perverse complementarity: political connections and the ......theories that predict the declining...

Perverse Complementarity: Political Connectionsand the Use of Courts among Private Firmsin China

Yuen Yuen Ang University of Michigan

Nan Jia University of Southern California

Using survey data of over 3,900 private firms in China, we examine whether—and how—political connectionspromote or undermine the use of formal legal institutions. We find that politically connected firms are moreinclined than nonconnected firms to use courts over informal avenues of dispute resolution. Furthermore, bycomparing the effects of political connections on dispute-resolution patterns across regional institutionalenvironments, we find that ‘‘know-who’’ (political influence over adjudication) dominates ‘‘know-how’’(knowledge of navigating courts) in linking political connections to the use of courts. Contrary to canonicaltheories that predict the declining significance of connections following the expansion of courts, our study suggeststhat informal networks and formal laws are more likely to share a relationship of perverse complementarity intransitional and authoritarian contexts. Political connections are positively linked to the use of legal procedures,and the primary mechanism behind the link is ‘‘know-who’’ over ‘‘know-how.’’

The influence of informal institutions on thedevelopment of formal legal institutions is acentral concern in political economy (North

1990).1 Canonical institutional theories posit a substitu-tive relationship between formal laws and informalnetworks. In early Western Europe, private orderinginitially compensated for deficiencies in the formalsystem; however, the rise of formal institutions, par-ticularly third-party enforcement via courts, eventu-ally replaced the earlier substitutes, paving the wayfor modern market economies (Greif 2006; North1991; Stiglitz 2000). Thus, to many, the juxtapositionof informal networks against formal law seemsuniversal, even in contemporary China. As Potternoted, ‘‘[the] closely held notion that formal and legalinstitutions operate in contradistinction to the rule ofguanxi [personal connections] means that anyone whoattempts to relate the two runs the risk of professionalbeheading’’ (2002, 180). Indeed, some sociologists

have predicted the ‘‘declining significance of guanxi’’and political connections in business (Guthrie 1998,2002) as Chinese society modernizes and expandingmarket mechanisms erode the traditional base ofpower among Communist officials (Nee 1989).

Others, however, disagree. Informal institutionsdo not necessarily substitute for formal institutions; theformer may complement or accommodate, dependingon the strength of the latter (Helmke and Levitsky2004). Moreover, the form and effectiveness of formalinstitutions cannot be assumed in transitional and de-veloping economies, where formal institutions are stillbeing constructed and shaped by preexisting networks(Grzymala-Busse 2010; Tsai 2006). Some maintain thatguanxi, particularly personal connections to politicalauthorities, remain indispensable in China for a varietyof endeavors, including operating businesses (Kennedy2005; Kung and Ma 2011; Pearson 1997), obtainingbank loans (Li et al. 2008; in Vietnam, see Malesky and

The Journal of Politics, Vol. 76, No. 2, April 2014, Pp. 318–332 doi:10.1017/S0022381613001400

� Southern Political Science Association, 2014 ISSN 0022-3816

1Both authors contributed equally to this article. Support was provided by the Greif Center for Entrepreneurial Studies at the Universityof Southern California. Data necessary to reproduce the results are available upon request from Nan Jia at [email protected] on the personal webpage of Yuen Yuen Ang. An online appendix with supplementary material for this article is available athttp://dx.doi.org/10.1017/S0022381613001400. An earlier version was titled ‘‘Political Connections & Use of Courts in DisputeResolution: A Survey Analysis of Chinese Private Firms’’ and posted on SSRN on August 29, 2011.

318

Taussig 2008), expressing civic grievances (Michelson2006, 2007b; Michelson and Read 2011; Tsai andXu 2012), and avoiding bureaucratic harassment(Michelson 2007a). These findings echo Walder’sinfluential claim about the persistent value and returnsto political office in China’s transitional economy(Walder 2002).

We revisit the debate surrounding formal institu-tions and informal networks by interrogating whether—and how—political connections promote or underminethe use of formal legal institutions among firms. Inrecent years, Chinese firms have become increasinglywilling to use courts to resolve disputes (Clarke, Murrel,and Whiting 2008; Whiting 2010). However, with onlya few exceptions in Russia (Frye 2006; Hendley, Murrell,and Ryterman 2000), little remains known about thecharacteristics of firms that employ formal over in-formal modes of dispute resolution in transitional andauthoritarian countries. This is particularly worthy ofinvestigation among China’s private firms, which aretraditionally dependent upon personal ties with bureau-cratic patrons for economic survival (Tsai 2007; Wank1996; Yang 2002). By employing survey data for over3,900 private firms sampled nationwide, our paperaddresses two questions. First, are politically connectedfirms more or less inclined than unconnected firmsto use courts when facing disputes? Appropriate tothe context of China, we define politically connectedentrepreneurs as those who are (1) congressionaldelegates or (2) former party-state officials. We findthat, all things being equal, politically connectedfirms are more likely than others to use courts overinformal means of dispute resolution. Put colloquially,although having political connections is normallyassociated with exploiting ‘‘back door’’ strategies, wefind that connections may paradoxically empowermarket actors to ‘‘use the front door.’’

Our primary finding motivates a critical andmore challenging question: Are politically connectedfirms more likely to use courts because of theiradvantages in ‘‘know-how’’ (knowledge of navigatingthe courts) or ‘‘know-who’’ (political influence overadjudication)? Although those familiar with the polit-ical embeddedness of Chinese courts may not be sur-prised by our main finding (i.e., politically connectedfirms use courts more), the divergent mechanisms bywhich connections induce firms to use courts remainuntested. On the perverse end, political connectionsevoke images of unscrupulous individuals who pullstrings behind the scenes. On the positive end, however,others contend that ‘‘political connections are not alwayscorruption’’ (Tsai 2011). Politically connected actorsare generally more capable of maneuvering complex

bureaucratic and legal institutions. Political connec-tions can provide an empowering resource for civicand legal participation in democracies and nondemoc-racies (Fung 2010; Tsai and Xu 2012). Both positiveand perverse mechanisms are likely to link politicalconnections to legal recourse. Therefore, the empiricalchallenge we address is disentangling the two mecha-nisms and assessing their relative weights.

It is difficult, if not altogether impractical, to askpolitically connected firms directly for their primarymotives in using courts, particularly in large-n surveys.To circumvent data restrictions, we apply simple infer-ential logic: if politically connected firms use courtsprimarily to exploit ‘‘know who’’ (political advantage)over ‘‘know how’’ (knowledge advantage), then weshould predict variance in correlational patternsbetween political connections and the use of courtsas the operating environment varies. To employ ananalogy, we can infer the dominant mechanism offriendship between two individuals (e.g., money versusloyalty) by observing how their bond manifests acrossdifferent environments. Specifically, in our regressionanalyses, we first employ interaction effects withregional legal-services capacity as a moderator andthen simulate the effects. Our analytical strategyyields evidence suggesting that political advantagedominates knowledge advantage in linking politicalconnections to the use of courts. Whereas others haveasserted that political connections can either facili-tate or obstruct legal systems, or do both, we presentthe first empirical attempt to adjudicate these claims.

On a broader level, we highlight core differencesin interactive patterns between formal and informalinstitutions in early Western Europe and contempo-rary developing and transitional countries such asChina. Studying patterns of legal development inearly Western Europe, many economic historiansfind formal laws and informal networks substitutive.However, this observation is not universal. Amongdeveloping countries today, the development of law istightly compressed, and where authoritarian partiesrule, law is subordinated to politics. Contrary to canon-ical theories that predict the declining significance ofconnections following the expansion of courts, ourstudy suggests that informal networks and formallaws are more likely to share a relationship of perversecomplementarity in transitional and authoritariancontexts. Political connections are positively linkedto the use of courts, and the primary mechanismbehind the link is ‘‘know-who’’ over ‘‘know-how.’’

