Fam ily f rm research–A review☆

Qiang Cheng

Singapore M anagement University,Singapore

Fam ily f rm research–A review☆

Qiang Cheng

Singapore M anagement University,Singapore

A R T IC L E I N F O

Article history:

Received 6 February 2014

Accep ted 8 M arch 2014

Availab le online 5 August 2014

JEL classif cation:

G 30

G 32

G 38

M 20

M 41

Fam ily f rm s

Ownership structure

Agency p roblem s

Contro lling owners

M inority owners

Thisarticle reviews fam ily f rm studies in the fnance and accounting literature, primarily those conducted using data from the United States and China.Family owners have unique features such as concentrated ownership,long investment horizon,and reputation concerns.Given the distinguishing features of fam ily ownership and contro l,fam ily f rm s face unique agency conf icts.W e discuss the agency p roblem s in fam ily f rm s and review the fndings of recent fam ily f rm studies.W e call formore research to understand the unique fam ily ef ects and encou ragem o re research on Chinese fam ily f rm s.

Part Io f the article discusses the fundaments o f fam ily f rms:the p revalence of and theagency con f ictswithin fam ily f rms.Part IIsummarizes the f ndings of recent U.S.fam ily f rm studies.It reviews the evidence on the fam ily f rm prem ium(how,which,and when fam ily f rm s are associated with a valuation prem ium),the manifestation of the agency con f ict between majority and m ino rity shareho lders in fam ily f rm s,earnings quality and co rporate d isclosure,and the determ inants o f fam ily ownership and contro l.Part IIIdiscusses the prevalenceand characteristicso f Chinese fam ily f rm sand reviews the fndings o f related studies.The article concludeswith some suggestions for future research.

?2014 Production and hosting by Elsevier B.V.on behalf o f China Journalo f Accoun ting Research.Founded by Sun Yat-sen U niversity and City University o f Hong Kong.

1.Fundamentals of fam ily f rms

1.1.The prevalence and uniqueness of family f rms

A fam ily f rm isa f rm in which the foundersor descendantsof the founding fam ily continue to hold positions in the topm anagement,serve on the board,or are b lockholders.1Th is defnition is w idely used in fam ily f rm studies conducted using U.S.data(e.g.,Anderson and Reeb,2003;Villalonga and Am it, 2006;A lietal.,2007).Using this commonly used defnition is im portant for several reasons.First,it facilitates com parison of the resu lts across studies.Second,it is less subject to researchers’discretion in the classif cation of fam ily.Third,to the extent that the fam ily owners in som e of the classifed fam ily f rm s have weak infuence in the f rm s,it would introduce a conservative bias to the resu lts.A more restrictive defnition(e.g.,requiringm ultip legenerations to be involved in the f rm)m akes the results less generalizable.One should keep in mind the tradeof between using a more restrictive versus a more lenient defnition and themost appropriate defnition is likely to be country specif c.See Villalonga and Am it(2010,pp.866–867)for a further discussion of fam ily f rm defnition.Asan importantorganizational form, fam ily f rms account for 44%of large f rms in W estern Europe(Faccio and Lang,2002),over two-thirds o f f rms in East Asian countries(Claessens et al.,2000),and 33%and 46%o f the Standard and Poor(S&P) 500 and 1500 index companies,respectively(e.g.,Anderson and Reeb,2003;Chen et al.,2008a).Fam ily f rms also operate in a broad range of industries.Am ong the S&P 1500 com panies,fam ily f rm s account fo r twothirdso f f rm s in the high-tech industries(e.g.,pharm aceu tical products,electronic equipm ent),wholesale and retail,transpo rtation,and p rin ting and pub lishing.Even am ong cap ital-in tensive industries(steel wo rks, machinery,autom obile,petroleum,and naturalgas),regu lated industries(banking and insurance companies), and thebusiness supp lies industry,which are the least likely to be fam ily contro lled,over 30%are fam ily f rms (Chen et al.,2008a).

Compared to nonfam ily owners and other blockho lders,fam ily ownershave some unique characteristics. First,founding fam iliesho ld poorly diversifed portfo lios due to their concentrated ownership in fam ily f rms. W ithin fam ily f rms in the S&P 1500 index,founding fam ilieshold 17%o f the shares in their f rmson average. M oreover,69.5%o f founding fam ilies hold more than 5%ownership in their f rms,and 24.7%of them ho ld more than 25%.Due to their high ownership and low diversif cation,founding fam ilies enjoy the benef t o f good corporate decisions and at the sam e tim e bear the consequences of bad corporate decisions,and thus fam ily owners have strong incen tives to increase f rm value.

Second,fam ily owners have longer investment horizons than other shareho lders.They generally regard their ownership as an asset to pass on to future generations.For examp le,when W illiam Lauder,grandson of the founder of Estee Lauder,stepped down,hemade the follow ing comment:“Iam comm itted to the company.It’s thevastmajority o fmy personalwealth andmy fam ily’spersonalwealth–and we fully expect to be actively involved w ith thiscompany going forward”(The Wall Street Journal Nov 9,2007,‘Lauder ScionW ay Out,P&G ExecutiveW ay In’).Such long-term comm itment imp lies that fam ily owners care about the longterm value of the f rm,rather than the short-term gain.

Third,fam ilym embersareactively involved in themanagemento f their f rms,either as top executivesor as directors.On average,found ing fam ilies ho ld the CEO position in 62%of fam ily f rm s w ithin the S&P 1500. M o reover,98.4%o f founding fam ilies appoint at least one fam ily m em ber to their boards,54.6%of them appoint two fam ily m em bers,and 22.9%of them appoin t three or m ore fam ily m em bers.Found ing fam ilies’substantial invo lvement in their f rms’managem ent teams ensures that their preferences are ref ected.

Notall fam ily f rmsare thesame.Themost important classif cationw ithin fam ily f rms is the identity of the CEO.Depending on the identity of the CEO,fam ily f rms can be classif ed as(1)founder CEO f rms,(2) descendant CEO f rms,and(3)other fam ily f rms,usually referred to as pro fessional,or hired,CEO fam ily f rms.2The name of the third category does not imp ly that founder CEOs or descendant CEOs are not professional;nor does it im p ly that founder CEOs and descendan t CEOs are not hired.Founder CEOsare usually charismatic and visionary leaderswith greatmanagem ent skills.They also tend to have a strong w ill and an undisputed and powerfu l status in their f rms.In contrast,descendantsare often criticized for being spoiled bratsand lessskilled.This classif cation iscritical to the understanding of the agency p rob lem s in fam ily f rms,as discussed in the next section.

The unique characteristics o f fam ily owners and fam ily f rm sm otivate m any of the in teresting topics in fam ily f rm research.This is thekey reason for theemerging lineof research on fam ily f rms in thepast decade. This review focuses primarily on fam ily f rm studies published in leading accounting and f nance journals.

1.2.Agency problems in family f rms

The foundation on which most fam ily f rm studiesare built is the agency con f ict framework.The unique characteristicso f fam ily f rmsafect thenatureand extent of agency problems,which also varyw ith the typeo f fam ily f rm.There are two main agency problems in public companies:the con f ict between managers and shareholders and the conf ict between majority and m inority shareholders.Below we discuss the two types o f agency problems that app ly to fam ily f rms relative to non-fam ily f rms.

1.2.1.The conf ict between managers and shareholders

In the classic owner-m anager conf ict,as described in Jensen and M eck ling(1976),the separation o fm anagers from shareholdersm ay lead to m anagers no t acting in the best interest o f the shareho lders.W e refer to this type o f agency prob lem as the Type Iagency p rob lem.However,the extent of Type Iagency problem s is reduced in fam ily f rms for several reasons.

