What Drives People to Protest in an Authoritarian Country? Resources and Rewards vs Risks of Protests in Urban and Rural China

Abstract

What drives people to protest in an authoritarian country? Drawing from a rich set of individual-level data from the China General Social Survey 2010, we address the question of protest participation by focusing on the factors of resources, and rewards vs risks, that might be unique to protestors in an authoritarian state. We find strong evidence for education, typically conceived as a key enabling resource in protests, to be negatively associated with likelihood of participation. There are, however, significant differences between political behavior in urban and rural samples. We find some, though rather weak, evidence to suggest that as urban residents become wealthier over time, they will increasingly turn to protests as a form of political participation, demanding greater accountability of government and corporate actions.

Keywords

protest authoritarianism resource mobilization China

Introduction

While the existing literature on social movements has provided plenty of inspiration for the motivating factors of political participation in Western liberal democracies, we know relatively little about the extent to which the literature is applicable to authoritarian states. There are several grounds for skepticism. First, there are some key distinctions between institutionalized and non-institutionalized forms of political participation, such as voting and protest. Voting is relatively effortless and incurs lower costs compared to protest. However, in authoritarian countries where voting is conspicuously absent, protests can become an important form of political participation. Second, protests incur considerably higher costs and pose higher risks for participants in authoritarian states compared to their counterparts in liberal democracies. Protestors in illiberal states face high risk of reprisal and state repression that may result in loss of income or even incarceration and loss of freedom. Third, in many authoritarian states where governments tightly control non-governmental organizations (NGOs), the mobilizing structures of collective action are palpably lacking, posing significant challenges for would-be protestors to organize and mobilize their actions. Social movement organizations in democracies that are funded by members’ dues and run by professional activists can openly mobilize supporters for their causes. In contrast, the aggrieved in illiberal states have to rely on informal social networks and friendship to bring people together (Beinin and Vairel, 2013; Della Porta, 1988; Loveman, 1998; McAdam, 1986; Zuo and Benford, 1995).

With a few notable exceptions, the literature on contentious politics in China largely consists of microanalyses of individual or groups of cases. These qualitative studies with rich contexts have helped to inform on protests in China from multi-dimensional perspectives. They include framing (Chen, 2003; Hurst, 2008; Lee, 2007; Mertha, 2008; Thornton, 2002), political opportunity (Cai, 2010; O’Brien and Li, 2006; Wallace and Weiss, 2015; Wright, 1999), and mobilizing structure (Cai, 2002; Zhao, 2001). O’Brien and Li (2006) note that the Chinese state is not a “monolith,” but a “hodgepodge of disparate actors,” which gives rise to different political opportunities. Political opportunity may also arise when local governments, acting under pressure to maintain social stability, yield to the threat of disruption by activists (Cai, 2010). Aggrieved state-owned enterprise workers invoked the language of class struggle and the implicit social contract between state workers and the government to make their claims (Lee, 2007). Perry (2002) argues the repertoires of contention in contemporary China have strong historical and cultural roots. Lacking formal mobilizing structure, protestors in China have to use informal networks to mobilize support for their actions. These networks are often of a clandestine nature, and their viability and momentum depend critically on the quality of leadership (Cai, 2002; Li and O’Brien, 1996; Shi and Cai, 2006).

Drawing on survey data, various scholars have studied how cultural and social values may explain political participation and trust (Shi, 2001; Tang, 2005; Tao et al., 2014; Wong et al., 2011). Furthermore, Su and Feng (2013) study whether the social connections of urban residents raise or reduce their propensity to protest. They find urban residents with better social networks or guanxi are more inclined to protest. Wu and Dong (2014) find a positive association between one’s experience of being unfairly treated and protest participation.¹ Michelson (2007, 2008) examines how the social and political connections of rural residents affect their probability of using legal mobilization when grievances arise. That proposition is relevant to this study because the legal system is an institutional and an alternative channel of dispute resolution. Other studies have looked at why the authoritarian regime may tolerate dissent: protest is used as an information-gathering mechanism (Lorentzen, 2013) and the government allows for only selective online criticisms while silencing collective expression (King et al., 2013).

Given existing works on what explain political participation in China and the availability of data from the China General Social Survey (CGSS) 2010 upon which this study is drawn, we set out two objectives for this article. First, drawing upon biographical data in the CGSS 2010, we address the question of protest participation by focusing on resources and reward vs risk calculations that may be unique to protestors in an authoritarian state. A value-based approach has been taken up by studies elsewhere. Resources, income, and education, which are enablers of social activism in democratic societies, may behave differently in illiberal states. The monetary cost of protest participation may rise for those with higher income and education levels as they have more to lose if they face reprisals for their actions. We may also observe a non-linear relationship between these resource variables and protest actions as citizens become more active at low resource level but less active as the level of resources rises. Between income and education, we are more interested in the effect of education on protest participation since obtaining information about one’s educational attainment is straightforward while income levels are grossly underreported and notoriously difficult to measure in China.

In an authoritarian state where protests incur high costs for participants, we are interested in how the rewards and risks of actions feature in one’s calculation to become involved. Rewards are measured in terms of those incurred to the individual participants rather than any intangible benefit arising from altruistic concerns. The degree of risks varies between different types of protest actions, such as those against private entities and the state or government officials. Anti-regime dissent, such as human rights activism, ethnic, and religious protests, carries much higher risks than actions against private companies over issues of food safety or owed wages. Even though protest participation is likely to be underestimated in survey data conducted in authoritarian countries, the response rate for the CGSS is comparable to other similar nationwide surveys. In addition, CGSS is among the best survey data available for China given the sensitivity of the information.

To be sure, protest is not the only form of political participation. Yet, in an authoritarian regime, such as China, formal institutional channels for the expression of grievances are systematically lacking. Legally speaking, Chinese citizens could go to court and submit petitions to local and higher government authorities. Nevertheless, access to these institutional channels is not afforded to everyone, justice is often denied, and they carry risks of reprisals too as petitioners are routinely rounded up, put into “black jails,” and suffer physical abuses (Fu, 2010; Minzner, 2010; Ong, 2018). Technically speaking, we cannot address these alternative means of conflict resolution in this study because the CGSS 2010 contains no specific questions that allow us to do so.

