Political connections,mediamonitoring and long-term loans☆

Dem ing Yang,Zhengfei Lu,Danglun Luo

aM anagement School,Jinan University,China

bGuanghua School of M anagement,Peking University,ChinacLingnan College,Sun Yat-Sen University,China

Political connections,mediamonitoring and long-term loans☆

Dem ing Yanga,*,Zhengfei Lub,Danglun Luoc

aM anagement School,Jinan University,China

bGuanghua School of M anagement,Peking University,ChinacLingnan College,Sun Yat-Sen University,China

A R T IC L E IN F O

Article history:

Received 14 September 2012 Accep ted 12 August 2014

Availab le online 17 Sep tember 2014

JEL classif cation:

G 34

D 21

L51

Po litical connections

W e analyze data on Chinese non-state-listed f rm s and f nd that it is easier for f rms w ith political connections to obtain long-term loans w ith extended debtmaturities than it is for f rmswithout po litical connections.Our investigation indicates that this phenomenon is signif cantly less common w ith increased media monitoring.Houston et al.(2011)f nd strong evidence that the state ownership o fmedia is associated w ith higher levels o f bank corruption in China,but our study shows that,to a certain extent,mediamonitoring can curb corruption.

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

1.Introduction

Som e researchers have found that non-state-owned f rm sw ith po litical connections can secu re preferential access to fnancing and tax b reaks(e.g.Johnson and M itton,2003;C laessens et al.,2008;Liet al.,2008;Luo and Zhen,2008;W ang and Wang,2013;Yu and Pan,2008).Yet there are also disadvantages to f rms settingup po litical connections.Som e em p irical studies have shown that po litical connectionsm ay im p rove or reduce f rm value and performance(Liu et al.,2010).For exam ple,Fisman(2001),Johnson and M itton(2003),and Faccio and Parsley(2009)fnd that political connections can improve f rm performance and value.However, other studies come to the opposite conclusion,such as Fan et al.(2007),who f nd that the accounting and market performance o f f rmsw ith political connectionsare signif cantly lower than those o f other f rmsafter initialpub lic of erings(IPOs),due to a lack ofmanagerial capabilities.Thus,further study is clearly needed to determ inewhy these empirical results diverge and to identify the internalmechanismsof po litical connections that afect f rms.

Thereareat least two theories that can exp lain theexistenceandmechanismsof political connections.First, repu tation theory em phasizes the im portance of relationships,using a repu tation enhancem en t argum ent that suggests that the po litical connections of f rm execu tives serve as an alternative channel fo r establishing f rm repu tation when quality d isclosu re is absent(Sun et al.,2005;Yu and Pan,2008).Second,ren t-seeking theory also explainspolitical connectionsby arguing that f rmsuse them to engage in activities that inf uence the government’sapprovaldecisionsand governmentof cials then show partiality to f rmswhose executives prom ise these of cials personal favors.In otherwords,the app rovaldecision process is inf uenced by of cials desire to seek rents(e.g.,Fan et al.,2007).

Which theorymoreefectively exp lains theexistenceandmechanismso f po liticalconnections in theChinese stock market?In this study,we attempt to answer this question from amediamonitoring perspective.Based on the extant research,we investigate how m ediamonitoring af ects the relationship between po litical connections and long-term loans.W e anticipate that f rm sw ith po litical connections can more easily obtain longterm loansw ith extended debtm aturities.On the su rface,the phenom enon can be exp lained by both theo ries, so determ ining which ism o re signif cant requires deeper study.If the reputation theo ry ism ore signif cant, then we would expect the relationship between po litical connections and long-term loans to be rein forced by imp roved media monitoring,because f rms’reputations can be strengthened by media exposure.If the rent-seeking theory ismore signif cant,we expect the relationship to be reduced by imp rovedmediamonitoring becausemost rent-seeking behavior is un law ful and irrational,and thus the ro le of po litical connections shou ld beweakened bymedia exposure.

Using data for Chinese non-state-owned listed f rms from 2006 to 2012,we fnd that f rms with po litical connections have easier access to long-term loans w ith extended debtmaturities.This result suggests that political connections facilitate f rms’f nancing.We also f nd that the relationship between political connections and long-term loans is signif can tly reduced w ith im proved m ed ia m onito ring.Overall,ou r f nd ings support the rent-seek ing theory.