First,as discussed above,fam ily owners tend to ho ld concentrated and under-diversifed ownership of their f rms.Asa result,fam ily owners are likely to have strong incentives to monitormanagers,reducing the free rider prob lem that is p revalent among other f rms.[The benef t ofmonitoring does not outweigh the cost o f m onitoring for smallatom istic shareho lders,and assuch,they tend to free rideon others’monitoring(Shleifer and Vishny,1986).]Given the under-diversif cation o f their portfolios,fam ily owners bear the idiosyncratic risk associated w ith the f rm and are thus concerned w ith the cash fows it generates.Founding fam ilies’long tenure and substantialinvolvement inmanagement imply that they areknow ledgeableabout their f rms’activities,which in turn enables them to provide better m onito ring o fm anagers.

Second,found ing fam ilies tend to have m uch longer investm en t ho rizons than other shareholders.Their long-term p resence in the f rm im p lies that fam ily owners are w illing to invest in long-term p rojects.Thus, fam ily owners can help to m itigate them anagerialm yop ia problem(Stein,1988,1989).Because the founding fam ily views the f rm asan asset to passon to futuregenerations rather than aswealth to be consumed during their lifetimes(James,1999),f rm survival isan important concern.Hence,fam ily owners have even stronger incentives to monitor than other large and long-term shareho lders.

Third,founding fam ilies are concerned w ith the fam ily’s reputation.They aremore w illing to build and p rotect their reputation,which is likely to have long-term ef ectson third parties,and hence the fam ily business.Founding fam iliesare likely to dealw ith other stakeho lders,such asbanks,suppliers,and customs,for longer periods.Thisalso gives fam ily f rms stronger incentives to execute ef ectivemonitoring than other large shareholders.

Lastly,in founder and descendan t CEO f rm s,the owner and the CEO are one and thus there isno incen tive m isalignment and no Type Iagency problem.Recall that Type Iagency p rob lemsarisewhen the owners’and themanagers’interest are not aligned.

In summary,compared to non-fam ily f rms,fam ily f rms face less severe Type I agency problems arising from the separation o f ownership and contro l.

1.2.2.The conf ict between majority andminority shareholders

The second type of agency prob lem is the conf ict between majority and m inority shareho lders.As they hold substantialownership and have contro lling positions in the f rm,majority shareholdersmay seek private benef ts at the expense ofm inority shareholders(Shleifer and Vishny,1986).We refer to this type of agency p rob lem as the Type IIagency problem.Fam ily f rmshavea largeshareholder(the fam ily owner)and a fringe o f sm all shareho lders.As such,fam ily f rmsare sub ject to severeagency problemsbetween fam ily ownersand m inority shareho lders.

The prim ary sou rce for this type o f agency p rob lem is found ing fam ilies’concen trated equity holdings and substantial con trol in their f rm s,which gives them the opportunity to extract private benef ts at the expense o f other shareholders.Privatebenef tsmay beboth themonetary and thenon-monetary benef ts from running af rm.For exam p le,when discussing the CEO turnover decision in Ford Corporation,BusinessW eek(August 21–28,2006)comments that“[given his poor performance,]CEO Bill Ford would have been f red by now by mostboards if hisnamewere Sm ith.”Fam iliesare also capableo f exp rop riatingwealth from the f rm through excessive compensation,related-party transactions,or special dividends(Burkart et al.,2003).

Another important sourceo fpotential fam ily entrenchment is thediferencebetween their control rightsand cash-f ow rights.Villalongaand Am it(2009)show that founding fam iliesare theprimary typeo fb lockho ldersto hold control rights in excessof their cash-fow rights in U.S.corporations.Based on 3006U.S f rm-year observations from 515 f rmsbetween 1994 and 2000,they f nd that founding fam ilieson average own 15.3%of the shares(cash f ow rights),butcontrol18.8%of thevotes in those f rms.Thewedgeisprimarily due to theissuance of dual-class shares.Fo r exam p le,Google’sco-founders,Sergey Brin and Larry Page,own super-voting class B shares,which have10 votespershare.O therhigh-tech f rm s,such asFacebook,havesim ilardual-classstructures. Founding ownersalso obtain dispropo rtionally higher controlvia d isp roportionateboard represen tation,voting agreem ents,and pyram id ownership structures.Such awedgeprovides them w ith theincentiveand ability to pursueprivatebenef ts.While fam iliesmay takeactions thatmaxim ize their personalbenef t,many of theseactions can lead to suboptimal corporatedecisions that reduce thevalue tom inority shareholders.

In summary,compared to non-fam ily f rms,fam ily f rms facemore severe agency conf ictsbetweenmajority and m inority shareholders.

1.2.3.Tension within the family

Compared to Type Iand Type IIagency p roblems,the intricate relationshipswithin fam iliesare evenmore unique and interesting.Fam ily feuds can exist between founders and descendants and,more commonly, between descendants,due to dif erences in their interest and vision.

Oneexamp leof a fam ily feud between foundersand descendants is theRedstone fam ily.Sumner Redstone, an American media magnate who is the controlling shareholder and chairman o f the National Am usement theater chain,shoved a succession of‘heirs apparent’ou t of the doo r.H is daugh ter,Shari Redstone,was set to assum e his role.However,in 2007,when Sum ner Redstone was 84 years old,they feuded publicly over issues of co rpo rate governance and the fu ture o f the cinem a chain.3Jenn Abelson,“Redstone says he relies on his instinct,”The Boston Globe,Sep tember 19,2007.

Another example of a fam ily feud is the discord in the Georgina Rinehart fam ily.Georgina Rinehart’s farther,Lang Hancock,established the Hope M argaret Hancock Trust before his death,nom inating Georgina Rinehart as the trustee and his four grandchild ren asbenef ciaries.4Hall,Louise and Pennells,Steve.“Rinehart’s children w in f rst round,”The Sydney M orning Herald,October 8,2011.However,in 2011,Rinehart’s three o lder child ren brought legal action in the Supreme Court o f New South W ales,Australia,to have Georgina Rinehart removed as so le trustee due to a commercial dispute.5H all,Lou ise.“Fam ily feud details to stay secret for at least f vem ore weeks,”The Sydney M orning H erald,February 2,2012.,6Am y D ale.“Days away from being billionaires,Gina R inehart locks trust fo r half a century,”The Daily Telegraph,M arch 13,2012.

The tensionswithin the fam ily havenot yetbeen fu lly investigated.A casestudy,rather than a largesample study,is better suited for such investigations.

2.What havewe learned from recent family f rm research?

2.1.Valuation premium

One fundam ental fnding in the fam ily f rm research is that fam ily f rm s on average perfo rm better than non-fam ily f rm s.Anderson and Reeb(2003)docum ent in their sem inal work that fam ily f rm s have higher returns on assets and Tobin’s Q than non-fam ily f rm s.This f nding is con f rm ed by later stud ies(e.g., Villalonga and Am it,2006;M aury,2006;And res,2008)and suggests that fam ily ownership is an ef ective organizational structure.The questions then are how,which,and when fam ily f rms create value.

First,how do fam ily f rms createmorevalue?Anderson and Reeb(2003)attribute the fam ily f rm valuation prem ium to the reduction in agency prob lems associated w ith managerial opportunism,or Type I agency prob lems.They and others fnd that fam ily f rms perform better when fam ily members serve as the CEO(Anderson and Reeb,2003;M au ry,2006;Andres,2008),suggesting that active fam ily invo lvem ent and contro l can readily align managers’interestwith that o f the fam ily.Anderson et al.(2003)further investigate the issue from theangleof the costo f debt.They f nd that founding fam ily ownership isassociated w ith lower cost o f debt.This resu lt is consistentw ith the notion that agency conf ictsbetween equity and debt claimantsare also lower in fam ily f rms,and/or the notion that fam ily owners’reputationsand long horizonsare benef cial and can reduce the transaction costs of dealing w ith other stakeholders.