The second objective of the article is to delineate the drivers of protests in urban and rural China. If we see protest as a form of political participation among others, existing scholarship has suggested that demographic factors play a role in voter preference and turnout. Citizens in rural areas are more likely than those in urban areas to turn out at the polls and to vote for pro-government candidates. Among the factors at work are lower income level and stronger kinship or community ties that facilitate channeling of selective benefits (Magaloni, 2006). In China, rural and urban regions are structurally dissimilar in terms of economic structure, citizen’s resources, including levels of income and education, legal knowledge, and media access. In addition to citizen-activist’s personal characteristics, protests in urban and rural China also differ in terms of grievance type. In the Social Unrest in China dataset, which consists of more than 2500 cases from 2003 to 2013, the top three categories of urban protests and their respective shares in total regional incidents are land-related incidents, such as housing demolition (24.0%), state-owned enterprise disputes (18.2%), and migrant labor disputes (15.1%), respectively. On the other hand, the top categories in rural protests are land expropriation (58.8%), environment (9.9%), and state-owned enterprise labor disputes (6.6%), respectively.² In relative term, urban protests are more claims of material interests over owed wages or relocation compensation, whereas rural disputes tend to be subsistence in nature with citizens fighting for claims for farmland, a major source of income for rural residents, and environmental pollution that endangers people’s lives. To the extent that these regional protests are driven by dissimilar reasons and staged by diverse actors, this study contributes to enhancing our understanding of the reasons for their variations.

Liberal economic reforms since 1979 have produced a growing and sizable middle class in China particularly in the urban areas. Drawing upon the relationship between economic development and democratization as depicted in modernization theory, we are interested in discerning whether demand for democratization is more likely to originate from rural or urban China. As the income and/or education level of urban citizens rises, are they more or less likely to participate in protest? Is this necessarily a linear relationship? This question is particularly intriguing as the existing literature has indicated that the rising middle class in China may not necessarily demand democracy as modernization theory has predicted (Chen and Lu, 2011; Dickson, 2003). Overall, this study aims to contribute to studies of contentious politics in China and to the literature on social mobilization in illiberal states in general.

The rest of the article is organized as follows. Section “Theoretical Framework: What Drives People in Authoritarian Countries to Protest?” looks at the theoretical literature on the drivers of protests and raises a few testable hypotheses given the data in hand. Section “Research Design, Data Source, and Methods” discusses the research design, data source, and methods. Section “Analysis and Results” presents the findings and discusses their significance for existing theories. Section “Conclusion” concludes the study.

Theoretical Framework: What Drives People in Authoritarian Countries to Protest?

This article draws upon social movement theories to analyze what drives citizens in an authoritarian state to protest. Given the systemic differences between democracies and non-democracies, we focus our attention on the drivers of protest from the perspectives of citizen’s resources in an authoritarian state, their calculation of potential rewards vs risks, and how these factors may play out differently in urban and rural areas.

Resource mobilization theories seek to explain not what motivates people to take to the streets but why some successfully mobilize while others do not. While resource mobilization in the social movement literature typically refers to the organizational or network basis of mobilization, studies conducted in a wide range of countries have consistently shown levels of income and education to be positively associated with higher rates of protest participation at the individual level (Dalton et al., 2009; McVeigh and Smith, 1999; Pattie et al., 2004; Schussman and Soule, 2005; Verba et al., 1995).

Nevertheless, some research on illiberal states and low-income countries finds individual resources to have a weaker or even negative impact on political participation (Loveman, 1998; Schock, 1999). In non-liberal states where the organizational basis of collective action, such as the presence of autonomous NGOs, is feeble or non-existent, people with resources may still face challenges in mobilizing action (Earl, 2006). Separately, higher cost of reprisal in authoritarian states implies that those with higher income and education levels have considerably more to lose than the impoverished and poorly educated. In other words, the opportunity cost of participating in protest actions, a high-risk activity in authoritarian states, rises as one progresses up the social ladder. The cost could take the form of forgone income, loss of social status, or even loss of personal freedom (Campante and Chor, 2014).³

In a high-risk environment, the rewards for action must be significantly high to justify participation. Put simply, why would a citizen risk his job or life to participate in contentious action if he could ride free on other people’s actions (Olson, 1965)? Therefore, understanding of reward vs risk calculations is key to understanding why people in an authoritarian state participate in contentious actions.

Urban vs Rural Protests

The existing literature also leads us to expect a priori the drivers of protest participation in urban and rural China to differ. First, urban residents generally have significantly higher income levels than rural residents. Scholars have attributed the increase in urban protests, such as Not In My Back Yard (NIMBY) environmental protests and homeowners’ resistance, to the rise of the middle class in China (Liu, 2013; Wang et al., 2013). We borrow inspiration from modernization theory (Lipset, 1959; Przeworski and Limongi, 1997), which asserts that economic development will bring about social changes, such as higher levels of education and a larger middle class, that will result in greater demand for democracy. However, we are aware of the need for caution because existing studies indicate that the middle class in China has not demanded greater democracy, primarily because they are dependent on the state and are the main beneficiary of the present system (Chen, 2013; Chen and Lu, 2011; Dickson, 2003; Tsai, 2007). Given the higher cost of reprisal, the urban middle class in China may not be willing to “risk” what they have to participate in contentious politics.

Second, separate from higher income and education levels, urban residents in China also have enhanced access to resources, such as the media, and some though very few civil society organizations in the cities, such as environmental NGOs (Steinhardt and Wu, 2015; Tang and Zhan, 2008; Yang, 2005) and homeowners’ associations (Merle, 2014; Yip and Jiang, 2011), which play an effective role in disseminating information, if not mobilizing the public. Armed with better information or enhanced channels to network, would urban residents exhibit a greater likelihood of participating in protest compared to their rural counterparts, ceteris paribus?

Research Design, Data Source, and Methods

Data Source

Our data come from the CGSS 2010, a nationwide representative survey, administered by the National Survey Research Center at the Renmin University of China. The 2010 survey used a four-stage stratified sampling scheme to sample 11,783 adults age 18 years and above in Mainland China. The survey includes 22 provinces, four autonomous regions, and four central government-designated municipalities. Tibet, Hong Kong, Macau, and Taiwan are not included in the survey.⁴ CGSS 2010 has been used to study a wide range of issues in China, such as poverty and income inequality (Xie and Zhou, 2014; Zhang et al., 2014).

Dependent Variable

Our dependent variable is a binary variable measuring individual’s participation in collective contention. It is coded as participate = 1 and did not participate = 0. The original question is “In daily life, we always observe some collective contention or activities, such as boycotting unreasonable charges, protests against demolition or land requisition, collective resistance against government projects, collective petitions, collective strikes, gatherings, and demonstrations. In the past 3 years, have you ever taken part in any of these activities?” Multiple-choice answers are “1. I have been an organizer; 2. I have been a participant; 3. I have never participated but have provided material support; 4. I have never participated but have provided moral support; 5. Other; 6. I have never attended.” For the sake of simplicity, we combine selections no. 1 (organizer) and no. 2 (participant) and recode them as participate = 1. The rest of the selections is coded as did not participate = 0 in the regression analysis.⁵

Independent Variables⁶ and Hypotheses

H1: The higher the income or education level an individual has, the more likely he has taken part in collective action.

