The bloody suppression of the student movement at Tiananmen Square in 1989 gave the Chinese government a well-deserved reputation for repression. Going into the 1990s, protests against state policies – such as compensation in housing demolitions – rarely succeeded and could result in harsh repression. As one evictee whose shop was being demolished told a reporter in 1997, “I am not protesting although I am unhappy … it is useless to protest – if you do, you go to jail and the others get the best apartments.”Footnote 1 By the early 2000s, protests were on the rise, but the popular perception of China remained one of a highly repressive state. The New York Times artfully captured this popular image in a 2005 article entitled, “Land of 74,000 Protests (But Little Is Ever Fixed).”Footnote 2 This lack of responsiveness would hardly surprise scholars of authoritarian politics. China, after all, lacks institutions that commonly promote responsiveness in other authoritarian states, such as national elections and powerful opposition parties.
The Puzzle
This popular understanding misses an important change that has taken place in China over the past decade. Authoritarianism in China has become much more responsive. By the early 2000s, local governments frequently bought off protesters with cash in order to quiet them down. Repression did not disappear, but concessions became a much more prominent response than they had been previously. The central government introduced a raft of populist policies that addressed protesters’ grievances. In the countryside, the central government greatly increased the compensation given to farmers who were dispossessed of their land in government-led land expropriations and called on provincial governments to establish a social security system for landless farmers. Some provinces quickly followed suit, although others dragged their feet for years. In urban areas, the central government abolished administratively set compensation for home owners whose houses were demolished, mandating instead that home owners be given market compensation. Similarly populist policies emerged in a variety of other areas, including taxation, pensions, and labor relations. State responsiveness to protests was not uniform, however. In other areas, such as policies toward benefits for community and substitute teachers, no populist policies have emerged.
In sum, the Chinese government has been responsive to protests at the local, provincial, and national levels. This responsiveness raises a series of puzzles. Under what conditions can citizens in authoritarian regimes influence policy making through protest? Why have local governments been much more responsive to the demands of protesters in recent years? Why have some provinces aggressively promoted populist policies in response to protests, while others have stalled? Why has the central government responded with extensive policy changes in response to protests from some social groups and with only moderate or no policy changes to protests from others?
Protest and Policy Making in Authoritarian Regimes: Available Explanations
Does protest influence policy making in authoritarian regimes? There are good reasons to believe that influence is unlikely. Recent studies suggest that even social movements in advanced industrial democracies achieve only some level of influence 50 to 70 percent of the time.Footnote 3 Closed authoritarian regimes, moreover, often try to ban or prevent virtually all forms of popular protest.Footnote 4 When protests do arise, “the quintessential governance strategy in closed autocracies is to reward loyalists and repress independent citizens and movements.”Footnote 5
A small but growing body of literature, however, suggests that China is much more responsive to protests than most closed authoritarian regimes.Footnote 6 In this book, I argue that China represents a case of responsive authoritarianism.Footnote 7 I adopt a definition of responsiveness drawn from studies of accountability in democracies. As Manin, Przeworski, and Stokes note, “a government is ‘responsive’ if it adopts policies that are signaled as preferred by citizens. These signals may include public opinion polls; various forms of direct political action, including demonstrations, letter campaigns, and … elections.”Footnote 8 By responsive authoritarianism, I refer to a regime that proactively monitors citizen opposition to state policies and selectively responds with policy changes when it gauges opposition to be particularly widespread. Responsiveness, moreover, is intended to strengthen the state and avoid the development of a revolutionary opposition rather than being a sign of state weakness.
Compared to other aspects of protest in China, the influence of protest on policy making has received comparatively little attention. Indeed, as Tarrow recently noted, “scholars of Chinese politics have not yet systematically examined relations between protest and policy response.”Footnote 9 In order to identify the gaps in our understanding of the relationship between protest and outcomes in China, it is helpful to approach the question from the perspective of the policy-making process. The policy-making process can be broadly conceived of as occurring in three stages: agenda setting, policy formulation and adoption, and policy implementation. Because this book explores only the agenda-setting and policy formulation and adoption stages, I will limit my discussion to these two.Footnote 10 In so doing, I show that while scholars have explored the agenda-setting stage to a certain extent, the policy formulation and adoption stage has been largely neglected.
The Agenda-Setting Stage
Unsurprisingly for an authoritarian regime, scholars have suggested that it is difficult for petitions and protests to influence high-level politics.Footnote 11 Chen, for example, argues that while it is possible for petitions to lead to policy changes, the petitioning system is “deeply flawed and severely inefficient in channeling interest articulation.”Footnote 12 Instead, Chen and Xu found that the support of a mass organization led by an official with close ties to the top Party leaders was necessary to place protesters’ demands on the agenda.Footnote 13 Mertha, in his study of mobilization surrounding hydropower policy in China, found that protest was actually counterproductive. Through his compelling case study of the Pubugou dam, where tens of thousands of landless farmers protested against low compensation, Mertha argues that “protests had absolutely no effect on the dam project.”Footnote 14 Instead, Mertha found that lobbying by nongovernmental organizations (NGOs) and critical reporting by activist journalists were much more effective at influencing the agenda. In particular, NGO and media activists were able to effectively change the “issue frame” away from economic development and toward environmental protection and cultural preservation.