We will proceed as follows. We present thebackground of China’s legal development and thendiscuss whether and how political connections influence

perverse complementarity 319

the use of courts among private firms, specifying ob-servable implications for testing. Following this dis-cussion, we describe our data, variables, and statisticalmethods. The fourth section reports the results of ourregression analyses and simulation, and the fifth spec-ulates on why politically connected firms might exercisegreater advantage in courts than in informal forums.Finally, we conclude with comparative implications anddirections for future data collection.

Political Connections andUse of Courts

China’s legal system has undergone tremendousgrowth and reform in the last 30 years. The EconomicContract Law was first promulgated in 1981, with anew version passed in 1999. From 1983 to 2001, eco-nomic disputes grew 18.3% each year on average,twice the rate of civil disputes and four times that ofcriminal cases (Clarke, Murrel, and Whiting 2008).Contract-related disputes took the lion’s share ofeconomic disputes during this period (Lubman1999, 254–55). Besides litigation, alternative informalmechanisms of dispute resolution include self-enforcingcontracts, negotiation and mediation, and arbitration(Whiting 2010). Another option is to seek interventionfrom local or higher level governmental authorities(Michelson and Read 2011, 12–13). Existing studiesreport that private mediation remains the most popularmechanism (Tang 2009; Whiting 2010). However, thelegal system is an increasingly important avenue forresolving corporate disputes (Clarke, Murrel, andWhiting 2008; Whiting 2010). Consistent with others,we find a steady rise in the use of courts in our sample,even though mediation is still the most frequentlyemployed method.

Main Test: Are Politically ConnectedFirms More Inclined to Use Courts?

We review two divergent hypotheses regarding howpolitical connections relate to legal utilization.Following the substitutive view, political connectionsand the use of courts are expected to be negativelycorrelated. As Stiglitz asserted, ‘‘network[s] of in-terpersonal relationships get dispensed with and de-stroyed’’ with the emergence of formal marketinstitutions such as courts (2000, 64, italics added).Likewise, in the policy world, international assistanceprojects typically assume that expanding the use oflaw equates to leveling the playing field between

individuals with and without connections (Jensen2003).2

However, the reverse, that political connectionsare positively associated with the use of courts, couldbe true if one takes into account the developmentaland political context in China. In any developing ortransitional country, the legal system is new, evenintimidating, to ordinary citizens and entrepreneurs.As observed in Russia, despite the government’s supplyof laws and courts, firms shied away from using thelegal avenue because they lacked knowledge andtrust (Hendley 1999). Furthermore, in an authori-tarian system, courts are not politically independent(Ginsburg and Moustafa 2008). Thus, compared toordinary firms, we can expect that politically con-nected (PC) firms,3 whose owners formerly served asofficials or congressional delegates, may be moreempowered to use formal institutions. Indeed, anumber of studies in China focusing on citizens (butnot firms) find that former cadres are more likelythan ordinary citizens to employ litigation, petitions,and even protests to resolve civic grievances (Li andO’Brien 2008; Michelson 2006; Michelson and Read2011; O’Brien and Li 2006).

H1a: Political connections and use of courts aresubstitutive: PC firms are less likely to use courts overinformal means of dispute resolution than non-PC firms.

H1b: Political connections and use of courts are com-plementary: PC firms are more likely to use courts overinformal means of dispute resolution than non-PCfirms.

Which is the Dominant Mechanism:Political Advantage orKnowledge Advantage?

The complementary view, which predicts a positivecorrelation between political connections and the useof courts, motivates a challenging empirical question:In what ways do political connections empower firmsto use courts more than those lacking connections? Isit because political connections endow ‘‘know-who’’(political advantage) or ‘‘know how’’ (knowledgeadvantage)?4 Although there could be several pathwaysthrough which political connections are positively

2This view has been criticized as ‘‘naıve formal legalism’’(Michelson 2007a, 355; Suchman and Edelman 1996).

3To abbreviate, in the rest of the article, we will refer to politicallyconnected firms as PC firms and nonconnected firms as non-PCfirms.

4The terms ‘‘know how’’ and ‘‘know who’’ earlier appeared inMichelson (2006, 2007a).

320 yuen yuen ang and nan jia

linked to the use of courts, ‘‘know who’’ and ‘‘knowhow’’ are the two most widely discussed mechanismsin the literature. Below, we discuss each mechanism.

First, we discuss political advantage. PC firms mayturn to courts primarily because they are (or believethey are) more effective at influencing courts for theirpersonal advantage. By design, Chinese courts aresubordinated to party and state authorities (Lubman1999; Michelson 2006). Judges and court personnelin China are appointed by the party’s organizationdepartment and financed by local governments.Political intervention of judicial processes ‘‘continuesto be a legitimate action’’ despite top-down efforts atprofessionalizing courts (Liebman 2007, 626). Havingpolitical connections provides an entrepreneur privi-leged access to court personnel and supervisory offi-cials. A recent study on commercial disputes suggeststhe prevalence of political influence. Although themajority of respondents refused to answer questionsduring interviews relating to influence, those whodid revealed attempts to ‘‘[invite judges] to banquets,[call] judges in their homes, and [send] emissaries tomeet with the judges’’ (Pei et al. 2010, 225).

Second, we discuss knowledge advantage. Politicalconnections may empower firms to seek legal recoursefor a positive reason: they possess more knowledgeabout how to use courts. In regard to litigation, usershave to learn the legal system before they can approachit. Entrepreneurs who formerly served as officials ordelegate members (our measures of PC) have excep-tionally direct exposure to laws and policymakingcompared to ordinary firms. To quote Michelson,‘‘political connections are more than influence-peddlingand shady backdoor dealings . . . It is [also] aboutadvantages in learning how to navigate new andcomplex institutions’’ (2006, 27). Connections canenhance ‘‘understanding of and adherence to legal rulesand procedures; awareness of legal rights, informationabout legal processes’’ (27). More generally, some polit-ical scientists have argued that political connections canfunction as a form of social capital that empowerscivic participation. For example, Fung argued that‘‘empowered participation’’ was made possible by newinitiatives that gave underclass communities inChicago more direct access to city agencies andpowerful political actors (Fung 2010). Likewise, draw-ing on the survey results of Chinese residents, Tsai andXu concluded that ‘‘political insiders are more likely tohave more resources for participation and find it easierto overcome barriers to political action in nondemo-cratic and transitional contexts’’ (2012, 8).

So how can we disentangle the two mechanisms—political advantage versus knowledge advantage—and

test which is more influential? Answering this questionis the central task and contribution of our analysis.Our data, as well as other survey sources of whichwe are aware, do not reveal whether firms soughtto influence judges when they used courts. Nor didthe survey we use directly measure the firms’ legalknowledge or consciousness. Thus, our analyticalstrategy is to work with the existing data and analyzethe moderating effect of the local institutional envi-ronment in which firms operate. We label this LegalServices Capacity (LSC). We measure LSC as theavailability of legal professionals and the efficiencyof courts in processing cases by province (more detailsin the data section). Our logic is that in areas with moretechnically efficient courts and where professional legalservices are more widely available, non-PC firms canmore easily navigate the legal system and learn to usecourts. In other words, a higher level of legal servicescapacity approximates a smaller ‘‘knowledge gap’’between PC and non-PC firms.

