Geographical relationships and CEO compensation contracts

Junli Yu,Wei Xu,Ping Zhng

aAntai College of Economics and Management,Shanghai Jiao Tong University,China

bBusiness School,Nanjing University,China

cAccounting School,Nanjing University of Finance and Economics,China

Geographical relationships and CEO compensation contracts

Junli Yua,Wei Xub,*,Ping Zhangc

aAntai College of Economics and Management,Shanghai Jiao Tong University,China

bBusiness School,Nanjing University,China

cAccounting School,Nanjing University of Finance and Economics,China

ARTICLE INFO

Article history:

Accepted 15 March 2017

Available online 25 April 2017

JEL classi fi cation:

G39

M49

Geographical relationships

Culture

CEO compensation

Pay-performance sensitivity

In this paper,we empirically analyze the ef f ects that the geographical relationships between chairman and CEO have on the latter’s compensation contracts, based on samples of listed A-share private f i rms from 2005 to 2014.We f i nd thatgeographicalrelationshipsarerelatedtolowerpay-performance sensitivity,and that the correlation mainly exists in poor performance periods, suggestingthatgeographicalrelationshipsweakentheef f ectivenessof compensation contracts.We also f i nd that geographical relationships can be substituted by external formal institutions.

?2017 Sun Yat-sen University.Production and hosting by Elsevier B.V.This is an open access article under the CC BY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/4.0/).

1.Introduction

CEO compensation contracts lie at the core of f i rm governance.Ef f ective contracts relieve the agency problems that stem from a separation between ownership and management(Jensen and Meckling,1976;Jensen and Murphy,1990).However,the compensation contract is not a perfect tool for situations involving information asymmetry and limited rationality.CEOs tend to take opportunistic actions in their pursuit of private benef i ts,and shareholders rarely know that it is happening.From a Western perspective,agency problems are thought to be solved by external institutions.Yet while perfect property protection systems and legal mechanisms can improve contract enforcement(Williamson,1985),emerging and transitioning environments suchas China cannot provide complete external institutional control,which can lead to inadequate property protection and legal penalty mechanisms(Chalos and O’Connor,2004).As such,some studies treat social relationships as an alternative mechanism,as they provide motivation and reduce information asymmetry.The possible positive ef f ects of social relationships are referred to as the substitution hypothesis(Wang,2005; Zhao and Lv,2015).Other research f i nds that social relationships reduce boards’supervision ef f ectiveness. The possible negative ef f ects of social relationships are referred to as the weakening hypothesis(Core et al., 1999;Hwang and Kim,2009).In this paper,we test which of the aforementioned hypotheses is deterministic.

This paper focuses on representative social and geographical relationships to test their ef f ects on compensation contracts.Geographical relationships are generated by one’s proximity to another.From a sociological perspective,Chinese society is constructed through classif i cations and relationships.Classif i cations are the most fundamental informal social construct upon which relationships are built(Pan,2000).‘‘Countryman”is a common classif i cation.As Fei(1948)says,Chinese social relationships form concentric circles,with home in the center.In addition to genetic relationships,geographical relationships and clanship are also important in China(Ma,2008).In contrast to Zhao and Lv(2015),who focus on genetic relationships,we doubt the universality of altruism in genetic relationships(Wang et al.,2014;Wei and Chen,2015).Compared with academic and colleague relationships,for which there is relatively little information,the ef f ects of geographical relationships on contracts have been recorded(Cai et al.,2008)and are common in practice(Lu and Hu, 2014).Some studies fi nd that geographical relationships in fl uence economic behavior.For instance,informal fi nancial organizations in Wenzhou built a credit network using geographical relationships to ultimately lower credit rates(Guo and Liu,2002).However,geographical relationships can also increase fi rm risk(Lu and Hu, 2014)and reduce the ef f ectiveness of internal control(Yu et al.,2017).Thus,the ef f ects of social relationships on compensation contracts deserve to be explored.

