining variables; most evidently, family firms are smaller than non-family firms (median market cap US $557m vs. US $790m), they hold more cash and higher collateral. Concerning our proxies of asymmetric information, family firms have higher idiosyncratic volatility and are followed by fewer analysts than non-family firms. We then analyze whether and how the 777 sample companies issued debt and equity in the years We start by considering all equity and debt issued announcements reported in Thomson Financial Securities Data s Equity and Debt Databases over the five years between January 1998 and December We use SDC to collect information about the issues. After matching the data we obtained from SDC, we find that the 777 sample firms made 603 new issues in the period Among these issues, sample firms issued equity 240 times and debt 363 times. In our sample, we have 216 firms that made at least an issue during the five years period considered (27.80% of our sample firms). Equity and debt issues were made by 154 and 120 firms, respectively (19.81% and 15.44%). Only 58 firms made both equity and debt issues in the period considered. Overall, these firms made 240 equity issues and 363 debt issues throughout the period examined. At firm-year level, there are 216 observations both for equity issues and debt issues (397 in total considering both types of issues). Table 2 (Panels A and B) 11

12 summarizes these data. Panel C shows that family firms made less equity and debt issues than non-family firms. Moreover, the proceeds raised from these issues are smaller for family firms. Panels D to F compare issuers to non-issuers. Important differences arise. Issuers have ownership structures different from those of non-issuers. Issuers have less concentrated ownership, and they also tend to have smaller second largest block holders. Moreover, equity issuers use less control enhancing mechanisms (like pyramids, dual class shares) and have a smaller wedge. Issuers are on average larger, have more liquidity, and less debt than nonissuers. Surprisingly, debt issuers are more leveraged than non-debt issuers, indicating that probably these firms have not reached their full debt capacity yet. Interestingly, issuers, including both debt and equity issuers, have extremely good stock price performance in the calendar year preceding the issue. Their market-to-book ratio is significantly larger than the ratio of non-issuing firms. For equity issuers, the difference is quite significant from an economic point of view. The average market-to-book is 2.51 for non-equity issuers and 4.10 for equity issuers (medians: 1.68 and 2.72). This finding gives support to the story that firms issue equity when they are overvalued. The difference is less accentuated for debt issuers, but still significant. Issuers are followed by more analysts than non-issuers. This is consistent with the fact that issuers are larger companies. Idiosyncratic volatility only matters in the case of equity issuers. However, surprisingly, equity issuers are the ones with the highest idiosyncratic volatility. [Please insert Table 2 about here] 12

13 Overall, the table indicates that there are several factors that may affect the decision of raise funds externally. Multivariate logit regressions will help us discover if the presence of a family as controlling shareholder is one of them. 4. Propensity to raise capital with debt and equity issues Table 3 presents the results of logit regressions where the dependent variable takes value one if in year t the firm makes at least one issue of any type (Columns I and II), at least one equity issue (Columns III and IV), and at least one debt issue (Columns V and VI). These dummies are then regressed on ownership structure and financial variables for the year t-1. The unit of analysis is at firm-year level. As it is clear from Table 3, being controlled by a family is particularly important in the decision to issue equity. In fact, in Columns III and IV the coefficient for family is negative and statistically significant. The negative coefficient indicates that the presence of a family at the helm of the firm reduces the probability of issuing new equity, even once we control for ownership and financial variables. On the other hand, the coefficient for family firms is not significant for debt issues. These results support the view that families are very reluctant at giving up, even partially, control. No other ownership variable is significant, with the only exception of the stake owned by the second largest shareholder in the debt regressions. A second block holder with a large stake decreases the probability of a debt issue. As suggested by the descriptive statistics, size matters when it comes to make new issues. The coefficient for the log of the size, measured as the market value of equity of the company, is positive and significant in all the regressions. Having cash reserves negatively 13