This study con tribu tes to the literatu re in several ways.First,previous stud ies have no t d istinguished between these two theories,such that research based on one cannotexclude theother.Thisstudy diferentiates between these two theoriesand analyzeswhich isbetter suited to theChinesestockmarket.Second,the ro leo f themedia in China ishighly suspect due to a higher degree of government intervention.Houston et al.(2011) note a strong correlation between state-fundedmedia and banking corrup tion.According to their study,it is dif cult for the Chinesemedia to p lay a role in supp ressing corruption,but we observe that it can p lay an important ro le in inhibiting rent-seeking,which can be explained by the market-oriented media reform o f recent years.Our empiricalevidence thereforep rovidessome support for the roleof themedia in China,which subsequent studies can further investigate.

2.Institutional background,theoretical analysis and hypothesis development

2.1.Institutional background

Due to ideological factorsand compared with state-owned f rms,non-state-owned f rms faceunfairmarket conditions,also known as“tilted playing feld”p rob lem s.The constraints on non-state-owned f rm developm entm ain ly include legalobstacles,governm ent intervention or adm inistration and fnancing dif cu lties(Bai et al.,2003).Another problem for non-state-owned listed f rm s is related to stock m arket developm ent.The Chinese stock m arket is a burgeoning traditionalp lanned econom y and socialistm arket econom ic system pu rsuing state-owned enterprise(SOE)reform.The central and local governm ents have rapid ly recognized thestock m arket’s potential to facilitate SOE reform.In the early 1990s,the stock m arket’s SOE reform bu rden was the result of the SOEs general inef ciency,which led to the stock-issuing system and the comp lication o f SOE reforms.This indicates that SOEs inevitab ly dom inate the Chinese stock market.A lthough non-SOEs can list through mergersand acquisitions(M&A)and management buyout(M BO)markets,SOEs still dominate thestockmarket because theM&A and M BOmarketsare largely contro lled by governmentsand SOEs. Thus,themajority o f Chinese listed f rms are ultimately under governm ent contro l(Liu et al.,2003).Hence, non-state-owned listed f rms naturally face tilted p laying feld p rob lems such as those listed below.

First,there are several legal and adm inistrative barriers for non-state-owned listed f rms.According to Chen et al.(2008),the p roportion of non-state-owned listed f rms is less than 20%in high-barrier industries such asautom otive,m arine and road and rail transportation—far below the propo rtion of SOEs.The propo rtion o f non-state-owned listed f rm s in m oderate-and low-barrier industries is abou t 30–50%and/or exceeds that of SOEs,respectively.These resu lts ref ect the ubiquity o f barriers,such as legal and adm inistrative,fo r non-state-owned listed f rms in high-barrier industries.

Second,non-state-owned listed f rms in China are developing serious fnancing dif culties in that they are unable to gain adequate external fnancing in the stockmarketor from thebanking system,the latter ofwhich exhibits credit discrim ination.Banks,especially those that are state-owned,prefer to provide loans to SOEs rather than to non-state-owned listed f rms.In the past 10 years,while non-state-owned f rms contribution to Chinese GDP exceeded 70%,they obtained less than 20%of the bank loans(Lu and Yao,2004).On the stock market,in the early 1990s,non-state-owned f rms could not easily achieve listing directly through IPOs due to the stock market’s burden in serving the SOE reform.Given the reform,some non-state-owned f rms had to initially list through the takeoverm arket o r buy shell resou rces1Shell resources refer to listed f rm s that are sm all in size or exhibit poor perform ance or operational dif culties.from SOEs,but the quality of such resources was typically low because SOEs and governm ents preferred no t to sell high-quality shell resources (W ang et al.,2001).These details explain why non-state-owned listed f rms pro f tability is generally not as good as thato f SOEs,despite the former’ssuperior system design,corporategovernmentand ownership structure.Gaining a low-quality shell doesnot solve non-state-owned listed f rms f nancing p rob lem s,and neither does equity ref nancing,according to Chen et al.(2008),due to low benef ts and government intervention.

