Opportunistic Bargaining: Negotiating Distribution in China

Abstract Using a detailed case study of house eviction in peri-urban China as well as original data from an online survey experiment, this article explores the opportunistic bargaining phenomenon in China in which citizens leverage the policy priorities of authorities with tactics that are not approved by the state to bargain for goals beyond those promised by the state. We find that opportunistic bargaining is widely accepted by Chinese citizens and that such an inclination is encouraged by successful precedents and clear signals of an opening through which to leverage government policy priorities; however, it is dampened by unclear signals and failed precedents. We also find that opportunistic bargainers tend to hold more negative perceptions of the current regime and are less likely to abide by state rules or social norms. The characteristics of opportunistic bargaining appear to be the opposite of the dominant “rightful resistance” framework.

however, murky signals of any opening and failed precedents dampen the inclination. Further, citizens who are opportunistic bargainers hold more negative views of the regime and are less likely to abide by state rules or social norms than those who lean towards rightful resistance. Thus, the phenomenon is more detrimental to authoritarian rule than other types of resistance, which essentially reaffirm regime legitimacy.
Rather than focusing on the institutional factors in citizen participation, 2 we look into the timeand place-sensitive power negotiations between the state and society through non-institutional channels. The bargaining process and outcome are contingent on the interaction between both parties largely deviating from the formal institutions and state policies. By revealing how the bargaining process plays out, we broaden the spectrum of empirical studies on participation in strong authoritarian regimes, showing that disadvantaged social actors do not necessarily rely on state rules or social norms to bargain. 3 Instead, they can expand the political opening by forcing local authorities to concede. Since the ways in which citizens perceive and make sense of political openings shape their will and strategy to engage in opportunistic bargaining, this research adds to studies on political opportunity structures by highlighting the interaction between structural factors and actors' agency.
We reveal a different aspect of the debated paradoxical coexistence of popular contention and authoritarian resilience in China. Opportunistic bargaining, despite falling short of open civil resistance or rebellion, defies the righteousness or benevolence of the regime and refuses to play according to its rules. The temporary stability it allows local authorities to buy is only achieved through the informal, inefficient and uncontrollable brokerage between local authorities and citizensa mechanism defined by Prasenjit Duara as "state involution"which erodes state power and legitimacy. 4 Widespread opportunistic bargaining also disables the policy-feedback function, 5 thus weakening the state's control over both society and its own local agents. The challenge is even further exacerbated given its contagiousnessnot because its success encourages opportunistic bargainers to deviate further but because it motivates citizens who had previously been compliant to join the game.

Bargaining with the Authoritarian State
Strong authoritarian states such as China, with their institutional adaptability, 6 coercive apparatus 7 and time-and place-sensitive control capacity, 8 not only effectively contain popular contention but also turn it into policy input to perpetuate authoritarian rule. 9 Accordingly, contention in repressive settings is often constrained, with citizens either relying on everyday forms of resistance 10 or resorting to in-between forms of resistance that comply with rather than challenge state-sanctioned goals and rules. 11 Featuring a "hegemonic state" 12 with strong "despotic" and "infrastructural" power, 13 China serves as an ideal case to study how citizens negotiate with a high-capacity repressive regime. Kevin O'Brien and Lianjiang Li find that Chinese villagers often deploy "rightful resistance" by making claims based on official rhetoric and commitments, exploiting divisions within the state 2 Brownlee 2007;Gandhi 2010;Magaloni 2006;Manion 2016;Truex 2016. 3 Turton 1986Straughn 2005;O'Brien and Li 2006. 4 Duara 1988. 5 Lorentzen 2013. 6 Nathan 2003Shambaugh 2008;Gandhi 2010. 7 Bellin 2004. 8 O'Brien and Deng 2015Lee and Zhang 2013;Stern and Hassid 2012;Truex 2019;Read and Michelson 2008;Diamant 2000;Deng andO'Brien 2013. 9 Lorentzen 2013;Tsai 2015. 10 Scott 1985. 11 Turton 1986Straughn 2005;O'Brien and Li 2006. 12 Blecher 2002. 13 Mann 1984 and operating on the periphery of authorized channels. 14 Yongshun Cai confirms this point, showing that protestors typically behave in accordance with prevailing statutes and gain leverage by appealing to higher authorities or the media. 15 Studies further find that Chinese citizens have effectively used the quasi-democratic institutions to their advantage 16 and can go beyond "moderate forms of collective action" 17 by employing disruptive, opportunistic or disguised tactics. 18 The plethora of studies helps to map how protestors exploit political opportunities beyond everyday forms of resistance yet implies a specific set of political opportunity structures, bargaining tactics and state-society dynamics, which we argue fall in the "regime-affirming resistance" paradigm. 