14 affects the probability of making a debt issue, as expected, but it does not impact the probability of having an equity issue. Leverage has a positive coefficient in the debt issues regressions, confirming the results of the univariate tests. The market-to-book ratio affects the probability of having an equity or debt issue differently. As expected from market timing considerations as well as from the necessity to finance new growth opportunities, a high market-to-book ratio increases the likelihood that the firm will issue new equity. However, the evidence in Columns V and VI appears to be consistent only with market timing story. In fact the coefficient for the market-to-book ratio is negative coefficient. This result supports the view that firms want to exploit the opportunity to sell overpriced equity, it will make an equity issue, not a debt issue. Firms are more likely to issue equity when their stock price performance is good. A good operating performance (measured as ROA) decreases the probability that firms need to raise outside capital. This result is in line with the pecking order theory (Myers and Majluf, 1984). Finally, the two proxies for information asymmetry give different results: while idiosyncratic volatility positively affects the likelihood of an equity issue, the number of analysts that issue recommendations about a firm is only significant in the debt regressions (positive coefficient). [Please insert Table 3 about here] In Table 4, we run the same logit regression model but we divide family firms into five groups, based on the role of the founder and heirs. The first one is composed by firms where the founder is the CEO (or chairman) of the company; the second is still characterized by the presence of the founder in the board of directors, but without being CEO or chairman; the third 14

15 group is made up by firms where the founder is no longer present but an heir is CEO (or chairman); in the fourth group, heirs sit in the board of director, but they are not involved in executive tasks. Finally, no family member is either an executive or a board member in the last group of firms. When we include these variables, we can observe that the presence of the founder has a negative effect on the probability of issuing equity, consistent with the fact that founders are the most attached to the firm, but a positive one on the probability of using new debt issues. Heirs in the board of directors reduce the likelihood of an equity issue. All other results are similar to those of Table 3. [Please insert Table 4 about here] To take into account the fact that not all issues have the same importance, we also run Tobit regressions where the dependent variable is the ratio between the proceeds obtained with the issue and the market value of the company s equity. 5 Tables 5 and 6 present the results for equity and debt issues, respectively. Results in Table 5 are similar to those of the logit models. The family dummy is significant and negative, indicating once again that being controlled by a family firm reduces equity issues. Among the types of family firms, firms with the founder still alive and in the company issue less equity. Firms with an heir in the board of directors issue less new equity, too. The findings in Table 6 for debt issues also closely match those obtained in Tables 3 and 4. Here, the coefficient for family firms is not significant. However, founder-led family firms positively affect the amount of debt issued by the firm. 5 We use a Tobit model instead of a standard OLS model because the dependent variable cannot take values less than zero. 15

16 [Please insert Table 5 about here] [Please insert Table 6 about here] Finally, in Tables 7 and 8 we present the results of robustness checks. As widely known, leverage and market-to-book ratios vary considerably across industries. Thus, it could be that the results we obtained before are due to industry effects. In order to mitigate these concerns, we compute the industry-adjusted leverage and market-to-book ratio for each firm-year in our sample. We compute these industry adjusted measures using the 12-industry classification proposed by Fama and French. 6 As it appears clear in Table 7, our results are not due to the fact that firms operate in industry with high leverage or market-to-book ratio. Results are remarkably similar to those presented in Table 3. More important, the dummy for family control is still significant and negative in Columns III and IV, confirming the negative effect that families have on equity issues. [Please insert Table 7 about here] Table 8 takes into account the effect of growth in GDP. Dittmar and Dittmar (2008) document that that periods of economic expansion favor new equity issues because they reduces the cost of equity relative to the cost of debt. To include the effect of the growth in GDP, we use the following approach. Using data for real GDP growth rates from the World Bank World Development Indicators, we identify expansionary and contraction periods for each country in our sample over the period 1980 to First, we identify each year in this period 6 Nothing changes if we use the Fama and French 48-industry classification. 16

17 as low, middle and high growth year. Low growth years are those that have real GDP growth rate in the first quartile and high growth years are in the forth quartile of the empirical distribution of the country s growth rate. The remaining years are classified as middle growth years. Second, we label each year as the first, second or third years of low, middle, and high growth years. 7 There is no year L2 or L3 in our sample, but we have L1, H1, H2, and H3. 8 The findings of Table 8 again confirm the importance of family control in deterring firms from issuing new equity. Differently from Dittmar and Dittmar (2008), we do not find a significant effect due to growth in GDP, probably because of the short length of our sample period. To test the hypothesis that family firms are more reluctant to exploit periods in which equity is overvalued by investors, we include in our regression models the interaction between the family dummy and the two variables that should capture overvaluation: the market-to-ratio and the stock price performance in the previous calendar year. As Table 9 shows, the interaction variables are not significant in the equity regressions. 9 Thus, the evidence is inconclusive about the whether or not family firms exploit less overvaluation than other types of firms. As expected, the interaction variables are not significant in the debt issue regression. We also run these regressions using the industry adjusted market-to-book instead of the firm s market-to-book. Results are similar to those presented in Table 9, and not reported in the sake of brevity. 7 For example if 2001 is a low growth year and 2000 is not a low growth year then, 2001 is labeled as L1 year. if 2000 is also a low growth year, then 2001 is labeled as L2 year. if this is so and then 2002 is a high growth year then 2002 is labeled as H1, etc. 8 In the regressions, years M are used as numeraire. 9 We obtain the same results if we use only the interaction of M2B and family. 17