Personal Resources

Income level is based on self-reported annual personal income in 2009. As previously mentioned, the impact of income level on protest participation may be positive or negative depending on whether the enabling or opportunity cost effect dominates. We do not expect either effect to dominate a priori. Therefore, we have included both income and income-squared in the model to capture a possible reversal of relationship after income reaches a certain level.

In this case, we have decided to use personal instead of family income because we think it is more appropriate in capturing the opportunity cost of participation. If an individual were fired from his job because of participation, only his personal income would be affected.

Education level is measured in two ways, both linear and non-linear. The linear measurement is by years of schooling, while the non-linear way is by levels of education, including junior school and below, middle school, college diploma, and university bachelor’s degree and above.⁷ Similar to income level, education could have a positive or negative effect on protest. Overall, we expect education to be a more reliable predictor of participation because of unreliability of self-reported income and difficulty in measuring it accurately.

We have also included an interaction term between education and income to test whether increases in income level will result in dissimilar effects on protest participation for respondents with different levels of education. This allows us to project into the future as income per capita rises over time.

Rewards vs Risks of Action

H2: One is more likely to have taken part in protest when the action is perceived to preserve or increase his self-interest.

Based on Olson’s collective action theory, we ask whether or not the contentious action preserves or increases the respondent’s self-interest. It is coded “1” if it did. If it reduced or had no impact on one’s interest, it is coded “0.”

H3: When non-government entities become the targets of contentious action, individuals are more likely to have been involved than in situations where actions are targeted at the government.

If the targets of contention are non-state entities, it is coded “1,” and “0” otherwise. In authoritarian settings, a protestor’s action is particularly risky if it is targeted at the government, state institutions, or government officials. On the contrary, if the action is aimed at non-state institutions, such as a private enterprise or individuals, the risks are considerably diminished. In China, while human rights lawyers and rights activists who target the state regularly face the risks of incarceration and loss of freedom, migrant workers employed by private firms who go on strike are most likely to lose their jobs in the worst-case scenarios.

Urban Middle Class

H4: Increased income and/or education levels of urban residents will lead to a greater likelihood of protests in urban China.

Since the existing evidence on the effects of a rising urban middle class on their likelihood of protest participation in authoritarian countries is inconclusive, we have no clear expectation of whether the relationship will be positive or negative (Table 1).

Table 1.

Hypotheses about the Variables Affecting Protest Participation in China.

Variables	Effect on likelihood of protest Participation
Income	?
Education	?
Self-interest	+
Lower risk (non-state entities as targets)	+
Urban middle class	?

Control Variables

In all models, we include variables to control for “personal constraints that may increase the costs and risks of movement participation” (McAdam, 1986). Various studies suggest that one’s age, gender, and marital status can have a bearing on the propensity to participate.

In all models, we include marital status (married or cohabiting = 1; single, divorced, or widowed = 0) and the logged form of age. The other control variables included in the models are gender (female = 1; male = 0), ethnicity (ethnic minorities = 1; Han Chinese = 0), and urban vs rural location of survey (urban = 1; rural = 0) in the full-sample regression.

Like income, legal knowledge can also be viewed as a resource that enables and facilitates political participation. In our model, legal knowledge is measured by a series of self-evaluated questions on one’s legal knowledge. These include basic legal knowledge, familiarity with the functions of the public security bureau, the procuratorates, the courts, and knowledge of how to enlist a lawyer, seek legal assistance, file a lawsuit, and find the petition bureaus. The questions adopt a 5-point Likert-type scale, ranging from completely unknown = 1 to completely known = 5. Legal knowledge is scaled from 8 to 40. The Cronbach’s alpha test generates a score of 0.927, indicating a high level of reliability of the indicator. We expect one’s legal knowledge to have a positive effect on protest participation.

We have included two variables to control for the effects of organizational involvement on participation: communist party membership and trade union membership. The existing theoretical literature suggests that individuals’ organizational involvement matters.⁸ However, both the communist party and trade unions are state-sanctioned institutions that help indoctrinate members of state ideologies. The All-China Federation of Trade Unions (ACFTU), which retains a complete monopoly on trade unionism, is co-opted by the state. While the ACFTU has been credited for its pursuit of enhanced legislative protection of workers at the national level in recent years, its effectiveness at the enterprise level is disputable (Friedman and Lee, 2010; Gallagher, 2014).

Descriptive Statistics

Descriptive statistics are summarized in Table 2. Survey respondents consisted of 7222 people (61.9%) from urban areas and 4561 people (38.1%) from rural areas. In all, 261 respondents or 2.2% of the total sample reported that they participated in protests at least once in the past 3 years, as shown in Appendix 2. Among them, 15 people reported being organizers, and 246 claimed to be participants. Granted, protest participation is most likely underreported in an authoritarian state as the action carries considerable risks. The CGSS does not use Global Positioning System (GPS) sampling and is likely to underestimate the number of migrant workers in the sample (Landry and Shen, 2005). However, other similar surveys in China have also yielded levels of protest participation of around 2%, which provides us with assurance that our response rate is not an anomaly.⁹

Table 2.

Summary Statistics of Key Variables.

Variables	Mean			SD	Min	Max
Variables	Urban	Rural	Full	Full	Full	Full
Male	0.475	0.492	0.481	0.499	0	1
Age (years)	46.441	48.667	47.303	15.677	17	96
Years of schooling	10.248	6.302	8.721	4.648	0	20
Level of education
Junior school and below	0.503	0.892	0.653	0.475	0	1
High school	0.256	0.089	0.191	0.393	0	1
College diploma	0.123	0.012	0.080	0.271	0	1
University and above	0.118	0.007	0.075	0.262	0	1
Income (10,000 yuan)	3.088	1.254	2.377	7.974	0	600
Media consumption	9.171	5.020	7.564	4.473	0	24
Married or cohabiting^a	0.781	0.842	0.804	0.396	0	1
Ethnic minority	0.074	0.128	0.095	0.293	0	1
CCP membership	0.166	0.057	0.124	0.329	0	1
Trade union member	0.178	0.027	0.119	0.325	0	1
Legal knowledge	22.747	17.791	20.829	7.604	8	40
Urban	–		0.613	0.487	0	1

Source: CGSS 2010.

SD: standard deviation; CCP: Chinese Communist Party.

Includes all types of relationships: never married, cohabiting, married, separated, divorced, and widowed. All data are unweighted.

Most of the protests were targeted at the government or government-related entities: 53.4% were aimed at the government, 13.3% at government officials, 10% at public institutions, and 6% at government policies.¹⁰ Appendices 2 and 3 contain figures of the distribution of respondents’ involvement in collective action and targets of protests, respectively.