Another reason fam ily f rms perform better is that they are less likely to destroy value throughmergersand acquisitions.Caprio et al.(2011)investigate 777 large European companies during the 1998–2008 period and show that fam ily f rms are less likely to make acquisitions and that their acquisitions are of higher quality. Focusing a sam p le o f Japanese f rm s,Shim and Okam u ro’s(2011)analysis also show s that fam ily f rm s are less likely to conductm ergers than non-fam ily f rm s.

Second,who creates value in fam ily f rm s?Som e stud ies show that founders create value and descendan ts destroy value(Villalonga and Am it,2010).Founder CEOscan bring innovativeand value-enhancing expertise to the f rm(M orck et al.,1988).Villalonga and Am it(2006)f nd that fam ily ownership creates value on ly when the founder serves as the CEO of the fam ily f rm or as the Chairman w ith a hired CEO.Fahlenbrach (2009)documents that11%o f the largest U.S.public f rmsareheaded by founder CEOs,and these f rms invest m ore in research and development,have higher capital expenditures,makemore focused mergersand acquisitions,and earn a benchmark-adjusted annual return o f 8.3%during the 1993–2002 period.

Pe′rez-Gonza′lez(2006)observes that compared w ith p rofessional CEO fam ily f rms,descendant CEO f rms underperform in terms of operating p rof tability and market-to-book ratio.Bertrand et al.(2008)also docum ent lower f rm perform ance associated w ith greater invo lvem en t by founders’sons,especially after the founders pass away.Cucculelli and M icucci(2008)f nd sim ilar results based on a sam p le of Italian fam ily f rm s.

The underperformance by descendant CEO f rms is largely attributed to their lack ofmanagerial skills. These descendants are chosen as the CEO not because they are the best candidates for the position,but because they are the descendants of the founders.To useW arren Bu fett’s analogy,picking executives from the small pool of fam ily heirswould be like“choosing the 2020 Olympic team by picking the eldest sons o f the go ld-medalwinners o f the 2000 Olympics.”7D avid C.Johnston,“Dozens of R ich Am ericans Join in Fight to Retain the Estate Tax,”New York Times,February 14,2001.

The third question is,when do fam ily f rmscreatevalue?As founding fam iliescan use their power to accrue p rivate benef tsat the expense ofm inority shareho lders,the checksand balanceson founding fam ilies’power can help to reduce Type IIagency p rob lems.Anderson and Reeb(2004)f nd that fam ily f rms perform best when the power of the found ing fam ily is balanced by the presence o f independen t d irectors.Fam ily f rm s,on average,only ou t-perform when there is no dual-class structu re o r over represen tation on their boards (Villalonga and Am it,2009).Anderson et al.(2009)fu rther docum ent that pub licly traded f rm s exp loit opacity to extract p rivatebenef tsat theexpenseofm inority shareholders.They f nd that founder-and descendantcontro lled f rmsoutperform other f rms on ly when the information environment is o f high quality.Evidence from Western European countries shows that fam ily contro l is associated w ith higher f rm valuemain ly in econom ies with better shareholder p rotection and country-level legal infrastructure,which reduce Type II agency p rob lem s(M aury,2006).

2.2.Evidence of Type IIagency problems in family f rms

Because founding fam iliesen joy substantial contro lasa resulto f their concentrated equity holdings,fam ily f rmshavemoresevere Type IIagency prob lems than non-fam ily f rms,asdiscussed above.Thevaluation p rem ium enjoyed by fam ily f rms isconsistentw ith thenotion that thebenef t from the reduction in Type Iagency p rob lem s outweighs the cost from them ore severe Type IIagency problem s.However,several recen t studies p rovide som e direct evidence that fam ily f rm s are sub ject to m ore severe Type II agency p rob lem s.

Villalonga and Am it(2006)f nd that the use o f control-enhancingmechanisms,e.g.dual-class structures, pyram ids,and voting agreements,reduce founder CEO f rms’valuation prem ium.Sim ilarly,Anderson and Reeb(2004)f nd that in f rms with concentrated founding-fam ily ownership and relatively few independentdirectors,f rm perfo rm ance is signif cantly wo rse than in non-fam ily f rm s.These results indicate that the Type IIagency problem is particularly severe for fam ily f rmswith contro l-enhancingmechanismsand lessef ective monitoring o f fam ilym embers.

M ore direct evidence comes from the investigation of specif c corporate decisions.Anderson et al.(2012) exam ine the in formation content of short sales in fam ily and non-fam ily f rmsand fnd that founding fam ilies are involved inmoreaggressive in formed trading than other large shareho lders.M oreover,Chen etal.(2013a) observe that both founder and descendant CEO fam ily f rmsare less likely to let theCEO go after their f rms perform poorly.CEO turnover–performance sensitivity,ameasure o f agency problems in the CEO turnover setting,is lower for fam ily CEO f rms than for pro fessional CEO fam ily f rms.When the fam ily ownership is higher,fam ily CEO tu rnover is even less sensitive to f rm perform ance.Their evidence suggests that the Type IIagency problem m anifests itself in the CEO tu rnover setting and the p rob lem increases w ith fam ily ownership and control.They also fnd that CEO turnover–perform ance sensitivity is higher in pro fessional CEO fam ily f rms than in non-fam ily f rms.Overall,they fnd that fam ily ownership/control is a doub le-edged sword:it leads to Type II agency prob lems when fam ily ownership is combined w ith fam ily contro l and reduces Type Iagency p rob lemswhen fam ily ownership is separated from contro l.The ef ects increasew ith the level of fam ily ownership.

2.3.Accounting phenomenon

Theuniqueownership structure of fam ily f rmshas important imp lications for their fnancial reporting and disclosure practices.In term s of earningsquality,W ang(2006)argues that the incen tives for found ing fam ilies to exp ropriatewealth from other shareholders leads to lower earningsquality(the en trenchm ent ef ect),while greater m onito ring by found ing fam ilies im p lies higher earnings quality(the alignm en t ef ect).He show s that founding fam ily ownership isassociated w ith lower abnormal accruals,greater earnings informativeness,and less persistence of transitory loss components in earnings,suggesting higher earnings quality in fam ily f rms. A li et al.(2007)further suggest that fam ily f rms can address agency p roblems through directmonitoring rather than rewarding managers based on accounting performance,such that managers are less likely to manipulate theearnings.Consistentw ith thisnotion,they fnd that fam ily f rms reporthigher earningsquality, including lower discretionary accruals,higher p redictability of cash fows,and higher earnings response coeff cients,thus corroborating the results reported by W ang(2006).

In addition to the quality o fmandatory reporting,studies also exam ine the vo luntary disclosure po licies adopted by fam ily f rm s.A liet al.(2007)argue thatm ain taining the opacity o f corporate governance p ractices m ay facilitate fam ily m em bers’entrenchm ent.They fnd that fam ily f rm sm ake fewer disclosures abou t their corporate governance p ractices.Chen et al.(2008a)further argue that fam ily owners have a longer investm en t horizon than other shareho lders;hence,they do not en joy the benef t of timely disclosure but bear the costo f disclosure due to their high ownership.Consistentw ith their prediction,they fnd thatmanagersare less likely to p rovideearnings forecastsand hold fewer conference calls.However,w ithho lding bad news iscostly for the f rm and shareholders because o f the potential litigation costs.Given founding fam ilies’concentrated and under-diversif ed equity ho ldings,they arem ore concerned w ith the litigation and reputation costs.Consistent with their conjecture,Chen et al.(2008a)document that fam ily f rms providemore earnings warnings than non-fam ily f rms.

