NBER WORKING PAPER SERIES FAMILY FIRMS. Mike Burkart Fausto Panunzi Andrei Shleifer. Working Paper 8776

Transcription

1 NBER WORKING PAPER ERIE FAILY FIR ike Burkart Fausto Panunzi Andrei hleifer Working Paper NATIONAL BUREAU OF ECONOIC REEARCH 1050 assahusetts Aenue Cambridge, A 0138 February 00 We are grateful to Julian Franks, Denis Gromb, imon Johnson, Rafael La Porta, Enrio Perotti, Jeremy tein, Daniel Wolfenzon, and seminar partiipants at Bologna, London Business hool, tokholm hool of Eonomis, and Zurih for helpful omments and disussions. Finanial support from Uniersità Booni (Riera di Base), from the Bank of weden Terentenary Foundation, and from the Gildor Foundation is gratefully aknowledged. Robin Greenwood proided exellent researh assistane. This paper is produed as part of a CEPR projet on Understanding Finanial Arhiteture: Legal Framework, Politial Enironment and Eonomi Effiieny, funded by the European Commission under the Human Potential - Researh Training Network program (Contrat No. HPRN-CT ). The iews expressed herein are those of the authors and not neessarily those of the National Bureau of Eonomi Researh. 00 by ike Burkart, Fausto Panunzi and Andrei hleifer. All rights resered. hort setions of text, not to exeed two paragraphs, may be quoted without expliit permission proided that full redit, inluding notie, is gien to the soure.

2 Family Firms ike Burkart, Fausto Panunzi and Andrei hleifer NBER Working Paper No February 00 JEL No. G3, K, 13 ABTRACT We present a model of suession in a firm ontrolled and managed by its founder. The founder deides between hiring a professional manager or leaing management to his heir, as well as on how muh, if any, of the shares to float on the stok exhange. We assume that a professional is a better manager than the heir, and desribe how the founder's deision is shaped by the legal enironment. peifially, we show that, in legal regimes that suessfully limit the expropriation of minority shareholders, the widely held professionally managed orporation emerges as the equilibrium outome. In legal regimes with intermediate protetion, management is delegated to a professional, but the family stays on as large shareholders to monitor the manager. In legal regimes with the weakest protetion, the founder designates his heir to manage and ownership remains inside the family. This theory of separation of ownership from management inludes the Anglo-axon and the Continental European patterns of orporate goernane as speial ases, and generates additional empirial preditions onsistent with ross-ountry eidene. ike Burkart Fausto Panunzi Andrei hleifer tokholm hool of Eonomis Uniersità di Bologna Harard Uniersity P.O. Box 6501 Piazza arailli 9 Littauer Center tokholm, E , Bologna Cambridge, A 0138 and CEPR and CEPR and NBER mike.burkart@hhs.se fpanunzi@eonomia.unibo.it ashleifer@harard.edu

3 I. Introdution ost firms in the world are ontrolled by their founders, or by the founders families and heirs. uh family ownership is nearly uniersal among priately held firms, but also dominant among publily traded firms. In Western Europe, outh and East Asia, iddle East, Latin Ameria, and Afria, the ast majority of publily traded firms are family ontrolled (La Porta et al. 1999, Claessens et al. 000, European Corporate Goernane Network 001, Faio and Lang 00). But een in the United tates and the U.K., some of the largest publily traded firms, suh as Wal-art tores and Ford otor, are ontrolled by families. A ruial issue in the disussion of family firms from the perspetie of orporate goernane and finane is suession. For nearly eery entrepreneurial firm that does not fail, there omes a moment when the founder no longer wishes to manage it. This an happen from the ery beginning, when founders seek professional managers to run their firms, as is the ase in high tehnology startups in the United tates. Alternatiely, suession takes plae later in a founder s life, when he retires or uts his workload. The new manager an be either a professional or the founder s heir. If a professional is hired, the founder also needs to deide whether to stay on and monitor him, or to sell out. Unless the founder appoints his heir or the professional manager buys the ompany outright, ownership and management beome separated. The patterns of separation of ownership and management ary aross ountries. In the United tates, founders often hire professional managers early on. By the time the founder retires, he and his family retain only marginal ownership. In suh Berle and eans (193) orporations, professional managers exerise nearly full ontrol. In Western Europe, signifiant ownership typially stays with the family after the founder retires. His hildren either hire a 3

4 professional manager, as in BW or FIAT, or run the firm themseles, as in Peugeot. In emerging markets, both management and ownership tend to stay with the family when the founder retires. When a professional manager is badly needed, he oasionally marries into the family. In this paper, we attempt to understand theoretially these different patterns of separation of ownership and management. The prinipal benefit of hiring a professional is that he is likely to be a better manager. The prinipal ost is that now the professional manager, rather than the founder, ontrols the ompany and so an expropriate inestors or, in more polite language, onsume the priate benefits of ontrol (Jensen and ekling 1976, Grossman and Hart 1988). We argue that a ruial fator shaping the attratieness of delegated management is the degree of legal protetion of outside shareholders from expropriation (or tunneling) by the insiders. Earlier researh shows that legal protetion of outside shareholders aries sharply aross ountries, and that this ariation predits the differenes in finanial deelopment and ownership strutures (La Porta et al. 1997, 1998, 000, Johnson et al. 000). In this paper, we examine the osts and benefits of delegating management from this perspetie. This allows us, in partiular, to examine the osts and benefits of keeping the suession of management inside the family. We present a model of a founder looking for a manager to sueed him. We assume that there is no superior manager aailable with suffiient resoures to buy the firm outright. When suh a manager (or a ompany) is aailable, the firm is simply sold to them as often happens. Absent an outright buyer, the founder hooses among three options. He an sell out ompletely in the stok market and reate a widely held firm run by a professional manager. He an hire a professional manager but stay on as a large shareholder to monitor him. He an also keep the firm inside the family by either staying on as a less than ideal manager or passing management to 4