Given non-state-owned listed f rms tilted playing f eld p rob lem s,they should use the fo llow ingmeasures to avoid constraints.First,they should seek political connections to avoid the negative ef ects of laws and governmental regu lations(Fan et al.,2007).Second,they shou ld solve their f nancing dif culties through political connections and the leakage efect to seek informal fnancial support(Lu and Yao,2004;Yu and Pan,2008). 2.2.Theoretical analysis and hypothesis development

The two theories considered in this study—repu tation and ren t-seeking—are used to analyze the relationship between political connections and long-term loans.

Based on reputation theory,Sun et al.(2005)suggest that f rm s’political connectionsare importantaspects o f their reputations that are likely to p romp t government intervention.M oreover,while itmay help f rms avoid troub le,government intervention is not always to a f rm’s benef t,nor does it line up w ith the f rm’s socialgoalsor o f cial interests(Shleifer and Vishny,1994).La Porta et al.(2002)demonstrate that f rmsw ith political connections easily gain governm ent subsidies when they are not facing dif cu lties in their businessoperations.

Against the background o f the pub lic p roperty system,political connections af ect f rms long-term loans and debtmaturities in the fo llowingways.First,f rmswith po litical connections can easily acquire long-term loans because they can use subsidies to reduce the possibility of defau lt.Likew ise,banks prefer to give them loans because the possibility of default is relatively low.

Second,f rmsw ith political connections can directly in fuence bank lending decisionsand gain more longterm loans to avoid the uncertainty inherent in the rotation of of cials(Sun et al.,2005).Thus,our f rst hypothesis is:H ypothesis 1.Non-state-owned listed f rm sw ith po litical connectionsarem ore likely to gain long-term loans and have longer debtmaturities than non-state-owned listed f rmsw ithout po litical connections.

Sim ilarly,the rent-seeking theory also supports H 1,but un like the reputation theory,it stresses thatmanagers shou ld estab lish and use political connections to gain excess pro f ts,which is comm only regarded as a non-productive activity and asunfair socialbehavior.Rent-seek ing behavior leads to governm ent intervention and non-state-owned listed f rm sw ith political connections can gain m o re long-term loans through their ren tseeking for governments or banks,especially state-owned banks.

A lthough H 1 is supported by both theories,the reputation theory emphasizes political connections as a positive reaction tomarket and legal failures,and asan efectivealternative to the formal system.In contrast, the rent-seeking theory suggests that f rmsw ith po litical connections are likely to take advantage of government intervention and slip through policies,such that there is a close relation between political connections and corruption.

Weadd resswhich theorym oreef ectively exp lainspo litical connectionsand their functionalmechanismsby distinguishing between them from a media supervision perspective.First,we introduce the background o f m ediam onitoring in the Chinese stockm arket.In recent years,them ed ia asan im po rtan t externalgovernance m echanism has received w idesp read atten tion for its inf uence over the stock m arket in China.The fo llow ing are themost im portant characteristics ofmedia monitoring in the Chinese stock market.

First,the in fuence of state-owned media is signif cantly higher than that o f othermedia(Yang and Ling, 2011)because the fourmajor securities newspapers—regarded as themost inf uential—are state-owned.

Second,mostmedia,especially state-owned media,involvemulti-tasking—the p resence o fwhich creates a m ixtureo f adm inistration andmarketing.M edia coveragenotonly needs to obey government control requirements,but also has to adap t to themarket’s developmentwhilemeeting readers needs.

Third,the depth and breadth ofmedia reportsare improving as the stock market develops.

A lthough themedia in Chinamustoperateunder government contro ls,they can stillpursue their own utility m axim ization fo rm arket pu rposes,typ ically by exhibiting a p reference fo r infamm ato ry new s,social ho t spots and con tentious phenom ena.Them ed ia’s behavior has been repo rted as in fuencing co rporations and corporate governance(D yck et al.,2008;M iller,2006;Houston et al.,2011;Li and Shen,2010;Yang and Zhao,2012).According to Dyck et al.(2008),themedia inf uences corporate governance through supervision or reputation m echanisms.Under the supervision mechanism,themedia reveals improper behaviorwith the aid o f adm inistrative and external supervision and internal governance that u ltimately constrain f rms’m isconduct(Liand Shen,2010).Under the reputation mechanism,themedia inf uencesmanagers’reputations, which is considered an important governance function.M anagers considering future emp loyment and wages actively respond to media reports thatm ight af ect their reputations.