Even among scholars who argue that protests have influenced policy making, much of the emphasis has been on the role of public opinion and the media in mediating the impact of protests. Reilly has argued that nationalist protests provide information to the regime about public opinion on foreign policy. Although this information about public opinion is heavily skewed toward the views of a negative and engaged segment of the public, it is precisely this highly mobilized minority that authoritarian leaders fear. Reilly examines the role of several state institutions in monitoring public opinion, particularly the Ministry of Foreign Affairs, the Statistical Bureau, the Public Security Bureau, and the Propaganda Ministry.Footnote 15 Cai has argued that larger and more frequent protests strain the legitimacy of the state by turning the private grievances of the protesters into public knowledge. The media, in turn, sometimes support protesters by publicizing their plight. Once the public is aware of protesters’ grievances, the state can no longer claim ignorance and must adjust policies or risk losing legitimacy.Footnote 16 Scholars have also noted that petitions provide a source of information to the central government.Footnote 17 In particular, petitions can provide information to the government on where policies have been implemented poorly.Footnote 18
The Policy Formulation and Adoption Stage
Scholars have adopted three approaches to the policy formulation and adoption stage: fragmented authoritarianism, advocacy by mass organizations, and cost-benefit analysis. Each approach offers a different explanation for why officials might support policy changes. The key insight of the fragmented authoritarianism model is that by integrating the interests of implementation agencies into policy making itself, the policies that emerged out of bureaucratic bargaining were often significantly different from the initial goals of policy makers at the top.Footnote 19 Early research by O’Brien and Li uncovered a “structural opening” between central and local governments whose interests diverge that allows protests to influence policy implementation.Footnote 20 Mertha builds on the notion of a structural opening, arguing that official organizational mandates can create “disgruntled officials” who oppose a particular policy. In order to defend their organizational interests, disgruntled officials form coalitions of broad-based support. For example, environmental and cultural protection agencies banded together with NGOs and media outlets to oppose large hydropower projects.Footnote 21
Chen and Xu have shown that mass organizations also advocate policy changes on behalf of protesters. Designed as “two-way transmission belts,” mass organizations in theory both channel the demands of the masses to Party leaders and assist the state in policy implementation. In practice, mass organizations tend to neglect the former role in favor of the latter. The Chinese Disabled Persons Federation (CDPF), however, took up the cause of disabled taxi drivers because the protests of these drivers brought the CDPF’s representation role into conflict with its policy implementation role. As protests strained its legitimacy, the CPDF was compelled to argue on behalf of its constituents.Footnote 22 The All China Federation of Trade Unions, another mass organization, has similarly advocated policy changes in response to worker protests.Footnote 23 Chen and Xu thereby highlight an additional reason why officials might support demands mobilized by citizens: because they have an institutionalized representation role.
Cai, by contrast, adopts a cost-benefit approach to protest and policy making, conceptualizing the cessation of protest as a benefit that preserves regime legitimacy. Cai argues that the central government cares more about protecting the legitimacy of the state than local governments, making the central government more inclined toward policy changes.Footnote 24 Policy changes are costly if they require government expenditures or if they result in a loss of revenue.Footnote 25 As a result, “the cost determines the pace and degree of policy adjustment.”Footnote 26
Gaps in the Existing Literature and the Puzzle of Responsiveness in Land Takings and Demolitions
As the sections that follow will show, this book builds on insights from the existing literature and particularly the fragmented authoritarianism approach. Nonetheless, there are significant gaps in the existing literature. First, while scholars have long noted that petitions and protests convey information to the regime, the link between petitions signaling information and change in formal policies has largely been inferred rather than demonstrated through policy studies that follow the policy-making process from start to finish.Footnote 27 In particular, we know little about how the petitioning system processes information transmitted by protests and petitions and how this influences the agenda. Does the state respond equally to all information transmitted by protests and petitions? If not, what influences decisions about how the petitioning system allocates attention?
The second issue is the level of aggregation of the state. While O’Brien called for future research to more fully “unpack the state” a decade ago, our progress since has been somewhat limited.Footnote 28 This book represents an effort to move beyond the central–local dichotomy of previous studies by showing that provincial governments have varied dramatically in their support for policy changes in land takings. Some provinces adopted policy changes even before the central government mandated that they do so, while other provinces delayed adoption of such policies for a decade or more. Moreover, this book also unpacks the central government, showing that central ministries have not been completely united in their support for policy changes. While the Ministry of Land Resources has supported policy changes in land takings, for example, the Ministry of Railroads has opposed them. This suggests that both the central and the local state must be further disaggregated to advance our understanding of protest and policy response.
The third issue is why state officials advocate for or oppose policy changes. As this book will show, the Ministry of Land Resources and the Ministry of Housing and Urban-Rural Development were the strongest advocates for policy changes in land takings and demolitions, respectively. Yet, far from being harmed by the existing policies, these ministries benefited considerably from them because land takings and demolitions brought considerable funds and authority to these two ministries.Footnote 29 Unlike mass organizations, moreover, these ministries lacked an institutionalized representation function. Likewise, local governments often benefit even more from land takings and demolitions, yet some provincial governments have advocated for policy changes. In many cases, moreover, the provinces that adopted social security for landless farmers were among the provinces where the population of landless farmers was the greatest and the costs of the programs therefore the highest. This suggests that the effect of fiscal costs on policy responsiveness may be more complex than previously thought.