19 In this paradigm, the central government or the regime as a whole appears as benevolent and responsive while local state agents are often villainous; citizens, whether they sincerely support the regime or are simply being instrumental (to avoid repression or increase protest effectiveness), tend to pursue causes sanctioned or even promoted by the state, abide by state rules or normsat least on the surfaceand seek state, public or media approval. Even when citizens use law-breaching tactics, their ostensible goal is to help the party-state improve policymaking and implementation, making the protest "constructive." 20 In short, by framing negotiation "in officially approved terms," whether out of intrinsic trust in the regime's benevolence or a strategic calculation to bargain effectively, 21 citizens arguably affirm rather than challenge the regime by confirming its legitimacy, 22 restraining from more threatening mobilization forms 23 and providing the state with policy feedback. 24 Implying that the state prescribes the goals and means of bargaining in both normative and instrumental senses, such bargained authoritarianism is self-perpetuating. 25 However, citizens' hands and minds are not completely tied by the state. China scholars also find that there can be opportunities for citizens to bargain aggressively with the state. For instance, Ching Kwan Lee and Yonghong Zhang identify three micro-foundations of Chinese authoritarianism, namely protest bargaining, legal-bureaucratic absorption and patronclientelism. In particular, protest bargaining absorbs social unrest using the logic of market exchange, i.e. by buying-off protesters. 26 The specific mechanism exemplifies the state's ability to control, but it also implies an opportunity for citizens to hold the state's priorities hostage to bargain for goals beyond state promises with a broader array of tactics. However, while Lee and Zhang have described a system of incentives from the state's perspective, they have not explored the agency and motives of citizens. 27 Inspired by these studies, the following sections first develop the concept of "opportunistic bargaining" and then explore its dynamics empirically as well as its implications for authoritarian rule.
Conceptualizing Opportunistic Bargaining "Opportunistic bargaining" is the act of citizens leveraging the policy priorities of authorities with tactics that are not approved by the state to bargain for goals beyond state promises. The term "opportunistic" is not a moral assessment but rather a bargaining logic that differs from "regime affirming." More specifically, opportunistic bargainers bargain because they are able to secure concessions from the authorities; they harness the state's weak points instead of betting on the regime's benevolence; they make threats instead of filing rights-or rules-based petitions; they are not morally or instrumentally bound by state rules and social norms but are open to all tactics; and, consequently, they do not affirm the legitimacy of authoritarian rule. Table 1 compares opportunistic bargaining to the two major in-between forms of resistance in the China studies field with regard to the causes, opportunity structures, bargaining tactics and implications.
In their seminal work on the causes of rebellion, Paul Collier and Anke Hoeffler find that the opportunity and ability to build a rebellion group provide more explanatory power than severe grievances such as ethnic and religious divisions, political repression and huge inequality. 28 Similarly, opportunistic bargainers are more driven by the opportunity and their ability to win concessions than by their sense of being wronged, and they are thus more opportunistic and more sensitive to risks but care less about the righteousness of causes.
Opportunistic bargainers not only leverage one state actor against the other (for example, central versus local authorities as in rightful resistance or constructive noncompliance) but also take the state's policy agenda hostage. While local authorities may selectively implement central policies and abuse their power, 29 citizens can (threaten to) block local authorities' policy priorities, thereby incurring extra costs for the latter (such as by delayed policy implementation and the risk of revealing official incompetence to upper levels), which in turn may force the authorities to agree to the citizens' terms. This means a bigger opening, richer bargaining tactics and more mobilization resources than those offered by the state.
While rightful or constructive resisters may opportunistically use trouble-making tactics, as Xi Chen has documented, they often try to ally with the central government and signal to higher authorities or the public. 30 At least, they would pretend to do so to make the protest more effective or less risky. Opportunistic bargainers, however, work with local authorities by putting forth their own price for cooperation 31 and focus on private gains instead of policy changes. Subsequently, they do not seek state approval for their causes or justify their actions with state promises or social norms. Thus, they often do not pursue what Charles Tilly terms as WUNC (i.e. worthiness, unity, numbers and commitment) displays. 32 And, precisely because opportunistic bargainers do not count on the regime (central government) to be on their side, or at least pretend to be so, they do not have to project a benevolent image of the regime and thus do not affirm the regime's legitimacy.