Analysis and Results

We first adopt logistic regression modeling to investigate factors that may affect protest participation in China. In Table 3, Model 1 is the baseline model, which includes a limited set of independent variables, such as years of schooling, income, and income-squared, and control variables, such as rural/urban location, gender, age (logged), media consumption, party membership, ethnic minority, trade union membership, and marital status. Model 2 changes the linear form of education, years of schooling, into four levels of education—primary school and below (base), middle school, college diploma, and university bachelor’s degree and above. Model 3 adds two additional predictors, namely, self-interest and non-state protest targets. Models 4 and 5 add legal knowledge and interactions between education and income to the baseline model. The pseudo R² values of all the models are between 0.009 (Model 1) and 0.509 (Model 5), indicating the better fit of the more complex models.

Table 3.

Binary Logit and Heckman Selection Regression Results (Dependent Variable: Protest Participation).

						Heckman selection model
						Probit	Selection
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Years of schooling	−0.025
	(0.023)
Middle school		0.230	−0.124	−0.214	0.230	−0.157	0.285^***
		(0.180)	(0.264)	(0.263)	(0.180)	(0.170)	(0.051)
College diploma		−0.949^*	−0.750	−0.900^†	−0.949^*	−0.683^*	0.420^***
		(0.394)	(0.487)	(0.470)	(0.394)	(0.341)	(0.068)
University and above		−1.354^**	−1.698^**	−1.730^**	−1.354^**	−0.952^*	0.452^***
		(0.484)	(0.622)	(0.608)	(0.484)	(0.384)	(0.073)
Income (10,000 RMB)	−0.036	−0.013	0.034	0.016	−0.019	−0.037	−0.001
	(0.057)	(0.056)	(0.071)	(0.071)	(0.092)	(0.046)	(0.006)
Income2 (100 Mil RMB)	−0.008	−0.012	−0.020	−0.014	−0.031	−0.019	−0.000
	(0.023)	(0.023)	(0.027)	(0.025)	(0.033)	(0.025)	(0.000)
Self-interest			4.848^***	4.838^***	4.833^***	1.467^***	7.682
			(0.240)	(0.240)	(0.240)	(0.142)	(300.108)
Non-state target			1.472^***	1.505^***	1.515^***	0.300^*	7.724
			(0.326)	(0.325)	(0.327)	(0.130)	(313.9)
Legal knowledge				0.044^**	0.044^**	0.018^*	0.014^***
				(0.015)	(0.015)	(0.007)	(0.003)
MidSchoolXincome					0.098	0.069
					(0.116)	(0.057)
CollegeDipXincome					0.082	0.08
					(0.145)	(0.084)
UniversityXincome					0.093	0.066
					(0.136)	(0.069)
Control
Rural	−0.208	−0.204	0.312	0.359	0.347	0.224^*
	(0.174)	(0.168)	(0.227)	(0.228)	(0.233)	(0.113)
Female	−0.413^**	−0.372^*	−0.397^*	−0.299	−0.302	−0.137
	(0.148)	(0.147)	(0.201)	(0.205)	(0.205)	(0.099)
Age (logged)	−0.357^†	−0.500^*	−0.038	0.030	0.012	0.048
	(0.207)	(0.220)	(0.310)	(0.313)	(0.312)	(0.153)
Media	0.016	0.018	−0.026	−0.042	−0.043	−0.031^*
	(0.021)	(0.021)	(0.030)	(0.031)	(0.031)	(0.014)
Party member	−0.011	0.129	0.518^†	0.438	0.432	0.086
	(0.237)	(0.236)	(0.288)	(0.286)	(0.289)	(0.160)
Ethnic minority	−0.351	−0.339	0.484	0.472	0.471	0.321
	(0.282)	(0.284)	(0.323)	(0.325)	(0.327)	(0.197)
Trade union member	−0.221	−0.173	−0.609^†	−0.650^†	−0.652^*	−0.376^*
	(0.262)	(0.263)	(0.337)	(0.333)	(0.332)	(0.156)
Married		0.437^†	−0.162	−0.184	−0.183	0.046
		(0.240)	(0.279)	(0.276)	(0.277)	(0.134)
cons	−1.902^*	−2.004^*	−4.844^***	−5.893^***	−5.759^***	−1.875^**	−2.007^***
	(0.918)	(0.826)	(1.262)	(1.354)	(1.360)	(0.657)	(0.074)
N	11,783	11,783	11,783	11,783	11,783	11,783
Rho							−0.033 (0.092)
Pseudo R²	0.009	0.021	0.504	0.509	0.509
Log likelihood	−1272.456	−1256.607	−636.829	−630.650	−630.076	−2778.827
Chi-square	18.616	36.223	901.145	877.893	873.620	182.139

Numbers in parentheses are standard errors. Data are weighted according to the 2005 census.

†

p < 0.10; ^*p < 0.05; ^**p < 0.01; ^***p < 0.001.

However, given the risks associated with protest participation—and with providing an affirmative answer to that question—we used the Heckman two-stage method to correct for any potential selection bias (Heckman, 1979). Respondents who chose to answer the participation question may be self-selected. Rho’s from the various Heckman two-stage models suggest that the concern of self-selection is not warranted. However, we decide to take a cautious approach of reporting both sets of results and place greater confidence on the variables that are statistically significant in both models. Model 6 reports results from the Heckman selection model.

In selecting independent variables included in the selection model, we follow the advice of Sartori (2003) and Wooldridge (2002) by selecting a range of variables that may potentially affect a respondent’s decision to answer the participation question, such as the case that a married person may be more reluctant to answer compared to a non-married respondent because the former is more risk averse. However, the choice of variables in the selection equation does not appear to change the significance of the Heckman model much at all.

Generally, magnitudes of the coefficients are smaller in the Heckman model. Personal resources matter a great deal in explaining one’s involvement in contentious politics. There is evidence that education level is negatively related to participation. Furthermore, a higher level of education reduces the likelihood of participation more than a lower level of education does. This lends support to the opportunity cost proposition of higher education, and it outweighs the resource-enabler effect. It underlines the high cost of participating in contentious politics in China, where demonstrators and protestors are routinely rounded up and constantly face threats of losing their jobs and social status. Legal knowledge is another resource factor that is positively associated with participation. Income effect is insignificant which is not entirely unsurprising given its self-reported nature and measurement issues.

We also find strong evidence to support the hypothesis that individuals participate in a protest if it preserves or enhances their interest. Individuals are also more likely to protest if the actions are targeted at non-state entities, such as private enterprises or individuals, which entail considerably less risk compared to government or government-related institutions. In China, it is highly conceivable that workers in private firms who protest against their employers may lose their jobs, while those who contend against the state are confronted with both economic and political costs for their actions.