We empirically analyze the ef f ects of geographical relationships on the ef f ectiveness of compensation contracts(compensation-performance sensitivity,also known as pay-performance sensitivity).We show that the sensitivity is lower in f i rms with geographical relationships.To distinguish between the substitution hypotheses and weakening hypotheses,we test the compensation stickiness and performance.According to the weakening hypothesis,geographical relationships can increase compensation stickiness.If geographical relationships act as umbrellas for CEOs’self-serving behavior,then compensation-performance sensitivity should only decrease in declining performance periods.According to the substitution hypothesis,CEO supervision does not rely on compensation contracts,and as such the reduction in compensation-performance sensitivity should be bi-directional.Our empirical result shows that the weakening ef f ects of geographical relationships are only signif i cant in declining periods,which supports the weakening hypothesis.We also test for the crosssectional dif f erences in institutions and f i nd that the weakening ef f ect is only signif i cant in poor external institutions,indicating that governing by relationships may not be as powerful as doing so by institutional constraints.The abovementioned results remain robust after eliminating alternative mechanisms and endogeneity.

Our research makes several contributions to the literature.First,it supplements the relevant work being conducted in emerging and transitional markets.Studies on the ef f ects of social relationships on compensation are largely based on developed markets(Core et al.,1999;Hwang and Kim,2009;Fracassi and Tate,2012). We also distinguish between two possible hypotheses and show how geographical relationships weaken CEO supervision.Second,unlike the research that focuses on clanship(Zhao and Lv,2015),we explore geographical relationships to achieve a more universal conclusion.Finally,we emphasize the ef f ectiveness of formal institutions to help guide regulators.

The remainder of this paper is organized as follows.The second section features a literature review,the third presents our theory and hypotheses,the fourth covers the research design,the f i fth shares the empirical results and the f i nal section concludes the paper.

2.Literature review

2.1.Board of directors and manager compensation

Early research is characterized by its focus on the ef f ect of board structure on compensation contract ef f ectiveness(Cyert et al.,1997),with board size as a deterministic factor of CEO variable compensation.When theCEOs also serve as chairpersons,their compensation is typically 20-40%higher than average.Moreover,CEO compensation is negatively related to board shareholdings,and Brickley et al.(1997)conf i rm the positive relationship between duality and higher payment.Core et al.(1999)measure board ef f ectiveness using an eightstructure index with items such as CEO-director separation and board size.They f i nd that board ef f ectiveness is negatively related to CEO payment.Cordeiro and Veliyath(2003)show that the ratio of independent directors is positively related to CEO cash payment.Chhaochharia and Grinstein(2009)reveal that following the Sarbanes-Oxley Act(2002),increases in independent director ratios lead to reduced CEO compensation.

Some Western studies explore the ef f ect of boards on managerial compensation from a social network perspective.They suppose that every economic organization and person exists within a social network,such that all decisions are inf l uenced by others and a f i rm’s governance is inf l uenced by its network.Core et al.(1999) show that in f i rms where the directors have more extensive external social relationships,CEO compensation is excessive,and ultimately damages managerial supervision.Larcker et al.(2005)conf i rm that in f i rms with social directors,CEO compensation is signif i cantly higher,but future business performance is poorer.Some studies discuss how the private relationships between directors and CEOs af f ect managerial compensation. Hwang and Kim(2009)measure the private relationships between directors and CEOs based on army service experience,graduation from the same college,shared hometowns and shared major acquaintances.In f i rms where the directors have no private relationships with the CEOs,the latter’s compensation is lower(reduced by$3.3 million on average).In f i rms where the directors and CEOs share private relationships,the latter’s compensation-and layof f-performance sensitivity are lower.Engelberg et al.(2013)f i nd that the presence of private relationships between CEOs and directors is related to higher CEO payment(increased by $17,000 on average),based on 2700 CEOs of large listed f i rms from 2000 to 2007.Faleye et al.(2011)conf i rm that private relationships between directors and managers increase managers’compensation.Armstrong et al. (2006)show that in f i rms where directors and managers share private relationships,compensation is higher, but future f i rm performance is poor.