Because tax avoidanceactivitiesare comp lex and opaque,they can be used to hide lossesor rentextraction by them ajority shareholders(Desaiand D harm apala,2006).Build ing on this line o f logic,Chen et al.(2010) fnd that fam ily f rm sare less tax aggressive than their non-fam ily coun terparts.They suggest that fam ily ownersarew illing to forgo tax benef ts to avoid the non-tax costo f a potentialprice discount.As fam ily f rmsare characterized by a unique agency conf ict between dom inant and small shareholders,fam ily owners enjoy greater share price discounts if they aremore tax aggressive,which can be viewed asmasking rent extraction activities.In addition,due to theirmuch larger equity ownership and theirmuch longer investment horizons, fam ily ownersaremore concerned w ith thepotentialpenalties imposed by the IRSand the reputation damage from being invo lved in a tax-related lawsuit.

Overall,the extant literature fnds that fam ily f rms generally have better fnancial reporting quality,are more likely to issue earningswarnings,and are less tax aggressive.

2.4.Determ inants of fam ily ownership

Because the decision to control a f rm is endogenous,a crucial issue in the study of fam ily f rm s iswhat determ ines the ownership and contro lstructureo f fam ily f rms.The currentevidence suggests that bothm icro and macro factors shape the fam ily ownership concentration and the wedge between fam ily control and ownership.

Themajority of f rms start as fam ily f rm s.W hether the founders or later generations decide to maintain fam ily ownership and controldependson the costsand benef tsof doing so,and on thestructureof the founding fam ily.Bennedsen et al.(2007)f nd that the gender of the departing fam ily CEO’s eldest child strongly in fuences the decision to appoin t a fam ily m em ber or an outsider as the CEO.The frequency o f appointing a fam ily CEO is 29%when the eldest child is fem ale,bu t 39%when the eldest child ism ale,suggesting that founding fam ilies arem o re likely to m ain tain contro lwhen they have sons.Bertrand et al.(2008)docum en t a positive association between fam ily size and fam ily ownership and contro l.Such evidence is corroborated by anecdotalevidenceand ref ects the general cu ltureo f letting theson takeover the fam ily business.In addition, Bertrand and Schoar(2006)suggest that the fam ily values in a culture p lay an important ro le in shaping how businesses are organized and their ef ciency.

A t themacro level,fam ily ownership is found to be related to the institutional and market environment. Burkart et al.(2003)show that fam ily ownership isnegatively correlated with the levelo f investor p rotection in a country:when investor protection isweak,fam ily ownership ishigh.In their international study,Franks etal.(2012)observe that fam ily f rmsgradually evolve into w idely held companies in countriesw ith developed f nancialm arketsand strong investor p rotection.In contrast,fam ily con tro l is prevalent in countriesw ith less developed fnancialm arkets and weak investor p rotection.

Studieso f how fam ily businessgroupsarestructured come to a sim ilar conclusion:pyram idsare formed to allow a fam ily to achieve controlof the businessgroup using only a small cash fow stake(e.g.,Claessensetal., 2000;A lm eida and Wo lfenzon,2006).In an empirical study of fam ily-contro lled businessgroups in 45 countries,M asulisetal.(2011)document that fam ily groups,especially those structured aspyram ids,arem orep revalent in marketsw ith low capital availability,suggesting that group structures emerge not only to maintain contro l,but also to alleviate fnancing constraints.

Additional evidence from K ing and Peng(2009)shows that fam ily ownership concentration varies w ith industry characteristics.Founding-fam ily f rms in cyclical,capital intensive,and grow th businesses have signif can tly sho rter con trol spans than other fam ily f rm s.

In sum,whether the founding fam ily can m ain tain con trol and high ownership depends on the level o f investo r p rotection at the country level,the characteristics of the business at the industry level,the capital needs at the f rm level,and the structure of the founding fam ily at the fam ily level.

3.Family f rms and family f rm research in China

The above d iscussion focuses p rim arily on stud ies using U.S.data and som e international studies.In this section,we brief y d iscuss the characteristicso f Chinese fam ily f rm s and review the literatu re on these f rm s in the area of accounting and fnance.

3.1.Family f rms in China

Compared to fam ily f rms listed on theU.S.exchanges,Chinese fam ily f rmsarestillatan early stage.Since China started to transfo rm from a p lanned econom y to a socialistm arket econom y in 1992,fam ily f rm shave grown rapid ly in both p revalence and size.To obtain a basic understanding of the characteristics o f Chinese fam ily f rms,we exam ine those listed on the Shenzhen and Shanghai Stock Exchanges in the 2003–2012 period.8W e only include A-share fam ily f rm s here.There are only about eigh t B-share fam ily f rms and m ost of them also issue A-shares.Because Chinese listed f rmshave been required to disclose their control structure and ultim ate ownership since 2003,our sampleperiod starts in that year.D isclosureo f the contro lstructuremakes it possib le toidentify whether a listed f rm is controlled by the government,an institution,or a fam ily.W e use the same def nition as discussed in Section 1 to classify fam ily f rms.9Th is def nition is diferent from others in that it extends the defnition o f“fam ily”to include no t only found ing fam ilies bu t also individual investors or fam ilies that are not(related to)the founder.See Villalonga and Am it(2010,p.867).

Table 1 The distribution of Chinese fam ily f rm s.

3.1.1.The distribution of Chinese family f rms

Table 1 presents the distribution of Chinese fam ily f rms.Several features emerge.First,themajority o f Chinese fam ily f rms are founded by mainland Chinese citizens.A small p roportion o f them are contro lled by citizens from Hong Kong,M acao,Taiwan,or overseas(Table 1,Panel A).10Note that based on the data available from the database,it is im possible to comp letely separate f rm s into single-fam ily f rm s,multifam ily f rm s(where founders are not related to each other),and m ulti-generation fam ily f rm s(held by multip le members of the sam e fam ily).It is also d if cult to identify f rms controlled by the founder or the descendent.Such data need to be hand-collected.

Second,the num ber of fam ily f rm s was sm all com pared to the num ber o f state-owned f rm s in the early years.This is consistentw ith the notion that the Chinese stock market was established to help state-owned f rms to obtain capital and to reform,and that itwas dif cu lt for fam ily f rms to obtain IPO app roval from the regu lators in the early years.From 1992 to 2002,17%o f IPOsare fam ily f rms(not tabulated).The number o f fam ily f rm IPOs gradually increased after M ay 2004 when the small-and medium-sized enterprise board was established and increased dramatically after October 2009 when the grow th enterprise board was founded.As shown in Panel A of Tab le 1,the number o f Chinese fam ily listed f rms increased from 170 in 2003 to 1373 in 2012.Given thestrict IPO criteria for fam ily f rms in theearly years,mostof them went pub lic by taking over a listed f rm and then relying on the seasoned equity of erings to obtain capital.

Third,although fam ily f rm soperate in a w ide range o f industries,m ost are in industries that have en joyed grow th in recen t years,such as petro-chem ical,electronic,iron/steelm ills,m achinery,pharm aceutical,telecomm unication,and real estate(see Panel B o f Table 1).