Finally and most important, there is the issue of the policy formulation and adoption stage itself and how institutional arrangements influence responsiveness. Mertha’s excellent study of the policy-making process in dam building has highlighted the continuing relevance of the fragmented authoritarianism model, but he does not examine formal legislative policy making of regulations and laws. While Cai’s study highlights the importance of costs as a constraint, the role of formal policy-making institutions – such as the State Council and the National People’s Congress – in mediating the impact of protest has remained unexplored.
Protest Signals and Agenda Setting
Protests signal information to policy makers. Scholars of democracies disagree about whether the information environment is rich or poor for policy makers.Footnote 30 The overwhelming consensus in authoritarian politics, however, is that dictators are starved for information. As Policzer put it, “[D]ictators may be powerful, but they are often also information-poor.”Footnote 31 This is particularly the case for politically closed communist regimes, which “are not well equipped to respond to the changing demands and needs of society – precisely because they are intrinsically top-down ‘mobilization’ regimes rather than regimes that possess the feedback mechanisms to hear and respond to aggregated social needs and demands.”Footnote 32
The multiple levels of officials between rulers and the citizenry in China exacerbate monitoring problems.Footnote 33 Marketization and decentralization during the reform period, moreover, have only caused monitoring problems to become more acute.Footnote 34 Making matters worse, local officials frequently manipulate information.Footnote 35 Wallace, for example, has shown that provincial officials are more likely to “juke the stats” by inflating gross domestic product (GDP) growth figures during times of leadership turnover when there are possibilities for promotion.Footnote 36 As one top Chinese Communist Party (CCP) leader lamented, “[T]he most difficult thing for a leadership unit to do is to collect accurate information at the basic level.”Footnote 37 Consequently, studies have found that “lower-level officials in Leninist systems have a strong incentive to lie to their superiors … the quality of information available to leaders in such systems is generally poor.”Footnote 38
Protests and petitions are important for precisely this reason.Footnote 39 Studies of social movements in the United States have argued that citizens can use protests to signal information on their policy preferences to politicians.Footnote 40 Arguments about the influence of protest on policy making are almost always grounded (at the very least implicitly) in the theory of democratic representation.Footnote 41 The theory holds that legislators seek first and foremost to win reelection. As a result, they will support or oppose policies on the basis of “the number of votes that they think their actions will win or lose them at election time.”Footnote 42 Consequently, the number and size of protests provide information on the extent of support for policy changes among the electorate. Applying this approach to a closed authoritarian regime (which, by definition, lacks elections) presents obvious difficulties. I will return to this issue shortly.
Scholars have only very recently begun to apply the signaling approach to the study of protest in China.Footnote 43 Much of the research has focused on signals emanating from the state. Weiss has shown that the state selectively facilitates or represses nationalist protests in order to signal information to foreign governments about the Chinese state’s resolve, hawkish commitment, or credible reassurance.Footnote 44 Stern and O’Brien note that the state signals information to citizens through two main channels: direct experiences with state agents and indirect communication of official preferences.Footnote 45 Repression is the most direct experience with state officials and a powerful signal. Hurst has shown that by repressing protesting state-owned enterprise (SOE) workers, cash-strapped local governments in the North-Central and Upper Changjiang regions were able to deter future mobilization by signaling that protests would not be tolerated.Footnote 46 In effect, instances of repression in these regions created “stories about repression” that indirectly signaled the boundaries of acceptable behavior.Footnote 47
Yet, as Stern and O’Brien note, “A fully interactive account of signaling will also require special attention to moments at which bottom-up initiative induces the state to respond.”Footnote 48 In this vein, scholars have begun to examine the role of protest in signaling information from protesters to the Chinese state. Scholars have identified several kinds of information that protests can signal to the state. First, protests and petitions can signal information to the central government about corruption by local officials.Footnote 49 Second, protests can signal to the central state the location of discontented communities, who the state can then buy off with ad hoc transfers.Footnote 50 Finally and most relevant for our purposes, petitions and protests can signal information about citizen satisfaction with policy implementation.Footnote 51
An important consideration – and an issue often overlooked in the protest signaling literature in the United States – is the capacity of the regime to receive the signals sent by protesters. Most quantitative studies of agenda setting rely on event counts of media reports on protests.Footnote 52 As Koopman notes, “[A]uthorities will not react to – and will not even know about – protests that are not reported in the media.”Footnote 53 Yet, if press censorship ensures that the media do not report most protests, how will the state receive the signal of protests?
In monitoring social unrest, China relies most heavily on the petitioning system (xinfang zhidu). The petitioning system serves an information-gathering function.Footnote 54 As O’Brien and Li have noted, the petitioning system can be understood as a form of “fire alarm,” alerting state officials to emerging problems.Footnote 55 The petitioning system has officials at all levels of the government, from the State Bureau of Letters and Visits in Beijing down to the lowliest county in the most remote areas of Guangxi. As Chapter 4 will show, petitioning offices at all levels receive petitions and meticulously sort and classify petitions. The petitioning system then regularly generates reports that process this information and further transmit it to leaders and relevant departments at every level.