2.2.Geographical relationships and f i rm behavior

In China,relationship culture is deeply rooted in the public psyche and thus tends to dominate behavior. Lin and Sun(2005)f i nd that in situations lacking formal institution,geographical relationships can reduce search and trust costs between organizations and improve informal f i nance development.Guo and Liu (2002)f i nd that geographical relationships provide informal f i nancial institutions in Wenzhou with adequate information about the operating conditions,backgrounds and credit of mid-and small-sized f i rms,which reduces the credit risk.This also relieves the small-to-medium enterprise f i nancing problem.Close relationships between shareholders and managers also largely reduce supervision costs and the likelihood of negative behavior.Liu and Chen(2012)also show that informal f i nancial institutions in Wenzhou use social relationships to reduce credit risk.Industry clusters formed from geographical relationships benef i t from the continuous interaction between f i rms and economy of scale(Li,2008).However,there has been little research on internal geographical relationships’ef f ects on internal behavior.Lu and Hu(2014)discuss the geographical relationships between directors and CEOs at the fi rm risk level and fi nd that fi rms with them su ff er higher fi nancial risk and more takeovers.

Some researchers are aware of the e ff ects that geographical relationships have on compensation contracts. Some studies show that such relationships have wide-reaching and profound e ff ects when external institutions are inadequate.However,there is not enough research in the Chinese context,which makes our research both valuable and necessary.

3.Theory and hypotheses

3.1.Geographical relationships and the ef f ectiveness of compensation incentives

Research on geographical relationships and how they inf l uence the ef f ectiveness of compensation incentives can be conducted from two sides:supervision and incentive.

From the supervision side,social relationships weaken contract enforcement.Social norms replace formal regulations in guiding behavior(Uzzi,1996).Social relationships can reduce directors’independence and CEOs’sense of responsibility to maximize shareholders’benef i ts.For instance,Hwang and Kim (2009)f i nd that in f i rms where the directors and CEOs have private relationships,the payment-and layof f-performance sensitivities are lower.Kramarz and Thesmar(2013)also f i nd that CEOs with private relationships are less likely to be replaced for poor performance.Lu and Hu(2014)show that in f i rms with geographical relationships,director supervision is relaxed,increasing f i rm risk.These results suggest that social relationships can reduce directors’supervision incentives,which may reduce the ef f ectiveness of compensation contracts.

From the incentive side,both compensation contracts and social relationships can provide incentives for CEOs to reduce agency costs.Becker(1974,1976)proposes altruism incentives.When family members serve as CEOs,the clanship can provide incentives to reduce dependence on compensation contracts(Zhao and Lv,2015).Social relationships can also protect reputations.Pan(2000)shows that the social relationship between two individuals is also an indication of individual existence in certain groups.For example, in a geographical relationship,both parties are from the same hometown group,and any opportunistic behavior may damage a member’s reputation in the group(Standif i rd and Marshall,2000).Finally,social relationships reduce information asymmetry(Adams and Ferreira,2007),as directors’are less likely to demand accounting numbers by which to judge CEO performance(Yang et al.,2014).Thus,from the incentive side,social relationships reduce the demand for compensation contracts,along with their ef f ectiveness.

Both the weakening hypothesis and the substitution hypothesis suggest that social relationships reduce the ef f ectiveness of compensation contracts.Hence,we propose the f i rst hypothesis:

H1.Ceterisparibus,compensation-performancesensitivityislowerinf i rmswithgeographical relationships.

3.2.Weakening or substitution?