5 a family member, who is generally not as talented as a professional manager. The founder makes this deision to maximize the alue of the firm, whih initially he owns ompletely. We study the trade-off between superior management by the professional outsider and his disretion to expropriate shareholders. If the founder stays on as a large shareholder and monitors, he an ontrol expropriation to some extent. In our framework, both the law and the monitoring redue managerial expropriation. We show that, when legal protetion of outside inestors is ery good, there is no need for monitoring in equilibrium, and the best arrangement is a widely held professionally-managed firm. When legal protetion of outside inestors is moderate, the benefits of professional managers are still high enough for the entrepreneur to surrender ontrol, but it pays for him or his family to remain as large shareholders and monitor the manager. Finally, with suffiiently weak shareholder protetion, the founder s ability to ontrol expropriation is too limited and he does not delegate management. In this situation, management stays with his family een when someone else an run the firm better. In general, this analysis leads to a predition of a negatie relationship between inestor protetion and ownership onentration, onsistent with a range of empirial eidene. We onsider two ersions of the model, one in whih the large shareholder when he detets expropriation fores the manager to stop it and pay diidends, and another in whih he and the manager just share the spoils. In the first ersion, as in hleifer and Vishny (1986), the large shareholder proides a publi good to the minority shareholders by monitoring or jawboning. In the seond ersion, monitoring is no longer a publi good, the benefits of whih are shared by all shareholders. To the ontrary, there is a pererse inentie for monitoring, namely to share the loot. The basi results we desribe hold in both speifiations, but the seond ersion also yields the empirially aurate predition of a positie premium paid for a 5

6 ontrolling blok of shares. At a theoretial leel, the model ombines in one unified framework the twin onflits essential to understanding orporate goernane: that between the manager and the outside shareholders, and that between the large shareholder and the minority shareholders. By doing so, the model sheds light on the different patterns of ownership and management among ountries. It shows, for example, why Anglo-axon patterns of orporate goernane, with widely-held firms and traditional onflits between professional managers and dispersed shareholders (Berle and eans 193), are likely to be a feature of ountries with ery good legal protetion of minority shareholders. It explains why family firms, in whih the founder s family is a signifiant shareholder, or een the manager, oer seeral generations are suh an enduring phenomenon in ountries with less effetie legal protetion of shareholders (La Porta et al. 1999, Claessens et al. 000). Indeed, it explains how, in suh ountries, the twin onflits between the manager and the large shareholder and between the two of them ombined and the minority shareholders, oexist. The model is moreoer onsistent with the growing body of eidene that family management is generally inferior to professional management (ork et al. 000, Perez-Gonzales 001). The basi trade-off between the benefits of delegated management, and the osts of giing up ontrol espeially when legal protetion is poor appears onsistent with a great deal of data. Our paper joins a growing theoretial literature on orporate goernane in the regimes of poor inestor protetion. Bebhuk (1999) shows that poor legal protetion renders dispersed ownership struture unstable, beause it allows extration of signifiant priate benefits. La Porta et al. (00) and Himmelberg et al. (001) study theoretially and empirially the determination of ownership struture when firms raise funds to finane inestment. Burkart and 6

7 Panunzi (001) and hleifer and Wolfenzon (00) analyze the impat of legal shareholder protetion on the optimal ownership struture. hleifer and Wolfenzon (00) onsider ownermanagers and examine the relationship between legal protetion and inside equity. Burkart and Panunzi (001) assume that the professional manager and the large shareholder are distint parties and analyze the relationship between the law and outside ownership onentration. By making the separation of ownership and management a hoie ariable, the present model extends and generalizes these two papers. The paper is organized as follows. etion outlines the model. etion 3 examines the founder s deision to hire a professional manager and to float shares when he annot extrat priate benefits. etion 4 analyses the ase with ollusion between founder and professional manager and deries impliations for share alue, blok premium, and ageny osts of separating ownership and management. etion 5 onludes. athematial proofs are in the appendix. II. The model Figure 1 presents the model s timeline. We onsider a firm initially fully owned by its founder. At date 0, the founder deides whether to appoint a professional manager to run the firm or keep management in the family. imultaneously, he deides what fration 1-α of the shares to sell to dispersed shareholders. The family keeps the remaining fration α. All shareholders are risk-neutral. If management stays in the family, there is no separation of ownership and management. If the founder appoints a professional manager, ownership and management are separated. In this ase, the founder may also offer a wage and the professional manager aepts or rejets the offer to run the ompany at date 1. 7