18 5. Analysis of Equity Issues We now take a closer look at the 240 equity issues made by our sample firms. We were able to obtain data concerning private placements and rights offerings by the Thomson One Banker s Equity Issues Database. Eckbo (2008) argues that the cost of a rights offering can be prohibitively high in larger companies with a fragmented ownership structure. Family firms are well known for their concentrated ownership structures, thus they might be the ideal candidate to make such an offer. Moreover, Cronqvist and Nillson (2005) find that in Sweden family firms are twice as likely as non-family firms to avoid SEOs method that can put their control at risk, like for example a private placements. Table 10 reports the number of rights offerings, private placements, and the abnormal return in the event window (-2, +2) around the issue announcement. 10 First of all, the results indicate that neither rights offerings nor private placements are very common in our sample. In fact, Thomson One Banker classifies as rights issue only 52 new issues out of the 240 we have in our sample. There are also five issues in which rights were used only for a part of the offering. 11 The relatively small fraction of rights offers (less than 24% of the offers involves at least partially a rights issue) is consistent with the trend observed by Eckbo (2008) that, even in Europe, firms are using less and less rights issues. Concerning family firms, the results show that family firms are marginally more likely to use rights issue, but the difference is not statistically significant. Differently from the Swedish sample used by Cronqvist and Nillson (2005) where the 10 Abnormal returns are computed using a market model to adjust for systematic risk. 11 Thomson One Banker records as separate issues parts of a single issue that go to different markets. In a few cases, rights are used only for the shares that are offered in a particular market. The same is true for private placements. 18

19 number of private placements was similar to that of rights offerings, in our sample private placements are very rare. In fact, only 20 issues involve at least partially a private placement. The market reacts positively to a rights issue by a family firm (+1.48%), while the reaction to a rights issue by a non-family firms is negligible (+0.29%). However, the difference between the two groups is not statistically significant. When rights issues are not used, there is negative reaction to the news of the issues, both for family (-1.05%) and non-family firms (- 0.50%). The difference between a rights issue and an issue without rights is significant at the 10% level for family firms, but not for non-family firms. This result gives some support to the view that the market perceives more negatively the news that a family firm is issuing new equity to new investors than for non-family firms. The explanation may be related to the fact that families are the most reluctant to give up control, and thus, a non-rights issue may indicate serious problems or overvaluation. Consistent with this story, the market reaction to a private placement by a family firm is negative (-3.21%). Conversely, private placements by non-family firms are well received by investors (+1.95%), as usually find in the literature (Eckbo, Norli, Masulis, 2008). While relatively small, the market reaction is broadly consistent with the average market reaction for international SEOs reported in the survey by Eckbo, Masulis, and Norli (2008). The only exception concerns the already discussed negative abnormal returns for private placements of family firms. [Please insert Table 10 about here] Table 11 presents the results of logit regressions for rights offerings (Columns I and II) and private placements (Columns III and IV) to control for factors that may affect the univariate 19

20 results presented above. As it turns out, there is only very weak evidence that family control increases the likelihood of a rights offer in Column II, confirming the univariate results. The findings of the private placements regressions show that family control is not significant. Finally, in Columns V and VI we regress the announcement returns in the event window (-2, +2) on the same set of covariates used in previous Tables. Again, there is no evidence that family control affects the market reaction around the announcement of a new equity issue. [Please insert Table 11 about here] 6. Conclusions The paper presents evidence on the decision to issue new equity and debt for a comprehensive sample of European firms in the period We find that family firms make less equity issues and to raise less outside capital than non-family firms. However, family firms are as likely as non-family firms to issue debt. The lack of equity issues by family firms can be explained by the well-documented fear of the controlling family to dilute or relinquish control. We also found that firms who are still led by their founders make very few equity issues, but they are not averse to raising capital with debt issues. While our results confirm previous evidence that firms issue equity after period of strong stock price performance, our tests do not find support to the view that families tend to exploit overvaluation more than other types of firms. We also document that, consistent with Eckbo (2008), rights offerings are not very common in our sample. In fact, less than 24% of the issues are, at least partially, rights offerings. Private placements are even less common than rights offerings. We do not find 20