Both the weakening hypothesis and the substitution hypothesis refer to lower compensation-perfor mance sensitivity,but the ef f ects of geographical relationships on compensation contract ef f ectiveness dif f er between them.Compensation stickiness can be a good entry point to distinguish between the two hypotheses.Under the supervision hypothesis,geographical relationships reduce director supervision,resulting in lower compensation-performance sensitivity during poor performance periods because the inef f ectiveness makes CEOs more likely to behave opportunistically.When f i rm performance is poor,CEOs tend to protect their own benef i ts f i rst,keeping compensation high despite the drop in performance.When f i rm performance is good,CEOs may pursue higher compensation,creating a positive relationship.Under the substitution hypothesis,geographical relationships provide CEOs with incentives to reduce their selfserving behavior.Then,the payment-performance sensitivity is low,regardless of f i rm performance. Meanwhile,if geographical relationships become the protection system in a hometown group,the reputation-pursuing incentive drives CEOs to decrease their self-serving behavior.Finally,as Ouchi (1980)and Adams and Ferreira(2007)mention,geographical relationships can reduce information asymmetry and make performance judgments independent of f i nancial numbers,reducing payment-performance sensitivity independent of f i rm performance.

Accordingly,we propose two alternative hypotheses:

H2a.Ceteris paribus,the negative connection between geographical relationship and compensation-perfor mance sensitivity only exists in poor performance periods.

H2b.Ceteris paribus,the negative connection between geographical relationship and compensation-perfor mance sensitivity exists both in good and poor performance periods.

4.Research design

4.1.Sample and data sources

In terms of CEO compensation,we focus on CEOs’salaries because stock-based incentives are not common at present,and salaries are still the primary composition of executive compensation(Xin et al.,2007; Fang,2009).

We use the private listed companies that issued A-shares from 2005 to 2014 as the initial sample.We exclude(1)companies with incomplete board chairman and CEO information;(2)companies with a dual chairman/CEO;(3)f i nancial and insurance f i rms;(4)special treatment(ST)or particular transfer(PT)f i rms; (5)companies where the chairman and the CEO are the same person;and(6)companies for which f i nancial data are missing.Our f i nal sample includes 4017 f i rm-year observations for the 2005-2014 period.To avoid the ef f ects of extreme values,all of the continuous variables are winsorized at both the top and bottom percentiles.

Using executive information(name list,tenure,etc.)from the China Stock Market and Accounting Research(CSMAR)listed company database,we manually collect personal information from prospectuses, annual reports,company Websites,the SINA Finance website and other public channels.To characterize the strength of the geographical relationship,we quantify the distance(proximity)between the chairman and the CEO using the latitude and longitude of his or her birth locationusing Google Earth.We then obtain external governance circumstance data from the Fan et al.(2011)marketization index.Other f i nancial data are from the CSMAR databases.

4.2.Empirical model and variable constructions

Following the compensation model developed by Xin and Tan(2009)and Fang(2009),we estimate the following regression model:

The dependent variable,Comp,refers to the CEO’s compensation.The independent variables refer to the geographical relationships,and we proxy for these connections using two variables:Province and Geodist. Province is an indicator variable that is equal to 1 if the birthplaces of the board chairman and the CEO are in the same province,and 0 otherwise.Geodist is the negative distance between the birthplaces of the board chairman and the CEO.All of the variables are as def i ned in Table 1.

5.Results

5.1.Descriptive statistics

Table 2 presents our descriptive data.The mean and standard deviation(SD)of CEO compensation (Comp)are 12.63 and 0.86,respectively.The mean and SD of geographical relationship(Province)are 0.46 and 0.50,respectively.Geodist,another indicator of geographical relationships,has a minimum value of 11.89,a maximum value of 0,a mean of 0.54 and an SD of 0.71,indicating a larger dif f erence in the geographical relationship between board chairpersons and CEOs.In addition,the mean of return on assets(ROA)is 0.04,with a median ROA of 0.03,suggesting that the majority of the f i rms are prof i table.

5.2.Empirical results

Based on Model 1,we use a regression and a two-way cluster(in the f i rm and two standard error dimensions)to test our hypotheses,and the results are as follows.

Table 1Variable def i nitions.

Table 2Descriptive statistics.