Fourth,most Chinese fam ily f rmsare clustered in themore developed provinces,as shown in Panel C o f Table 1.Over 60%o f Chinese fam ily f rm s are located in coastal provinces and regions:Guangdong(250), Zhejiang(188),Jiangsu(169),Beijing(94),Shandong(88),and Shanghai(79).Very few fam ily f rms are located in interior p rovinces such as Qinghai(5),N ingxia(6),Yunnan(6),and Guizhou(6).The heterogeneousmarketization across p rovinces providesa good opportunity to study how the institutional and market environmentaf ectsChinese fam ily f rms.A re fam ily f rms themain d river o f growth in the region,or does the in frastructure in the region help thegrow th o f fam ily f rms?Does the social and commercial cu lture afect the formation and evolvement of fam ily f rms?

3.1.2.Ownership structure,corporate governance and f nancial characteristics

Table 2 p resents some basic descrip tive statistics for Chinese fam ily f rms.Several points are noteworthy.

First,Panel A of Table 2 shows that Chinese fam ily f rmshaveaverage assetsof US$393.3m illion(median o f US$206.7m illion)and an averagemarket valueo f US$594.2m illion(median of US$324.3m illion).11W e use the of cial 2012 exchange rate of US$1=RM B6.39.Chen etal.(2008a)report that during the 1996–2000 period,S&P 1500 f rmshad average assetsof US$1,152.9m illion(m ed ian of US$982.4m illion)and an averagem arket value o f US$6266m illion(m edian o f US$1150 m illion).In term sof p rof tability,the average retu rn on assetsof Chinese fam ily f rm s is3.36%(m edian of 4.39%), while the average return on assets in the S&P 1500 fam ily f rm s is 5%(m ed ian of 6%).W ith respect to grow th opportunities,the averagem arket to book ratio is 4.32(median o f 3.19)for Chinese fam ily f rms and 4.31 (median o f 2.64)for S&P 1500 fam ily f rms.While not exactly comparable,Chinese fam ily f rms are smaller and less p rof tab le and have higher grow th than their U.S.counterparts.

Second,theownership concentration isvery high in Chinese fam ily f rms.Panel B of Table2 shows that the average ownership among fam ily owners is 34.59%(median o f 31.36%),which is about tw ice the p roportion among S&P 1500 fam ily owners(Chen et al.,2008a).Considering the pyram id structure,Chinese founding fam ilies own an average o f 28.32%(median o f 24.74%)of the cash f ow rights and 36.44%(median o f 33.33%)o f the voting right in their f rms,while Villalonga and Am it(2009)report that the founding fam ilies o f Fortune 500 com panies own an average of 15.3%of the shares and 18.8%o f the vo tes in their f rm s.12If a family owns 60%of f rm A,which then owns 40%of f rm B,which is listed on one of the exchanges,then the fam ily’s direct ownership in f rm B is considered to be 40%and the cash fow right in f rm B is considered to be 24%(=60%×40%).These term s are common ly used in the literature on pyram id structures.The direct ownership and cash fow right arenaturally the sam e in casesw ithout a pyram id structure.Cash fow righ ts difer from vo ting rights when some shares have diferent voting and cash fow rights.Chinese fam ily f rm s p resent am uch higher level of ownership concentration and higher degrees of separation between cash f ow righ ts and vo ting righ ts.13Theownership,cash fow rights,and voting rights reported here are thoseof the contro lling family owners.Because fam ily ownership consistsof controlling ownership and other fam ily ownership,the amountsof family ownership,fam ily cash fow rights,and fam ily voting rights would be even greater than those reported here.

Third,asPanel B of Tab le 2 shows,only 30%of CEOs in fam ily f rmsserveas theChairman o f theboard in Chinese fam ily f rms.14Recently,the founder of the A libaba G roup,M a Yun,resigned from the CEO position while still serving as the chairman of theboard.A typical board has ninemembers,three of whom are independent.This fraction is m uch lower than it is in U.S.fam ily f rms.Chen et al.(2008a)report that 62%o f the directorson the boardsof S&P 1500 fam ily f rms are independent.Furthermore,a mere 21%of Chinese fam ily f rms have implemented equity-based incentive plans for their CEOs.

Table 2 Descriptive statistics for Chinese fam ily f rm s.

Fourth,compared to non-fam ily f rms,fam ily f rms are smaller and have lower leverage,better performance,and higher grow th potential.Fam ily f rmsare alsomore likely to have CEOs serving as the chairman of the board,have sm aller boards,and arem o re likely to use incen tive-based com pensation fo r their CEOs.

These statistics suggest that Chinese fam ily f rm s are sm aller and younger than U.S.f rm s.A t this stage,the ownership,contro l,andmanagement aremain ly centralized in the handsof the foundersand their immediate fam ilymembers.The fam ilymembersof Chinese founding fam ilies are likely to havemuch greater inf uence on their f rms’corporategovernanceand fnancialdecisions than fam ilymembersof U.S.fam ily f rms.Hence, the“fam ily ef ect”is likely to bemuch more p rom inent in Chinese than in U.S.fam ily f rms.

3.2.Chinese fam ily f rm research

To date,there are very few stud ies of Chinese fam ily f rm s.A search on Google Scho lar fo r stud ies published in the internationalaccounting and f nancial journals15Theacadem ic accounting journals include The Accounting Review,Journalof Accounting&Econom ics,Journalof Accounting Research, Review of Accounting Studies,Contemporary Accounting Research,Journalof Accounting&Public Policy,Journalof Accounting,Auditing and Finance,and Journal of Business Finance and Accounting.The academ ic fnancial journals include Journal of Finance,Journal of FinancialEconomics,Review of FinancialStudies,Journalof Financialand Quantitative Analysis,Journalof Banking and Finance,Journalof Corporate Finance and FinancialM anagement.Because theA sia-Pacif c Journalof Accounting&Econom ics and China Journalof Accounting Research m ainly focus on the Asia-Pacif c and Chinesem arkets,we also include these two journals in the Google search.revealed on ly three papers that focusexclusivelyon Chinese fam ily f rm sand 27 papers that investigate the dif erences between state-owned en terprises(SOEs) and non-SOE f rms,most o f which are fam ily f rms.Research on fam ily f rm s is also lim ited in the Chinese academ ic journals.This lack of research is puzzling given how p revalent Chinese fam ily f rm s are and how diferent SOEs and fam ily f rms are in terms o f their ownership structures and managers’incentives.Below we discuss some important f ndings based on these studies and of er some suggestions for future research.

3.2.1.Corporate valuation and f rm performance

Sim ilar to those in theU.S.market,Chinese fam ily f rmsoutperform non-fam ily f rms in the Chinesemarkets.A llen et al.(2005)fnd that the p rivate sector in China growsmuch faster than other sectorsand is the p rimary d river behind most o f the country’s econom ic grow th.Sim ilarly,Chen et al.(2008b)document that when contro l o f a f rm is passed to a p rivate entity,the f rm’s performance imp roves.

Given that fam ily f rms perform better than non-fam ily f rms,one naturalquestion iswhy fam ily f rmsare m ore ef cient in an econom y where the legal and m arket environm ents are far from perfect.This question is particu larly interesting because SOEsen joymost o f the fnancing and po litical advantages.One possible reason is the better incentivemechanism in Chinese fam ily f rms,although the evidence is scarce at themoment. Some studiesargue that Chinese fam ily f rms estab lish incentive systems for p rof tmaxim ization.Firth et al. (2006)fnd that f rmsowned by p rivateb lockholders link theCEO’spay to stock and accounting performance. Chen et al.(2012b)fnd that fam ily f rms use relative performance evaluation for top executives’compensation.Cheng etal.(2008)fnd that turnovers in fam ily f rmsarenegatively associated w ith core earnings.M ore research can be done to exam inewhether there are other drivers o f fam ily f rms’outperformance.