21 evidence suggesting that family firms use rights offerings and private placements differently from non-family firms. 21

22 References Andres, Christian, 2008, Family Ownership, Financial Constraints and Investment Decisions, Working paper, University of Bonn. Andres, Christian, 2008, Large shareholders and firm performance: An empirical examination of founding-family ownership, Journal of Corporate Finance, doi: /j.jcorpfin Anderson, Ronald C., and David M. Reeb, 2003, Founding-family ownership and firm performance: evidence from the S&P 500, Journal of Finance 58, Baker, M., Wurgler, J., Market timing and capital structure. Journal of Finance 57, Barontini, Roberto and Lorenzo Caprio, 2006, The Effect of Family Control on Firm Value and Performance: Evidence from Continental Europe, European Financial Management 12, Ben-Amar, Walid, and Paul Andre, 2006, Separation of Ownership from Control and Acquiring Firm Performance: The Case of Family Ownership in Canada, Journal of Business Finance and Accounting, 33, Bennedsen, M., Nielson, K.M., Perez-Gonzalez, F., Wolfenzon, D., 2007, Inside the family firm: The role of families in succession decisions and performance. Quarterly Journal of Economics, 122, Bertrand, Marianne, and Antoinette Schoar, 2006, The role of family in family firms. Journal of Economic Perspectives 20, Bigelli, Marco, 1998, The quasi-split effect, active insiders and the Italian market reaction to equity rights issues. European Financial Management, Vol. 4, No. 2, 1998, Claessens, S., Djankov, S., Fan, J.P.H., Lang, L.H.P., Disentangling the incentive and entrenchment effects of large shareholdings. Journal of Finance 57, Caprio, Lorenzo, Ettore Croci, and Alfonso del Giudice, Ownership Structure, Family control and acquisition decisions. Working paper. Cronqvist, H., Nilsson, M., Agency costs of controlling minority shareholders. Journal of Financial and Quantitative Analysis 38, Demsetz, Harold, and Kenneth Lehn, 1985, The structure of corporate ownership: causes and consequences, Journal of Political Economy 93, Dittmar, Amy K., and Robert F. Dittmar, The timing of financing decisions: An examination of the correlation in financing waves, Journal of Financial Economics 90, Eckbo, B.E., Masulis, R.W., Adverse selection and the rights offer paradox. Journal of Financial Economics 32,

23 Faccio, Mara, and Larry H. P. Lang, 2002, The ultimate ownership of Western European corporations, Journal of Financial Economics 65, Franck, Murray Z., and Vidhan K. Goyal, 2007, Trade-off and pecking order theories of debt, Forthcoming in B. Espen Eckbo (ed.), Handbook of Corporate Finance: Empirical Corporate Finance,Volume 2 (Handbooks in Finance Series, Elsevier/North-Holland). Henderson, Brian J., Narasimhan Jegadeesh, Michael S. Weisbach, 2006, World market for raising new capital, Journal of Financial Economics 82, Hertzel, M., Smith, R.H., Market discounts and shareholder gains for placing equity privately. Journal of Finance 48, Holderness, C. G., and D. P. Sheehan, 1988, The role of majority shareholders in publicly held corporation: An exploratory analysis, Journal of Financial Economics 20, Klasa, Sandy, 2007, "Why do controlling families of public firms sell their remaining ownership stake?", Journal of Financial and Quantitative Analysis 42, La Porta, Rafael, Florencio Lopez-de-Silanes, and Andrei Shleifer, 1999, Corporate ownership around the world, Journal of Finance 54, Loughran, T., Ritter, J., The new issues puzzle. Journal of Finance 50, Maury, B., Family ownership and firm performance: empirical evidence from Western European corporations. Journal of Corporate Finance 12, Miller, Danny, Isabelle Le Breton-Miller, Richard H. Lester, and Albert A. Cannella Jr., 2007, Are family firms really superior performers? Journal of Corporate Finance, 13, Myers, Stewart, and Nicholas Majluf, 1984, Corporate financing decisions when firms have investment information that investors do not, Journal of Financial Economics 13, Pérez-González, F., 2006, Inherited control and firm performance, American Economic Review 96, Spiess, D., Affleck-Graves, J., The long-run performance following seasoned equity issues. Journal of Financial Economics 38,