Table 3 presents the regression results of the e ff ects of geographical relationships on compensation-perfor mance sensitivity.The second and fourth columns report the results after controlling for correlated variables. As the results show,the coefficient on the interaction terms between Province and ROA is negative and signi fi cant at the 5%level,regardless of controls or other correlated variables.Similarly,we fi nd that the coeffi cient on the interaction terms between Geodist and ROA is negative and signi fi cant at the 1%level.Taken together,these results provide some support for H1,that geographical relationships decrease compensationperformance sensitivity,consistent with the previous research.

We further fi nd that the adverse e ff ects of geographical relationships on compensation-performance sensitivity are caused by monitoring decreases in the e ff ectiveness of the board,or the alternative role played by geographical relationships in relation to compensation contracts.Table 4 reports the results.

Columns 1 and 2 of Table 4 present the e ff ects that geographical relationships have on compensation-per formance sensitivity when performance is good(higher than last year).As the results show,the coefficients onthe interaction terms between geographical relationships and ROA are also negative,but not signif i cant.In contrast,Columns 3 and 4 present the ef f ects of geographical relationships on compensation-performance sensitivity when performance is poor(lower than last year).As the results show,the coefficient on the interaction terms between geographical relationships and ROA is negative and signif i cant(p＜0.01 and p＜0.05, respectively).

Table 3Geographical relationships and CEO compensation.

Thus,the negative ef f ects that geographical relationships have on compensation-performance sensitivity only appear in cases of poor performance,supporting H2a.We f i nd that in companies with geographical relationships where the directors’monitoring ef f ectiveness has been reduced,there is less performance-sensitive CEO compensation.

Table 4Performance,geographical relationships and CEO compensation.

5.3.Further analysis and robustness checks

5.3.1.The substitution ef f ect between formal institutions and geographical relationships

Our main analysis shows that geographical relationships have complex ef f ects on CEO compensation.Nevertheless,it is important to determine whether such an informal institution is necessary.Theoretically,if the external formal institutions are perfect,there is no need to resort to the informal institution.In this section,we further test the substitution ef f ect between formal institutions and geographical relationships,as it inf l uences the ef f ects that geographical relationships have on CEO compensation under dif f erent institutional environments.As Table 5 shows,geographical relationships only signif i cantly af f ect CEO compensation in cases ofpoor institutional environments,suggesting that formal institutions have a substitution ef f ect on geographical relationships.

Table 5Geographical relationships and CEO compensation under dif f erent institutional environments.

5.3.2.The ef f ects of other types of social connections

A number of recent papers emphasize how other types of social connections(relatives,colleagues,etc.) between executives or directors af f ect CEO compensation,and as such may af f ect our conclusion reliability. To account for such a possibility,in this section,we perform two tests.First,following Zhao and Lv (2015),we delete the samples with related connections between board Chairman and CEO.Second,we control for colleague connections by creating a new variable,Inside:an indicator variable that is equal to 1 if the CEOcomes from inside the company,and 0 otherwise.As Table 6 shows,after excluding other types of social connections,the results are similar to those in Table 3.

Table 6Control for other types of social connections.

5.3.3.Endogeneity

A concern with these regressions is endogeneity.To grant our analysis more generality,we extend our analysis by testing whether our results may be driven by omitted variables.We also apply other sample specif i cations-board chairman or CEO turnover-that lead to changes in geographical relationships.Table 7 presents the results of these tests,which address the concern that our results may be driven by an unobserved characteristic.

Table 7Changes in geographical relationships.

6.Conclusion

There is growing interest in CEO compensation,both in practice and in the academic literature.Most studies focus on formal institutions,yet due to the traditional social structure and the imperfection of external institutional environments,informal institutions such as GuanXi have had widespread inf l uence in corporate governance practices.In this study,we explore whether social connections among board chairpersons and CEOs af f ect CEO compensation.In particular,we inquire whether geographical relationships inf l uence CEO compensation.Using data on private listed companies that issued A-shares from 2005 to 2014,we f i nd that f i rms with geographical relationships typically exhibit lower levels of compensation-performance sensitivity.Further testing shows that the ef f ects of geographical relationships only weaken compensation-perfor mance sensitivity signif i cantly in cases of poor performance.These f i ndings suggest that geographical relation-ships reduce boards’monitoring ef f ectiveness.We also f i nd that geographical relationships only have signif icant ef f ects on CEO compensation in cases of poor institutional environments and SOE,suggesting that formal institutions have a substitution ef f ect on geographical relationships.Controlling for other types of social connection and endogeneity,our results are also statistically and economically signif i cant.