In sum, opportunistic bargaining conceptually differs from rightful resistance and constructive noncompliance in terms of whether citizens base their actions on grievances outside of state promises or their ability to bring authorities to their terms; whether they abide by state rules and/or social norms (no matter if sincerely or instrumentally); and whether their actions affirm authoritarian rule or not (for example, by relying on state rules or the regime's benevolence, sincerely or strategically).
Admittedly, "opportunistic bargaining" is similar to what Jennifer Pan finds in the contention over the "minimum livelihood guarantee" (dibao 低保) programme. Pan reveals that beyond rulebased resistance, citizens may bargain and win concessions by exploiting the state's preoccupation with stability. 33 Her conceptual framework echoes ours, although there are some nuanced differences. First, the conception of "rule-based resistance" does not capture contention forms like constructive noncompliance very well. Second, we believe opportunistic bargaining is a broader concept that can also capture Pan's bargaining mechanism because: (1) opportunistic bargaining allows citizens to harness any weakness of the state rather than just the state's obsession with stability, thus it implies a bigger opportunity structure, richer mobilization resources and potentially more tactical options; (2) although seemingly a system under which everyone has the opportunity to bargain (in ways similar to opportunistic bargaining with similar implications), dibao as a method to preempt disorder is primarily applicable to a "targeted population" already on the state's radar; and (3) bargainers in Pan's study rely on highly visible dissenting tactics to gain attention while we highlight that citizens can adjust tactics based on a perceived opening, make threats rather than actually be disruptive, and seek a tacit agreement with local authorities while avoiding the media or upper levels' attention.

Opportunistic Bargaining in Eviction: A Case Study
The phenomenon of "opportunistic bargaining" was first observed by one of the authors when conducting fieldwork in a village in City X, Jiangsu province, for another project in 2009. Since it was upgraded to prefectural level in 1996, City X has undergone a massive wave of industrialization and urbanization, with its urban population rising from 613,000 to 1.64 million from 2002 to 2014. 34 In this process, thousands of rural families on the urban fringe were evicted. When the author was there, villagers were talking excitedly about a coming eviction while taking a series of preparation steps to bargain for more compensation. No one was actually resisting the eviction, other than some elderly residents who were complaining about losing their traditional rural lifestyle.
The observation was surprising, since negotiating land dispossession is anything but easy in China. Indeed, studies show that land dispossession and eviction have become a major source of 30 Chen, Xi 2009b. 31 Chinese citizens often work with local authorities to seek "justice from below." See Michelson 2008. 32 Tilly 2004. 33 Pan 2020 Data from City X's Statistical Reports on Socio-Economic Development. Original sources are not provided for reasons of anonymity but are available upon request.
state-society conflict in China. 35 Among the scholars examining the "extent and patterns of dispossession, the complex dynamics driving it, the consequences for farmers, as well as the factors shaping resistance or compliance," 36 many have highlighted the disadvantaged bargaining position of citizens, who have very limited options particularly because the state and collective ownership provide political and legal justification for predatory land-grabbing rather than for the protection of evictees. 37 Moreover, when citizens attempt to bargain or protest, local authorities often resort to coercive methods, 38 or more innovative modes of control such as bureaucratic absorption, strategic favouritism and collective pressure. 39 Other than in rare cases, such as "minor property housing" (xiao chanquan fang 小产权房) when low-level authorities and entrepreneurs collude with peasants rendering eviction a non-zero-sum game for the latter, 40 the best citizens can hope for are fair and expeditious compensation payments, transparent processes and procedures, and public participation. 41 So why would villagers in City X welcome eviction and actually win concessions without being suppressed? They were not colluding with local authorities, as in a "minor property housing" situation. To solve this puzzle, one of the authors made three field trips between 2009 and 2014 to visit eviction sites, interview evictees and officials and observe the negotiations in several villages on the urban fringe of City X. 42 The fieldwork data are supplemented and triangulated by government sources such as the official compensation scheme and news reports. The observed bargaining mechanism differs from explanatory frameworks of rightful resistance or constructive noncompliance in the following significant ways.
