Large shareholders and firm performance —An empirical examination of founding-family ownership ☆
Christian Andres ⁎
University of Bonn,Department of Economics —BWL1,Adenauerallee 24-42,53113Bonn,Germany
a r t i c l e i n f o a
文胸英文b s t r a
c t
Article history:Received 14May 2007Received in revid form 14May 2008Accepted 14May 2008Available online 23May 2008Using panel data on 275German exchange-listed companies I examine the relationship
between founding-family ownership and firm performance.By parating the family effect
halloween什么意思from general blockholder effects,the paper shows that family firms are not only more pro fitable
than widely-held firms but also outperform companies with other types of blockholders.
However,the performance of family business is only better in firms in which the founding-
family is still active either on the executive or the supervisory board.The findings suggest
that family ownership is related to superior firm performance only under certain conditions.If
families are just large shareholders without board reprentation,the performance of their
companies is not distinguishable from other firms.In addition,the results indicate that other
blockholders either affect firm performance adverly or have no detectable in fluence on
performance measures.
幼片下载
©2008Elvier B.V.All rights rerved.JEL classi fication:G32G34Keywords:Family firms Ownership structure
Blockholders
Agency theory
1.Introduction
Recent empirical evidence suggests that founding-family ownership is associated with superior firm performance when compared to widely-held companies,both in terms of accounting performance and market valuation (Anderson and Reeb,2003a;Villalonga and Amit,2006;Barontini and Caprio,2006).This result is found to be particularly strong for founder-led companies.Therefore,families as large blockholders em to successfully mitigate the owner-manager agency problem described by Jenn and Meckling (1976).The positive effects thus em to outweigh the possible counter-argument that large shareholders may u their position to extract private bene fits at the cost of minority shareholders.However,empirical studies so far only show that family ownership leads to a better performance when compared to non-family firms.The existing literature fails to adequately address the question of whether families as blockholders are more successful than other controlling shareholders.As a result,previous studies on the performance of family firms might document a general (positive)blockholder effect,rather than provide evidence in favor of families as speci fic blockholders.
In order to examine the issue I analyze a sample of 275German listed companies from 1998through 2004.In my study,family business make up 37.5%of the sample.Families,on average,own voting rights of 63.0%in their company.As families frequently Journal of Corporate Finance 14(2008)431–445
☆I would like to thank Wolfgang Ausnegg,Mark J.Flannery,Marc Goergen,Joel F.Houston,Jeffry M.Netter (the editor),Jay R.Ritter,Lukas Roth,Erik Theisn,minar participants at the University of Florida,Rotterdam School of Management (RSM),Bocconi University,participants of the 2006European Finance Association Meeting in Zurich,the 2006European Financial Management Association Meeting in Madrid,the 2006Annual Meeting of the German Finance Association and an anonymous referee for their helpful comments.I also acknowledge the contribution of the Hoppenstedt Financial Information GmbH to the data collection process.
⁎Tel.:+49228739205;fax:+49228735924.
E-mail address:andres@uni-bonn.de
shoal.
0929-1199/$–e front matter ©2008Elvier B.V.All rights rerved.
doi:10.1016/j.jcorp fi
n.2008.05.003
Contents lists available at ScienceDirect
Journal of Corporate Finance
j o u r n a l h om e p a g e :w w w.e l v i e r.c o m /l oc a t e /j c o r pfi n
432 C.Andres/Journal of Corporate Finance14(2008)431–445
employ mechanisms that violate the one-share-one-vote principle,the average family share of cashflow rights is lower,at48.7%. The results of the multivariate analysis show thatfirms with family blockholders outperformfirms with other types of blockholders as well as widely-heldfirms.It thus appears that not only are the blockholder position and the monitoring incentives it entails relevant,but the identity of the blockholder is as well.In panel regressions of performance measures on different blockholder types,family block ownership is the only variable showing positive and significant coefficients.The economic impact of family ownership on return on asts is significant and lies within the range of3.1%and4.5%(depending on the model specification)per year.The results are robust to alternative econometric specifications and em to be robust to endogeneity concerns.
A further analysis reveals thatfirm performance is highest for family business in which the founder is still active as CEO. Bad on accounting measures,their performance is still superior compared to non-familyfirms when either a descendant or a hired professional has taken over the CEO position.However,the performance of familyfirms is only better infirms in which the family is still active,either on the executive or the supervisory board.The results point out that families are only able to induce positive performance effects if they have a deeper relationship with their business a
nd act as stewards of thefirm.If families are just large shareholders without board reprentation,the performance of their companies is not distinguishable from otherfirms.