Yet, while the capacity of the petitioning system to receive information about citizen policy preferences has been widely noted, scholars have paid much less attention to how the petitioning system allocates its attention and the impact this has on policy making.Footnote 56 The importance of attention allocation cannot be overstated. Throughout the 2000s, the petitioning system received 10 to 13 million petitions every year.Footnote 57 This means that in as much as the petitioning system is a fire alarm, it rings constantly. The petitioning system is overwhelmed with information. Instead of searching for information on social unrest, the petitioning system must winnow and prioritize information.Footnote 58
Strong protest signals from some groups have a “crowding-out effect” that drowns out weaker signals from other groups.Footnote 59 With regard to policy making, an agenda-setting process occurs in which “information is prioritized for action, and attention allocated to some problems rather than others.”Footnote 60 The petitioning system plays a role in the agenda-setting process by helping officials to determine the strength of the protest signals and to allocate attention to the most serious conflicts.
How does the petitioning system determine the intensity of the signal? The petitioning system differentiates petitions along three main dimensions, which collectively determine the strength of the signal. The first and most important component of the signal consists of the number of protest events. Cai has argued that the number of protests is important because as protests grow more numerous, they attract the attention of the general public and the media, mobilizing public opinion against the regime, challenging state legitimacy, and making it harder for the state to feign ignorance of protesters’ grievances.Footnote 61 Without denying the importance of legitimacy costs, I highlight here an institutional logic within the petitioning system that enhances the importance of the number of protest events. Each protest event provides information to the regime on citizen satisfaction with policies or policy implementation in a particular locality. From the perspective of rulers, protests may be caused by a variety of factors, including incompetence, poor policy implementation, and faulty policies.Footnote 62 When the number of protests is small, rulers have relatively limited information on which to base their decisions and are consequently much more likely to blame protests on incompetence or poor policy implementation. Each additional protest provides rulers with another information point. When the number of protests is exceptionally large, rulers and policy makers can be more confident that the protests are the result of faulty policies and not incompetence or poor implementation. For this reason, the institutional procedures of the petitioning system, which Chapter 4 will examine in greater detail, emphasize the investigation of the largest causes of petitions. This makes the number of petitions critically important.
Second, the petitioning system tracks and treats differentially the type of tactic used. Broadly speaking, the petitioning system divides tactics into conventional and disruptive tactics. Conventional tactics include petition letters and visits in person to petitioning offices. Petitions, however, are only legal (and therefore conventional) if they are delivered by no more than five people. Disruptive tactics include collective petitions, protests, sit-ins, traffic blockages, attacks on state agents, and petitions targeting higher-level authorities.Footnote 63
Disruptive tactics send a stronger signal to the state than conventional tactics largely because of differences in the costs and risks that the two types of tactics impose on participants. Under authoritarian rule, the state generally tolerates a much narrower range of protest tactics than in a democracy.Footnote 64 The risk of repression is also greater because authoritarian rulers are less inhibited in their use of repression.Footnote 65 Disruptive tactics therefore carry much higher risks.Footnote 66 One landless farmer I interviewed, for example, was sentenced to three years in prison for “disturbing social order” by organizing a protest of several thousand landless farmers.
Tactics also involve different costs. The costs of sending a petition letter to the township (a conventional tactic) are considerably lower than the costs of organizing a collective petition by dozens of petitioners to the provincial capital (a disruptive tactic). Violent tactics – as in the case of evictees who light themselves on fire – may cost activists their lives. If participants begin to reject lower-cost/risk tactics in favor of costlier and riskier tactics, this signals that their grievances against state policies are more severe. The costs and risks of more disruptive tactics prevent autocrats from thinking that protesters are engaging in “cheap talk.”Footnote 67
The third and final component of the signal is the target of contentious mobilization. The importance of the target has been overlooked in the literature on protest signaling in China. Yet, as Chapter 4 will show, the upward shift of mobilization toward Beijing played a critical role in putting social stability onto the agenda. From the perspective of the autocrat, mobilization targeted at local governments indicates that citizens are protesting against a localized problem that citizens feel local governments can address. In this sense, mobilization targeted against local authorities is less serious. Mobilization targeted at the central government is much more serious. For citizens, targeting mobilization primarily at the local government is less costly than traveling to the comparatively distant capital. When mobilization targets the central government, it may indicate citizens’ lack of confidence in the ability or willingness of local governments to resolve the problem. Social unrest concentrated at the seat of regime power is also far more likely to unseat the regime than scattered contention in the countryside.Footnote 68
In keeping with the analogy of the “fire alarm,” by examining these three factors, the petitioning system distinguishes the severity of grievances between a “one-alarm fire” and a “five-alarm fire.” Perhaps fittingly, the petitioning system designates the most serious problems as “burning issues” (redian wenti). The institutional procedures of the petitioning system emphasize burning issues over other less common grievances, and the petitioning system devotes considerable resources toward analyzing and advocating policies to respond to burning issues. In particular, key policy makers who exert considerable control over the policy-making agenda – provincial chiefs, ministers, state councilors, and most important, Politburo Standing Committee members – receive much more information about burning issues.