24 Appendix: Variables Definitions Variable VR UO CFR UO Wedge VR 2 nd LS Size Relative Size Sales Growth Collateral ROA Cash Holding Leverage M/B Stock Price Performance Run-up Idiosyncratic volatility Analysts Family No Founder/Heir Heirs Board Heirs CEO Founder Board Founder CEO L1 H1 H2 H3 Definition The ultimate owner s voting rights in the firm. The ultimate owner s cash-flow rights in the firm. the difference between cash-flow and voting rights held by the ultimate owner Voting rights held by the second largest shareholder in the company Log of the firm s market value of equity (Worldscope Item WC07210). Ratio between the deal value and the acquirer s size. Growth rate in total sales in the previous year (WC07240). Ratio of tangible assets to total assets (WC02501/WC02999). Return on assets (WC08326). Ratio of cash plus tradable securities over total assets (WC02001/WC02999). Ratio of book value of financial debt as a percentage of the book value of total assets (WC03251/WC02999). Ratio of market value of equity in US$ (WC07210) divided by common equity in US$ (WC07220). Stock price performance over the year before the issue (WC05070). Stock price performance over the event window (-240, -41) with respect to the announcement day of an acquisition. Standard deviation of daily stock returns over the year before the issue The number of analysts that issues recommendations for the company s shares. Dummy variable taking value 1 when the controlling shareholder is a family. Dummy variable taking value 1 when neither the founder nor an heir is a board member (or the CEO/Chairman) in a family firm. Dummy variable taking value 1 when heirs are board members in a family firm, but they are neither CEO nor Chairman. Dummy variable taking value 1 when an heir is CEO in a family firm. Dummy variable taking value 1 when the founder is only a board member in a family firm, but he or she is neither CEO nor Chairman. Dummy variable taking value 1 when the founder is the CEO (or Chairman) in a family firm. First low growth year. First high growth year. Second subsequent high growth year. Third subsequent high growth year.

25 Table 1: Descriptive statistics of sample companies The table presents descriptive statistics of the 777 sample companies. The variables are defined in the Appendix. Non-ownership variables (all variables except CF Ult. Owner, Wedge, VR 2 nd LS) are winsorized at 0.01 and The number of observations is in firm/years (at the starting date, 1997 January 1 st, 434 of the 777 companies are family-controlled according to our definition, 343 are not). In Panel B, the symbols ***, **, * denote statistical significance at the 1%, 5%, and 10% levels, respectively, for the tests of difference in means and the median tests between family and non-family firms. Panel A: Full Sample Mean Median N.Obs CF Ult. Owner Wedge VR 2 nd LS Size Collateral Cash holding Leverage M/B Stock Price Perf ROA Sales Growth Id. Volatility Analists Panel B: Family vs. Non-Family Firms Family Non-Family Mean Median N.Obs Mean Median N.Obs CF Ult. Owner 35.57*** 34.00*** Wedge 15.34*** 11.40*** VR 2 nd LS *** Size *** *** Collateral *** *** Cash holding *** *** Leverage ** M/B Stock Price Perf ROA ** Sales Growth * Id. Volatility ** ** Analysts *** 11***