With this inferential logic in mind, we test thetwo competing mechanisms as follows. If PC firmsuse courts more than non-PC firms primarily becauseof their knowledge advantage over non-PC firms,then we should observe a weaker positive correlationbetween political connections and the use of courtsin areas of higher Legal Services Capacity, i.e., wherethe ‘‘knowledge gap’’ is smaller (expressed in H2a).For example, Shanghai has more legal professionalsand more technically efficient courts than Guizhou,such that non-PC firms in Shanghai, compared tothose in Guizhou, can more easily compensate forknowledge disadvantage by hiring lawyers or evenhandling lawsuits themselves. If the knowledge factordominates and political advantages are less significant,then we should see a lower propensity for PC firms touse courts over non-PC firms in Shanghai than inGuizhou; i.e., the positive correlation between PC andlitigation should weaken.

Conversely, if PC firms use courts more becausethey can exercise influence and less because of knowl-edge differences, then a smaller ‘‘knowledge gap’’ shouldnot diminish their propensity to use courts. In fact,wider availability of professional services and efficientcourts should make it easier for PC firms to employ legalprocedures. In this case, expressed in Hypothesis 2b,we expect to see a stronger positive correlationbetween political connections and the use of courts.Figure 1 summarizes our hypotheses and testingstrategy.

H2a: knowledge advantage dominates political advantage—where legal services capacity is higher (i.e., knowledge gap


is smaller), the positive association between PC and theuse of courts is weaker.

H2b: political advantage dominates knowledge advantage—where legal services capacity is higher, the positive asso-ciation between PC and the use of courts is stronger.

To be clear, our objective is to exploit the predictedmoderating patterns in Hypothesis 2a and Hypothesis2b to draw inferences about the dominant mechanismlinking political connections to the use of courts. Ourtesting strategy makes no assumption that LSC causesfirms to exercise political advantage over knowledgeadvantage or vice versa. Nor do we assume any causaldirection between political connections and LSC. Werealize that the two are likely correlated. However, inthe absence of an experimental set-up, identifying sucha causal link is impossible, and more importantly, notthe goal of our analysis.

Data and Methods

Data Source

We use survey data on Chinese privately owned firmscollected by the Privately-Owned Enterprises ResearchProject Team in 1995, 1997, and 2000.5 Jointly fundedby state-linked organizations and academic institu-

tions, the surveys are part of an ongoing nationalproject to collect information of the Chinese privatesector to facilitate policymaking. The research teamgenerated a nationwide random sample of privatefirms using multistage stratified sampling across ad-ministrative regions and industries and conducteddirect interviews using a questionnaire with the majorowner of each POE in the sample. The survey covereda range of questions on firms’ performance, as wellas owners’ demographic and personal information.We use the National Economic Research Institute(NERI) database to generate indicators of province-level variation in LSC (see the online appendix foran elaboration of our use of the NERI indicators).

At the outset, we should emphasize that this is afirm-level dataset that measures firms’ most frequentresponses to disputes, not a dispute-level dataset.The survey asked the firms about the type of disputesmost frequently encountered in the year of survey(e.g., disputes with government, with businesses, andwith others), which we control for in the analyses.However, no questions were asked about each dispute,e.g., issues of conflict and monetary cost. Nor do weknow about the firms’ trading partners or detailsabout the process of dispute resolution.

We acknowledge the limitations in our datasetbut stress two points. First, information on the mostfrequently used method to resolve disputes essentiallysynthesizes all dispute-level incidents and captures an‘‘average’’ scenario for each firm. Second, large-sample dispute-level data of firms are rare not onlyin China but even in developed countries. In most ofthe existing literature, court case files were used tostudy dispute resolution dynamics. However, suchdata reflect only disputes that are elevated to courtsand provide no indication of alternative means bywhich disputes are handled (Siegelman and Donohue1990). To understand patterns of dispute resolution,

FIGURE 1 Hypotheses and Testing Strategy

5The survey is also available in the years of 2002, 2004, and 2006.Unfortunately, the questions related to dispute resolution werechanged considerably in each of these years, such that these threeyears of data are not comparable with one another and notcomparable with the years we used. It would be erroneous to pooldata from all of these years only to expand the dataset. Althoughour dataset captures an earlier time period (up to 2000), bydelimiting our scope of data to comparable years, we can be moreconfident of delivering robust results with clear interpretations toinform future data-collection efforts.


it makes sense to start at the base—that is, by askinghow subjects respond generally to dispute encounters.This method is used in a variety of survey-basedstudies on disputes across different countries (Frye2006; Gallagher and Wang 2011; Hendley, Murrell,and Ryterman 2000; McMillan and Woodruff 1999;Michelson 2007b; Michelson and Read 2011; Tsai2007, chap. 5). In this regard, one advantage of theprimary survey question we used to generate ourdependent variable is that it asked about a firm’sactual experiences in dealing with disputes and not ahypothetical situation that they had not encounteredin practice.

Dependent Variable

We generate our dependent variable based on a ques-tion that asked ‘‘Which is the most frequently methodused to resolve disputes encountered by your firm?’’Respondents were presented with four options: (1) usecourts, (2) negotiate and mediate, (3) seek interventionfrom government, and (4) other. We generate thedependent variable Use Courts based on this question.Use Courts equals 1 if the firm reported that it usedlitigationmost frequently to resolve disputes and 0 if thefirm reports any of the other methods as most fre-quently used. Firms may use a combination of methodsand negotiations before deciding to use courts(Michelson and Read 2011); thus, we acknowledgethat an expressed choice of litigation does not precludeother methods before or during the course of litiga-tion. That said, if a firm indicates ‘‘use courts’’ as themost frequent method of dispute resolution, whichonly 9.6% of the firms did, this response represents astrong signal of a firm’s willingness to use courts.

Measures of Political Connections

On a theoretical level, we conceptualize political con-nections as having the privilege of personal interactionswith governmental authorities and first-hand insightsinto processes of decision making. Empirically, weoperationalize politically connected firm owners as(1) Former Officials or (2) Delegates.

Delegate is a dichotomous variable indicatingwhether a firm’s owner is a delegate of the People’sCongress or the People’s Consultative Conference,which are two political organs at the national andlocal levels that parallel legislatures in democracies.The Congress drafts and approves laws, whereas theConference contributes opinion to policymaking.The former officially nominates and elects party-stateleaders, a role that congressional members have increas-ingly asserted in recent years (Manion 2008). From the

1990s onwards, the Communist Party stepped upefforts to recruit private entrepreneurs into the formalpolitical system, inviting them to serve as congressionaldelegates (Dickson 2008). Firms are also known toactively seek delegate positions to enhance their politicalstanding, as members advise party and governmentofficials through regular meetings (Li, Meng, and Zhang2006; Tsai 2007, chap. 5).

Former Official measures whether the firm ownerpreviously worked as a government official at thedirectorate (chu) rank, equivalent to a county mag-istrate, ministerial division chief, or higher rank.The directorate rank is a distinct cutoff point in theadministrative hierarchy that indicates membershipin China’s stratum of ‘‘political elites’’ (Walder 2004,195). We expect firm owners who previously held elitebureaucratic positions to hold extensive and deepcontacts in the party-state administration. This mea-sure has the theoretical value of isolating the effects ofthe most exclusive political connections, distinguishingour analysis from previous studies that measured‘‘former cadres’’ or ‘‘public employees’’ withoutdelimiting ranks (for example, Malesky and Taussig2008; Michelson 2006).

Being a delegate or former official constitutes twomain avenues by which firm owners gain direct andclose connections to the government in the Chinesecontext. The type of political connections we captureis distinguished from measures of indirect connections,such as whether one has ‘‘relatives working in thebureaucracy’’ or ‘‘friends in the government’’ (Kungand Ma 2011; Tsai and Xu 2012). It is also distin-guished from political connections acquired withoutregular face-to-face interaction with officials andwhich are less exclusive than delegate membershipor former official posts, such as Chinese CommunistParty (CCP) membership (Dickson and Rublee 2000).Although PC can be defined in infinite ways, we chooseto focus on elite and exclusive forms of politicalconnections because they usefully specify the effectsof the strongest type of PC.