Interestingly, there is some evidence to suggest a negative association between trade union membership and protest participation. The co-opted trade unions are not mobilizing structures for collective action. Instead, our results suggest that they may serve as conduits in channeling workers’ grievances to the authorities, thereby reducing the likelihood of workers expressing their discontent on the street. While it is more or less mandatory for large-scale enterprises to set up trade unions, union membership is optional. Therefore, we should expect those who become union members to belong to a self-selected group of political insiders, who are less likely to protest. Party membership does not appear to be significant in all models.

In the first few models, there is no evidence that rural or urban populations were more likely to protest. However, after taking into consideration interaction terms between income and education, the rural population appears more likely to protest compared to their urban counterparts in the Heckman selection model.

Protest Participation in Urban vs Rural China

Tables 4 and 5 present the binary logit and Heckman probit model regression results for the urban and rural samples, respectively. The split sample tests reveal a few interesting findings about the motivating factors of protests in the two areas. Preservation of self-interest is positively associated with participation in both the urban and rural samples, as they are in the full sample. The positive effect of non-state target is statistically significant in the urban sample but becomes insignificant in the Heckman model. Meanwhile, it remains positive in the rural sample, although the size of the coefficient falls significantly in the Heckman model. This suggests negative self-selection among the respondents whose protest actions were targeted at the state. When one’s action was targeted at the state, one is less likely to answer the participation question because of the higher risks involved. Once this selection bias is taken into consideration, the positive effect of non-state targets declines or disappears altogether in the case of our urban sample.

Table 4.

Binary Logit and Heckman Selection Regression Results, Urban Sample (Dependent Variable: Protest Participation).

						Heckman selection model
						Probit	Selection
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Years of schooling	−0.043
	(0.027)
Middle school		0.294	0.139	0.021	0.511	−0.183	0.196^***
		(0.210)	(0.284)	(0.288)	(0.417)	(0.207)	(0.059)
College diploma		−0.783^†	−0.605	−0.824^†	−1.514^*	−0.821^*	0.362^***
		(0.413)	(0.450)	(0.452)	(0.638)	(0.352)	(0.072)
University and above		−1.184^*	−1.352^*	−1.444^*	−2.285^**	−1.112^**	0.400^***
		(0.492)	(0.585)	(0.575)	(0.838)	(0.397)	(0.077)
Income (10,000 RMB)	−0.063	−0.051	−0.040	−0.050	−0.242^*	−0.115^†	−0.004
	(0.041)	(0.037)	(0.050)	(0.052)	(0.119)	(0.063)	(0.005)
Income2 (100 Mil RMB)	0.001	0.001	−0.000	0.001	−0.020	−0.014	0.000
	(0.001)	(0.001)	(0.004)	(0.003)	(0.029)	(0.024)	(0.0005)
Self-interest			4.812^***	4.825^***	4.834^***	1.512^***	8.150
			(0.310)	(0.312)	(0.313)	(0.172)	(2383.232)
Non-state target			0.902^**	0.947^**	0.950^**	0.188	8.212
			(0.328)	(0.330)	(0.332)	(0.154)	(2602.248)
Legal knowledge				0.051^**	0.052^**	0.025^**	0.008^*
				(0.018)	(0.018)	(0.009)	(0.004)
MidSchoolXincome					0.279^†	0.128^†
					(0.148)	(0.075)
CollegeDipXincome					0.308^*	0.156^†
					(0.157)	(0.094)
UniversityXincome					0.302^†	0.140^†
					(0.157)	(0.094)
Control
Female	−0.356^†	−0.325^†	−0.310	−0.220	−0.241	−0.098
	(0.185)	(0.184)	(0.236)	(0.239)	(0.242)	(0.121)
Age (logged)	−0.356	−0.509^†	−0.107	−0.076	−0.134	−0.001
	(0.254)	(0.279)	(0.361)	(0.366)	(0.366)	(0.187)
Media	0.014	0.010	−0.027	−0.043	−0.043	−0.033^*
	(0.027)	(0.026)	(0.035)	(0.035)	(0.036)	(0.016)
Party member	0.053	0.180	0.652^†	0.575^†	0.563^†	0.156
	(0.275)	(0.280)	(0.339)	(0.335)	(0.339)	(0.181)
Ethnic minority	0.140	0.179	0.989^*	0.983^*	0.981^*	0.361
	(0.330)	(0.334)	(0.391)	(0.382)	(0.391)	(0.241)
Trade union member	−0.254	−0.247	−0.672^*	−0.726^*	−0.742^*	−0.461^**
	(0.288)	(0.290)	(0.323)	(0.318)	(0.315)	(0.167)
Married		0.405	−0.177	−0.195	−0.035
		(0.277)	(0.320)	(0.321)	(0.158)
cons	−1.734	−1.890^†	−4.382^**	−5.489^***	−4.963^**	−1.636^*	−1.750^***
	(1.134)	(1.063)	(1.428)	(1.546)	(1.561)	(0.773)	(0.090)
N	7222	7222	7222	7222	7222	7222
Rho							−0.083 (0.117)
Pseudo R²	0.010	0.024	0.497	0.504	0.508
Log likelihood	−743.659	−733.121	−378.022	−372.908	−369.985	−2021.149
Chi-square	16.204	28.577	560.194	543.550	537.674	130.923

Numbers in parentheses are standard errors. Data are weighted according to the 2005 census.

†

p < 0.10; ^*p < 0.05; ^**p < 0.01; ^***p < 0.001.

Table 5.

Logit Regression Results, Rural Sample (Dependent Variable = Protest Participation).