Thefindings are relevant for countries in which a high percentage offirms have a dominant shareholder.In fact,veral studies show that the image of the publicly traded company with a disperd ownership structure,free-rider problems,and the classical owner-manager conflict is not appropriate for most countries(Claesns et al.,2000;Faccio and Lang,2002;La Porta et al., 1999).Widely-held corporations are primarily prent in the U.S.,the U.K.and in Japan,whereas concentrated ownership is typical for Western Europe and the rest of Asia.Family-controlledfirms in particular are found to account for up to60%-70%of all companies in the countries.Bad on the study by La Porta et al.(1999),family control can be regarded as the most common ownership structure around the world.
Given thefindings,it is not surprising that previous studies on the U.S.do not specifically control for the identity of other blockholders;in a capital market that is characterized by disperd ownership structures,this is not much of an issue.In Germany, about85%of the listedfirms have at least one blockholder who holds voting rights of more than25%.Therefore,the German market provides an ideal environment to gain deeper insight into the performance of family-controlledfirms as compared
to companies controlled by other types of blockholders.This study further points out the necessity to not only control for blockholdings,but also for the identity of the blockholder when analyzing the effects of concentrated ownership.For example, Thomn et al.(2006)find a negative effect of blockholder ownership onfirm value and accounting profitability in Continental Europe.However,my results suggest that suchfindings might probably be driven by an insufficient distinction between different blockholder types.For some blockholders,concentrated ownership indeed ems to have a negative influence onfirm performance.In line with theoretical models on the privatization of state-owned Perotti,1995),firms with government blockholders show negative and highly significant coefficients in all regressions with accounting measures as the dependent variable.In contrast,families do not em to u their controlling position at the expen of minority shareholders, resulting in a significantly better performance of familyfirms compared to both widely-held corporations andfirms with other blockholders.
珠海翻译公司The remainder of the paper is organized as follows.The next ction provides a brief outline of the literature related to the performance effects of large shareholdings and points out what distinguishes families from other blockholders.Section3focus on the construction of the datat and prents descriptive statistics as well as univariate results.Section4contains the multivariate analysis bad on different panel regressions.Section5concludes.
2.Large shareholders andfirm performance
The influence of ownership structures onfirm performance has been rearched extensively in the theoretical and empirical literature.In exchange-listed corporations,ownership and control are usually parated(with the exception of managerial ownership),leading to conflicts of interest betweenfinanciers and the managers who run a company(Fama and Jenn,1983; Jenn and Meckling,1976).The agency-conflicts can be mitigated by monitoring.In widely-heldfirms,however,shareholders will most likely not be informed well enough and refrain from investing their personal resources in monitoring activity.This is referred to as the free-rider problem(Grossman and Hart,1980;Holmstrom,1982).
In contrast to small shareholders,large investors have a big enough stake that it pays for them to spend private resources to monitor management.Large investors thus provide a solution to the free-rider problem(Shleifer and Vishny,1986).In addition to a higher incentive to actively decrea agency costs,it is much easier for large shareholders to coordinate their actions and put pressure on managers since voting power is not split among a highly gmented group of investors.If managers repeatedly act against the wishes of the large investor,they are likely to be displaced soon.Therefore,large blockholders differ from small shareholders in that they do not only have the in
centive to decrea agency costs,but also the power to do so(Shleifer and Vishny, 1997).
On the other side,concentrated ownership can also imply potential drawbacks.First and foremost,large shareholders primarily reprent their own interests and not tho of other shareholders or employees of the company they invested in.This means that they will u their control rights in order to maximize their own utility,which might come at the expen of other shareholders. The probability of minority shareholder expropriation is particularly high if large investors hold voting rights in excess of cashflow rights(Faccio et al.,2001).In the cas,they have an incentive to pay out a larger proportion of company cashflows to themlves
instead of evenly distributing funds among all shareholders.One possibility to do so could be to redirect funds to other companies they control.Dyck and Zingales (2004)—among others —have empirically documented the bene fits of control over corporate asts.