We analyze the ef f ects of geographical relationships on CEO compensation.In doing so,our work complements the literature by adding a new dimension to our understanding of the factors that af f ect CEO compensation.Our results highlight the inf l uence that informal institutions can have on f i rm governance,and the importance of improving external formal institutions.

Acknowledgments

The authors are grateful to Prof.Donghua Chen,Qingquan Xin and the anonymous referees for their helpful comments and suggestions.The view is of the authors,as of course are any errors.The authors gratefully acknowledge f i nancial support from China Postdoctoral Science Foundation(2016M601606)and Humanity and Social Science Youth foundation of Ministry of Education of China(15YJC790137).

Adams,R.B.,Ferreira,D.,2007.A theory of friendly boards.J.Finance 62(62),217-250.

Armstrong,C.S.,Jagolinzer,A.D.,Larcker,D.F.,2006.Timing of Employee Stock Option Exercises and the Valuation of Stock Option Expense.Stanford University working paper.

Becker,G.S.,1974.A theory of social interactions.J.Polit.Econ.82(6),1063-1093.

Becker,G.S.,1976.Altruism,egoism,and genetic f i tness:Economics and sociobiology.J.Econ.Lit.14(3),817-826.

Brickley,J.A.,Coles,J.L.,Jarrell,G.,1997.Leadership structure:separating the CEO and chairman of the board.J.Corpor.Finance 3 (3),189-220.

Cai,H.H.,Zhou,L.A.,Wu,Y.Y.,2008.The clan rule,the faith of merchants and the management of merchant organizations:a comparative study on merchants of Anhui province and Shanxi province in the period of the Ming and Qing dynasty.Manage.World 8,87-99(in Chinese).

Chalos,P.,O’Connor,N.G.,2004.Determinants of the use of various control mechanisms in US-Chinese joint ventures.Acc.Organ.Soc. 29(7),591-608.

Chhaochharia,V.,Grinstein,Y.,2009.CEO compensation and board structure.J.Finance 64(1),231-261.

Cordeiro,J.J.,Veliyath,R.,2003.Beyond pay for performance:a panel study of the determinants of CEO compensation.Am.Busin.Rev. 21(1),56.

Core,J.E.,Holthausen,R.W.,Larcker,D.F.,1999.Corporate governance,chief executive officer compensation,and f i rm performance.J. Financ.Econ.51(3),371-406.

Cyert,R.,Kang,S.,Kumar,P.,et al.,1997.Corporate Governance and the Level of CEO Compensation.Working Paper,Carnegie Mellon University.

Engelberg,J.,Gao,P.,Parsons,C.A.,2013.The price of a CEO’s rolodex.Rev.Financ.Stud.26(1),79-114.

Faleye,O.,Hoitash,R.,Hoitash,U.,2011.The costs of intense board monitoring.J.Financ.Econ.101(1),160-181.

Fan,G.,Wang,X.L.,Zhu,H.P.,2011.NERI Index of Marketization of China’s Provinces 2011 Report.Economic Science Press,Beijing (in Chinese).

Fang,J.X.,2009.Is top management compensation of public companies sticky?Econ.Res.J.3,110-124(in Chinese).

Fei,X.T.,1948.Fei Xiaotong Collection.QunYan Press(in Chinese).

Fracassi,C.,Tate,G.,2012.External networking and internal f i rm governance.J.Finance 67(1),153-194.

Guo,B.,Liu,M.L.,2002.Private f i nance and small and medium-sized enterprises development:empirical evidence from Wenzhou region. Econ.Res.J.10,351-382(in Chinese).