3.2.2.Type IIagency problems in Chinese family f rms

Some recent studiesexam ine Type IIagency p rob lem in Chinese fam ily f rms.Compared to SOEs,fam ily f rms appear to be less likely to use related party transactions for“tunneling”(Peng et al.,2011;Ying and W ang,2013).However,Chinese fam ily owners have a signif cant wedge between vo ting rights and cash fow righ ts,which can m o tivate the found ing fam ily to tunnel corporate resou rces at the expenses of m ino rity shareholders.

To date,m ost o f the studies exam ining Type IIagency problem s in fam ily f rm s are published in Chinese journals.The general conclusions are as follows:(1)founding fam ilies use group companies,internal capital m arkets,and pyram id structures to tunnel resources from their listed companies(Shao and Liu 2007;Liu et al.,2008;Jiang et al.,2010);(2)the investment ef ciency is low in fam ily f rmswhere the founding fam ily hasexcessive control rightsand low cash f ow rights(Han et al.,2007;Chen etal.,2012a);and(3)fam ily f rms have a lower cash dividend payout ratio and a lower p ropensity to pay dividends than non-fam ily f rms(Wei etal.,2011).Extant studiesalso exam ine theconsequenceso f such agency problemsand they f nd that(1)family f rmsw ith a higher degreeof separation between controland cash f ow rightsareassociated w ith lower f rm value(Su and Zhu,2003;Zhang et al.,2004;Yeetal.,2007;Yang and Su,2009);(2)fam ily f rmsw ith amu ltip le-layer corporate structure and a higher degree of separation between contro land cash fow righ ts are associated w ith higher bid-ask spreads(Leiet al.,2013);and(3)auditors charge higher fees for fam ily f rm s that have a higher degree of separation between contro l and cash fow rights,a disp roportionally high num ber o f fam ily directors on the board,and a fam ily member as the CEO(Hu et al.,2012;Liu and Subramaniam, 2013).

3.2.3.Financial reporting quality

In contrast to their counterparts in the U.S.market,Chinese fam ily f rmsexhibit lower fnancial reporting quality than non-fam ily f rms.Huang and Zhang(2011)and Wang and Yung(2011)document that Chinese fam ily f rmshavehigher abnormalaccruals,lower earningsp redictability,and lower conservatism.They attribute their f ndings to theweak investor p rotection in China and Type IIagency problems in fam ily f rms.In contrast,Xu and Lv(2011)indicate that when fam ily members are appointed as top executives,f nancial reporting quality is higher.

Som e stud ies exam ine how m arket interm ediaries af ect the extent o f earningsm anagem en t.Chen et al. (2011a)fnd that the efects o f aud it quality on earningsm anagem ent and the cost o f equity capital arem o re p ronounced for non-SOEs(p rimarily fam ily f rms)than for SOEs.Chen et al.(2013b)also f nd a signif cantlynegative relationship between underw riter reputation and pre-IPO earningsm anagem ent fo r non-SOE issuers. Thisnegative relationship is due either to reputab leunderw riters’selection p rocessor to their activemonitoring and intervention in the p re-IPO stage.

3.2.4.Financing decision and the governance structure of family f rms

The Chinese capitalmarketsare characterized by(1)an uneven levelofmarketization across provinces;(2) weak investor p rotection and legal in frastructure;(3)strong governmental intervention;(4)uneven resource allocation between SOEs and non-SOEs;and(5)a lack o f trust amongmarket participants.These characteristics af ect the form ation and developm ent of Chinese fam ily f rm s.

In terms o f f nancing,Chinese fam ily f rms achieve tremendous grow th despite the lim ited support from state-owned banks.This fnding providesa uniquesetting to testhow f rms in a countryw ith poorly developed fnancial institutions fund their grow th opportunities.Ge and Qiu(2007)fnd that compared to SOEs,non-SOEs usem o re trade credit for fnancing.Lu et al.(2012)f nd that non-SOEs have a greater propensity to hold signif cant ownership in commercial banks,and such ownership helps them to reduce interest expenses and obtain short-term loanswhen the government’smonetary policy is tight.

The uneven resource allocation system in China also motivates Chinese fam ily f rms to establish po litical connections.Chen et al.(2011c)f nd that in regionswhere the localeconomy is lessmarket-oriented orwhere the governm ent hasmore discretion in allocating econom ic resources,non-SOEsarem ore likely to estab lish political connections.W u et al.(2012)also show that fam ily f rm sw ith politically connected managers enjoy tax benef ts and outperform thosew ithout such managers.

In termso f governance structure,Liet al.(2008)fnd that the discrim ination of state-owned banksagainst non-SOEs encou rages the founding fam ilies to build a pyram id ownership structu re to reduce their f nancial constrain ts.Chen et al.(2011b)f nd that the layers o f pyram ids and thewedge between vo ting and cash fow rights decrease w ith the quality of governm ent services.The early stage o f Chinese fam ily f rm s and the poo r legal infrastructu re in China p rovide great opportunities for fu ture research to exam ine how the governance structure of fam ily f rms evolves.

3.2.5.The allocation and ef ects of fam ily authority

A lthough it isw idely accepted that the invo lvement o f fam ilym embers in the fam ily business is themost unique feature of fam ily f rms(Chua etal.,1999),little research hasbeen done in thisarea.Data availability is the biggest challenge.Nevertheless,as suggested above,Chinese fam ily f rmsare still young and fam ilymembers still hold large ownerships and are invo lved in their f rm s’boardsand managem ent teams.These unique attributesp rovidegreatopportunities to explorehow the founding fam ily controls the f rm,how theallocation of fam ily authority afectsoperating ef ciency and corporate governance,and how the con f ictsamong fam ily members af ect f rm operations.Field research and case studiesm ight bemore app rop riate for such research than a large sample study.

Som e recent studies have started to exp lore these issues.He et al.(2010a,b)and He and Lian(2009)examine the composition of fam ily f rms in termsof core fam ilymembers,close relatives,and distant relatives.Lian etal.(2011)fnd that ownership of the f rm ismore likely to beheld by core fam ilymembers,buton ly capable fam ilymembersor p rofessionalCEOsareassigned to themanagement team.Heand Lian(2009)and Heetal. (2010b)show that conf ictsam ong fam ilym embersare induced when a fam ilymember’sownership andmanagement position do notmatch,and such conf icts reduce f rm value.He et al.(2010a,b)find that the con f ict among core fam ily members is the lowest,and the majority of con f icts are between distant relatives or between close and distant relatives.These studies co llectively provide some initial evidence on how the characteristics o f the founding fam ily afect the ownership,corporate governance,and valuation o f a fam ily business.

4.Concluding remarks and opportunities for future research

Fam ily f rm s are very comm on around theworld.Found ing fam ily ownership exhibits unique characteristics.Fam ily owners have concentrated and under-diversif ed ownership.They also have longer investm en t horizons than other shareholders,including other blockho lders.Fam ily owners are usually actively involvedin f rm s’m anagem ent,serving as d irecto rsand/orm anagers.These characteristics im p ly that fam ily f rm shave less severe agency con f icts between owners and managers,butmore severe agency con f icts between fam ily owners and non-fam ily m inority shareholders.Fam ily f rms are also afected by the tensions or con f icts among fam ilymembers.Theseagency con f ictsaf ect theoperations,fnancial reporting,and valuation o f family f rms.

The low-hanging fruitshavebeen picked in the last decade,although opportunities for further research still bound.First,itwou ld be of particu lar interest to explore the potentially“disguised”fam ily ef ects.Some phenomena docum ented in prior research m ight be due to fam ily f rm efects.For examp le,many studies have exam ined theefectof a dual classstructure.However,asmost dualclasssharesare owned by founding fam ily m em bers,is it possible to separate the dual class efect from the fam ily f rm efect?