If the rent-seeking theorymore efectively explains the presence and function of political connections,then wewould expectmediamonitoring to reduce the role played by political connections due to the close relation between co rrup tion and rent-seek ing behavio r.As a non-p roductive activity,rent-seek ing should decrease as m edia coverage increases.2A ccording to Dyck et al.(2008)and M iller(2006),media coverage is considered to be a proxy variable fo rm ed ia monitoring.

Hypothesis 2a.The relationship between po litical connections and long-term loans decreases as media coverage increases.

In contrast,if political connections are better exp lained by reputation theory,then we can d raw an alternative hypothesis to H 2a,because a f rm’s political relationships are seen as its reputation.The behavior o f establishing and using political relationships is,in essence,market oriented.We argue that such behavior is not restricted bymedia coverage.

Hypothesis2b.The relationship between political connectionsand long-term loans isnot signif cantly reduced by increased media coverage.

3.Data and research design

3.1.Data

We d raw our sample from A-share non-state-owned listed f rms in the 2006–2012 period,duringwhich the numbers of non-fnancial listed companies provided by the W IND database were 1420,1521,1575,1721, 2072,2301 and 2422,respectively.On this basis,after deleting those in the grow th enterprises market (GEM),those that were state-owned and thosem issing themain variab le,our fnal sample is comp rised o f 5215 f rm s that met the data requirements for our hypotheses testing.The observations by year are 500, 562,619,673,846,978 and 1037,respectively.The data in this study consist o f political connections thatwere m anually collected by reading annual repo rts(the def nitions fo r po litical connections appear in Table 1)and them ed ia m onito ring o r coverage details and other f nancial data,which were m anually collected from the W IND database.TheW IND database includeseveryday news reportson Chinese listed f rms from more than 100 important newspapers.According to Yang and Zhao(2012),thedata onmediamonitoring or coverage in theW IND databasemainly cover the fnancialmedia reports of Chinese listed f rms.

3.2.Research design

We use the fo llow ing model to test H 1:

where subscript i is the sample f rm,subscript t is the year in the samp le period and Bank isa dependent variab le comp rising Bank1 and Bank2.According to Yu and Pan(2008),the f rst dependent variab le is Bank1, which equals long-term loans.Sun et al.(2005)argue that the second dependen t variable is Bank2,which equals deb tm atu rities.

ZG is an independent variab le thatm easures po litical relations,and com p rises ZG 1,ZG 2,ZG 3 and ZG 4, which are calculated as follows.ZG 1 equals 1 if the chairman or the CEO of the f rm isa current or former government of cial,and 0 otherw ise.ZG 2 equals 1 if the chairman or the CEO of the f rm is a current or former governm ent of cial,or a current or former National People’s Congress(NPC)o f cial or the Chinese People’s Political Consultative(CPPC)of cial,and 0 otherw ise.ZG 3 equals3,2 or 1 if the chairm an or theCEO of the f rm is a curren t o r fo rm er central,provincialo r city and coun ty governm ent of cial,respectively, and 0 o therw ise.ZG 4 equals 3,2 o r 1 if the chairm an or the CEO o f the f rm is a curren t o r fo rm er central, provincialor city and coun ty governm en t of cial(including NPC o r CPPC o f cials),respectively,and 0 otherwise.This design facilitates our ability to quantize po litical relations.

Table 1 Variable defnitions.

Table 2 Descriptive statistics for political connections.

Table 3 Descriptive statistics for other variables.

If the regression coef cientsof ZG 1,ZG 2,ZG 3 and ZG 4 are signif cantly positive inmodel(1),then H 1 is supported.

In model(1),we choose contro l variab les according to Sun et al.(2005)and Lu et al.(2008).The specif c def nitions o f the control variables are shown in Table 1.

We use the fo llow ing model to test H 2:

wheresubscrip t i is thesample f rm,subscript t is theyear in the sampleperiod,Bank’sdef nition is identical to that inmodel(1),M C isan independent variab le refectingmediamonitoring fo llow ing Dyck etal.(2008)and Coreetal.(2008)and MCZG 1,M CZG 2,MCZG 3 and M CZG 4 rep resent interactionsbetween MC and ZG 1, ZG 2,ZG 3 and ZG 4,respectively.The control variables in models(1)and(2)are identical.

If H 2a(H 2b)is supported,the regression coef cientso f MCZG 1,MCZG 2,M CZG 3 and M CZG 4 shou ld be signif cantly negative(positive)in model(2).