Moderator: Legal Services Capacity

To generate the moderator, Legal Services Capacity,we use province-level indices developed in the NERIdatabase (for details, see Fan and Wang 2000), whichis available only at the province level. A higher valueof LSC indicates that the legal system in the pro-vince has greater technical efficiency. LSC is a com-posite index based on Legal Services Developmentand Court Capacity. Legal Service Development is astandardized index measuring the availability of


professional legal services in a province based on twosubindices: the number of lawyers divided by theprovince’s population and the number of accountantsdivided by the province’s population. Court Capacity isa standardized index measuring the capacity of courtsin the province to handle cases based on two sub-indices: the number of business lawsuits filed in thesurvey year divided by the province’s GDP and thenumber of business lawsuits concluded in the surveyyear divided by the province’s GDP.

We elaborate on the merits and shortcomings ofLegal Services Capacity. As a measure of province-levelvariation in the availability of legal professionalsand court efficiency, we use LSC to proxy for the‘‘knowledge gap’’ between PC and non-PC firmswithin each province. Intuitively, one may questionif a province-level proxy is too coarse given the hugeinstitutional complexities across China. In terms ofdata constraint, we do not know of an equivalentmeasure of LSC at the subprovincial levels, and evenif there is, there would be limited observations ineach city or county in our survey to merit a robustanalysis. More importantly, on a theoretical level, wesubmit that province is actually an appropriate levelof analysis. In the era of market reform (post-1979),the central government granted tremendous autonomyto province-level governments, resulting in a well-documented structure of market fragmentationand local protectionism along provincial borders(Donnithorne 1972; Wedeman 2003). This institu-tional set-up implies that firms do not face disputesonly or primarily with firms within subprovincialboundaries. Instead, it is far more likely that firmsare grouped at the provincial level for two reasons:restrictions on market entry are imposed by provincialgovernments, and firms shy away from going to courtoutside of their home province because state andjudicial protectionism of local firms is common(Howson 2009).

Another theoretical leverage of LSC is that itprovides a rationale for both Hypothesis 2a andHypothesis 2b. Hypothetically, an ideal measure ofthe ‘‘knowledge gap’’ would come from questionsin the survey that directly measure firms’ legalknowledge, but these were not asked. However, evenif such a firm-level proxy was available, it cannotaccount for why the positive correlation shouldbecome stronger, as we hypothesized in Hypothesis 2b.A positive moderating effect is possible because ahigher level of LSC not only approximates a smallerknowledge gap, but it also measures an institutionalenvironment where legal professionals abound andcourt efficiency is higher. These conditions render it

easier for PC firms to employ litigation or defendcases in courts. In fact, what we later report, usingLSC as the moderator, is that the positive correlationbetween PC and the use of courts indeed becomesstronger even as the knowledge gap narrows, consistentwith Hypothesis 2b.

Lastly, our moderator of LSC is an objectivemeasure of judicial technical capacity (availability oflegal professionals and court efficiency), which hasadvantages over attitudinal measures of judicial fair-ness. There have been notable efforts to survey citizens’subjective evaluation of court performance. However,as acknowledged, the vast majority of citizens surveyedhad never interacted with courts (Michelson and Read2011), and evaluations are captured via vignettes ratherthan actual encounters (Wang 2013). Hence, thesesurveys are less appropriate in approximating theobjective knowledge gap between PC and non-PCfirms that encounter actual disputes. In sum, despitesome limitations of LSC, it is not only the best availablemoderator but also compatible with our theoreticalaccount.

Other Firm-Level Variables

We included a battery of controls in the regressions.We include firm-level variables, controlling for yearof survey, firm’s age, industry, provincial location,number of employees, profits, firm owner’s age, andowner’s education level. We also control forthe type of disputes that firms most frequentlyencountered in the survey year: disputes withgovernment bureaus, firms, and others. To rem-edy the possibility that LSC may be confoundedwith other province-level characteristics, we alsoinclude three indices from the NERI database thatmeasure the development of the private sector ineach province: Private Sector Industrial Output,Private Sector Fixed Asset Investment, and PrivateSector Employment.

In the final sample of 3,980 firms, the average firmowner was 44 years old and received approximately11 years of formal education. Approximately 42.6%of firm owners held positions in either the Congressor Conference, and 1.3% were former officials. Theaverage firm had been in business for over six years,hired 114 employees, and made a profit of 115 millionYuan. Approximately 74.9% of the firms reported thatthey most frequently experienced disputes with otherbusinesses, and only 9.5% with government bureausand officials. Only 9.6% of the firms reported that theymost frequently used courts, over other options, toresolve disputes.


Methods

We estimate logit models to predict the probability ofUse Courts based on firm-level variables that measurepolitical connections and other characteristics, provincedummies, and industry dummies.6 To test the moder-ating effects, we begin by reporting the interaction termsof the logit models in the conventional way. Then, weexamine the true interaction effects, using the simu-lation approach developed by King, Tomz, andWittenberg (2000) and Zelner (2009). We employsimulation because estimates of the interaction termbetween two variables of interest in a nonlinearmodel, such as a logit model, does not represent thetrue interaction effect (Norton, Wang, and Ai 2004).(In an online appendix, we elaborate on the econo-metric logic and mathematical proof.) We reportrobust standard errors throughout our analyses.

Empirical Results

Descriptive Patterns of Dispute Resolution

Before discussing the results of statistical analysis, wedescribe the patterns of dispute resolution for differenttypes of disputes (disputes with businesses, govern-ment, and other parties). On average, 84.6% of firmsreport using private mediation and negotiation as themost frequent method of dispute resolution for allthree types of disputes. Using courts is the second—but distant second—most popular resolution method,accounting for only 9.6% across all three types ofdisputes. This is followed by seeking governmentintervention informally, chosen by 4.4% of firms.Like others (Clarke, Murrel, and Whiting 2008;Whiting 2010), we observe that the use of courtshas become more popular over time. In the onlinesupplement, Table A1(a) reports the percentage ofdifferent resolution methods used in all three types

of disputes, and Table A1(b) reports the summarystatistics of variables used in the empirical analyses.

Main Effects

We now examine the characteristics of firms that preferusing courts over informal means of dispute resolution,focusing on political connections. Table 1 reports theresults of the logit models. The results should be inter-preted as the probability of using courts as the mostfrequent mode of dispute resolution vis-a-vis alternativemethods. Model 1 contains all control variables, andModel 2 adds Delegate and Former Official.

Supporting Hypothesis 1b, the results in Table 1show that the probability of using formal legalprocedures to resolve disputes increases with politicalconnections. Delegate is a positive and statisticallysignificant (p , 0.05) predictor. Holding all othervariables at their median values, and after logistic trans-formation, being a Delegate increases the probability ofusing courts by 2.2 percentage points. Former Official isalso a positive and statistically significant (p , 0.10)predictor. Holding all other variables at their medianvalues, prior political office increases a firm’s proba-bility of using courts by 8.6 percentage points.

The substantive effects of political connectionsshould be interpreted in view of the small likelihoodthat firms would report Use Courts as the most fre-quent method of dispute resolution. Given that onlyless than 10% of firms reported litigation as the mostfrequent means of resolving disputes in our sample, ifall the sampled firms previously held office, the meanvalue of Use Courts would increase by 85.6% (from10 to 18.6 percentage points). Likewise, if the sampledfirms were all delegates, the predicted mean value ofUse Courts would increase by 22.4 % (from 10 to 12.4percentage points).