Second,the found ing fam ily is treated as a b lack box and is assum ed to be hom ogeneous.As indicated by Bertrand and Schoar(2006),“m uch can be learned by investigating the‘fam ily’part of‘fam ily f rm s’.”Understanding the nexuswithin the fam ily and its inf uence on f rm operationswould move the literature forward. Such studieswou ld also be useful for fam ily owners in improving the governance,succession planning,and operations o f their f rms.

Third,more data are needed to p rovide additionalinsights.An interesting examp le along this line isa study by Bunkanwanicha et al.(2013).They compile a unique and comp rehensive data set of fam ily f rms in Thailand and fnd that fam ily f rms’stock prices increasewhen one fam ilymembermarriesamember from a prominent business or po litical fam ily.

Fourth,case studiesare quite lim ited in the fam ily f rm literature and m ight p rovide themost value-added in future research.W hile such studies arem ore dif cult to publish due to the lack of generalizability,the collective evidence from such studies can provide m uch needed insigh ts into the operation o f fam ily f rm s and p rovide guidance for fam ily ownerswho are concerned w ith governance,operation,and succession.

Finally,internationalstudiesarehelp ful in understanding how dif erent culturalnormsand legal in frastructuresaf ect the evo lvemento f fam ily f rms.M orck et al.(2011),M asu liset al.(2011),and Frankset al.(2012) are some recent examp les.Of course,both authors and readersmust bem ind ful of the heterogeneity across countries,which imposes challenges for research design and m ight af ect the interpretation of resu lts.

A li,A.,Chen,T.Y.,Radhakrishnan,S.,2007.Corporate disclosu res by fam ily f rms.J.A ccount.Econ.44,238–286.

Allen,F.,Qian,J.,Qian,M.,2005.Law,fnance,and econom ic grow th in China.J.Financ.Econ.77,57–116.

Almeida,H.,Wolfenzon,D.,2006.A theory of pyram idal ownership and fam ily business groups.J.Financ.61,2637–2681.

Anderson,R.C.,Duru,A.,Reeb,D.M.,2009.Founders,heirs,and corporate opacity in the United States.J.Financ.Econ.92,205–222.

Anderson,R.C.,Duru,A.,Reeb,D.M.,2012.Investm ent policy in fam ily controlled f rm s.J.Bank.Finance 36,1744–1758.

Anderson,R.C.,M ansi,S.,Reeb,D.M.,2003.Founding fam ily ownership and the agency cost of debt.J.Financ.Econ.68,263–285.

Anderson,R.C.,Reeb,D.M.,2003.Founding-fam ily ownership and f rm performance:evidence from the S&P 500.J.Finance 58,1301–1328.

Anderson,R.C.,Reeb,D.M.,2004.Board com position:balancing fam ily infuence in S&P 500 f rm s.Adm in.Sci.Quart.49,209–237.

Andres,C.,2008.Large shareholders and f rm performance:an empiricalexam ination of found ing-fam ily ownership.J.Corp.Finance14, 431–445.

Bennedsen,M.,N ielsen,K.M.,Pe′rez-Gonza′lez,F.,Wolfenzon,D.,2007.Inside the family f rm:the roleo f fam ilies in succession decisions and perform ance.Quart.J.Econ.122,647–691.

Bertrand,M.,Johnson,S.,Samphantharak,K.,Schoar,A.,2008.M ixing fam ily w ith business:a study of Thaibusiness groups and the fam iliesbehind them.J.Financ.Econ.88,466–498.

Bertrand,M.,Schoar,A.,2006.The role of fam ily in family f rms.J.Econ.Perspect.20,73–96.

Bunkanwanicha,P.,Fan,J.P.H.,W iwattanakantang,Y.,2013.The value of m arriage to fam ily f rms.J.Financ.Quan t.Anal. (fo rthcom ing).

Burkart,M.,Panunzi,F.,Shleifer,A.,2003.Family f rms.J.Finance 58,2167–2202.

Chen,S.,Chen,X.,Cheng,Q.,2008a.Do fam ily f rm s p rovidemore or less voluntary disclosure?J.Account.Res.46,499–536.

Chen,S.,Chen,X.,Cheng,Q.,Shevlin,T.,2010.A re fam ily f rm smore tax aggressive than non-fam ily f rm s?J.Financ.Econ.95,41–61.

Chen,H.,Chen,J.Z.,Lobo,G.J.,Wang,Y.,2011a.Efectsof auditquality on earningsmanagement and costof equity capital:evidence from China.Contemp.Account.Res.28,892–925.

Chen,X.,Cheng,Q.,Dai,Z.,2013a.Fam ily ownership and CEO turnovers.Contem p.Account.Res.30(3),1166–1190.

Chen,G.,Firth,M.,X in,Y.,Xu,L.,2008b.Control transfers,privatization,and corporate perform ance:ef ciency gains in China’s listed companies.J.Financ.Quant.Anal.43,161–190.

Chen,D.,Jin,X.,Liu,X.,2011b.Governm ent quality,social capital and endogenous determ inants of pyram id structure of fam ily business.China Ind.Econ.7,129–139(in Chinese).

Chen,D.,Li,S.,Lei,G.,2012a.Government governance,ownership structure and investment decision:evidence from Chinese fam ily listed f rms.Journal of Financial Research 3,124–137(in Chinese).

Chen,C.J.P.,Li,Z.,Su,X.,Sun,Z.,2011c.Rent-seeking incentives,corporate political connections,and the con tro l structure of private f rm s:Chinese evidence.J.Corp.Finance 17,229–243.

Chen,D.,Liang,S.,Zhu,P.,2012b.Relative performance evaluation and executive compensation:evidence from Chinese listed companies.China J.Account.Res.5,127–144.

Chen,C.,Sh i,H.,Xu,H.,2013.U nderw riter reputation,issuer ownership,and pre-IPO earningsmanagement:evidence from Ch ina. Financ.M anage.(fo rthcom ing).

Cheng,P.,Li,J.,Tong,W.H.S.,2008.What triggers top turnovers in management China?J.Contemp.Account.Econ.4,50–87.

Chua,J.H.,Chrisman,J.J.,Sharma,P.,1999.Defning the fam ily businessby behavior.Entrep.Theory Practice 23,19–39.

C laessens,S.,Fan,J.P.H.,Lang,L.H.P.,2000.The separation of ownership and control in East Asian corporations.J.Financ.Econ.58, 81–112.

Caprio,L.,Croci,E.,Giudice,A.D.,2011.Ownership structure,fam ily contro land acquisition decisions.J.Corp.Finance 17,1636–1657.

Cucculelli,M.,M icucci,G.,2008.Family succession and f rm performance:evidence from Italian family f rm s.J.Corp.Finance14,17–31.

Desai,M.,Dharmapala,D.,2006.Corporate tax avoidance and high-powered incentives.J.Financ.Econ.79,145–179.

Faccio,M.,Lang,L.H.P.,2002.The ultimate ownership of W estern Eu ropean corpo rations.J.Financ.Econ.65,365–395.

Fahlenbrach,R.,2009.Founder-CEOs,investment decisions,and stock market performance.J.Financ.Quant.Anal.44,439–466.

Firth,M.,Fung,P.M.Y.,Rui,O.M.,2006.Corporate performance and CEO com pensation in China.J.Corp.Finance 12,693–714.

Franks,J.R.,M ayer,C.,Volpin,P.F.,W agner,H.F.,2012.The life cycle of fam ily ownership:Internationalevidence.Rev.Financ.Stud. 25,1675–1712.

Ge,Y.,Qiu,J.,2007.Financial development,bank discrim ination and trade credit.J.Bank.Finance 31,513–530.