8 Figure Founder hooses the anagers job Founder hooses Deision on manager, sells aeptane monitoring intensity m diidends and 1-α shares, offers deision priate benefits. professional manager Payoffs are wage w m realized At date 3, the firm generates reenues that depend on the identity of the manager. If ontrol remains inside the family, total reenues generated are F. If a professional manager runs the firm, total reenues are. The professional has an outside option that gies him utility. For simpliity, the outside option of the founder or the family is normalized to zero. Assumption 1: > F There are two interpretations of the model. Under the first, the hoie is between the founder himself, who is beoming outdated or relutant to manage, and a professional outsider. Under the seond, the founder definitely retires from management, and hooses as his suessor either a professional manager or his heir. In both interpretations, retaining management inside the family redues the profitability of the firm relatie to hiring a professional. 1 Assumption 1 is onsistent with reent study by ork, tangeland, and Yeung (000) of orporate ontrol of Canadian firms. They find that heir-ontrolled firms hae lower returns on sales and assets than omparable firms. Furthermore, firms with founder ontrol hae earnings 1 In this model, we abstrat from the possibility that appointing a hild as the CEO may be an additional priate benefit to the founder. Introduing this feature would not hange our results. 8

9 that are lower than those of widely held firms but higher than those in heir-ontrolled firms. Perez-Gonzales (001) proides eidene on firm performane following inherited ontrol by studying 16 family transitions in the United tates. In 38% of these ases, family members inherit the CEO position. These family CEOs are promoted to the post an aerage of 9 years earlier than professional managers and are detrimental to firm performane- the return on assets falls by 16% within two years of transition and 5% ompared with unrelated CEOs. In this model, if the founder is the best manager himself, there is no reason for him eer to sell equity. He stays on as the manager, keeps 100% of the firm, and there are no ageny problems or onflits. This assumption distinguishes the model from the papers of La Porta et al. (00), hleifer and Wolfenzon (00), and Himmelberg et al. (001), where equity is raised to finane inestment projets, and therefore the size of the firm is endogenous. The problems arise when the founder is no longer the best manager. He must then hoose between hiring a more qualified outsider to manage, or alternatiely staying on or (equialently) naming a mediore son as a suessor. Importantly, we assume that the ompetent professional outsider has neither the resoures nor the external funds to just buy the firm himself. As we show below, the outsiders inability to raise external funds is onsistent with the assumptions of the model, sine to buy the firm, he has to pay for the priate benefits aruing to the founder, whih he annot pledge to inestors. Unless the superior manager is himself rih, he has to work for the family. Hiring a professional gies rise to a separation of ownership from management. At date, shareholders an monitor the professional manager and thereby deprie him of at least some priate benefits. The monitoring tehnology is disussed below. At date 3, the reenues an either be paid out to all shareholders proportionally to their 9

10 ownership stakes or dierted to generate priate benefits. In ountries with weakest shareholder protetion, suh priate benefits take the form of outright theft. ore ommonly, they take the form of transations with related parties, expropriation of orporate opportunities, transfer priing, exessie salaries and perquisites, and so on (see Johnson et al. 000). Whoeer manages the firm hooses the leel of expropriation, subjet to being monitored and partially impeded by the law. The non-ontratible expropriation deision is modeled as the hoie of φ [0,1] suh that seurity benefits (diidends) are ( 1 φ)i and priate benefits are φ, i =, F. Expropriation of shareholders is limited by the law. To model legal shareholder i protetion, we assume that the law sets an upper bound φ [0,1] on the fration of reenues that an be (at no ost) dierted by the party in ontrol. tronger legal protetion orresponds to lower alues of φ. The law is not the only determinant of the fration of resoures dierted for priate benefits. The other is monitoring, whih ours when a professional manager is hired. Although in priniple all shareholders an monitor the manager, the free-rider problem preents small shareholders from hoosing to inur the ost. In equilibrium only the large shareholder monitors to redue the fration of resoures appropriated by the manager. 3 Reall that the legal upper bound on priate benefits of ontrol is φ. Following Pagano and Röell (1998), we assume that the large shareholder an at a ost m k redue priate benefit extration by m where φ The amount i is the legal upper bound that an be extrated as priate benefits of ontrol, irrespetie of the form in whih those benefits are enjoyed. In partiular, wages in exess of market alue are already inorporated in i φ. 3 We assume that there is no reason for the founder to sell his shares to another large shareholder who would monitor, sine, if anything, the founder would hae a omparatie adantage at monitoring beause of his knowledge. 10

11 m [0,1] and k > 0. Priate monitoring and the law are alternatie mehanisms for reduing expropriation of shareholders. When k is stritly positie, monitoring is ostly for the founder whereas reliane on the law is free. The publi benefits of the law may not be free from the iewpoint of soiety, but they are to the founder. The interation between these two mehanisms of poliing the manager shapes the equilibrium goernane struture. Importantly, we assume that orporate and other law goerning inestor protetion matters, and that firms annot opt into more protetie legal regimes by ontrat. This assumption is onsistent with the eidene that legal rules goerning inestor protetion in different ountries hae signifiant onsequenes for finanial deelopment (La Porta et al. 1997, 1998, 000). In this model, if a founder ould effetiely opt into a more protetie legal regime, he would do so sine in equilibrium he pays for all the priate benefits of ontrol out of his own poket. If a professional manager is hired, the question arises whether a monitoring founder an enjoy (part of) the priate benefits. We onsider both the ases of ollusion and no ollusion between the professional manager and the founder. Exluding the founder from the spoils of extration is tantamount to assuming that his interests and those of the small shareholders are perfetly ongruent. This ase is most appropriate when the legal duties of the large shareholder, perhaps as a board member, bar him from ompliity with the manager in expropriating shareholders. In ontrast, when the founder and the professional manager an share the priate benefits, they may ollude at the expense of minority shareholders. This assumption might be more suitable for weaker legal regimes. The seond assumption ompliates the model, in that rent-seeking monitoring, intended to apture some of the priate benefits rather than sere all 11