Table 4 Regression resu lts for m odel(1).

We also use subsamp le regressions to test H 2a and H 2b.Based on themedia coveragemedian,our sample can be divided into two groups:high(higher than the sample’smedian)and low media coverage.

If H 2a is suppo rted,the relationship between po litical connections and long-term loansand extended deb t m aturities shou ld be stronger(weaker)in the low(high)m ed ia coverage subsam p le.If H 2b is suppo rted,the relationship between po litical connectionsand long-term loansand extended debtm aturities should beweaker (stronger)in the low(high)media coverage subsample.

3.3.Descriptive statistics

Table 2 p rovides the descriptive statistics for political connections:18.7%of the f rms had a chairm an or CEO that was a curren t o r fo rm er governm en t o f cial—45.27%when NPC and CPPC o f cials are included. This suggests that having political connections is fairly popular.

Table 3,which p rovides the descriptive statistics of the other variab les,shows that non-state-owned listed f rmsw ith political connectionshavesignif cantlymore long-term loansand longer debtmaturities than those w ithout political connections,in addition to lower MC and better performance(ROA).A ll of the continuous variab lesarew insorized at the 1st and 99th percentiles.

4.Empirical analysis

Table 4 repo rts the regression resu lts ofm odel(1).A fter contro lling fo r o ther facto rs,the results show that politically connected f rmshave access to extended debtmaturitiesand more long-term loans.The regressionresults in colum ns(1)–(8)of Tab le 4,in which the exp lanatory variables are Bank1 and Bank2,show that the ZG(including ZG 1,ZG 2,ZG 3 and ZG 4)coef cientsare positive at the 10%signif cance level,w ith the exception of co lum ns(3)and(6).However,the resu lts in Table 4 can be used to support the reputation theo ry;specif cally,that f rmsw ith political connections tend to be regarded as having good reputationsand thusbanks arew illing to p rovide them w ithmore long-term loans.The resultscan also beused to support the rent-seeking theory;specif cally,that f rmsw ith political connections gain more long-term loans and extended debtmaturities through rent-seeking.

Table 5 Regression results for m odel(2).

Two methods are used to test H 2.The f rst ismodel(2),and the relevant regression results are shown in Table 5.The regression coef cientso f ZG 1,ZG 2,ZG 3 and ZG 4 are positive,which indicate that po litically connected f rmsgetmore long-term loansand extended debtmaturities.The coef cientso fMCZG 1,M CZG 2, MCZG 3 and MCZG 4 are negative and nearly all pass the signif cance test.

Table 7 Regression results for m odel(1)(using lagged variables).

The second m ethod is subsam p le regressions based on them edian o f M C.If M C is greater(less)than the m edian,it is in the group w ith strong(weak)m ediam onito ring.Then,we perfo rm regressionsusingm odel(1) for the diferentgroups.Tab le 6 p rovides the relevant regression resu lts,which indicate that in theweakmedia monitoring group(columns(1),(3),(5),(7),(9),(11),(13)and(15)),the regression coef cientsof ZG 1,ZG 2, ZG 3 and ZG 4 are signif cantly positive.Comparing the resultsbetween the groups(columns(1)vs(2),(3)vs (4),(5)vs(6),(7)vs(8),(9)vs(10),(11)vs(12),(13)vs(14)and(15)vs(16)),we fnd the regression coef cien ts of ZG in theweak m edia m onitoring group are nearly all signif can tly higher than in the strong m edia m onito ring group.

The results in Tab les 5 and 6 verify H 2a.Using po litical connections to obtain long-term loans shou ld be understood as rent-seeking behavior.Given that rent-seeking behavior can be unreasonable or even illegal,in a strongmedia monitoring environment,the use of po litical connections to obtain long-term loans is significantly reduced.

5.Robustness tests

5.1.Variables

Fo llow ing Fan et al.(2007),we consider whether only the CEO of the f rm is a current or former governm ent o f cial and the conclusions are not substan tially changed.