Bad on theoretical considerations,it is uncertain as to whether large investors act in favor of all shareholders or if the overall effect of their in fluence on the firm is negative.In addition,empirical studies have not been able to provide unambiguous evidence concerning the effect of ownership concentration on firm performance.Morck et al.(1988)investigate the relationship between manageria
l ownership and market valuation and document a non-linear pattern:market value first increas,then declines (for ownership stakes between 5%and 25%),and finally increas again as ownership by the board of directors ris.For a large sample of firms from eight East Asian economies,Claesns et al.(2002)find a positive and signi ficant relationship between firm value and the cash flow ownership of the largest shareholder,a result interpreted as evidence for a positive incentive effect of large shareholdings.In contrast,studies by Demtz and Lehn (1985)and Demtz and Villalonga (2001)show no relation between ownership concentration and firm performance.In a cross-country study,Thomn et al.(2006)find a negative impact of blockholder ownership on firm performance for Continental Europe.They interpret the findings as evidence of a con flict of interest between large investors and minority shareholders.
2.1.Family ownership and control
At first sight,founding families are just one among different types of blockholders.Like other large shareholders (such as the state,financial institutions,and other companies),families have a strong incentive to decrea agency costs and increa firm value.However,veral characteristics of family ownership give reason to assume that they differ from other types of blockholders.
2.1.1.Potential bene fits of family ownership
As mentioned above,concentrated shareholders have strong economic incentives to monitor managers and decrea agency costs (Demtz and Lehn,1985).In the ca of founding-family ownership,this incentive should be particularly strong since families usually have invested most of their private wealth in the company and are not well-diversi fied.Therefore,families are a unique type of investor who has exceptional concerns over firm survival and strong incentives to monitor management cloly.If the monitoring activity requires the knowledge of a firm-or market-speci fic technology,families might also have an advantage due to their long-term prence in the firm.
bombardier
In many cas,owner-manager con flicts will fail to ari in the first place since family members are part of the executive board.In addition,knowledge and experience are more likely to be pasd on within families as oppod to shared with outsiders.If descendants grow up clo to the firm they will also be able to build trust with employees and develop long-term relationships with suppliers and other external stakeholders.According to Ward (1988)families successfully create a working environment that fosters trust and loyalty,leading to lower turnover and recruitment costs.The long-term nature of family shareholdings also suggests that family firms build a reputation that affects their relationship with customers and external suppliers of capital.Anderson et al.(2003)show that family firms enjoy a lower cost of debt financing compared to non-family firms.They attribute this findi
ng to the unique interests associated with the long-term family commitment and suggest that bondholders regard founding-family ownership as an organizational structure that decreas the con flicts between shareholders and themlves and conquently better protects their interests.
In other words,the arguments suggest that family ownership increas a firm's credibility to commit to implicit contracts.Bad on the rationale that long-term contracts between shareholders and stakeholders are necessary in order to promote firm-speci fic investments by stakeholders (Williamson,1979)and given that complete contracting is costly,implicit contracts are a desirable solution.However,firms may have a strong incentive to renegotiate the implicit agreements and might therefore lack credibility when making promis to other stakeholders.As a conquence,shareholders must gain the trust of potential stakeholders in order to realize the possible gains of implicit contracts.Due to their long-term prence in the firm,families might have an advantage in credibly committing to the agreements.In line with the arguments,Tagiuri and Davis (1996)argue that family relationships generate loyalty and increa trust.
Since most families regard their company as an ast that should be pasd on from generation to generation,rather than consumed during a lifetime (Casson,1999;Chami,1999),their investment decisions are bad on a long-term pro fit maximization (James,1999).Stein (1988)prents a formal
model in which firms with a longer investment horizon suffer less from managerial myopia and will not sacri fice long-term interests in order to boost current earnings.This should lead to more ef ficient investment decisions that are primarily bad on net prent value considerations.