12 shareholders, beomes attratie. We sole the model under the first assumption in the next setion, and under the seond in setion IV. III. Owner-anager or Professional anager We analyze the founder s deision whether to hire a professional manager in steps. We begin by onsidering the founder s maximization problem for the ases of non-separation and separation of ownership and management. In eah ase, we sole the model by bakward indution, going from the date 3 expropriation deision, to the founder s date monitoring intensity, to the manager s date 1 job aeptane hoie. We then an determine the optimal number of shares that the founder retains in ases of separation and non-separation. Haing done that, we an ompare the entrepreneur s welfare for different legal enironments, i.e., different alues of φ, whih enables us to infer under what irumstanes he hooses to separate ownership from management. A) No eparation of Ownership and anagement Due to the simpliity of the model, the ase of no separation does not yield preise preditions, notably for the ownership struture. At date 3 the founder deides how to alloate the reenues subjet to the limits set by the law. By law, he annot diert more than φ of the reenues as priate benefits. Unless he owns all the shares, in whih ase he is indifferent between any φ [0, φ], he extrats the legal upper bound φ. Absent a professional manager, there is neither date monitoring nor a date 1 job aeptane deision. Hene, we moe diretly 1

13 to the founder s date 0 deision as to whih fration of shares to sell to outside inestors. He maximizes the sum of his date 3 blok alue α(1 φ) F + φ F and of the proeeds from selling 1 α shares at date 0. We refer to this payoff as the founder s welfare N V. ine priate benefit extration is effiient and sine the founder is by assumption neither finanially onstrained nor risk aerse, the optimal ownership struture is indeterminate when ownership and management are separated. Lemma 1: For any φ [0,1], α = and [ 0,1] N V ( ) F α. The founder s welfare is equal to the total reenues under his management. Een though priate benefit extration dereases with the quality of the law, the founder s welfare is independent of the legal enironment. ine the extration of priate benefits is effiient, eah dierted dollar redues the seurity benefits by a dollar. The sum of seurity and priate benefits and hene the founder s welfare is onstant. B) eparation of ownership and management What happens when ownership and management are separated? On the one hand, the professional manager prefers to diert orporate reenues as priate benefits to himself rather than pay them out as diidends to the shareholders. While the law onstrains diersion, the founder an further limit priate benefit extration through monitoring. On the other hand, this possibility may indue opportunisti behaior by the founder een when he does not share in the priate benefits. One the professional manager has aepted to run the firm and reenues are 13

14 realized, the founder has an inentie to redue the professional manager s priate benefits by monitoring more. Antiipating high leels of monitoring, the professional manager may rejet the offer to run the firm. That is, the founder may oer-monitor in the sense of the ex post optimal monitoring leel exeeding the ex ante optimal amount (Pagano and Roell 1998). To indue the manager to aept running the firm, the founder has to ommit himself not to monitor exessiely. He an do so by dispersing (some of) the shares to small inestors beause the atual monitoring intensity is determined by the size of the founder s equity stake (Burkart, Gromb, and Panunzi 1997). In addition (or instead), the founder may offer the professional manager monetary inenties to onine him to run the firm. We sole the game by bakward indution, beginning with the date 3 resoure alloation deision. Total reenues under the professional manager are. The law stipulates that ( 1 φ) must be paid out either to shareholders as diidends or to the professional manager as salary. What fration of the remaining φ is atually dierted depends on monitoring. The founder monitoring with intensity m an ontrol the use of an additional fration m (or at most φ ) of. Being exluded from sharing priate benefits, the founder fores the professional manager to disgorge all of them as diidends. The professional manager then has disretion oer max { 0,( } φ m) in resoures. He stritly prefers to extrat them as priate benefits, unless he is the sole shareholder. ine priate benefit extration is effiient, there are no gains to shareholders from using monetary inenties to resole the onflit oer resoure alloation. To indue the manager to abstain from extrating an additional dollar, shareholders hae to offer him this dollar as a transfer. onetary inenties, heneforth alled the wage, an, howeer, play a role in induing 14

15 the manager to aept the job of running the firm. Let w denote the wage paid to the professional manager when he aepts the job offer from the founder. 4 At date, the founder hooses the monitoring intensity. For a gien blok α and for a gien wage rate w, the founder maximizes α (1 w φ + m) k m.5 He reeies a fration α of the seurity benefits net of the wage bill less his monitoring osts. ine the law already shields ( 1 φ) of the reenues from priate benefit extration, the founder neer monitors more than φ. Hene, m = min φ,α k and weakly inreases with the blok size. At date 1, the manager aepts to run the firm if the sum of the wage and the priate benefits exeeds his outside utility. 6 The ondition ( w + φ m ) an be rewritten as m m = w + φ. High leels of monitoring and strit legal rules redue the professional manager s priate benefits and may thus disourage him from running the firm. Offering him a higher wage an sway him to aept the job. Higher ownership onentration and better legal protetion make it more diffiult to satisfy the professional manager s partiipation onstraint, whereas higher wages make it easier. This is the basi trade-off when ownership and management are separated. At date 0, the founder hooses the ownership struture and the wage to maximize his m = φ m k subjet to the manager s partiipation onstraint. s welfare V [ 1 w + ] 4 The subsequent analysis impliitly assumes that φ + w < 1, whih holds in equilibrium. 5 The range m [0,1] implies that k. 6 An alternatie interpretation of the model is one where the manager has to exert an effort to generate reenues and is the disutility of the effort. 15