26 Table 2: Issues by Sample firms Panel A breaks down the acquisitions made by our sample firms according to the type of acquisition during the period. Panels B and C document the acquisitions made by our sample firms according to the type of acquiring company (family/non-family controlled). Panel D documents the different characteristics of acquiring and non-acquiring firms. Panel E documents control changes and bankruptcies experienced by our sample firms according to the type of controlling shareholders. *The number of firms with control changes is different from the sum of the control change events because one firm may experience more than one control event during the fiveyear period examined. In panel B and D, the symbols ***, **, * denote statistical significance at the 1%, 5%, and 10% levels, respectively, for the difference in the percentage of acquirers between the two subsample. In Panels E and F, the symbols ***, **, * denote statistical significance at the 1%, 5%, and 10% levels, respectively, for the tests of difference in means and the median tests between family and non-family firms. Panel A: Issues in the period # of Issuer # of non- Issuer Equity Debt All Panel B: Issues by sample Firms throughout the years Year # Equity Issues Amount of Equity Issued # Firms issuing Equity # Debt Issues Amount of Debt Issued # Firms issuing Debt Total # Equity Issues Panel C: Family vs. non Family Amount of # Firms Equity Issued issuing Equity # Debt Issues Amount of Debt Issued # Firms issuing Debt Non Family Family

27 Table 2: Issues by Sample firms (Cont.) Panel D: Issuing Firms vs. Non-Issuing Firms Non-Issuing Firms Issuing firms Mean Median N.Obs Mean Median N.Obs CF Ult. Owner *** *** Wedge VR 2 nd LS ** * Size *** *** Collateral Cash holding ** Leverage *** *** M/B *** *** Stock Price Perf *** *** ROA Sales Growth *** Id. Vol *** N. Analysts *** *** Panel E: Equity Issuers vs. Non-Equity Issuers Non-Equity Issuers Equity Issuers Mean Median N.Obs Mean Median N.Obs CF Ult. Owner *** *** Wedge ** VR 2 nd LS Size *** *** Collateral Cash holding Leverage * M/B *** *** Stock Price Perf *** *** ROA Sales Growth ** Id. Vol *** ** N. Analysts *** *** Panel F: Debt Issuers vs. Non-debt Issuers Non-Debt Firms Debt Issuers Mean Median N.Obs Mean Median N.Obs CF Ult. Owner *** *** Wedge VR 2 nd LS *** *** Size *** *** Collateral Cash holding *** Leverage *** *** M/B ** *** Stock Price Perf ** * ROA * Sales Growth * Id. Vol N. Analysts *** ***

28 Table 3: Determinants of the Propensity to Issue In Columns I and II The table presents the results of logit regressions where the dependent variable is a binary variable that takes value one if the sample firm makes at least one issue (debt or equity) in year t. In Columns III and IV (V and VI) are the results of logit regressions where the dependent variable is a binary variable that takes value one if the sample firm makes at least one equity (debt) issue in year t. The variables are defined in the Appendix. Market and accounting variables are winsorized at 0.01 and Robust standard errors are in parenthesis. The symbols ***, **, * denote statistical significance at the 1%, 5%, and 10% levels, respectively. All Issuers Equity Issuers Debt Issuers I II III IV V VI Constant *** *** *** *** *** *** [0.6712] [0.7788] [0.7942] [0.9728] [0.9295] [1.0289] CFR Ult. Own [0.0032] [0.0034] [0.0040] [0.0043] [0.0042] [0.0045] Wedge UO [0.0042] [0.0043] [0.0068] [0.0067] [0.0050] [0.0054] VR 2nd LS ** * [0.0077] [0.0082] [0.0088] [0.0094] [0.0120] [0.0124] Size *** *** *** *** *** *** [0.0412] [0.0613] [0.0464] [0.0789] [0.0553] [0.0794] Collateral [0.3846] [0.3947] [0.5146] [0.5383] [0.4705] [0.4704] Cash Holding ** ** *** *** [0.7158] [0.7388] [0.8598] [0.8844] [1.0206] [1.0403] Leverage *** *** *** *** [0.5147] [0.5267] [0.6524] [0.6753] [0.6649] [0.6690] M/B ** *** *** *** [0.0227] [0.0235] [0.0236] [0.0252] [0.0386] [0.0373] Stock Price P ** *** ** ** ** [0.1272] [0.1360] [0.1486] [0.1672] [0.1676] [0.1680] ROA *** *** *** *** *** ** [0.0133] [0.0143] [0.0164] [0.0183] [0.0173] [0.0164] Sales Growth [0.2765] [0.2935] [0.3106] [0.3340] [0.3680] [0.3839] Id. Volatility * *** [8.5943] [ ] [ ] Analysts *** *** [0.0103] [0.0142] [0.0139] Family * *** ** [0.1441] [0.1482] [0.1873] [0.1923] [0.1991] [0.2057] Pseudo R Observations