First, evictees overall did not invoke state promises or claims of rights, instrumentally or sincerely, to legitimize their bargaining. This might be because City X offered relatively generous compensation. Besides being incorporated into the urban social security network of unemployment insurance, pension and medical care in exchange for the loss of farmland, evictees in City X were also compensated for their houses. There were two options for house eviction: cash only or property exchange. Property exchange was more popular as relocation apartments were heavily subsidized. During the fieldwork period, relocation apartments were priced at around 700-800 yuan per square metre while commercial apartments were at least 4,000-5,000 yuan per square metre. Families who chose the property exchange option would typically receive two relocation apartments plus additional cash for their houses and miscellaneous items like trees in the courtyard, and a compliance bonus. 43 From an outsider's view, the compensation might still look unfair considering the appreciation in the value of the land when it is turned into commercial property; the local residents, however, deemed it to be decent. Given the spiralling housing prices and low farming output (which was at the subsistence level owing to the small land size per capita), they were essentially offered an opportunity to gain from land that they technically did not own and which had little cash value, in addition to the social security benefits that would come with urban residency. City X offered a more generous compensation package than nearby cities. According to an official document, before 2014 City X compensated evictees at a rate equivalent to or higher than new commercial apartments, 35 Chen, Chih-Jou Jay 2020; Luo, Andreas and Li 2016. 36 Andreas et al. 2020, 1110. 37 Fu, Hualing, and Gillespie 2014Piles 2014;Cai, Yongshun 2003;Ho 2005;Hsing 2010;Rozelle andLi 1998. 38 Sargeson 2013;Andreas 2020. 39 Chuang 2014;Deng 2017;Zhang 2020. 40 Paik andLee 2012. Note that this implies an unequal power relationship in which evictees depend on local officials and elites, thus showing less agency than that prescribed by opportunistic bargaining. 41 Cai, Meina, et al. 2020. 42 The opportunities, bargaining strategies and outcomes of eviction negotiation vary dramatically depending on the location. For an excellent study on the different dynamics of urban expansion in the urban core, urban fringes and rural fringes, see Hsing 2010. 43 See Appendix 1 in the online supplementary material for a sample compensation scheme.
while two adjacent cities only provided compensation of up to 70 per cent and 80 per cent, respectively. Plus, City X provided an extra 18 per cent eviction bonus. 44 Second, while evictees generally welcomed the eviction and did not complain about the official compensation scheme, they did not simply accept the offer and move either. Instead, they made extra efforts that often deviated from state rules and social norms for extra compensation. One villager explained their shared rationale: "it is the CCP's money, only a fool would not take it!" 45 Since the compensation depended on the size and quality of a villager's house, evictees upgraded, expanded or furnished their homes upon hearing about the eviction. Because local authorities had stopped issuing permits, these unauthorized projects were technically illegal and thus not eligible for compensation. Nevertheless, most families rushed to complete these projects. Evictees also planted large numbers of ginkgo and other fruit trees in their courtyards, sometimes so densely that the tree saplings looked like bamboo forests. While planting trees in a courtyard needed no authorization, it was prohibited in the official compensation schemes. Entrepreneurial families even changed their dwellings into commercial or industrial properties for a better compensation offer. As it was difficult to obtain licenses via the normal legal avenues, some evictees simply borrowed or fabricated business certificates with or without the acquiescence of local cadres. Some more creative families expanded or set up new (practically defunct) industrial enterprises for extra compensation. Such tactics were in breach of the regulations, or at the very least dubious, as evictees rarely obtained proper licenses and they did not start real businesses.
Since compensation also depended on the size of a family, some large families split into smaller ones (usually a married couple with a child), allowing each new family to construct their own dwellings to gain extra benefits such as a compliance bonus for moving ahead of the deadline and rental subsidies if the relocation apartments were not ready to move into. Although these actions are legal, they are not entirely compatible with the social norms that value family ties and filial piety. For small families, adding an extra member to the family book could bring in more compensation. The practice is not quite lawful and requires the complicity of local officials, thus it was less openly discussed. Yet these practices are by no means rare, as confirmed by open reports across China. 46 Some of the tactics employed by City X evictees, although seemingly individualistic and covert, were quite widespread among the evictee communities. This is logical since the evictees, who were living within the same community and had close social ties, learned from each other. 47 The ethnographic evidence shows that almost all evictee families engaged in opportunistic bargaining, albeit to different degrees with various tactics. Their preferred course of action was primarily shaped by resource constraints, including having access to (1) the monetary investment needed for projects such as upgrading homes and planting trees, and (2) the social resources affecting the viability of certain options such as concocting a business. As monetary resources were more readily available to villagers in City X, we observed more families upgrading their homes and planting trees. In fact, as many families started preparing for eviction simultaneously, the prices of tree saplings, construction materials and labour all rapidly increased. 48 This group behaviour made the evictees feel more assured in using these tactics, as their neighbours were all doing the same. It also increased their bargaining leverage by (1)  authorities to turn them down, as refusing one family might trigger the anger of an entire village. After all, "the law fails where violators are legion" ( fa bu ze zhong 法不责众).