						Heckman selection model
						Probit	Selection
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Years of schooling	0.003
	(0.037)
Middle school		0.115	−0.795	−0.851^†	−0.914	−0.182	0.340^**
		(0.347)	(0.514)	(0.499)	(0.623)	(0.390)	(0.112)
College diploma						−4.893	0.078
						(11817.942)	(0.288)
University and above						−4.489	0.096
						(4772.316)	(0.366)
Income (10,000 RMB)	0.147	0.153	0.351^*	0.345^*	0.341^†	0.143	0.063
	(0.141)	(0.140)	(0.170)	(0.168)	(0.176)	(0.126)	(0.042)
income2 (100 Mil RMB)	−0.143	−0.142	−0.211	−0.209^†	−0.208^†	−0.115	−0.030
	(0.124)	(0.119)	(0.131)	(0.123)	(0.124)	(0.137)	(0.034)
Self-interest			5.270^***	5.247^***	5.247^***	1.590^***	31.455
			(0.387)	(0.386)	(0.386)	(0.282)	(615.806)
Non-state target			2.961^***	2.958^***	2.962^***	0.573^*	8.490
			(0.715)	(0.702)	(0.709)	(0.264)	(3349.337)
Legal knowledge				0.028	0.028	0.005	0.023^***
				(0.029)	(0.029)	(0.013)	(0.006)
MidSchoolXincome					0.036	−0.031
					(0.223)	(0.189)
CollegeDipXincome						0.016
						(4160.752)
UniversityXincome
Control
Female	−0.443^†	−0.421^†	−0.552	−0.468	−0.466	−0.248
	(0.233)	(0.242)	(0.375)	(0.377)	(0.378)	(0.185)
Age (logged)	−0.238	−0.344	0.235	0.287	0.286	0.220
	(0.340)	(0.335)	(0.559)	(0.552)	(0.551)	(0.292)
Media	0.027	0.039	−0.039	−0.056	−0.057	−0.057
	(0.032)	(0.030)	(0.054)	(0.061)	(0.061)	(0.061)
Party member	−0.103	−0.007	−0.192	−0.234	−0.235	−0.195
	(0.461)	(0.445)	(0.471)	(0.501)	(0.501)	(0.367)
Ethnic minority	−1.031^†	−1.034^*	−0.030	−0.080	−0.079	0.163
	(0.527)	(0.525)	(0.563)	(0.574)	(0.575)	(0.369)
Trade union member	0.235	0.253	0.682	0.667	0.667	0.467
	(0.614)	(0.612)	(0.853)	(0.856)	(0.852)	(0.533)
Married		0.508	−0.056	−0.053	−0.051	−0.051
		(0.456)	(0.502)	(0.489)	(0.488)	(0.488)
cons	−2.946^*	−3.053^*	−6.173^**	−6.827^**	−6.822^**	−2.356^†	−2.387^***
	(1.437)	(1.219)	(2.284)	(2.322)	(2.314)	(1.297)	(0.144)
N	4561	4475	4475	4475	4475	4549
Rho							0.109 (0.169)
Pseudo R²	0.019	0.023	0.550	0.551	0.551
Log likelihood	−521.641	−517.025	−238.456	−237.523	−237.516	−731.536
Chi-square	17.647	19.142	231.753	235.639	240.801	43.581

Numbers in parentheses are standard errors. Data are weighted according to the 2005 census.

†

p < 0.10; ^*p < 0.05; ^**p < 0.01; ^***p < 0.001.

Trade union membership is negatively associated with participation in the urban sample, as it is in the full sample. The relationship is not statistically significant in the rural sample. Trade unions are typically found in urban cities where large-scale factories are located. Individuals with more legal knowledge in urban areas are more likely to protest, while the same cannot be said about the rural sample. Ethnic minorities are more likely to protest in urban not rural areas. This confirms the popular perception that ethnic minorities, such as the Tibetans and Uighurs, face widespread discrimination in the cities.¹¹ The other variables, such as marital status, gender, age, media consumption, and party membership, are not consistently significant in either sample.

Income is marginally significant in the rural sample, but its significance disappears after the Heckman selection model is introduced. It is not significant in the urban sample, as in the full sample.

Education is negatively associated with participation in both the rural and urban samples, as it is in the full sample. The more educated one is in urban China, the less likely one is to protest. That said, it is noteworthy that the interaction terms between income and different education level are positive and statistically significant at the 10% level in the urban sample. This suggests that even though education level, on its own, has a negative effect on participation, at successively higher income levels, the negative likelihood is successively reduced. We take this as evidence that the well-educated and high-income urban residents in China are more likely to be involved in protest. In other words, the enabler effect may begin to outweigh the opportunity cost effect in high-income brackets.

To further investigate this important issue of the effect of education and income on protest, we simulate the 95% confidence interval for the marginal effect of income given various education levels in the urban sample, with and without the Heckman selection model. See Figures 1 and 2, respectively, with standard units on the Y-axis. At successively higher levels of income, respondents with a given level of education (middle school, college diploma, and university education) exhibit a decreasing likelihood of non-participation. When income rises, the slope of each curve becomes less steep, and the confidence interval becomes narrower. In the Heckman models, the income effects for all three education levels are positive, even though the 95% confidence intervals are of a larger range. This leads us to conclude that there is some—though weak—evidence to support the hypothesis of increasing protest actions by a rising middle class in urban China.

Figure 1.

Urban-Marginal Effects of Income on Protest Given Various Education Levels (Logit Regression) with 95% Confidence Intervals.

Figure 2.

Urban-Marginal Effects of Income on Protest Given Various Education Levels (Heckman Selection Models) with 95% Confidence Intervals.

Conclusion

Using China as a case, our study makes a few important points regarding the motivators of protest participation in authoritarian countries. The cost of participation is considerably higher for citizens of authoritarian states than those of liberal democracies. We find strong evidence to support a negative association between education level and likelihood of protest, which stands in contrast to the positive effect of education found in democratic societies. The negative effect is successively stronger for those with higher levels of education. This suggests that the opportunity cost of a higher education level in terms of higher forgone income far outweighs the positive enabling effect.

However, among the respondents with each level of education in the urban sample, the negative effect declines at successively higher levels of income. For those with middle school education, the declining negative effect never enters into positive territory, while for those with a college diploma and university education, the positive effect starts to manifest at high-income levels. We take this as evidence, albeit weak, that as urban residents become wealthier over time, they will increasingly turn to protests as a form of political participation. The findings for income and education combined suggest that the two variables that are typically resource enablers for individual protestors in democratic societies are not necessarily the case for citizens in authoritarian countries.

While this suggests that urban residents in China may exhibit political behavior resembling those in liberal democracies as their income level rises, this may not necessarily bring about democratization in the country. Protests in China are conventionally about material grievances rather than a demand for democratization (Ong and Gobel, 2012). At best, this weak evidence lends support to a reduced form of modernization theory that predicts increased income and education levels leading to demand by the urban population for greater accountability and transparency (manifested in a greater number of protest actions) rather than democracy. This is consistent with the emergence of broad-based protests in the cities supported by a wide spectrum of urban population, such as those against construction of chemical plants and incinerators, as existing studies have suggested (Johnson, 2010; Li and Liu, 2012; Lu and Chan, 2016).

Chinese citizens behave rationally in that they are more likely to take part in protest actions when they are perceived to be able to bring personal benefits. When actions are targeted at the government or government officials, which are of higher risks compared to those aimed at private entities and individuals, they are likely to attract fewer participants. We also find the organizational basis for contentious politics (measurable in our models) to be lacking in China, which is completely unsurprising given the nature of its political system. Trade union membership has a negative association with participation. Trade unions in China are rather unlike their counterparts in Western societies that advocate for workers’ rights but are subsumed within the ACFTU, which is co-opted by the state (Friedman and Lee, 2010). They serve as conduits in channeling workers’ grievances to the authorities, thereby reducing the likelihood of workers spilling their discontent to the streets.

Protestors in China as in other authoritarian countries have a different benefit–cost calculus compared to those in democratic societies. We believe the logic of political participation, such as the effects of income and education, rewards vs risks calculation of protestors, and organizational basis for contentious politics, such as trade unions, could be similarly applied to other illiberal states where fewer mobilizing structures are available, and aggrieved citizens face considerably higher costs when they take their grievances to the street.