W e summ arize the cu rrent o r ex-governm en t bu reaucrat situations of both the CEO and chairm an to obtain new ZG 1,ZG 2,ZG 3 and ZG 4 variables.Fo r exam p le,ZG 1 equals2 if both the CEO and the chairm an are cu rrent o r fo rm er governm en t of cials.Sim ilarly,ZG 4 equals 5 if the CEO of the f rm is a cu rrent o r fo r-m er central governm ent of cial and the chairman of the f rm is a current or former p rovincial government o f cial.W e rebuild ZG 1,ZG 2,ZG 3 and ZG 4 and then test H 1 and H 2,and the conclusionsare not substantially changed.

Table 8 Regression resu lts for m odel(2)(using lagged variables).

Wealso contro l for other variab les such as the levelo f regionalmarketsand earningsmanagement,w ith no substantial changes.

5.2.Endogeneity

W e address the possibility o f an endogeneity p rob lem in two ways.First,we adop t lagged variab les to perfo rm regressions usingm odels(1)and(2).Obviously,a f rm’s long-term loans in the presen t year do no t af ect the po litical connections o f the p revious year.The regression resu lts reported in Tab le 7 are consisten t w iththose in Table 4.The regression coef cients of ZG 1_lag,ZG 2_lag,ZG 3_lag and ZG 4_lag are signif cantly positive at the 10%signif cance level,w ith the exception of co lumns(1)and(3).

Table 9 Regression results for m odel(1)(using instrumental variables).

The regression results reported in Table 8 are also consistentw ith those in Tab le 5.The regression coefcientso f M CZG 1_lag,MCZG 2_lag,M CZG 3_lag and M CZG 4_lag)are negative and nearly all pass the signif cance test.

Second,we perform a two-stage regression based on instrum ental variab les.A ccord ing to Sun et al.(2005), the institutional environm en t can afect a f rm’s political connections and access to long-term loans and deb t m aturities(Sun et al.,2005;Yu and Pan,2008).Thus,we use a regionalm arketization index asan instrum entalvariab le for political connections,and whileweexpect it to afect long-term loansand debtmaturities,it is exogenous,such that long-term loans and debtmaturities do not afect the regionalmarketization level.

Fo llow ing Hung etal.(2012),we perform an analysisusing a two-stage regression.We begin by estimating a f rst-stagemodeland regressing the endogenouspolitical connectionsvariab les(including ZG 1 and ZG 2)on our instrum entsand controls.The exp lanatory variable is ZG 1(ZG 2)and the regression variab les include the regionalmarketization index and the other control variablesofmodel(1).We then use the p redicted value o f ZG 1(ZG 2)from the f rst stage as the instrumentalvariab le for thisvariable in thesecond stage.Tab le 9 shows the results and hypothesis 1 is still supported.

6.Conclusion

This study uses a media supervision perspective to extend the research on political connections.W e f nd that f rms w ith po litical connections aremore likely to gain long-term loans with extended debtmaturities, relative to other non-state-owned listed f rms.The resu ltssuggest that it ispopular for f rm sw ith po litical connections to use them to access debt fnancing.W e further fnd that this behavior decreases w ith increasing media coverage,indicating thatmediamonitoring restricts theuse of political connections to a certain degree.

This research also has theoretical signif cance.Bo th the repu tation and rent-seek ing theories can be used to analyzepo litical connections,and despite their apparent diferences,previous research hasnotmanaged to distinguish between them.Our study shows that the rent-seeking theory ismore suitable for explaining the p resence and functionalmechanisms o f political connections.That is,it ismore reasonab le to consider f rms’political relations as rent-seeking behavior than as a reputation efect in the Chinese stock market.

In addition,this research has p ractical signif cance.In China,themedia’s role ishighly suspect due to the degree o f government intervention.Houston et al.(2011)argue that the Chinesemedia should not p lay an important ro le in curbing corruption,but we show that,to a certain extent,media monitoring can curb corrup tion.

Bai,C.,Li,D.D.,Wang,Y.,2003.Thriving in Tilted Playing Field:An Analysis of Chinese Non-state Sectors.Working Paper.

Chen,B.,She,J.,W ang,X.,Lai,J.,2008.Em pirical Research on Development of Private Listed Com pany in Our Country.Research Report(in Chinese).

C laessens,S.,Feijen,E.,Laeven,L.,2008.Political connections and preferential access to fnance:the role of cam paign contributions.J. Financ.Econ.88,554–580.

Core,J.E.,Guay,W.,Larcker,D.F.,2008.The power of the pen and executive com pensation.J.Financ.Econ.88,1–25.