To further contextualize our results, it must benoted that it takes tremendously strong motivationand supporting factors to induce any firm to reportlitigation over other options as the most frequentmethod of dispute resolution. As emphasized, 84.6%of firms report private mediation as the primarymechanism. It is well established in both game-theoretical and empirical literature that firms preferto avoid litigation because of high costs and delays,turning to courts only as a last resort (Macaulay1963). Another sharp empirical illustration can befound in Indonesia, where despite major campaignsled by the World Bank and IMF to encourage firms toresolve contractual disputes in courts, firms refusedto use them (Andrews 2013, 35). Although litigationis on the rise, China remains a nonlitigious society.

6In our models that examine the moderating effect of LSC,we use indices that leverage variation at the province level.To avoid multicollinearity in testing for province-levelmoderating effects, we replace province dummies with geo-graphic region dummies to control for unobserved effectsshared by geographically adjacent provinces. The six regionsinclude the East Region (four provinces), the Central Region(six provinces), the North West Region (six provinces), theNorth East Region (three provinces), the South West Region(six provinces), and the Coastal Region (six provinces). Webase the categorization of regions on the China StatisticalYearbook.


A recent study asserts that ‘‘firms all avoid courts’’(Wang 2013, 46), noting that some firms even includestipulations in their contracts to seek mediation overlitigation in the event of disputes. Once an overallnonlitigious context is taken into account, it becomesapparent that the substantive effects of Delegate andFormer Official are not trivial.

Some results of the control variables in Table 1are worth noting. Education shows no statisticallysignificant effect, indicating that it is not formalschooling or literacy per se that empowers the use of

courts. Older firms and older owners are more likelyto use courts. More experience might allow one toaccumulate legal knowledge and/or political capital,both of which could enhance the probability of usingcourts. This observation partially supports Galanter’s(1974) seminal argument that the American legalsystem privileges ‘‘haves’’ with legal experience(e.g., large corporations). However, reflecting a non-democratic and state-dominant setting, our articlehighlights political connections as one dividing linebetween ‘‘haves’’ and ‘‘haves-not.’’

TABLE 1 Main Results

Dependent Variable:Use of Courts (1) (2) (3) (4)

Delegate position 0.248**(0.126)

0.233(0.278)

0.261**(0.122)

Former Official 0.684*(0.379)

0.519(0.382)

-0.340(0.950)

Delegate position 3 Legal Services Capacity 0.007(0.055)

Former Official 3 Legal Services Capacity 0.151(0.145)

Legal Services Capacity – – 0.103**(0.046)

0.103***(0.038)

Dispute with businesses 0.785***(0.199)

0.785***(0.199)

0.784***(0.200)

0.782***(0.200)

Dispute with government 0.489*(0.263)

0.484*(0.263)

0.435(0.265)

0.435(0.265)

Firm age 0.041***(0.013)

0.035***(0.013)

0.034***(0.013)

0.034***(0.013)

Employee number 0.000(0.000)

-0.000(0.000)

-0.000(0.000)

0.000(0.000)

Profit 0.010(0.022)

0.008(0.022)

0.022(0.021)

0.022(0.021)

Donation 0.013(0.025)

0.010(0.026)

0.003(0.025)

0.003(0.025)

Owner age 0.017***(0.006)

0.016***(0.006)

0.015***(0.006)

0.015***(0.006)

Education -0.007(0.023)

-0.011(0.023)

-0.014(0.023)

-0.014(0.023)

Year 0.084**(0.040)

0.071*(0.041)

0.074(0.048)

0.075(0.048)

Industry dummies Y Y Y YProvince dummies Y Y N NProvince controls N N Y YRegion dummies N N Y YConstant -170.136**

(79.822)-145.843*(81.456)

-152.326(95.811)

-154.013(95.803)

Observations 3,980 3,980 3,980 3,980

Note: Our dependent variable measures whether firms most frequently use courts to resolve disputes vis-a-vis alternative methods.Robust standard errors in parentheses. The province controls in Models (3) and (4) contain the following control variables: PrivateSector Industrial Output, Private Sector Fixed Asset Investments, and Private Sector Employment.*p , 0.1; **p , 0.05; ***p , 0.01.


Moderating Effects

In this section, we examine the moderating effects ofLegal Services Capacity. We present the interaction termsin Models 3 and 4 in Table 1. Although the interactionterms are not statistically significant in Table 1, it mustbe noted that because we used nonlinear parametriclogit estimation, these results alone do not indicate theabsence of statistically significant true interaction effects.As Norton, Wang, and Ai (2004) demonstrated, thestatistical significance of true interaction effects cannotbe determined by the z-statistic reported in the non-linear regression output of the interaction terms.Therefore, in the next step, we follow the approach ofKing, Tomz, and Wittenberg (2000) and Zelner (2009)to simulate the interaction effects of the logit models.

Recall our strategy for testing Hypotheses 2a and2b: political advantage versus knowledge advantage.We use higher levels of Legal Services Capacity toproxy for a smaller ‘‘knowledge gap’’ between PC andnon-PC firms. If the positive correlation betweenpolitical connections and the use of courts is strongereven in provinces where the ‘‘knowledge gap’’ issmaller, the result would support Hypothesis 2b:political advantage dominates knowledge advantage.The reverse would support Hypothesis 2a: knowledgeadvantage dominates political advantage.

In the remaining empirical discussion, we willpresent four figures in two sets that simulate theresults of the true interaction effects between Legal

Services Capacity and our two measures of politicalconnections: Delegate and Formal Official. Each set offigures follows the same logic. We begin with LegalServices Capacity and Delegate, as shown in Figure 2.Figure 2(a) shows the change in the predictedprobability of Use Courts (y-axis) as values of LegalServices Capacity (x-axis) rise for Delegates andnon-Delegates.

Three points stand out in Figure 2(a). First, forall firms, the probability of using courts increases asLSC increases, as evident from the upward slopinglines, which suggests that regardless of firm type,using courts is more prevalent in provinces wherecourts are more technically efficient and legal servicesare more widely available. Second, at any given levelof LSC, Delegates (solid line) are always more likely touse courts than non-Delegates (dash line), further con-firming the first-order positive relationship betweenpolitical connection and the use of courts, as hypoth-esized in Hypothesis 1b. Third, as LSC increases, theprobability of using courts increases at a steeper ratefor Delegates than non-Delegates. That is, Delegatesare more likely than non-Delegates to use courts evenin areas with higher legal-services capacity (where the‘‘knowledge gap’’ is smaller).

Figure 2(b) provides an alternative representationof the results in Figure 2(a). It plots the difference ofthe predicted probability of Use Courts betweenDelegates and non-Delegates (y-axis) against LegalServices Capacity (x-axis). The dotted region indicates

FIGURE 2 Moderating Effect of Legal Services Capacity on the Relationship between Delegate and theProbability of Using Courts to Resolve Disputes


a statistically significant association between Delegateand Legal Services Capacity at a 90% confidence level.(All but one set of our results show statistical signif-icance at the 95% level, which we will later discuss.)Again, consistent with Figure 2(a), the upward-slopingcurve in Figure 2(b) indicates that the difference in theprobability of using courts between Delegates and non-Delegates increases where LSC is higher. This effect isstatistically significant at medium levels (second andthird quartile) of LSC. Put differently, even in areaswhere there is a smaller ‘‘knowledge gap’’ between PCand non-PC firms, Delegates are more likely to usecourts than non-Delegates. This pattern supports theinteraction Hypothesis 2b: political advantage domi-nates knowledge advantage.