Third, evictees in City X aimed at reaching a tacit agreement with local authorities instead of drawing the attention of the upper levels or public to local conditions. This was probably why the evictees, despite asking for much more than what the local government was offering, not justifying their claims with rightful causes and often using law-breaching tactics, were still accommodated rather than suppressed. This was not because local officials were colluding with evictees (as in the "minor property housing" case) or were sympathetic to them (deeming their claims or bargaining tactics to be rightful). On the contrary, local officials were clearly unhappy about the evictees "demanding exorbitant prices." 49 It was evictees winning a deal with local authorities through opportunistic bargaining. Evictees knew they could ramp up the costs for local authorities to achieve policy goals such as progressing local development projects and stability maintenance (the fact that City X is the hometown of a top national leader makes this even more crucial for local authorities). Although local authorities could have refused the evictees' demands given that they asked for far more than what the government had promised and the law-breaching tactics were not sanctioned by the state, these policy priorities helped to convince local authorities that concession was more appealing than suppression. Meanwhile, in order to reach a tacit agreement of exchanging cooperation for extra compensation, evictees generally did not attempt to attract the attention of the public or upper levels.
The case of City X is not idiosyncratic. In addition to the cases recorded by scholars, 50 searching two specific law-breaching tactics, "unauthorized building" (weijian 违建) and "fake divorce" ( jialihun 假离婚) with "eviction" (chaiqian 拆迁) in the China Core Newspapers Full-Text Database yields over 7,200 reports between 2000 and 2017 across China (Appendix 2). While this does not mean evictees succeed in all these cases, it does imply that many evictees expect such tactics to work. As one local official from Chongqing put it: Villagers want to push their luck. They believe that as long as they make it a fait accompli, the government would have to compensate them regardless. More building means more compensation; less building means less compensation; no building means zero compensation. With such a rationale, villagers compete with each other, leading to massive illegal buildings. 51 The Survey Experiment While the analysis of eviction negotiation in City X and elsewhere reveals the mechanism of opportunistic bargaining nicely, it suffers from limited external validity. Therefore, we conducted a survey experiment, which we first piloted among college students in Beijing in March 2018 and then fully implemented online from 15 November to 15 December 2018 using a survey company to recruit respondents across China. 52 With 1,998 complete responses, the survey yields an analytical sample of 1,824 observations after dropping rushed survey responses (completed in less than four minutes, n = 97) and any that failed manipulation check questions (n = 77).
The survey experiment is primarily designed to test if opportunistic bargaining as an eviction negotiation strategy is indeed broadly accepted and, if so, which conditions encourage or dampen the inclination. For this purpose, we designed a house eviction vignette, randomly exposing 49 News source is available upon request. respondents to eight hypothesized scenarios in which their relatives were facing eviction. 53 We measured the opportunistic bargaining inclination in eviction (hereafter OB E ) by asking if respondents would recommend such a strategy (definitely not = 1; maybe not = 2; maybe yes = 3; definitely yes = 4). We probe the conditions that may dampen or encourage OB E by manipulating signals of the opening to leverage local authorities and the perceived prospect of success. We expect that clearer signals of the opening and successful precedents will increase OB E while murkier signals of opening and failed precedents will reduce the inclination. In addition to the experiment, we also asked further questions in our survey to explore whether OB is applicable beyond eviction, whether it is associated with certain socioeconomic and attitudinal traits, and whether it differs empirically from other patterns of resistance, as discussed in Table 1. We measured the general inclination to opportunistic bargaining, rightful resistance, and constructive noncompliance. As Table 2 shows, respondents were asked to rate four general statements, two on opportunistic bargaining (OB 1 and OB 2 ), and one each on rightful resistance (RR) and constructive noncompliance (CNC).