Footnotes

Appendix 1

Variables	Measurement
Dependent variable
Participation in collective action	In daily life, we always observe some collective contention or activities, such as boycotting unreasonable charges, protests against demolition or land requisition, collective resistance against government projects, collective petitions, illegal gatherings, demonstrations, and collective strikes. In the past 3 years, did you take part in any of these activities? Participate = 1; did not participate = 0
Independent variables
Urban	Survey location: urban = 1; rural = 0
Age	Self-reported age
Gender	Self-reported gender
Years of education	1–20
Level of education	Primary school and below = 1; middle school = 2; college diploma = 3; university bachelor’s degree and above = 4
Income	Annual personal income in 2009 (10,000 RMB)
Ethnic minorities	Ethnic minorities = 1; Han Chinese = 0
CCP membership	CCP member = 1; non-member = 0
Trade union membership	Trade union member = 1; non-member = 0
Marital status	Married or cohabiting = 1; single, divorced, or widowed = 0
Media consumption	How often did you use the following media (newspapers, magazines, broadcast, television, Internet, mobile phone customized news)? Never = 0; rarely = 1; sometimes = 2; often = 3; very often = 4
Self-interest	Did these contentious activities or actions affect your interest? Preserve or increase my self-interest = 1; hurt my interest or no impact = 0
Legal knowledge	To what extent do you have an understanding of the following? (1) Basic knowledge of the law; (2) the functions of the public security bureau; (3) the functions of the public prosecutor’s office; (4) the functions of the court; (5) how to hire a lawyer; (6) how to apply for legal aid; (7) how to litigate (file a lawsuit); (8) how to find the petition bureaus. Coding method: completely do not understand = 1; do not understand = 2; neither understand nor do not understand = 3; understand = 4; completely understand = 5
Non-state target	Who were the targets of these contentious activities or actions? Private companies, individuals and entities = 1; others = 0

Appendix 2

Appendix 3

Appendix 3 shows that among those had protested, 38.3% had “non-state” targets, including private firms and persons, totaling 554 cases. The other 61.7% in “others” include government or government-related entities, government officials, public institutions, government policies, and non-response.

Acknowledgements

The authors express their sincere gratitude to the journal editor, Charles Pattie, and anonymous reviewers for helpful comments. They are also grateful for useful feedback from Carolina de Miguel Moyer, John Ravenhill, Jeremy Wallace, and the organizers and participants of the following workshops and seminars where earlier versions of this article were presented: the International Symposium on Social Resistance in Non-democracies at Lund University in 2014, the Balsillie School of International Affairs at University of Waterloo in 2015, and the School of International Relations and Public Affairs at Fudan University in 2016.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: research fundings from the School of International Studies, Renmin University of China.

Notes

Author biographies

Lynette H Ong is an Associate Professor at the Department of Political Science, and Asian Institute, Munk School of Global Affairs, the University of Toronto. She writes about authoritarian politics, contentious politics and the political economy of development. She is the author of Prosper or Perish: Credit and Fiscal Systems in Rural China (Cornell University Press, 2012). Her publications have appeared or are forthcoming in Perspectives on Politics, Comparative Politics, International Political Science Review, China Quarterly, China Journal, Journal of Contemporary Asia, among others.

Donglin Han is an Associate Professor at the School of International Studies, Renmin University. His research interests include Chinese politics, public opinion and research methods. His publications have appeared in Democratization, China Quarterly, The Chinese Journal of International Politics, Journal of Contemporary China, among others.

References

Beinin

Vairel

(eds) (2013) Social Movements, Mobilization, and Contestation in the Middle East and North Africa. Stanford: Stanford University Press.

Cai

(2002) The Resistance of Laid-off Workers in the Reform Period. China Quarterly 170: 327–344.

Cai

(2010) Collective Resistance in China: Why Popular Protests Succeed or Fail. Stanford, CA: Stanford University Press.

Campante

Chor

(2014) “The People Want the Fall of the Regime”: Schooling, Political Protest, and the Economy. Journal of Comparative Economics 42 (3): 495–517.

Chen

(2003) Industrial Restructuring and Workers’ Resistance in China. Modern China 29 (2): 237–262.

Chen

(2013) A Middle Class without Democracy: Economic Growth and the Prospects for Democratization in China. Oxford: Oxford University Press.

Chen

(2011) Democratization and the Middle Class in China: The Middle Class’s Attitudes towards Democracy. Political Research Quarterly 64 (3): 705–719.

Dalton

Sickle

Weldon

(2009) The Individual-Institutional Nexus of Protest Behaviour. British Journal of Political Science 40 (1): 51–73.

Della Porta

(1988) Recruitment Processes in Clandestine Political Organizations: Italian Left-Wing Terrorism. International Social Movement Research 1: 155–169.

10.

Dickson

(2003) Red Capitalists in China: The Party, Private Entrepreneurs, and Prospects for Political Change. Cambridge: Cambridge University Press.

11.

Earl

(2006) Special Focus Issue of Mobilization on Repression and the Social Control of Protest. Mobilization: An International Journal 11 (2): 129–280.

12.

Friedman

Lee

(2010) Remaking the World of Chinese Labor: A 30-Year Retrospective. British Journal of Industrial Relations 48 (3): 507–533.

13.

(2010) Access to Justice in China: Potentials, Limits, and Alternatives. In: Gillespie

Chen AHY (eds) Legal Reforms in China and Vietnam: A Comparison of Asian Communist Regimes New York: Routledge, pp.163–187.

14.

Gallagher

(2014) China’s Workers Movement and the End of the Rapid-Growth Era. Daedalus 143 (2): 81–95.

15.

Gerlach

Hine

(1970) People, Power, Change: Movements of Social Transformation. Indiana, IN: Bobbs-Merrill Company.

16.

Grant

(2017) “Don’t Discriminate Against Minority Nationalities”: Practicing Tibetan Ethnicity on Social Media. Asian Ethnicity 18 (3): 371–386.

17.

Heckman

(1979) Sample Selection Bias as a Specification Error. Econometric Society 47 (1): 153–161.

18.

Hurst

(2008) Mass Frames and Worker Protest in Popular Protest in China (ed. O’Brien

). Cambridge, MA: Harvard University Press.

19.

Johnson

(2010) Environmentalism and NIMBYism in China: Promoting a Rules-Based Approach to Public Participation. Environmental Politics 19 (3): 430–448.

20.

King

Pan

Roberts

(2013) How Censorship in China Allows Government Criticism but Silences Collective Expression. American Political Science Review 107 (2): 326–343.

21.

Klandermans

(1997) The Social Psychology of Protest. Oxford: Blackwell.

22.