2.1.2.Potential costs of family ownership
confirmationOn the other side,there are veral potential costs associated with founding-family ownership.The combination of management and control might lead to sub-optimal investment decisions since the interests of the family are not necessarily in line with tho of other shareholders (Fama and Jenn,1985).Instead of maximizing firm value,entrenched families might have an incentive to exchange pro fits for private rents and thereby expropriate minority shareholders (Faccio et al.,2001).Families can also incur costs that do not directly come at the expen of pro fits.Demtz and Lehn (1985)suggest the term “amenity potential ”,standing for nonpecuniary private bene fits of control.For example,a founder might derive pleasure from eing his offspring run the company he established.In addition,families tend to hand over executive positions to family members and thereby restrict the 433
C.Andres /Journal of Corporate Finance 14(2008)431–445
vivian是什么意思
434 C.Andres/Journal of Corporate Finance14(2008)431–445
rockylabor pool to a very small group.Morck et al.(2000)argue that entrepreneurial spirit and experti are only partly inherited and conclude that descendants gradually regress towards average talent and affectfirm performance negatively.The family's role in lecting managers and members of the supervisory board will also increa entrenchment and may lowerfirm value since external parties can hardly capture control over thefirm.In line with the arguments,the high family stake reduces the probability of bidding by other external investors and should lead to a lower market valuation(Barclay and Holderness,1989).The entrenchment of family executives also might cau founders to remain active in thefirm although they are no longer competent. Shleifer and Vishny(1997)suggest that this is one of the largest costs that large shareholders can impo.Their argument implies that the performance of familyfirms gets wor with increasingfirm age.
In addition to the arguments,families as large and undiversified investors might pursue risk reduction strategies.Basically, families can reducefirm risk in two ways.First,firm investments might be channelled towards projects that create uncorrelated cashflows relative to thefirm's core business.Such diversification strategies,although in the best interest of the controlling family, are not necessarily beneficial to small shareholders who might wish to diversify their portfolios independently.
Second,the family might ek less risky(in the n of a higher default probability)forms offirmfinancing and is likely to u less debt in thefirm's capital structure.Thereby,the family will hamper thefirm's ability to rai external funds for investment projects and also forgo the(potential)advantage of a higher debt tax shield.Both risk reduction strategies impo costs on well-diversified minority shareholders.In contrast to the arguments,Anderson and Reeb(2003b)find a lower degree of diversification and similar leverage ratios between family and non-familyfirms.
2.1.
3.Empirical evidence
The conflicting ideas have recently evoked a number of empirical examinations of the relationship between family ownership andfirm performance.In a panel study on S&P500firms,Anderson and Reeb(2003a)find that familyfirms perform better than non-familyfirms,both in terms of market and accounting measures.Their results point in the same direction asfindings by McConaughy et al.(1998).Morck et al.(2000)show contradictory evidence for Canada,arguing that family ownership leads to poor financial performance.Their results suggest that family control by heirs leads to slower growth becau of inefficiencies that are due to entrenchment,high barriers against outside control and low investment in innovation.
When investigating familyfirm performance more cloly,a so-called“founder effect”can be identified.Founders em to have a special influence and put forth unique value-adding skills that lead to better performance.Bad on accounting performance measures,Anderson and Reeb's(2003a)results indicate that familyfirms only perform better when a family member is CEO. Founder descendants as CEOs do not em to affect market performance.
For a sample of Fortune500(the500largest U.S.firms,as measured by sales)firms,Villalonga and Amit(2006)find that family ownership creates value only if the founder rves as CEO or as chairman of the board of directors with a professional CEO. Contrary to Anderson and Reeb(2003a),theyfind thatfirm value even declines when descendants rve as CEOs.The u of control mechanisms like multiple share class,pyramids,cross-holdings or voting agreements has a negative effect onfirm value, particularly in founder-led companies.
In another recent study on the U.S.,Miller et al.(2007)u alternative definitions of the term“familyfirm”and show that results are nsitive to this definition.In particular,they distinguish“lone founder”business with no other family of the founder in the firm with“familyfirms”that do include multiple family members as major owners or executives(either contemporaneously or over time).Miller et al.find that familyfirms do not outperform,even during thefirst generation.In contrast,“lo
ne founder”business consistently outperform otherfirms,irrespective of different model specifications.
In terms of the European evidence,Sraer and Thesmar(2007)show that for a sample of French stock-market listed companies, familyfirms outperform widely-held corporations.Their results hold for founder-CEOfirms as well as for heir-managedfirms.They explain thisfinding through implicit insurance contracts with the labor force in heir-managedfirms:employment is less nsitive to industry shocks and as a conquence heirs pay lower wages.
Finally,in a cross-country study of Continental Europeanfirms,Barontini and Caprio(2006)confirm thefinding that market valuation and operating performance are higher in founder-controlled corporations and at least not wor in descendant-controlledfirms.
Turning the focus to Germany,empirical evidence is scarce.Ehrhardt et al.(2004)investigate a sample of62family and62non-familyfirms.Theyfind that family business outperform non-familyfirms in terms of operating performance.However,the generality of their results is unknown since they compare thefirms over a100year time-span and thus require allfirms to survive from1903till2003.