Based on the earlier discussion in Table 1, opportunistic bargaining, rightful resistance and constructive noncompliance imply different socio-political attitudes and behaviour. To examine the reasoning, we included the following variables in our survey in addition to regular demographic variables. First, we measured respondents' political attitudes such as political trust and confidence in the government performance. Second, opportunistic bargaining implies readiness to deviate from social norms or state rules while rightful resistance suggests the opposite; constructive noncompliance shall fall in-between. We asked a set of questions to measure respondents' inclination to abide by social norms and the inclination to play with state rules. Finally, aiming at a tacit agreement with local authorities, opportunistic bargainers shall avoid media or public attention rather than try to attract it. We asked whether respondents would advise their relatives to accept a media interview request if local cadres were found to be gaining extra compensation with unlawful tactics. For a full list of the variables and summary statistics, see Appendix 6.

Treatment design in the experiment
The case study and the literature suggest that opportunistic bargaining may be conditional on two types of factors: the example of others and the perceived opening to leverage government policy priorities. Thus, we constructed the treatments by citing one or more of the following five conditions: the neighbours of the respondents' relatives were engaging in opportunistic bargaining (neighbours); some opportunistic bargainers had failed (lose); some opportunistic bargainers had succeeded (win); local authorities intended to maintain social stability (government's stability 53 Appendix 5 (in the online supplementary material) shows that all covariates except "eviction experience" are perfectly balanced among the eight groups, indicating successful randomization. need); and local authorities hoped to complete the project sooner to generate economic benefits and avoid losses (project need).
In total, we have seven treatment groups. With Treatment 1 (including only the "neighbour" condition), we intend to see whether respondents will be emboldened upon learning about other opportunistic bargainers. We expect this to be the case because (1) opportunistic bargaining is often a group behaviour, and (2) more participants often signal richer mobilization resources and a greater opening for bargaining. 54 The impact may not be that strong because the control group was also exposed to a latent treatment as all respondents were asked if they would recommend opportunistic bargaining tactics to their relativeswe have to refer to such behaviour to measure a respondent's inclination. We deem this acceptable because while this problem weakens our treatment effects in all testing groups, it does not invalidate our findings. We also included the "neighbour" condition in all other six treatments because it makes little sense to say some failed or won without mentioning that there were people actually engaging in opportunist bargaining in the first place. In addition to "neighbour," Treatments 2 and 3 included information about the prospect of losing or winning, and we expect strongly negative (T2) and positive (T3) effects, respectively. In Treatments 4 and 5, besides the "neighbour" factor, we introduced local authorities' stability maintenance and project needs to test whether citizens will leverage authorities' policy priorities. We do not expect a strong impact from T4 or T5 because each of the two treatments alone sends only a murky signal of the opportunity. While the stability maintenance or project needs may incentivize authorities to pay citizens off, they may also induce suppression. 55 The murkiness is designed to imitate the reality that citizens do not possess explicit information about authorities' intentions or capacity. To reveal citizens' choices in different informational conditions, we combined local authorities' stability and project needs in Treatment 6, expecting a strong positive impact as the two conditions together make the signal of an opening clearer. We combine all conditions except the prospect of losing in Treatment 7, expecting a strong positive impact with both explicit signals of an opening and successful precedents. Table 3 summarizes the treatment assignments and the expected outcomes. For the detailed setup of control and treatment groups, see Appendix 7.