Following the same logic, Figure 3(a) and 3(b)present the interaction effects of Legal Services Capacityand the other measure of PC, Former Official.The patterns observed in Figure 2(a) are similarlyobserved in Figure 3(a). Specifically, the probabilityof using courts increases at a steeper rate for FormerOfficials than non-Officials even as Legal ServicesCapacity increases, consistent with Hypothesis 2b.Figure 3(b) graphs the difference in the probabilityof using courts between Former Officials and the re-maining firms. The difference increases with the valueof LSC. The correlation patterns are consistent withthose in Figure 2(a) and 2(b).

We acknowledge that a constraining part of ourtests is our theoretical choice to operationalize PC asFormer Officials, which has limited variation (1.4% of

firms). However, including Former Officials in com-parison to Delegate usefully shows that exclusivity ofconnections does have a greater substantive effect onUse Courts. As a robustness check, we report regres-sions with an alternative measure of PC: Former PublicEmployees, which relaxes the rank criteria and providesmore variation across observations. In a total of fourtests, all of our results, except Former Officials*LegalServices Capacity (p , 0.10), are statistically signifi-cant at the 95% confidence level (for additional tests,see the online appendix).

In summary, our results show a positive associationbetween PC (i.e., Delegate Position or Former Official)and the use of courts. The probability that PC firms usecourts more than non-PC firms is higher in areas withhigher legal-services capacity, where the ‘‘knowledgegap’’ is smaller. In other words, even in areas where re-sources are more widely available for non-PC firms tocompensate their knowledge disadvantage, PC firms notonly continue but are even more likely to use courts thannon-PCfirms. Both sets of results supportHypothesis 2b:political advantage, rather than knowledge advantage, isthe dominantmechanism linking PC to the use of courts.

Why PC Firms Might ExerciseAdvantage in Courts Than in

Other Forums

Our finding that PC firms use courts more primarilybecause of political advantage, as opposed to knowledge

FIGURE 3 Moderating Effect of Legal Services Capacity on the Relationship between Former Official andthe Probability of Using Courts to Resolve Disputes


advantage, raises the question of why and under whatconditions PC firms might exercise political advantagein courts than in nonlegal forums. Our survey data arenot equipped to answer this question; however, we canspeculate on some explanations based on the Chineselegal context.

First, there may be disadvantages to using informalforums, such as asking government officials to person-ally intervene or participate in mediation processes.For government officials, it can be politically riskyto show their presence to the disputing parties. As aprivate firm owner explained, ‘‘A government officialwho personally intervenes may have to consult a highersuperior, and the networks are very complicated. So asmuch as possible, they do not wish to appear inperson. Besides, the media these days are sensitive tocorruption reports, so if their intervention is known,then the problem will get blown up.’’7 Conversely, wecan imagine that exercising political influence throughcourts is more covert, allowing any intervention to beshielded behind a legal apparatus in which judicialpersonnel are subordinated to the government.

A second possible reason is that transactingpartners are more likely to initiate litigation whenfacing disputes with PC firms and the latter acceptsthis option believing it to be to its advantage.8 As onejudge noted in a conversation, ‘‘When transactingwith a politically connected party, the other party,if lacking connections, may decline to negotiate ormediate, thinking the former would be advantaged[in an informal avenue]. So even though the costs oflitigation are higher than in mediation, they mayrather go to court.’’9 Even though it may initially seemodd why non-PC firms would be willing to confrontPC firms in courts, given evidence we have shown ofthe relative dominance of political advantage, thisbehavioral paradox is actually well-documented andconsistent with findings about Chinese citizens’ in-flated confidence of the legal system. Previous researchfound that although few citizens ever use the courts,most express significant optimism about the law(Gallagher 2006; see also Michelson and Read 2011).Gallagher finds that even those bruised from actuallegal experience do not surrender but instead becomemore tenacious (Gallagher 2006, 800). Thus, for polit-ically connected firms, in circumstances where they

perceive a reasonable cause, coupled with politicaladvantage, they could end up in court with parties ofdisputes.

As our data do not indicate whether the surveyedfirms are litigants or defendants on a case-by-casebasis, we cannot tell if PC firms are more likely tosue or be sued when they use courts. However, giventhe significant advantages of political connections, itis unlikely that PC firms only react passively to chargesfrom non-PC parties. The two possible scenariosconsidered suggest that PC firms might either exercisepolitical advantage in courts over nonlegal forums, ifthey perceive political risks to using the latter avenue,or, if sued, they might go along and confront non-PCparties in courts, which, contrary to the optimism ofinexperienced firms, remain open to manipulation. Inshort, regardless of whether PC firms sue or get sued,courts provide an avenue where PC firms can activelyor passively exercise political advantage.

Conclusion

In this article, we examined whether and how polit-ical connections shape the use of courts for disputeresolution among private firms in a transitional andauthoritarian setting. Our evidence affirms that PCfirms are more inclined than non-PC firms to employlitigation over alternative and informal means ofdispute resolution. Further evidence suggests thatthis pattern results more from the political advantageof PC firms in potentially influencing adjudication(a perverse mechanism) than from their knowledgeadvantage (a positive mechanism) in navigating thejudicial system. In short, we report a relationship ofperverse complementarity between political connec-tions and legal means of dispute resolution.

Our study underscores conditional differencesbetween early Western Europe and present-day tran-sitional and authoritarian countries such as China.The substitutive view of formal laws and informalnetworks is premised on the substantial passage oftime and absence of a strong authoritarian state inlegal development (Greif 2006; North 1991; Stiglitz2000).The edifices of law can be quickly built, but onecannot assume that norms and practices of impartialitywill follow, particularly when courts are subordinated topolitics by design. In institutional landscapes such asthose of China, we can expect a fusion of legality withpolitics and the informal with the formal.

Yet, peering further down the evolutionary pathof institutions in China, we highlight the possibilitythat even as political connections become enmeshed

7Interview with private firm, March 28, 2013.

8Note that the survey question used to generate the dependentvariable does not indicate whether a firm that reports ‘‘usingcourts’’ as the most frequent method of dispute resolution sues,defends, or does both.

9Interview with city-level judge, March 25, 2013.


with courts, the expansion of legal knowledge and use ofcourts may in turn reshape networks and even producenew legal entrepreneurs among those excluded fromguanxi to the state.10 As Padgett and McLean (2011)finds, during the course of capitalist developmentin premodern Europe, new formal transactions andoligarchic social networks influenced each otherand co-evolved over time. Our work captures animportant snapshot in a formative phase of China’slegal development, where political connections andcourts share a perversely complementary relationship,but our findings should not be interpreted as an im-mutable end in China’s continuous development.11

The limitations inherent in the data we use forour study invite future research in several directions.Like most statistical studies, our findings leave openquestions about the process of dispute resolution andlitigation, which can best be answered by qualitativemethods. Further, to test precise theories about firms’preferences of dispute resolution methods, a dispute-level dataset is necessary. However, collecting suchdata is challenging because one would have to trackeach dispute for each firm. The closest examplewe know of a dispute-level dataset is a study of 102dispute cases drawn from legal files at a French lawfirm (Lumineau and Oxley 2012). However, even thatdata reflect disputes that were already elevated toa law firm rather than the entire pyramid of disputesfor each firm. Although a firm-level dataset is limitedin many ways, it has the benefit of capturing largerpatterns, providing signposts for future work indispute-level data collection.

Finally, one might wonder if our study is limitedby the unique characteristics of the Chinese courtsystem, which, after all, is unabashedly subordinatedto political executives. However, as is well knownamong political scientists, there are many authoritariangovernments that exercise considerable control overcourts (Ginsburg and Moustafa 2008; Magaloni 2003).It is possible to replicate our study across countriesand compare national- and firm-level variances.Furthermore, our analytical strategy can be applied toa variety of analyses to tease out enmeshed mecha-nisms that are otherwise difficult to disentangle.Despite the limitations noted above, we hope thatour study sheds new light on the relationship between

formal laws and informal networks, as well as thepathway by which they interact.