In an analysis of German and U.K.initial public offerings(IPOs)in the period from1981to1988,Goerge
n(1999)compares the performance of62Germanfirms that werefloated by families to a sample of U.K.IPOs.For36German companies,the founding-family remains the largest shareholder over the entire sample period.The results show no statistically significant difference of post-IPO performance betweenfirms under family control and widely-held corporations.
In sum,papers examining the influence of ownership concentration onfirm performance have shown mixed evidence,while the majority of empirical studies on the effects of founding-family ownership found familyfirms to outperform.Taken together, thefindings suggest thatfirms with family blockholders should not only outperform widely-heldfirms,but alsofirms with other types of blockholders.However,it is important to note that the studies referred to above provide evidence on very different capital market environments with different institutional ttings,on different samples,and sample periods.None of the empirical papers on founding-family ownership so far has attempted to parate the effect of founding-family ownership from blockholder effects induced by other types of large shareholders.By doing so,this study combines the empirical literature on ownership concentration
and studies on family firms and thereby prents a comprehensive analysis of the effects of founding-family ownership compared to other blockholders.
3.Data
3.1.Data sources and sample lection
The sample for this investigation is bad on all companies listed on the of ficial market (Amtlicher Handel )on the Frankfurt Stock Exchange on December,311998.Banks and insurance companies are dropped due to problems calculating Tobin's q and a lack of comparability concerning other performance variables bad on EBIT or EBITDA.Furthermore,four companies have to be excluded becau they were already insolvent at the beginning of the sample period and liquidated only shortly afterwards.For the remaining companies,I collect data until the end of 2004.This procedure results in a final sample of 275firms and 1701firm-year obrvations.
In order to classify the companies as family or non-family firms,data on the composition of executive boards (Vorstand ),and supervisory boards (Aufsichtsrat ),as well as detailed information on the shareholder structure,are obtained from Hoppenstedt yearbooks.The books provide in-depth information about all market-listed German companies.Names of the board members are gathered for every other year,shareholdings on a yearly basis.
For some companies,the af filiation of board members to a family is not obvious at first sight.In partic
ular for families with a long prence in the company,last names can be different from the founder's name due to marriages.In the cas,the family af filiation has to be con firmed by at least two publicly available sources (wspapers)or one of ficial company publication (annual statement,ad hoc announcement,anniversary publication …).
Accounting and share price data as well as industry-classi fications are taken from Datastream databas.
3.2.Family firms
In order for a firm to qualify as a family business it has to meet at least one of the following two criteria:a)the founder and/or family members hold more than 25%of the voting shares,or b)if the founding-family owns less than 25%1of the voting rights they have to be reprented on either the executive or the supervisory board.2
鲍尔森The term ‘founder ’requires some remarks concerning its exact meaning:first,a person is considered the founder if he or she founded the sample company or the predecessor company (in ca of a change in the legal form and/or the company's name).
Second,when a person acquires a majority stake in a company and runs the company as CEO,he/she is treated as a founder if he/she changes the company's operational business signi ficantly.For instance,Stolberger Zink AG,formerly a mining company which gave up its business in the 1970s,was bought by Günter Minninger who then took over veral telecom companies and t up a telecom business.In 1999,the name was changed to Stolberger Telecom AG.In the opposite ca —a family business is taken over —it is no longer treated as a family firm,although the founding-family might still have a stake in the new company.
Third,if a firm was founded by more than one person,they are together treated as one family.Among the different possibilities,this approach makes the most n since the founders usually act coordinated and almost always even pool their votes.For example,the three founders of SAP,Hasso Plattner,Dietmar Hopp and Klaus Tschira only dissolved the contract pooling their votes becau of “international capital market conventions ”.3
1Holdings of more than 5%have to be registered with the German Financial Supervisory Authority (BaFin).Shareholdings of less than 5%–even when reported in Hoppenstedt –are excluded for reasons of data consistency.Thus,a family (or any other shareholder)has to hold at least 5%of the shares.
2As a robustness test,I also u the alternative “lone founder ”de finition propod by Miller et al.(2007).
3Of ficial SAP press relea,September 3rd,
2002.Fig.1.Distribution of CEO-types in family firms from 1998–2004.
435
C.Andres /Journal of Corporate Finance 14(2008)431–445