Hwang,B.H.,Kim,S.,2009.It pays to have friends.J.Financ.Econ.93(1),138-158.

Jensen,M.C.,Meckling,W.H.,1976.Theory of the f i rm:managerial behavior,agency costs and ownership structure.J.Financ.Econ.3 (4),305-360.

Jensen,M.C.,Murphy,K.J.,1990.Performance pay and top-management incentives.J.Polit.Econ.,225-264

Kramarz,F.,Thesmar,D.,2013.Social networks in the boardroom.J.Europ.Econ.Assoc.11(4),780-807.

Larcker,D.F.,Richardson,S.A.,Seary,A.,2005.Back Door Links between Directors and Executive Compensation.Working Paper.

Li,S.L.,2008.A study of geo-relation and industrial clusters cooperation.Acad.Res.6,80-84.

Lin,Y.F.,Sun,X.F.,2005.Information,informal f i nance and SME f i nancing.Econ.Res.J.7,35-44(in Chinese).

Liu,X.C.,Chen,L.H.,2012.Non-interest terms,occupation-based social network and interlinked contacts against credit risks:an empirical study on informal credit in Wenzhou.Finance Trade Res.5,104-111(in Chinese).

Lu,Y.,Hu,J.Y.,2014.The impact of hometown connectedness between the CEO and board of directors on corporate risk.Manage. World 3,131-138(in Chinese).

Ma,G.Q.,2008.Relationships in categories:the foundations of a civil society.J.Literat.,History Philos.4,44-55(in Chinese).

Ouchi,W.G.,1980.Markets,bureaucracies,and clans.Adm.Sci.Quart.25(1),129-141.

Pan,G.,2000.Pan Guangdan Collection.Peking University Press.

Standif i rd,S.S.,Marshall,R.S.,2000.The transaction cost advantage of guanxi-based business practices.J.World Bus.35(1),21-42.

Uzzi,B.,1996.The sources and consequences of embeddedness for the economic performance of organizations:The network ef f ect.Am. Sociol.Rev.61(4),674-698.

Wang,Y.Q.,2005.Reputation,commitment and organizational form:The comparative institutional analysis.Shanghai People’s Press, Shanghai(in Chinese).

Wang,M.L.,Xu,M.N.,Wang,H.,2014.Does altruistic behavior reduce the agency cost?An empirical study based on kin altruistic behavior in family business.Econ.Res.J.4,144-157(in Chinese).

Wei,C.Y.,Chen,L.,2015.Altruistic behavior of CEO turnover in family f i rm:A perspective on asset impairment.Manage.World 3,137-150(in Chinese).

Williamson,O.E.,1985.The Economic Institutions of Capitalism.Simon and Schuster.

Xin,Q.,Lin,B.,Wang,Y.,2007.Government control,executive compensation and capital investment.Econ.Res.J.8,110-122(in Chinese).

Xin,Q.,Tan,W.,2009.Market-oriented reform,f i rm performance and executive compensation in Chinese state-owned enterprises.Econ. Res.J.11,68-81(in Chinese).

Yang,Y.,Pan,F.,Zhang,C.,2014.Enterprises’performance evaluation under the perspective of dif f erential mode of association. Account.Res.10,66-73(in Chinese).

Yu,J.L.,Xin,J.,Liang,S.K.,2017.Does the geographical relationship between the chairman and the CEO af f ect internal control quality? China J.Account.Stud.,forthcoming

Zhao,Y.Y.,Lv,C.J.,2015.Family or money?The impact of family relationship on compensation contract in family f i rms.Account.Res. 8,32-40(in Chinese).

30 September 2015

*Corresponding author.

E-mail addresses:jlyuchina@gmail.com(J.Yu),will-nap@163.com(W.Xu),zhangp_9892@163.com(P.Zhang).

http://dx.doi.org/10.1016/j.cjar.2017.03.001

1755-3091/?2017 Sun Yat-sen University.Production and hosting by Elsevier B.V.

This is an open access article under the CC BY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/4.0/).