Before presenting the findings, it is important to address two concerns with the design. First, one may question if an online survey is sufficient to elicit a representative sample. While this concern is justified, the experimental design enables us to test the impact of treatments, so even nonrepresentative samples can generate valid findings. 56 Second, there is the question of whether respondents would answer the questions honestly, since using law-breaching tactics is sensitive. We believe this shall not invalidate our findings given the experimental design. We further alleviate this issue by introducing the scenarios in which not the respondents but rather their relatives are experiencing eviction. Figure 1 illustrates the experiment results. Merely knowing of the existence of OB (T1) did not significantly increase the tendency to engage in OB. Respondents exposed to failed cases (T2) are less likely to recommend opportunistic bargaining than the control group (a drop of 6.9 per cent). For respondents exposed to successful cases (T3), OB E increases by 5.3 per cent compared to the control group. Signals of an opening through which to leverage policy priorities work as expected, showing a positive impact when combined (T6) compared to the control group (increases by 4.6 per cent), but not separately (T4 and T5). When the opening signals are combined with winning cases (T7), OB E significantly increases by 7.3 per cent compared to the control group. As expected, respondents are more inclined to engage in opportunistic bargaining if they learn that others are doing so, if such a strategy works and when they can exploit local authorities' policy concerns. The results also show that it is important for respondents to receive clear and explicit signals about a political opening for bargaining. When respondents are only told that local authorities want to maintain stability (T4) or finish the project sooner (T5), in each scenario it was not selfevident to respondents whether the authorities would accommodate opportunistic bargaining or deal with it forcefully. Such uncertainty helps to amplify silence. 57 When the two conditions appear together in T6, the political opening signals are clearer to respondents, encouraging opportunistic bargaining. The inclination further increases when clear political opening signals are accompanied by successful precedents (T7). Overall, while citizens are quite risk averse, explicit opening signals and successful precedents can effectively mobilize them to choose opportunistic bargaining. To check the robustness of the experiment findings, we conducted regression analysis with additional data from our survey by controlling for demographics, political attitudes and the inclination to abide by social norms or state policies. The results remain consistent. For details, see Appendix 8.

Experiment results
Since we have multiple treatment groups, some of which are closely related, we compute Romano-Wolf adjusted p-values to correct for the probability of making a false discovery among multiple comparisons. 58 Our findings are generally robust to the test: respondents exposed to failed cases (T2) / successful cases (T3) are less / more likely to choose opportunistic bargaining than respondents in the control group; respondents are more inclined to engage in opportunistic bargaining if there are clear opening signals and successful precedents (T7). The only differences are that T3 and T6 now occasionally have positive but insignificant effects. 59 We have also explored if our treatments have heterogeneous effects across respondents' demographics. The findings suggest that more resourceful citizens are more willing to take risks, while those who are less resourceful are more risk averse in opportunistic bargaining. 60

Opportunistic Bargaining, Rightful Resistance and Constructive Noncompliance
The experiment confirms that opportunistic bargaining as an eviction-negotiation strategy is broadly accepted among Chinese citizens and is affected by signals of a political opening and precedents. Then, as a broader concept, does opportunistic bargaining go beyond eviction? How does it compare to rightful resistance and constructive noncompliance? To answer these questions, we turned to an exploratory analysis by running 70 ordinary least-squares regression models involving each of the 14 attitudinal and behavioural variables we collected with control variables on our five dependent variables. We restricted the sample of analysis to the control group to avoid posttreatment bias (N = 194). We have visualized the results in Figure 2 with each subplot showing the results of five models, i.e. the coefficients with 95 per cent confidence intervals of each attitudinal and behavioural variable on all five dependent variables. 61 The results confirm that opportunistic bargaining bears a different set of attitudes from rightful resistance while showing similar features to constructive noncompliance. As far as the political attitude variables are concerned, OB and CNC are in general associated negatively with trust in the central and local governments, preference for the regime, confidence in regime stability or its economic prospects, and belief in public interests over private interests. RR presents an opposite pattern on many of these variables. Those who agree more with RR have more trust in the central government and a higher evaluation of the regime's performance, while at the same time showing more confidence in regime stability and its economic prospects. Regarding the norms-deviating measures, OB and CNC are positively associated with evading fares, defaulting on reservations and paying bribes, while RR shows opposite traits consistently. Regarding political participation, RR is associated with more institutional participation like lawsuits and petitions while OB is more associated with protests (the result for CNC is not statistically significant). Finally, neither OB nor CNC welcome publicity, while RR is ready for a media interview.