16 If the founder were to hoose an ownership struture suh that φ < α, the professional k manager would be left with zero priate benefits. Consequently, the founder would hae to offer a wage w= to indue the professional manager to aept the job. Leaing some priate benefits to the professional manager in exhange for a lower wage saes on monitoring osts. Hene, the founder always hooses an ownership struture suh that φ > α and m = α. k k V Inserting the monitoring leel α ( α ) = (1 w φ + ) with k k m k = α into the founder s welfare yields dv / dα = (1 α) / k 0 and dv / dw = < 0. The founder s welfare inreases with ownership onentration and dereases with the wage, proided that the professional manager s partiipation onstraint is satisfied. A binding partiipation onstraint is obiously in the interest of the founder as any managerial rent omes at his expense. ometimes, howeer, the founder annot aoid leaing some rents to the professional manager. ore preisely, there are parameter alues for whih the partiipation onstraint ( w + φ m ) does not bind despite a fully onentrated ownership struture and a zero wage. This ours when φ > / k+ /. We want to allow for the possibility of legal regimes in whih the professional manager an extrat a rent. Assumption : + < 1 k ine φ 1, Assumption is a neessary ondition for φ > / k+ / to hold. 16

17 Lemma V i) For φ, α = 0, w / φ, m = 0, and ( α, w, φ) =. = ii) For, < φ k + k α = φ, w = 0, m = φ, V ( k,, ) α w φ = φ. iii) For φ > +, α = 1, w = 0, k m = k, and V ( α, w, φ) = (1 φ) +. k When legal protetion is strong (ase i), then, een in the absene of monitoring, priate benefits are insuffiient to indue the professional manager to run the firm. Consequently, ownership is ompletely dispersed and the professional manager is offered a wage equal to the differene between his outside utility and the priate benefits. The founder s resulting welfare V α φ is at its first best leel ( ) and does not depend on the quality of legal rules. (, w, ) When legal protetion is moderate (ase ii), expeted priate benefits exeed the outside utility. As a result, the founder has to monitor the professional manager to limit the size of his rent. etting the wage equal to zero minimizes the monitoring intensity that keeps the professional manager s partiipation onstraint binding. ine monitoring is ostly, this dominates all other ombinations of positie wage and monitoring leel that also leae no rent to the professional manager. A positie wage and onentrated ownership do not o-exist in 17

18 equilibrium. Due to the monitoring osts, the founder s welfare best leel. oreoer, V ( α, w, ) V ( α, w, ) φ is below its first φ dereases in both φ and k : less legal protetion entails a higher optimal leel of monitoring, and a higher k makes monitoring more expensie. When legal protetion is poor (ase iii) the founder annot aoid leaing a rent to the professional manager. Offering a zero wage and retaining all shares to implement a monitoring leel m= / k is all that the founder an do. The resulting rent to the professional manager is equal to R = (φ ). The founder s welfare k V ( α, w, ) φ is equal to the first best ( ) less monitoring osts and managerial rent. As in the range with moderate legal protetion, V (, w, ) α φ dereases in both φ and k. We now turn to the final step of determining the onditions under whih the founder hooses to hire a professional manager. The answer follows from omparing the founder s welfare under no separation N V (Lemma 1) to that under separation two Propositions desribe the oerall equilibrium outomes. V (Lemma ). The next Proposition 1 When legal shareholder protetion is strong ( φ / separated, and ownership is fully dispersed. ), ownership and management are When legal rules are ery protetie, the separation of ownership and management only brings benefits by allowing the founder to apitalize on the superior ability of the professional manager. The strong legal protetion also soles at no ost to the founder the ageny onflit oer the alloation of reenues. ore preisely, the law restrits priate benefit extration below 18

19 the professional manager s outside utility. Letting this manager diert orporate resoures is part of his ompensation pakage, whih needs to be supplemented by a wage. In this ase, selling all the equity and hiring a professional manager is the optimal hoie for the founder. In this model, a legal system with strong protetion of outside shareholders, i.e., with φ, ahiees the first best leel of soial effiieny. The best manager is hired to run the firm, and no resoures are wasted on monitoring. This onlusion is drien by the fat that, in this model, law enforement is free at least from the iewpoint of the founder. If better legal protetion imposes higher enforement on the soiety, we would hae to ompare the soial osts of priate monitoring with the soial osts of law enforement. 7 One inestor protetion falls below the threshold of Proposition 1, so φ > /, the professional manager an expet to appropriate orporate resoures in exess of his outside utility. As a onsequene, monitoring is needed to limit the size of his rent. In this ase, the separation of ownership and management inoles a trade-off: on the one hand, the firm is run by a more qualified manager; on the other hand, the founder has to inur monitoring osts (and possibly leae a rent to the professional manager). Denote by φ ( /,1) the unique alue of N φ suh that V ( α, w, φ ) = V. This alue exists if /k <. F Proposition i) If /k < holds, ownership and management are separated and the founder F retains a blok when legal protetion is moderate φ ( /, φ ]. When legal protetion is poor 7 The priate and soial alulations would be further ompliated when there are priate osts of ompliane with legal rules. For example, the osts of omplying with better aounting and dislosure standards might be higher. 19