The Nomenklatura System as Political Mediation
As noted previously, the lack of elections in closed authoritarian regimes means that signaling approaches – which are traditionally based on democratic representation – cannot easily explain why dictators would respond to the information signaled by protests with policy changes. Understanding the political interests of officials in authoritarian regimes, however, is crucial to understanding why the regime might be responsive to protests. At the very top, dictators seek first and foremost to stay in power.Footnote 69 Protesters in authoritarian regimes could potentially overthrow the government, a phenomenon scholars of authoritarian politics usually refer to as the revolutionary constraint. Dictators who fear popular overthrow can be held accountable if protesters can credibly threaten to stage a revolution that overthrows the regime. Under such circumstances, policy changes become much more likely.Footnote 70 The revolutionary-threat argument cannot be easily applied to China, where most protest is reformist and has limited aims.Footnote 71
Drawing on Amenta and colleagues’ work on political mediation in the United States, I argue that the impact of protest signals on policy making is mediated by the political interests of officials. These political interests provide the political context in which mobilization occurs. The political mediation approach seeks to account for this, arguing that protests are most likely to be influential when politicians and bureaucrats see political benefits in aiding protesters.Footnote 72 In order for this to happen, citizens must use protests to change the calculations of politicians and bureaucrats by causing them to see protesters as potentially facilitating or disrupting their own goals.Footnote 73 This approach emphasizes the importance of the fit between the political interests of politicians and bureaucrats – the political context – and the type of tactics used. More disruptive tactics are necessary when the interests of citizens, politicians, and bureaucrats diverge sharply.Footnote 74 This makes influencing policy particularly difficult in authoritarian regimes. As Amenta notes, “[W]here powerful systemic conditions work against challengers, as in underdemocratized polities, it may be impossible for challengers to exert much influence.”Footnote 75
Why have officials in authoritarian China been responsive to protests? Why do some officials favor greater responsiveness than others? Existing accounts have stressed the role of “disgruntled officials” who are disadvantaged by existing policies, mass organizations that have an institutional mandate to represent the interests of their constituents, and legitimacy costs.Footnote 76 By contrast, I argue that the answer to this question can be found in the nomenklatura system. The nomenklatura system is an institution of personnel control commonly found in communist states. In essence, the nomenklatura system is a top-down system of vertical accountability that grants officials the power to appoint, promote, and dismiss officials at the next lower administrative level.Footnote 77 At the highest level, the nomenklatura system gives the Politburo Standing Committee (PBSC) the authority to control the careers of provincial chiefs (Party secretaries and governors) and central government ministers. The nomenklatura system allows the PBSC to identify certain outcomes as more important than others and offer incentives for certain kinds of behavior. Because provincial chiefs and ministers are careerists who seek promotions, they have strong incentives to be responsive to the demands of the PBSC.Footnote 78 Promotions are more meritocratic than in the Mao era, and political performance strongly influences career prospects.Footnote 79
Since the 1990s, the PBSC has emphasized two outcomes as more important than all others: social stability and economic development.Footnote 80 The PBSC values economic growth, moreover, primarily because it sees a certain level of economic growth as necessary to maintain social stability. When the petitioning system reported a sharp increase in the level of social unrest at the turn of the century, social stability entered onto the agenda of PBSC. The petitioning system did not see the protests as revolutionary but rather as reformist. Party leaders worried, however, that actual opponents of the regime (especially Western countries and domestic human rights activists) might try to take advantage of the protests to overthrow the regime in the future.
In an attempt to defuse protests before this occurred, the PBSC began to use the nomenklatura system to give officials at all levels incentives to maintain stability. Their central strategy was to emphasize the responsibility of provincial chiefs for social unrest in their jurisdictions and ministers for unrest in their policy domains.Footnote 81 Within this context, policy reforms became tangible political accomplishments that officials could point to as evidence of their efforts to improve social stability. In brief, the nomenklatura system makes ministers and provincial chiefs responsive to the PBSC’s demands that they maintain stability. As a consequence, ministers and provincial chiefs who came under pressure from the PBSC to reduce protests had an incentive to push for policy changes that addressed citizen grievances. Simply put, the nomenklatura system mediates the policy impact of protest by giving officials political incentives to minimize protests through policy concessions.
Officials, however, may also have conflicting political incentives that dampen their support for responsiveness. Ministers have fairly narrow policy portfolios and tasks on which the Party leadership evaluates them. By contrast, provincial chiefs have a much broader scope of tasks on which the Party leadership evaluates them.Footnote 82 Provincial chiefs’ prospects for promotion are closely tied to the economic performance of their provinces, particularly in the areas of GDP growth and revenue generation.Footnote 83 This is important because, as we will see in Chapter 2, activities that generate economic growth also tend to cause social instability. The nomenklatura system thus encourages support for policy concessions to maintain social stability but also discourages support for policy changes that might constrain growth. As Cai has argued, the fiscal costs of policy responses constrain local government support for policy responsiveness.Footnote 84 Equally if not more important than the losses of revenue, however, were the attendant “political costs” of policy changes that could hinder career advancement.
As a consequence of the nomenklatura system, provincial chiefs and ministers often vary in their level of responsiveness to social unrest. Provincial officials – facing significant political and fiscal costs – are much more inclined toward preserving status quo policies that benefit them. Ministers, however, are relatively unencumbered by the costs of reforms and face few counterveiling goals. As a consequence, ministers typically favor greater responsiveness to social unrest and policy innovations that go far beyond what provincial chiefs are willing to support in national policy making. These distinctions, however, are not absolute. Provincial chiefs who govern localities where mobilization is especially high may favor moderate policy concessions. At the same time, ministers who face limited pressure to maintain social stability – either because mobilization is low or because mobilization occurs outside their bureaucratic portfolio – have no incentive to support policy responses, especially if those responses conflict with the priorities of their ministerial portfolio.