Dyck,A.,Volchkova,N.,Zingales,L.,2008.The corporategovernance roleo f themedia:evidence from Russia.J.Finance63,1093–1135.

Faccio,M.,Parsley,D.C.,2009.Sudden deaths:taking stock of geographic ties.J.Financ.Quant.Anal.44,683–718.

Fan,J.,Wong,T.J.,Zhang,T.Y.,2007.Po litically connected CEOs,corporate governance,and post-IPO performance of China’s new ly partially privatized f rms.J.Financ.Econ.85,330–357.

Fisman,R.,2001.Estimating the value of political connections.Am.Econ.Rev.91,1095–1102.

Houston,J.F.,Lin,C.,M a,Y.,2011.M edia ownership,concentration and corruption in bank lending.J.Financ.Econ.91,326–350.

Hung,M.,Wong,T.J.,Zhang,T.,2012.Political considerations in the decision o f Chinese SOEs to list in Hong Kong.J.Account.Econ. 53,435–449.

Johnson,S.,M itton,T.,2003.Cronyism and capital controls:evidence from M alaysia.J.Financ.Econ.67,351–382.

La Porta,R.,Lopez-D e-Silanes,F.,Shleifer,A.,2002.Government ownership of banks.J.Finance 57,265–301.

Li,P.,Shen,Y.,2010.The corporate governance role ofmedia:em piricalevidence from China.Econ.Res.J.109–121,14–27(in Chinese).

Li,H.,M eng,L.,W ang,Q.,Zhou,L.,2008.Political connections,fnancing and f rm performance:evidence from Chinese private f rms.J. Dev.Econ.87,283–299.

Liu,S.,Sun,P.,Liu,N.,2003.The u ltim ate ownership and its sharehold ing structures:does itmatter for corpo rate perform ance?Econ. Res.J.,51–62(in Chinese).

Liu,H.,Zhang,M.,Wang,Y.,W u,L.,2010.Politicalconnections,compensation incentive,and emp loyeeallocation ef ciency.Econ.Res. J.,109–121(in Chinese).

Lu,F.,Yao,Y.,2004.Legality,f nancial developm ent and econom ic grow th under fnancial repression.Soc.Sci.China,42–55(in Chinese).

Lu,Z.,Zhu,J.,Sun,B.,2008.Earningmanagements,accounting information and contract of bank loans.M anage.W orld,152–158(in Chinese).

Luo,D.,Zhen,L.,2008.Privately control,political connectionship and fnancing constraint:an em piricalevidence o f private enterprises listed in China.J.Financ.Res.12,164–178(in Chinese).

M iller,G.S.,2006.The press as a watchdog for accounting fraud.J.Account.Res.44,1001–1033.

Shleifer,A.,Vishny,R.,1994.Po liticiansand f rm s.Quart.J.Econ.109,995–1025.

Sun,Z.,Liu,F.,Li,Z.,2005.M arket developm ent,governm ent infuence and co rporate debtm aturity structu re.Econ.Res.J.,52–63(in Chinese).

Wang,X.,Wang,S.,2013.Chairman’s government background,excess em ployment and government subsidies:evidence from Chinese local state-owned enterprises.China J.Account.Res.6,51–74.

W ang,H.,Li,D.,Lei,D.,2001.W hy do governm ent dump state enterprises?Evidence from China.Econ.Res.J.,61–70(in Chinese).

Yang,D.,Ling,Y.,2011.W hy m edia report the lists’scandals?Secur.M arket H erald,17–23(in Chinese).

Yang,D.,Zhao,C.,2012.M ediamonitoring,media governance and managers’com pensation.Econ.Res.J.,116–126(in Chinese).

Yu,M.,Pan,H.,2008.The relationship between politics,institutional environments and private enterprises’access to bank loans. M anage.W o rld,9–21(in Chinese).

*Corresponding author.

E-mail address:yangdem ing20012005@126.com(D.Yang).☆This study was supported by the M ajor Project of the National Natural Science Foundation of China(Nos.71372168,71002110, 71132004 and 71332004).W e acknow ledge the help fu l comm ents and suggestions provided by Professor Liandong Zhang at City U niversity of H ong Kong and the anonym ous reviewers.

http://dx.doi.org/10.1016/j.cjar.2014.08.004

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.

Long-term loans

M ediamonitoring