20 φ > φ, there is no separation of ownership and management. ii) If /k F, ownership and management are always separated, and the founder retains a blok when legal protetion is moderate or poor. When ownership and management are separated, the founder s welfare V dereases with φ, beause weaker legal protetion entails higher monitoring osts and (possibly) an inreasing managerial rent. Aordingly, V reahes its minimum when the law does not proide any protetion (φ =1). In this ase, all diidend payments are exlusiely due to monitoring, fully onentrated ownership is optimal, and V = /k. In ontrast, when ownership and management are not separated, the founder s welfare is F and independent of the quality of legal protetion. Hene, there exists a unique threshold alue φ below whih monitoring osts and managerial rent are less than the gain in managerial effiieny ( ). Conersely, F for φ > φ the forgone effiieny loss assoiated with keeping ontrol in the family is smaller than the ageny osts of separating ownership and management. If, howeer, the disrepany between the managerial abilities of the professional and that of the founder (or his heir) is ery large, keeping management in the family is inferior irrespetie of the quality of the law. This holds when the reenues under family ontrol are smaller than the diidends resued from managerial expropriation by monitoring only, i.e., when /k F. In this ase the founder or family simply retains an ownership stake whose size depends on legal protetion. The model has a lear impliation for how the law shapes ownership struture: 0

21 Proposition 3 For α (0,1), more onentrated ownership strutures go together with weaker legal protetion, i.e., dα / dφ 0. The founder s objetie under separation of ownership and management is to pay the professional manager no more than his outside utility. Both legal protetion and monitoring restrit the professional manager s ability to extrat priate benefits. ine the law limits extration at no ost, the founder resorts to monitoring only to the extent that the law leaes the manager a payoff in exess of his outside utility. To restore a binding partiipation onstraint on the professional manager, the founder has to monitor more as legal protetion deteriorates. Thus, for φ ( /, / + / k] legal protetion and ownership onentration are inersely related under separation of ownership and management. 8 Propositions and 3 make strong empirial preditions, namely that family ownership should be ommon around the world, and relatiely more ommon in ountries with poor inestor protetion. Reent empirial work is onsistent with these preditions. Family ontrol is the dominant form of orporate ownership around the world. Looking at the 0 largest firms in 7 wealthy eonomies, La Porta et al. (1999) find that families or indiiduals ontrol 30% in number and 5% in alue of the top 0 firms in eah ountry. These numbers are muh higher for smaller firms. Family transitions are a frequent and important ourrene: only about one third of family ontrolled firms are run by their founders, the rest by desendants or by families 8 The (inerse) relationship between ownership onentration and legal protetion depends ruially on the absene of wealth onstraints of the founder and the assumed monitoring tehnology. In partiular, the positie relationship between optimal monitoring intensity and ownership onentration relies on the assumption that the inentie to monitor depends on the ownership stake but is independent of legal protetion. By ontrast, when legal protetion has a diret impat on the (marginal) return from monitoring the relationship between the quality of the law and ownership onentration is not monotone (Burkart and Panunzi 001). 1

22 that ame to own them later. In addition, LLV (1999) show that widely held firms are more ommon in ountries with good shareholder protetion- 34% ersus 16% in the ountries with a low leel of protetion. oreoer, ownership patterns tend to be relatiely stable. In short, when expropriation is a onern, firms remain family ontrolled. Claessens, Djanko and Lang (000) find that with the exeption of Japan, more than 50% of all publily traded firms in nine East Asian ountries are ontrolled by families and that the top 15 families ontrol signifiant shares of ountry wealth. 9 East Asian ountries outside Japan are indeed known for partiularly poor protetion of outside inestors. Faio and Lang (00) find higher inidene of family firms in ountries with inferior shareholder protetion in a large sample of West European orporations. The European Corporate Goernane Network (001) douments the prealene of onentrated orporate ownership in OECD ountries. Propositions 1 and analyze the legal irumstanes under whih professional managers of a gien quality are hired. Alternatiely, one an onsider the managerial effiieny gain neessary to hae separation of ownership and management in a regime with weak legal protetion. Corollary 1 The separation of ownership and management requires higher managerial skills in regimes with poorer legal shareholder protetion. The founder s welfare under separation V inreases with both the professional 9 In Taiwan, the top 15 families ontrol 0.1% of listed orporate assets; in Hong Kong, Indonesia, Korea, alaysia, the Philippines, ingapore and Thailand, the top 15 families ontrol more than 5% of listed orporate assets.