Policy-Making Venues and Veto Players
Diverging policy preferences among policy makers make it difficult to achieve consensus on policy concessions. These difficulties are compounded by the fact that policy-making authority is so decentralized that the policy-making process can be usefully characterized as fragmented authoritarianism.Footnote 85 As Mertha has argued, for policy change to occur, policy proposals must enjoy broad-based support within the government.Footnote 86 More specifically, proponents of policy change must be able to overcome their opponents within the policy-making venue. The number of veto players largely determines the extent of support required. Veto players are institutional actors whose agreement is necessary to change the status quo. States with a high number of veto players tend to experience much more policy stability than states with fewer veto actors. Although Tsebelis conceived of communist parties as “monolithic majorities” that act as a single veto player, the fragmented system of authority in China empowers multiple actors with veto powers.Footnote 87 As we will see, the number of veto players strongly influences regime responsiveness to protest.
Policy-making venues either facilitate or constrain the ability of provincial chiefs and ministers to act on their political interests. Legislative policy making in China may take place entirely within the State Council – China’s cabinet – or begin in the State Council and move to the National People’s Congress. Importantly, ministers are extremely influential in the State Council, which directly supervises ministries.Footnote 88 At the same time, many local officials, including provincial chiefs, serve as deputies in the National People’s Congress.Footnote 89 When policy making takes place in the State Council, the veto powers of provincial chiefs are heavily curtailed. This means that ministers – whose veto powers are enhanced in the State Council – are typically much more successful at pursuing their policy goals, resulting in policies that are more responsive to citizen protests.
Policy-making venues that give greater access to provincial governments – particularly provincial-level government agencies and the National People’s Congress – afford provincial chiefs greater veto power than they enjoy in the State Council. Consequently, reforms that emerge from these venues are likely to be limited in scope, assuming that the reforms threaten the interests of local governments. Indeed, because these venues offer greater opportunities for local governments to resist reforms,Footnote 90 policy responsiveness may be slow or not forthcoming at all. For these reasons, the policy-making processes of the National People’s Congress have become an obstacle to responsiveness. To be sure, in a hierarchically organized system such as China – with a strong nomenklatura system – the ability of provincial chiefs or ministers to veto policies is not absolute. Nonetheless, they possess considerable authority to obstruct, delay, and alter policy proposals.
An Alternative Explanation: Elite Politics
The most prominent alternative explanation to the emergence of populist policies has emphasized changes in elite politics. As one of the most prominent scholars of Chinese elite politics recently argued, Chinese leaders’ “political position and policy preferences are often shaped or constrained by their personal experience, leadership expertise, factional affiliation, and bureaucratic portfolio.”Footnote 91 Scholars have frequently attributed the shift to more populist policy making to the rise of President Hu Jintao and his populist faction. During the 1990s, policy making in China was heavily pro-urban and pro-growth, a phenomenon attributed to the leadership of Jiang Zemin.Footnote 92 Jiang’s factional power base was the “princelings,” who advanced their careers in coastal regions. During the Jiang era, policy making heavily favored the interests of entrepreneurs, real estate developers, and local governments over common citizens. This elitist policy-making approach was epitomized by Jiang’s signature ideology, the “three represents,” which welcomed entrepreneurs into the Party.Footnote 93
By contrast, Hu gained experience working in poorer inland provinces, as did many of his factional supporters. His experiences in local administration in underdeveloped areas gave him a greater appreciation for urban–rural inequalities and led him to develop a populist policy-making style.Footnote 94 According to this argument, the emergence of populist policies is the consequence of the emergence of a new leadership with different sources of factional support and consequently different policy preferences. As Naughton argues, “[S]ince Hu assumed the top job, he has presided over a systematic reorientation of economic and social policy to the left in nearly every respect.”Footnote 95 This book will show, however, that elite politics provide at best a partial explanation for the emergence of policy changes. The elite politics approach has difficulty accounting for why policy changes would be more extensive for some groups left behind by reforms (e.g. evictees) but less for others (e.g. landless farmers). More critically, this book will show that there was strong bipartisan support for policy changes among both factions. The emergence of populist policies should be understood primarily as an effort by elite politicians to reduce social unrest rather than as a consequence of the changing composition of elite politicians.
Methodology and Sources
Case Selection and Scope
This book focuses on two substantive policy areas: land takings and demolitions. To what extent can the arguments developed in this book extend to other policy areas? Broadly speaking, most petitions concern the socioeconomic interests of different social groups, such as SOE workers complaining about unemployment and pensions, private-sector workers with labor disputes, overtaxed farmers, residents suffering from environmental pollution, demobilized soldiers, community teachers facing termination, and so on.Footnote 96 My initial expectation is that the argument applies to most policy areas where citizens have commonly mobilized petitions complaining about central policies.