Contrary to our expectations, constructive noncompliance has shown highly similar results with opportunistic bargaining. This might be because our CNC variable failed to differentiate the two concepts properly because opportunistic bargainers may also agree with the CNC statement that "law-breaching protest is acceptable if one's lawful rights are violated." 62 Nevertheless, although the comparison between opportunistic bargaining and constructive noncompliance is inconclusive, the differentiation between opportunistic bargaining and rightful resistance is explicit and 58 Clarke, Romano and Wolf 2020. 59 For detailed results, see Appendix 9 (available online). 60 For details, see Appendix 11 (available online). 61 For the full tables, see Appendix 10 (available online). 62 Tsai 2015, 264-65. systematic. Opportunistic bargainers, while focusing primarily on private gains rather than openly contesting state policies or demanding democracy, are not regime-affirming as they hold negative views of the regime and deviate from state rules. Building on previous studies, 63 we can safely conclude that opportunistic bargaining situates at the intersection between open and hidden activism as well as that between regime-affirming and regime-challenging. 64

Conclusion
Since the 1989 Tiananmen movement, China has gradually created "a multiagency juggernaut" of stability maintenance with a formidable coercive apparatus, 65 a huge amount of resources and cadre  Table 2); each spike maps the coefficient and 95% CI of the independent variable (the title of the respective subplot); control variables included. incentives, 66 as well as highly innovative strategies and techniques to defuse social unrest. 67 This has created a dire situation for citizens who wish to bargain for benefits. However, we articulate that even in such a highly repressive setting, citizens can bargain for promises beyond those on offer by the state through opportunistic bargaining. Unlike regime-affirming resisters who sincerely count on or strategically exploit state rules or the regime's benevolent image, opportunistic bargainers do not seek state or social approval for their means and goals in bargaining. Instead, they take local authorities' policy priorities hostage in order to "extort" benefits. As Kevin O'Brien notes, citizens do not just "fill spaces that the state and its reforms create, but also push against boundaries in ways that cannot be read straight off an opportunity structure" and may create their own opportunities by persistently probing the limits. 68 Through a case study on eviction and a broader survey experiment, we have empirically confirmed that opportunistic bargaining is widely embraced by Chinese citizens, especially when there are clear signals of political openings and successful precedents. In addition, opportunistic bargainers demonstrate opposite traits to rightful resisters, as they harbour negative perceptions towards the regime and are ready to deviate from state rules or social norms. Our findings add to the burgeoning literature on China's bargained authoritarianism by probing the conditions and mechanisms through which citizens may aggressively bargain with government as well as the subsequent implications.
Without any intention to make a moral judgement, we deem opportunistic bargaining to be an important part of Chinese citizens' continuous endeavour to define, defend and demand their social, economic and political interests. For Kevin O'Brien, Chinese villagers are "occupying an intermediate position between subjects and citizens." 69 Greg Distelhorst and Diana Fu identify three types of citizenship in China: subjecthood (citizens positioning themselves as subalterns before benevolent rulers), authoritarian legal citizenship (citizens appealing to the state's formal legal commitments) and socialist citizenship (citizens appealing to officials' moral duties). 70 Since opportunistic bargainers make claims not out of "rules consciousness" 71 or "rights consciousness," 72 but because they have the ability to bargain, they are not compliant subjects that define their goals and the means within state rules, nor rights conscious actors pursuing a more complete citizenship. They are hardly "villains" or "victims"the stereotypical images of evictees and protesters at large that are often portrayed by the media, state and academics. 73 Rather, they are rational, cunning actors exploiting existing opportunities under the regime to the maximum. Also, as opportunistic bargaining does not intend to attract the media's attention or to signal local conditions to upper levels, it is more a form of "destructive superficial compliance." While it helps to buy stability for now, the process resembles the "state involution" of early 20th-century China where the state's fiscal and functional growth gave rise to the informal, inefficient and uncontrollable brokerage structure of tax collectors, clerks, middlemen and bullies, 74 which in turn eroded the state's capacity to rule. Indeed, opportunistic bargainers' negative perception of the regime and their reluctance to follow the state rules are already telling in this regard.
Admittedly, our research has some limitations. First, as mentioned above, further studies may look further into the conceptual differences between opportunistic bargaining and constructive noncompliance since our design has not achieved this goal. Second, we are not looking at more difficult spaces where citizens may not aim at winning private gains while both local and central authorities are 66 Wang and Minzner 2015;Xie 2013;Gao 2015. 67 O'Brien andDeng 2015;Hassid and Sun 2015;Kan 2013;Biddulph 2015. 68 O'Brien 2013, 1053. 69 O'Brien 2001, 426. 70 Distelhorst and Fu 2019. 71 Li 2010. 72 Perry 20082009. 73 Sargeson 2012. 74 Duara 1988 unlikely to make concessions (for example, Uyghur independence or Falun Gong). While our survey experiment has shown that opportunistic bargaining is widely observable among citizens out of the context of eviction and some existing studies echo our findings, 75 it will be productive for future studies to map out the realm where opportunistic bargaining happens and where it does not.
Supplementary material. To view supplementary material for this article, please visit https://doi.org/10.1017/ S0305741022001370