Han,Z.,Yang,S.,Shi,H.,2007.Study on the fnancing constraints and inef cient investments o f ultimate owners in private pyramid f rm s.Chin.J.M anage.Sci.15,143–148(in Chinese).

He,X.,Li,J.,Chen,L.,2010a.Fam ilym em ber com position and its efect on corporategovernance:An em pirical study in China.N ankai Bus.Rev.13,149–160(in Chinese).

He,X.,Li,X.,Lian,Y.,Zhang,Y.,2010b.Power deviation among family agentsand itsefecton f rm value:an em piricalstudy in China. China Ind.Econ.10,96–106(in Chinese).

He,X.,Lian,Y.,2009.Fam ily authority and fam ily f rm value:an em pirical study in China.Econ.Res.J.4,90–102(in Chinese).

Hu,N.,W ang,F.,W ang,P.,Yao,L.J.,Zhang,J.,2012.The im pactof u ltimate ownerships on audit fees:evidence from Chinese listed companies.Asia-Pacif c J.Account.Econ.19,352–373.

Huang,J.,Zhang,T.,2011.Theaccounting in formation quality in fam ily f rms:evidence from China.China Account.Finance Rev.13,1–62.

Jam es,H.,1999.Owner asm anager,extended horizons and the fam ily f rm.In t.J.Econ.Bus.6,41–56.

Jensen,M.,M eckling,W.,1976.The theory of the f rm:managerialbehavior,agency costs and ownership structure.J.Financ.Econ.3, 305–360.

Jiang,G.,Lee,C.M.C.,Yue,H.,2010.Tunneling through intercorporate loans:the Ch ina experience.J.Financ.Econ.98,1–20.

K ing,R.,Peng,W.Q.,2009.The efect o f industry characteristics on control longevity of founding-fam ily f rms.W orking paper.Hong Kong University o f Science and Technology.

Lei,Q.,Lin,B.,Wei,M.,2013.Types of agency cost,corporate governance and liquidity.J.Account.Public Policy 32,147–172.

Li,Z.,X in,X.,Yu,X.,2008.Financial developm ent,debt f nancing constraint and pyram idal structure.M anage.W orld 1,123–135(in Chinese).

Lian,Y.,He,X.,Zhang,Y.,2011.The allocation and efects of family authority:evidence from Chinese fam ily listed f rm s.M anage. W orld 11,105–117(in Chinese).

Liu,Q.,Li,Z.,Yao,Y.,2008.Investor protection,private benef ts of control and pyram id structure:a case study of the“G reen Cool”company.M anage.W o rld 12,139–148(in Chinese).

Liu,L.,Subramaniam,N.,2013.Governmentownership,audit f rm sizeand auditp ricing:evidence from China.J.Account.Public Policy 32,161–175.

Lu,Z.,Zhu,J.,Zhang,W.,2012.Bank discrim ination,holding bank ownership,and econom ic consequences:evidence from China.J. Bank.Finance 36,341–354.

M asulis,R.W.,Pham,P.K.,Zein,J.,2011.Fam ily business groups around theworld:fnancing advantages,controlmotivations,and organizational choices.Rev.Financ.Stud.24,3556–3600.

M au ry,B.,2006.Fam ily ownersh ip and f rm performance:em piricalevidence from W estern European corporations.J.Corp.Finance 12, 321–341.

M orck,R.,Sh leifer,A.,Vishny,R.,1988.M anagement ownership andmarket valuation:an empiricalanalysis.J.Financ.Econ.20,293–315.

M orck,R.,Yavuz,M.D.,Yeung,B.,2011.Banking system control,capitalallocation,and econom y performance.J.Financ.Econ.100, 264–283.

Peng,W.Q.,Wei,K.C.J.,Yang,Z.,2011.Tunneling or propping:evidence from connected transactions in China.J.Corp.Finance 17, 306–325.

Pe′rez-Gonza′lez,F.,2006.Inherited control and f rm performance.Am.Econ.Rev.96,1559–1588.

Shao,J.,Liu,Z.,2007.A re the internal cap ital m arkets o f group-holding companies ef cient?A case study of the H ongyi group corpo rations.M anage.W orld 6,114–121(in Chinese).

Shim,J.,Okamuro,H.,2011.Does ownership matter in mergers?A comparative study of the causes and consequences ofmergers by fam ily and non-family f rm s.J.Bank.Finance 35,193–203.

Shleifer,A.,Vishny,R.,1986.Large shareholders and co rporate con trol.J.Po lit.Econ.94,461–488.

Stein,J.,1988.Takeover threats and m anagerialm yopia.J.Polit.Econ.96,61–80.

Stein,J.,1989.Ef cient capitalmarkets,inef cient f rms:amodel ofmyopic corporate behavior.Quart.J.Econ.103,655–669.

Su,Q.,Zhu,W.,2003.Fam ily control and f rm value:evidence from Chinese listed com panies.Econ.Res.J.8,36–45(in Chinese).

Villalonga,B.,Am it,R.,2006.H ow do fam ily ownersh ip,control,and m anagem ent af ect f rm value?J.Financ.Econ.80,385–417.

Villalonga,B.,Am it,R.,2009.H ow are U.S.fam ily f rms controlled?Rev.Financ.Stud.22,3047–3091.

Villalonga,B.,Am it,R.,2010.Fam ily control of f rm s and industries.Financ.M anage.39,863–904.

Wang,D.,2006.Founding fam ily ownership and earnings quality.J.Account.Res.44,619–656.

W ang,L.,Yung,K.,2011.D o state enterprisesm anage earningsm ore than privately owned f rm s?The case of China.J.Bus.Finance A ccount.38,794–812.

Wei,Z.,Wu,S.,Li,C.,Chen,W.,2011.Fam ily contro l,institutionalenvironmentand cash dividend policy:evidence from China.China J.Account.Res.4,29–46.

Wu,W.,Wu,C.,Zhou,C.,W u,J.,2012.Politicalconnections,tax benef tsand f rm performance:evidence from China.J.Account.Public Po licy 31,277–300.

Xu,J.,Lv,C.,2011.Typesof CEOs in fam ily f rms and earningsquality:evidence from China.M anage.W orld 1,112–120(in Chinese).

Yang,S.,Su,K.,2009.U ltimate control,free cash fow constraintand corporateperformance:evidence from Chinesep rivately-contro lled listed companies.Accounting Research 4,78–86(in Chinese).

Ye,Y.,Liu,B.,H uang,L.,2007.U ltim ate control rights,cash f ow righ ts and f rm value:an em pirical study of Chinese listed companies based on recessive u ltimate control rights.Journal of M anagement Sciences in China 10,66–79(in Chinese).

Ying,Q.,Wang,L.,2013.Propping by controlling shareholders,wealth transfer and f rm performance:evidence from Chinese listed companies.China J.Account.Res.6,133–147.

Zhang,H.,Zhang,J.,Song,M.,2004.Theefect of separation between ownership and con tro lon f rm value:an em pirical study on private listed companies in China.China Econ.Quart.3,1–14(in Chinese).

E-mail address:qcheng@sm u.edu.sg

☆Thisarticle isbased on theauthor’s speech at the 2013China Journalof Accounting Research annual conferenceheld in Zhuhai,China, and isprepared fo r a special issueof CJAR.The authorw ishes to thank M inying Cheng for her excellent research assistanceand X ia Chen for help ful comments.

http://dx.doi.org/10.1016/j.cjar.2014.03.002

1755-3091/?2014 Production and hosting by Elsevier B.V.on behalf of China Journalof Accounting Research.Founded by Sun Yat-sen

U niversity and City University of H ong K ong.