There may, however, be some variation between policy areas. Mertha has shown, for example, that NGOs and the media are influential in hydropower policy making.Footnote 97 In land takings and demolitions, however, NGOs and the media have played a relatively limited role.Footnote 98 Diamant’s work, meanwhile, suggests that groups mobilizing “political” claims are less likely to succeed because a moderately sized wave of petitions and protests by demobilized veterans has completely failed to achieve recognition for an official mass organization for veterans.Footnote 99 In the concluding chapter I offer brief “shadow cases” of community teachers, demobilized soldiers, rural taxation, pensions for SOE workers, and labor contracts for migrant workers, all of which offer preliminary support for the notion that the argument is broadly applicable.
From the perspective of the empirical contribution to China studies and theory building, however, land takings and demolitions provide for a fruitful comparison. The politics of legislative policy making in land takings and demolitions has received far less attention than the shadow cases. In addition to providing a theoretical explanation of the link between petitions and policy making, therefore, this book also provides fresh new material on two relatively understudied cases. Land takings and demolitions provide a particularly fruitful basis for comparison because they are in many respects “most similar” cases.Footnote 100 First, in both policy areas, protesters have mobilized demands that are not only economic but also very similar. Landless farmers have demanded higher compensation, a halt to coercive and forcible land takings against their will, and provision of social security in their old age. Urban evictees have demanded higher compensation, a halt to forcible housing demolitions, and resettlement in apartments near their original homes. Second, both groups are relatively similar in terms of their limited financial and political resources because they are made up predominantly of farmers and working-class urbanites.Footnote 101 Third, in both cases, the demands mobilized by protesters have resulted in central–local conflicts between ministries and provinces that would entail substantial fiscal expenditures by local governments.
More important, the two policy areas provide crucial variation in the policy-making venue. Governed by a law (the Land Management Law), policy making in land takings took place in the National People’s Congress. By contrast, demolitions were governed by a regulation (the Urban Housing Demolition Regulation), making the State Council the policy-making venue. The two groups also varied in the policy outcomes they achieved. Protests and petitions by landless farmers resulted in moderate policy responsiveness at the national level. Protests and petitions by urban evictees, meanwhile, resulted in extensive policy changes at the national level.
Research Site Selection
In making this argument, I chose two main research sites: Zhejiang province and Hubei province. I initially chose Zhejiang as a research site because of my personal connections in the province. Zhejiang is a highly developed province on China’s east coast with a strong fiscal capacity and a high level of mobilization by both landless farmers and evictees. Zhejiang is also a crucial case for understanding provincial responsiveness to protests because the local and provincial governments in Zhejiang were the first in the nation to introduce several policy responses to land takings and demolitions. I selected Hubei province as a research site both because of my personal connections in the province and also because Hubei is an inland province with a weak fiscal capacity that has experienced relatively limited mobilization by landless farmers and evictees. Differences in fiscal capacity and level of mobilization, as Chapter 5 will show, influenced the pace of policy change at the provincial level. With the exception of provincial capitals (which are so large that they provide a degree of anonymity), I use pseudonyms for all the counties and districts where I conducted research.
Interviews
I conducted over 250 interviews for this project over the course of thirteen months in field sites in the two provinces and Beijing. Within Zhejiang province, I interviewed local government officials, landless farmers, and evictees in the provincial capital of Hangzhou, as well as several highly developed counties and districts, including Dancheng, Xiaoyang, Yongwu, Shaoxiang, Liangwang, and Niandu, as well as the middle-income counties of Hujiang and Tingjiang and the poor county of Sun’an. Within Hubei, I conducted interviews with urban evictees in the provincial capital of Wuhan. I conducted all interviews in Chinese, sometimes with the help of a Chinese research assistant or professor.
Many, but not all, of the local officials I interviewed were introduced to me by friends and colleagues. These personal connections made them much more frank and willing to talk about land takings and demolitions. Difficulties accessing officials, however, convinced me that I should focus my efforts on locating documentary sources as my main form of evidence on the “state” side of the argument. I used interviews primarily to assess the validity of my documentary sources.
On the “society” side, landless farmers and evictees were generally quite willing and eager to accept interviews. My interviewees included not only landless farmers and evictees who had resisted land takings but also many who had not. The length of interviews varied from fifteen minutes to several hours. I arranged interviews with landless farmers and evictees via two methods. First, I used a snowballing approach, relying on introductions from acquaintances and other interviewees. Second, I located landless farmers and evictees by finding resettlement housing sites and development zones, as well as finding urban neighborhoods currently undergoing demolitions. Over the course of multiple visits to these sites, I was able to establish relationships with many evictees, especially the activist evictees who led petitioning activities.
I also benefited from interactions with Chinese scholars who were familiar with the legislative history of the land taking and demolitions policies, as well as the broader policy-making process. As I explained my research to them, they helped me to corroborate accounts reported in the Chinese media and documentary sources and in many cases were kind enough to share new details with me.