Acknowledgments

We thank the Universities Service Centre for ChinaStudies at the Chinese University of Hong Kongfor providing the data of this study. For commentson earlier drafts, we thank Donald Clarke, BruceDickson, Mary Gallagher, Allen Hicken, NicholasHowson, John Kennedy, Andrew Kerner, Kyle Mayer,Dominic Nardi, Joanne Oxley, Yuhua Wang, andMariah Zeisberg. Previous drafts of this articlewere presented at the 2011 American Political ScienceAssociation Meeting, the 2012 Annual Conferenceof the International Society for New InstitutionalEconomies, the 2012 Workshop on Chinese LocalGovernance at the University of Pittsburgh, and the2013 Research in Politics Seminar at the University ofMichigan. We thank participants at these events fortheir feedback. Zhuming Yao and Kejia Wu providedvaluable research assistance.

References

Andrews, Matt. 2013. The Limits of Institutional Reform inDevelopment: Changing Rules for Realistic Solutions. Cambridge:Cambridge University Press.

Ang, Yuen Yuen. 2013 ‘‘Becoming Developmental: TheCo-Evolution of Economy and Bureaucracy in Local China.’’Presented at The Princeton-Oxford Workshop on State Buildingin the Developing World.

Clarke, Donald C., Peter Murrel, and Susan H. Whiting. 2008. ‘‘TheRole of Law in China’s Economic Development.’’ In China’sGreat Economic Transformation, eds. Thomas G. Rawski andLoren Brandt. New York: Cambridge University Press, 375–428.

Dickson, Bruce J. 2008. Wealth into Power: The Communist Party’sEmbrace of China’s Private Sector. Cambridge: CambridgeUniversity Press.

Dickson, Bruce J., and Maria Rost Rublee. 2000. ‘‘MembershipHas Its Privileges: The Socioeconomic Characteristics ofCommunist Party Members in Urban China.’’ ComparativePolitical Studies 33 (1): 87–112.

Donnithorne, Audrey. 1972. ‘‘China’s Cellular Economy: SomeEconomic Trends Since the Cultural Revolution.’’ ChinaQuarterly 52 (52): 605.

Fan, Gang, and Xiaolu Wang. 2000. NERI Index of Marketizationof China’s Provinces. Beijing, China: Economics Science Press.

Frye, Timothy. 2006. ‘‘Reputation and the Rule of Law in Russia.’’Workshop of Joint Research Study Group ‘‘EU Economy’’ andEUIJ Tokyo Consortium.

Fung, Archon. 2010. Empowered Participation. Princeton, NJ:Princeton University Press.

Galanter, Marc. 1974. ‘‘Why the ‘Haves’ Come out Ahead:Speculations on the Limits of Legal Change.’’ Law & SocietyReview 9 (1): 95–160.

10See Gallagher’s (2006, 807–08) discussion of the emergence of‘‘post-dispute activism’’ among ordinary citizens who lost in legalbattles against more connected and powerful parties, but whomorphed into ‘‘little experts’’ afterwards.

11For an elaboration of the co-evolution of economy and stateinstitutions in China since 1979, see Ang (2013).


Gallagher, Mary E. 2006. ‘‘Mobilizing the Law in China: ‘‘InformedDisenchantment’’ and the Development of Legal Consciousness.’’Law & Society Review 40 (4): 783–816.

Gallagher, Mary E., and Yuhua Wang. 2011. ‘‘Users and Non-Users:Legal Experience and Its Effect on Legal Consciousness.’’ InChinese Justice: Civil Dispute Resolution in Contemporary China,eds. Margaret Y. K. Woo and Mary Elizabeth Gallagher.Cambridge: Cambridge University Press: 204–33.

Ginsburg, Tom, and Tamir Moustafa. 2008. Rule by Law: ThePolitics of Courts in Authoritarian Regimes. Cambridge:Cambridge University Press.

Greif, Avner. 2006. ‘‘History Lessons: The Birth of ImpersonalExchange: The Community Responsibility System and ImpartialJustice.’’ Journal of Economic Perspectives 20 (2): 221–36.

Grzymala-Busse, Anna. 2010. ‘‘The Best Laid Plans: The Impactof Informal Rules on Formal Institutions in TransitionalRegimes.’’ Studies in Comparative International Development(SCID) 45 (3): 311–33.

Guthrie, Doug. 1998. ‘‘The Declining Significance of Guanxi inChina’s Economic Transition.’’ China Quarterly (154): 254–82.

Guthrie, Doug. 2002. ‘‘Informal Asymmetries and the Problem ofPerception.’’ In Social Connections in China: Institutions, Culture,and the Changing Nature of Guanxi, eds. Thomas Gold,Doug Guthrie, and David L. Wank. New York: CambridgeUniversity Press, 37–56.

Helmke, Gretchen, and Steven Levitsky. 2004. ‘‘Informal Institutionsand Comparative Politics: A Research Agenda.’’ Perspectives onPolitics 2 (4): 725–40.

Hendley, Kathryn. 1999. ‘‘Rewriting the Rules of the Game inRussia: The Neglected Issue of Demand for Law.’’ EastEuropean Constitutional Review 8 (4): 89–95.

Hendley, Kathryn, Peter Murrell, and Randi Ryterman. 2000.‘‘Law, Relationships and Private Enforcement: TransactionalStrategies of Russian Enterprises.’’ Europe-Asia Studies 52 (4):627–56.

Howson, Nicholas. 2009. ‘‘Judicial Independence and the CompanyLaw in the Shanghai Courts.’’ In Judicial Independence in China:Lessons for Global Rule of Law Promotion, ed. Randall Peer-enboom. Cambridge: Cambridge University Press, 134–54.

Jensen, Erik G. 2003. ‘‘The Rule of Law and Judicial Reform:The Political Economy of Diverse Institutional Patterns andReformers’ Responses.’’ In Beyond Common Knowledge:Empirical Approaches to the Rule of Law, eds. Erik G. Jensenand Thomas C. Heller. Stanford, CA: Stanford UniversityPress, 336–81.

Kennedy, Scott. 2005. The Business of Lobbying in China.Cambridge, MA: Harvard University Press.

King, Gary, Michael Tomz, and Jason Wittenberg. 2000. ‘‘Makingthe Most of Statistical Analyses: Improving Interpretation andPresentation.’’ American Journal of Political Science 44 (2):347–61.

Kung, James, and Chicheng Ma. 2011. ‘‘Does It Pay to Have‘‘Friends’’ in the Government? Insecure Property Rights,Political Connections, and Private Enterprise Growth inTransitional China.’’ Working Paper, Hong Kong Universityof Science & Technology.

Li, Hongbin, Lingsheng Meng, Qian Wang, and Li-An Zhou. 2008.‘‘Political Connections, Financing and Firm Performance:Evidence from Chinese Private Firms.’’ Journal of DevelopmentEconomics 87 (2): 283–99.

Li, Hongbin, Lingsheng Meng, and Junsen Zhang. 2006. ‘‘Why DoEntrepreneurs Enter Politics? Evidence from China.’’ EconomicInquiry 44 (3): 559–78.

Li, Lianjiang, and Kevin J. O’Brien. 2008. ‘‘Protest Leadership inRural China.’’ China Quarterly 193 (193): 1–23.

Liebman, Benjamin L. 2007. ‘‘China’s Courts: RestrictedReform.’’ China Quarterly 191: 620–38.