23 manager s ability (higher alues) and the quality of the law (lower φ alues). In ontrast, the founder s welfare under no separation N V is independent of φ. Ceteris paribus, the swith from keeping family ontrol to hiring a professional manager requires a more able professional manager when legal protetion is less effetie: uh legal regimes are assoiated with higher ageny osts of separation of management and ontrol. To make hiring a professional manager nonetheless worthwhile, the effiieny gain must be larger. The setion has analyzed how legal rules affet the trade-off between the benefits and osts of separating ownership and management. The model predits three different patterns of ownership and management: separation with dispersed ownership, separation with blok ownership, and no separation. ome impliations of our basi model do, howeer, lash with the empirial eidene. In partiular, the model implies that the founder s blok trades at a disount when ownership and management are separated. This predition is inonsistent with the empirial eidene on blok premia, whih are both positie and higher in ountries with weaker protetion of outside shareholders (Zingales 1994, Nenoa 001, Dyk and Zingales 00). The reason for this result in our model is that all shareholders benefit in proportion of their shareholding from monitoring, but only the founder bears the ost. The negatie blok premium follows from the assumption that the founder annot extrat any priate benefits, thereby ensuring that monitoring is a publi good. We relax this assumption in the next setion and allow the founder to benefit priately from his monitoring atiity. A final theoretial point needs to be made before we moe on. In the model, we hae assumed that ownership struture remains stable one the manager has aepted to run the ompany. If the founder had the opportunity and inentie to retrade ex-post, he may want to inrease his stake and extrat a higher fration of priate benefits from the manager. This, in 3

24 turn, would affet the deision of the manager to aept the offer to run the ompany. For the model in this setion, we an proe that, so long as trade in not anonymous, the purhase of an additional fration of shares is not profitable for the founder. Beause shareholders free ride on the entire alue improement implied by the founder s final holding, 10 the founder does not make a profit on the additional shares aquired in the retrading stage. oreoer, the inrease in the monitoring osts due to the larger final holding exeeds the inrease in the alue of the shares owned initially by the founder. Things are more ompliated when trade is anonymous (Dearzo and Urosei 001), or when the founder shares some of the priate benefits of ontrol with the manager, as in the next setion. In these instanes, we need to make the assumption that there is no retrading, or alternatiely that the manager s wage an be onditioned on the founder s final equity stake. IV. Transferable priate benefits of ontrol In the preious setion, the founder ould not by assumption extrat any priate benefits unless he managed the firm himself. The founder s interest and those of the minority shareholders are perfetly ongruent when a professional manager runs the firm. In this setion we onsider the other extreme, where the priate benefits are perfetly transferable, thereby aligning the founder s interest in expropriation with that of the professional manager. The possibility of sharing the spoils with the professional manager proides an additional rationale for monitoring. Besides proteting the founder against the threat of managerial expropriation, monitoring also seures him a (larger) share of the priate benefits, thereby making it a rent- 10 A proof of a similar result is ontained in Burkart, Gromb, Panunzi (1997) and Pagano and Roell (1998). 4

25 seeking atiity. Below we repeat the analytial steps of the preious setion and establish that the existene of the three patterns of separation of ownership and management extends to the ase of transferable priate benefits. We then explore the impliations of the model for the relationship between legal protetion, share alue, blok premium, and the ageny osts of separating ownership and management. Transferable priate benefits only matter in the presene of a professional manager, leaing the founder s welfare in the ase of no separation unhanged (Lemma 1). We therefore moe diretly to the ase of separation of ownership and management. At date 3, the law imposes that ( 1 φ) be paid out either as diidends to all shareholders or as the wage to the professional manager. The founder and the professional manager bargain oer how to alloate the remaining φ. The outside options in the bargaining depend on monitoring. If the founder monitors with intensity m, his outside option is min { α m,αφ } and that of the professional manager is max{ 0,( φ m) } shared between the two parties amounts to min{ (1 α ) m,(1 α) φ }. The surplus to be, the reenues that would arue to the minority shareholders if no agreement is reahed. Under the assumption of equal bargaining power, the founder reeies in the bargaining a payment equal to 1+ α 1+ α min φ, m and the manager gets 1 α max φ 1+ α, φ m. ine these payoffs exeed their outside options, the founder and the manager always agree to set φ = φ The question arises again whether there are any gains from using monetary inenties to indue the manager to abstain from extrating priate benefits. At first glane this seems possible, sine the manager reeies only half of the surplus. While the use of suh a ompensation sheme is in the minority shareholders interest, the founder prefers ex post to extrat priate benefits. ine suh extration is effiient, the founder is ex ante indifferent. All that matters is to keep the manager s partiipation onstraint binding, i.e., to leae him no rents. 5

26 At date, the founder hooses the monitoring intensity to maximize 1+ α m [ α (1 w φ) + m] k. ine the law already protets (1 φ) of the reenues from extration, there is no gain in monitoring more than φ. The founder therefore hooses m 1+ α min φ,. If φ > (1 + α) / k holds, monitoring does not depend on the blok size k = and is purely drien by the prospet of extrating a part of the priate benefits. Otherwise ( φ (1 + α) / k ), monitoring is a funtion of both priate benefits and the ownership stake. Indeed, deomposing the first order ondition into α( / k) + [(1 α)/]( / k) reeals the two moties for monitoring. The first term reflets monitoring in the absene of ollusion aimed at preenting managerial expropriation of the founder s stake α. The seond term aptures the additional monitoring that the founder undertakes to appropriate some of the priate benefits. While monitoring inreases as before with the blok size α, the seond term implies that monitoring is positie een when the founder retains no shares, i.e., m ( 0) > 0. This implies that the founder annot withdraw from the firm. To remoe this purely mehanial feature of the model, we impose the following assumption: Assumption 3: If the founder retains less than α > 0, he abstains from monitoring. Assumption 3 simply means that, to monitor effetiely, the founder must hae some power oer the manager, and for that he needs a minimum ownership stake. This stake may enable him to sit on the board, to hae enough shares to onene an extraordinary shareholder meeting, to hae standing in litigation, or to exerise power in other ways. When the stake of the 6