Documentary Primary Sources
I also made use of a significant number of documentary primary sources, which I accessed primarily at universities and libraries throughout China. My sources consist of both “internal” and “open” serials. In terms of internal serials, I rely primarily on two types of serials. The first is a series produced by provincial petitioning bureaus, as well as the State Bureau of Letters and Visits (SBLV). Provincial petitioning bureaus produce serials commonly titled Zhejiang Petitions (Zhejiang Xinfang), Hubei Petitions (Hubei Xinfang), and so on. Provincial petitioning bureaus in other provinces, however, have given their publications less self-apparent titles. Civil Servants and Petitions (Gongpu Yu Xinfang), Conditions of the People and Petitions (Minqing Yu Xinfang), and Petitions and the Conditions of the People (Xinfang Yu Minqing) are published by the provincial petitioning bureaus in Heilongjiang, Shaanxi, and Zhejiang, respectively. The SBLV publishes People’s Petitions (Renmin Xinfang). Provincial and state petitioning bureaus typically issue these publications on a monthly or bimonthly basis. While not as highly classified as some series, such as Petition Information (Xinfang Xinxi), circulation of these monthly publications seems to be restricted to those “within the system (tizhinei),” and the publication is not generally available to the public.Footnote 102
The contents of the serials from petitioning offices typically include important instructions (pishi) and speeches from leaders, notifications and regulations regarding petition work, and interviews with officials. Most of the content of the serials, however, is typically devoted to investigative reports written by local and provincial petitioning officials and occasionally officials from other bureaus who handle petition work. The reports vary widely in content and often address issues such as petitions in a particular geographic jurisdiction or policy area and petition tactics. For example, the second issue of Zhejiang Petitions from 2004 includes an article from an official in the Longquan City Land Resources Bureau entitled, “Discussion of Problems, Special Characteristics, and Countermeasures for Land Petition Work.” As the title suggests, the tone of the article is analytical, and the article provides an explanation for why land-related petitions had increased and also offers several possible responses to these problems.
While using data from the petitioning system might raise concerns of bias, as Chen noted in his study of state-generated data on petitions, “information distortion is not a big problem for most governmental data from the petitioning system since the 1990s.”Footnote 103 Provincial petitioning bureaus – as the publishers of the series – exercise editorial control over the content. In as much as the selection of articles is biased, the content likely reflects the priorities of the provincial petitioning bureau. Because this in itself provides some insight into the agenda of the petitioning bureaus, this is less of a problem. It is likely that any policy proposals falls within the realm of what the provincial petitioning bureau considers permissible. In terms of the authors of the articles – primarily local petitioning officials – the largest danger of bias is that the authors would present an overly rosy view of the situation in their own jurisdictions in order to make themselves look good in the eyes of higher levels. A tutorial on writing reports published in Tianjin Petitions in the late 1990s, however, demanded that petitioning officials “liberate themselves from ideology,” “dare to speak the truth,” and “emphasize practicality” by making recommendations to correct problems.Footnote 104 Petitioning officials appear to have largely heeded these demands because, in practice, the articles that appear in the series are often quite frank and critical of state practices and policies. Local officials’ incentives, therefore, would seem to be to provide accurate and insightful reports. As such, these articles provide quite valuable insights into how petitioning officials perceive social unrest. These sources, moreover, have been mostly overlooked in the literature on protest in China.Footnote 105 More important, my interviews with local officials corroborated many of the accounts of the problems with state practices and policies.
The second set of internal sources I use is an internal Xinhua news series from Zhejiang and Hubei known as Zhejiang Internal Reference (Zhejiang Neican) and Hubei Internal Reference (Hubei Neican). The circulation of these periodicals is restricted to government officials, and reporting often touches on topics deemed too sensitive for the “open” Xinhua news reports. Land takings, demolitions, and the petitioning system were covered in much greater detail than in the open media. Coverage of protests against land takings and demolitions was more extensive in Zhejiang than in Hubei, likely reflecting the higher level of mobilization in the province.
Documentary series published by the Ministry of Land Resources and Ministry of Housing and Urban-Rural Development (MOHURD) tended to be open sources. This includes serials from the Ministry of Land Resources, particularly China Land (Zhongguo Tudi) and Land Resources Communication (Guotu Ziyuan Tongxun). It also includes serials from provincial land resources and housing bureaus, especially Zhejiang Real Estate (Zhejiang Fangdichan), Zhejiang Land Resources (Zhejiang Guotu Ziyuan), and Shandong Land Resources (Shandong Guotu Ziyuan). These publications seemed to be the functional equivalent of the petitioning bureau serials discussed earlier. The reports were also written almost exclusively by land resources and housing officials, had similar content, and were equally openly critical of state policies and practices. Some series from land resources bureaus, particularly Beijing Housing Reform (Beijing Fang Gai), Chengdu Land Resources (Chengdu Guotu Ziyuan), and Guangdong Land Resources (Guangdong Guotu Ziyuan), are classified as “internal,” although their content does not seem to vary remarkably from equivalent open series. I also make frequent use of Beijing Fangdichan (Beijing Real Estate) and Zhongguo Fangdichan (China Real Estate). Although not published by the MOHURD, they frequently carried articles written by officials in the local MOHURD and demolition offices, as well as articles written by real estate developers. I also make use of contemporary press reports about the revisions of the Urban Housing Demolition Regulation and the Land Management Law, both of which have received wide coverage in the media.
Where possible, I test my qualitative arguments with quantitative data. In particular, I have relied extensively on petitioning data reported by provincial governments and central government ministries. Chapter 2 uses petitioning data from central government ministries to examine changes in mobilization over time. Chapter 4 uses newspaper data and reports from provincial petitioning bureaus to examine the prominence of land takings and demolitions on the agenda. Chapter 5, meanwhile, uses provincial-level data to test my argument about the influence of mobilization and fiscal capacity on the pace of policy change.