Lubman, Stanley B. 1999. Bird in a Cage: Legal Reform in Chinaafter Mao. Stanford, CA: Stanford University Press.

Lumineau, Fabrice, and Joanne E. Oxley. 2012. ‘‘Let’s Work ItOut (or We’ll See You in Court): Litigation and PrivateDispute Resolution in Vertical Exchange Relationships.’’Organization Science 23 (3): 820–34.

Macaulay, Stewart. 1963. ‘‘Non-Contractual Relations inBusiness: A Preliminary Study.’’ American Sociological Review28 (1): 55–67.

Magaloni, Beatriz. 2003. ‘‘Authoritarianism, Democracy and theSupreme Court: Horizontal Exchange and the Rule of Law inMexico.’’ In Democratic Accountability in Latin America, eds.Scott Mainwaring and Christopher Welna. Oxford: OxfordUniversity Press: 266–306.

Malesky, Edmund J., and Markus Taussig. 2008. ‘‘Where Is CreditDue? Legal Institutions, Connections, and the Efficiency ofBank Lending in Vietnam.’’ Journal of Law, Economics, &Organization 25 (2): 535–78.

Manion, Melanie. 2008. ‘‘When Communist Party CandidatesCan Lose, Who Wins? Assessing the Role of Local People’sCongresses in the Selection of Leaders in China.’’ ChinaQuarterly 195 (195): 607–30.

McMillan, John, and Christopher Woodruff. 1999. ‘‘DisputePrevention without Courts in Vietnam.’’ Journal of Law,Economics, & Organization 15 (3): 637–58.

Michelson, Ethan. 2006. ‘‘Connected Contention: Social Resourcesand Petitioning the State in Rural China.’’ Working paper,University of Indiana Bloomington.

Michelson, Ethan. 2007a. ‘‘Lawyers, Political Embeddedness, andInstitutional Continuity in China’s Transition from Socialism.’’American Journal of Sociology 113 (2): 352–414.

Michelson, Ethan. 2007b. ‘‘Climbing the Dispute Pagoda: Grievancesand Appeals to the Official Justice System in Rural China.’’American Sociological Review 72 (3): 459–85.

Michelson, Ethan, and Benjamin L. Read. 2011. ‘‘Public Attitudestoward Official Justice in Beijing and Rural China.’’ In ChineseJustice: Civil Dispute Resolution in Contemporary China, eds.Margaret Y. K. Woo and Mary Elizabeth Gallagher. Cambridge:Cambridge University Press, 169–203.

Nee, Victor. 1989. ‘‘A Theory of Market Transition: FromRedistribution to Markets in State Socialism.’’ AmericanSociological Review 54 (1): 663–81.

North, Douglass Cecil. 1990. Institutions, Institutional Change,and Economic Performance of Political Economy of Institutionsand Decisions. Cambridge: Cambridge University Press.

North, Douglass Cecil. 1991. ‘‘Institutions.’’ Journal of EconomicPerspectives 5 (1): 97–112.

Norton, Edward C., Hua Wang, and Chunrong Ai. 2004.‘‘Computing Interaction Effects and Standard Errors in Logitand Probit Models.’’ Stata Journal 4 (2): 154–67.

O’Brien, Kevin J., and Lianjiang Li. 2006. Rightful Resistance inRural China. of Cambridge Studies in Contentious Politics.Cambridge: Cambridge University Press.

Padgett, John F., and Paul D. McLean. 2011. ‘‘Economic Creditin Renaissance Florence.’’ The Journal of Modern History 83(1): 1–47.

Pearson, Margaret M. 1997. China’s New Business Elite: ThePolitical Consequences of Economic Reform. Berkeley: Universityof California Press.


Pei, Minxin, Guoyan Zhang, Pei Fei, and Lixin Chen. 2010.‘‘A Survey of Commercial Litigation in Shanghai Courts.’’In Judicial Independence in China, ed. Randall Peerenboom.New York: Cambridge University Press, 221–33.

Potter, Pitman B. 2002. ‘‘Guanxi and the PRC Legal System: FromContradiction to Complementarity.’’ In Social Connections inChina: Institutions, Culture, and the Changing Nature of Guanxi,eds. Thomas Gold, Doug Guthrie and David L. Wank.Cambridge: Cambridge University Press, 179–98.

Siegelman, Peter, and John J. Donohue. 1990. ‘‘Studying theIceberg from Its Tip: A Comparison of Published andUnpublished Employment Discrimination Cases.’’ Law &Society Review 24 (5): 1133–70.

Stiglitz, Joseph E. 2000. ‘‘Formal and Informal Institutions.’’In Social Capital: A Multifaceted Perspective, eds. ParthaDasgupta, Ismail Serageldin, and World Bank. Washington,DC: World Bank, 59–70.

Suchman, Mark C., and Lauren B Edelman. 1996. ‘‘Legal RationalMyths: The New Institutionalism and the Law and SocietyTradition.’’ Law & Social Inquiry 21 (4): 903–41.

Tang, Wenfang. 2009. ‘‘Rule of Law and Dispute Resolution inChina: Evidence from Survey Data.’’ China Review 9 (1): 73–96.

Tsai, Kellee S. 2006. ‘‘Adaptive Informal Institutions and EndogenousInstitutional Change in China.’’ World Politics 59 (1): 116–41.

Tsai, Kellee S. 2007. Capitalism without Democracy: The PrivateSector in Contemporary China. Ithaca, NY: Cornell UniversityPress.

Tsai, Lily. 2011. Connections Are Not Always Corruption: VerticalTies and Civic Participation in Rural China. Boston, MA:Lecture at Fairbank Center, Harvard University on May 9 2011.

Tsai, Lily, and Yiqing Xu. 2012. ‘‘Outspoken Insiders: WhoComplains About the Government in Authoritarian China?’’Presented at the Annual Meeting of the American PoliticalScience Association, New Orleans.

Walder, Andrew G. 2002. ‘‘Markets and Income Inequality inRural China: Political Advantage in an Expanding Economy.’’American Sociological Review 67 (2): 231–53.

Walder, Andrew G. 2004. ‘‘The Party Elite and China’s Trajectoryof Change.’’ China: An International Journal 2 (2): 189–209.

Wang, Yuhua. 2013. ‘‘Court Funding and Judicial Corruption inChina.’’ China Journal 69 (1): 43–63.

Wank, David L. 1996. ‘‘The Institutional Process of MarketClientelism: Guanxi and Private Business in a South ChinaCity.’’ China Quarterly (147): 820–38.

Wedeman, Andrew Hall. 2003. From Mao to Market: RentSeeking, Local Protectionism, and Marketization in China.Cambridge: Cambridge University Press.

Whiting, Susan H. 2010. ‘‘Contracting and Dispute ResolutionAmong Chinese Firms: Law and Its Substitutes.’’ In Dynamicsof Local Governance in China during the Reform Era, eds.Tse-Kang Leng and Yun-han Chu. Lanham, MD: LexingtonBooks, 181–224.

Yang, Mayfair Mei-Hui. 2002. ‘‘The Resilience of Guanxi andIts New Deployments: A Critique of Some New GuanxiScholarship.’’ China Quarterly 170 (170): 459–76.

Zelner, Bennet A. 2009. ‘‘Using Simulation to Interpret Resultsfrom Logit, Probit, and Other Nonlinear Models.’’ StrategicManagement Journal 30 (12): 1335–48.

Yuen Yuen Ang is an Assistant Professor in theDepartment of Political Science at the University ofMichigan, Ann Arbor, MI 48109.

Nan Jia is an Assistant Professor in the MarshallSchool of Business at the University of SouthernCalifornia, Los Angeles, CA 90089.


perverse complementarity: political connections and the ......theories that predict the declining...

Documents