27 founder is below the threshold α, he has no power is-à-is the manager. Put differently, owning a stake below α and dispersing the shares among small inestors enables the founder to ommit not to interfere through monitoring in the running of the firm. 1 When α α, the founder monitors in part to aoid expropriation by the professional manager, but he does so only to help himself. Indeed, from the minority shareholders perspetie, monitoring is a pure rent-seeking atiity. The founder and the professional manager agree to set φ = φ irrespetie of the monitoring intensity m. 13 This result illustrates an important differene between legal shareholder protetion and monitoring: while the law protets all the shareholders, monitoring in this model is a form of self-protetion by the founder that has either positie or negatie externalities for other inestors. At date 1, the professional manager aepts to run the firm if the wage and his share of the priate benefits exeed the outside utility. The ondition ( w+ φ m(1 + α)/) an be rewritten as ( φ + w) m m min φ, 1+ α The maximum leel of monitoring ompatible with the professional manager s partiipation onstraint dereases with the founder s stake. A larger blok inreases the founder s outside option in the bargaining, thereby reduing the share of priate benefits that the 1 In the absene of ollusion, fully dispersed ownership (α =0) ensures no monitoring (m(0)=0). Introduing a threshold α in etion III would hae ompliated the analysis without adding any insight. 13 If one introdues a dead-weight loss assoiated with priate benefit extration, monitoring also benefits minority shareholders beause the founder internalizes part of the ineffiieny and hene redues the leel of diersion (Burkart and Panunzi (001)). 7

28 professional manager obtains. Nonetheless, as (1 + α) / 1, the threshold m is higher for a gien wage w and legal protetion φ when founder and professional manager ollude than in the absene of ollusion. Withholding ( 1 α)m from the minority shareholders makes it more likely that the professional manager s partiipation onstraint is satisfied. welfare At date 0, the founder hooses the ownership struture α and wage w to maximize his V 1+ α m = (1 φ w) + m k. As monitoring weakly inreases with the blok size, the founder s welfare inreases in α and dereases in w. As in the absene of ollusion, the optimal ownership struture is as onentrated as possible and the optimal wage as low as possible proided that the manager s partiipation onstraint is satisfied. To simplify the analysis, we impose an upper bound on α. k Assumption 4: α 1 The threshold α > 0 14 (Assumption 3) and Assumption 4 imply that the equilibrium monitoring intensity m is gien by the first order ondition (1 α ) / m= + k. In the absene of the threshold α > 0 or of its upper bound, monitoring intensity for small φ alues would be determined by the legal threshold ( m = φ ), leaing the ownership struture indeterminate in that range of φ alues. Nonetheless, all our subsequent results on founder welfare, share alue, and 14 As α > 0, Assumption 4 implies that / k /. 8

29 separation of ownership and management are robust with respet to relaxing Assumptions 3 and 4. Lemma 3 i) For φ, α < α, w = φ, m = 0, and V ( α, w, φ) = (1 + α) ii) For < φ +, α < α, w = 0, m = 0, and 8k V = ( 1 φ) (1 + iii) For α ) (1 α) φ + + < +, α = α, 8k 4k w α (1 + ) 1+ α = φ +, m = 4k k, and V = (1 + α) 8k ( ) i) For (1 + α) + < φ 4k + k, α = k( φ ) 1, w = 0, m = φ k, and V φ ( α, w, φ) =. ) For φ > +, α = 1, w = 0, k m = k, and V ( α, w, φ) = (1 φ) + k Lemma 3 repliates Lemma with an added twist due to ollusion and the disontinuity 9

30 in the feasible monitoring leel. 15 In partiular, region i) oinides with region i) in Lemma and region ) with region iii) of Lemma. The differenes between transferable and non-transferable priate benefits appear in the intermediate range of legal protetion. In region i), weak legal protetion permits priate benefits to an extent that sharing them between the founder and the manager is ompatible with the partiipation onstraint and a zero wage. Indeed, to aoid leaing a rent to the manager, the founder has to retain α > α and monitor aordingly. In ontrast, in regions ii) and iii), where the law is more protetie, a zero wage, sharing of priate benefits, and satisfying the manager s partiipation onstraint are not ompatible with eah other. Beause of the disontinuity in monitoring, the founder faes a trade-off between under- and oer-monitoring. ore preisely, the founder has the option to either not monitor or to monitor at least m= (1 + α) / k. While abstaining from monitoring onedes a rent to the professional manager, monitoring with intensity m= (1 + α) / k requires a positie wage to satisfy the partiipation onstraint. In region ii), the sum of the wage and the monitoring ost when m= (1 + α) / k exeeds the rent that the professional manager an extrat. Hene, it is optimal to fully disperse ownership and to leae the manager with a rent of R = φ. The reerse holds in region iii): oer-monitoring and ompensating the professional manager with a wage to satisfy the partiipation onstraint is less ostly to the founder. Consequently, the optimal ownership onentration jumps to α = α. Another impliation of Lemma 3 is that the founder annot gain from selling the entire ompany to a penniless professional manager who would raise funds in the apital market. From 15 Irrespetie of Assumptions 3 and 4, the equilibrium outome with ollusion is more ompliated than Lemma. In the absene of Assumptions 3 and 4, the ompliation is due to the fat that monitoring may be determined by the law ( m = φ ) rather than by the first order ondition. 30