Landry

Shen

(2005) Reaching Migrants in Survey Research: The Use of the Global Positioning System to Reduce Coverage Bias in China. Political Analysis 13 (1): 1–22.

23.

Lee

(2007) Against the Law: Labor Protests in China’s Rustbelt and Sunbelt. Berkeley, CA: University of California Press.

24.

O’Brien

(1996) Villagers and Popular Resistance in Contemporary China. Modern China 22 (1): 28–61.

25.

Liu

(2012) Getting Their Voices Heard: Three Cases of Public Participation in Environmental Protection in China. Journal of Environmental Management 98: 65–92.

26.

Liu

(2013) China’s New “Middle Class” Environmental Protests. China Dialogue. Available at: https://www.chinadialogue.net/article/show/single/en/5561-China-s-new-middle-class-environmental-protests

27.

Lorentzen

(2013) Regularizing Rioting: Permitting Public Protest in an Authoritarian Regime. Quarterly Journal of Political Science 8 (2): 127–158.

28.

Loveman

(1998) High-Risk Collective Action: Defending Human Rights in Chile, Uruguay, and Argentina. American Journal of Sociology 104 (2): 477–525.

29.

Chan

(2016) Collective Identity, Framing, and Mobilization of Environmental Protest in Urban China: A Case Study of Qidong’s Protest. China: An International Journal 14 (2): 102–122.

30.

McAdam

(1982) Political Processes and the Development of Black Insurgency. Chicago, IL: The University of Chicago Press.

31.

McAdam

(1986) Recruitment to High-Risk Activism: The Case of Freedom Summer. American Journal of Sociology 92 (1): 64–90.

32.

McAdam

Paulsen

(1993) Specifying the Relationship between Social Ties and Activism. American Journal of Sociology 99 (3): 640–667.

33.

McPherson

Smith-Lovin

(1987) Homophily in Voluntary Organizations: Status Distance and the Composition of Face-to-Face Groups. American Sociological Review 52 (3): 370–379.

34.

McVeigh

Smith

(1999) Who Protests in America: An Analysis of Three Political Alternatives—Inaction, Institutionalized Politics, or Protest. Sociological Forum 14: 685–702.

35.

Magaloni

(2006) Voting for Autocracy: Hegemonic Party Survival and Its Demise in Mexico. New York: Cambridge University Press.

36.

Merle

(2014) Homeowners of Beijing, Unite! The Construction of a Collective Mobilisation. China Perspectives 2: 7–15.

37.

Mertha

(2008) China’s Water Warriors: Citizen Action and Policy Change. Ithaca, NY: Cornell University Press.

38.

Michelson

(2007) Climbing the Dispute Pagoda: Grievances and Appeals to the Official Justice System in Rural China. American Sociological Review 72: 459–485.

39.

Michelson

(2008) Justice from Above or Below? Popular Strategies for Resolving Grievances in Rural China. China Quarterly 193: 43–64.

40.

Minzner

(2010) Xinfang: An Alternative to Formal Chinese Legal Institutions. Stanford Journal of International Law 42 (103). Available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1548577

41.

O’Brien

(2006) Rightful Resistance in Rural China. New York: Cambridge University Press. Available at: http://link.library.utoronto.ca/eir/EIRdetail.cfm?Resources__ID=748647&T=F

42.

Olson

(1965) The Logic of Collective Action. Cambridge, MA: Harvard University Press.

43.

Ong

Gobel

(2012) Social Unrest in China. Europe China Research and Advice Network. Available at: https://www.chathamhouse.org/sites/files/chathamhouse/public/Research/Asia/1012ecran_gobelong.pdf

44.

Ong

(2015) Reports on Social Unrest in China. In: Goodman

(ed.) Handbook of Research on Politics in China. Cheltenham: Edward Elgar Publishing, pp.345–362.

45.

Ong

(2018) Thugs and Outsourcing of State Repression in China. The China Journal. Epub ahead of print 2 February. DOI: 10.1086/696156.

46.

Pattie

Seyd

Whiteley

(2004) Citizenship in Britain: Values, Participation and Democracy. Cambridge: Cambridge University Press.

47.

Perry

(2002) Challenging the Mandate of Heaven: Social Protest and State Power in China. Armonk, NY: ME Sharpe.

48.

Przeworski

Limongi

(1997) Modernization: Theories and Facts. World Politics 49: 155–183.

49.

Roche

Hillman

Leibold

(2017) Why Are so Many Tibetans Moving to Chinese Cities? China File. Available at: http://www.chinafile.com/reporting-opinion/viewpoint/why-are-so-many-tibetans-moving-chinese-cities

50.

Sartori

(2003) An Estimator for Some Binary-Outcome Selection Models without Exclusion Restrictions. Political Analysis 11 (2): 111–138.

51.

Schock

(1999) People Power and Political Opportunities: Social Movement Mobilization and Outcomes in the Philippines and Burma. Social Problems 46 (3): 355–375.

52.

Schussman

Soule

(2005) Process and Protest: Accounting for Individual Protest Participation. Social Forces 84 (2): 1083–1108.

53.

Shen

(2008) Zhongguo Gongmin Yishi Diaocha Shuju Baogao; Attitudes Toward Citizenship in China: Data Report of A National Survey. Beijing: Social Science Academic Press.

54.

Shi

Cai

(2006) Disaggregating the State: Networks and Collective Resistance in Shanghai. China Quarterly 186: 314–332.

55.

Shi

(2001) Cultural Values and Political Trust: A Comparison of the People’s Republic of China and Taiwan. Comparative Politics 30: 410–419.

56.

Steinhardt

(2015) In the Name of the Public: Environmental Protest and the Changing Landscape of Popular Contention in China. The China Journal. Epub ahead of print 20 November. DOI: 10.1086/684010.

57.

Feng

(2013) Adapt or Voice: Class, Guanxi, and Protest Propensity in China. Journal of Asian Studies 72 (1): 45–67.

58.

Tang

S-Y

Zhan

(2008) Civic Environmental NGOs, Civil Society, and Democratization in China. Journal of Development Studies 44 (3): 425–448.

59.

Tang

(2005) Public Opinion and Political Change in China. Stanford, CA: Stanford University Press.

60.

Tao

Yang

et al . (2014) Hoes Does Political Trust Affect Social Trust? An Analysis of Survey Data from Rural China Using an Instrumental Variable Approach. International Political Science Review 35 (2): 237–253.

61.

Thornton

(2002) Framing Dissent in Contemporary China: Irony, Ambiguity and Metonymy. China Quarterly 171: 661–681.

62.

Tsai

(2007) Capitalism without Democracy: The Private Sector in Contemporary China. Ithaca, NY: Cornell University Press.

63.

Verba

Schlozman

Brady

(1995) Voice and Equality: Civic Voluntarism in American Politics. Cambridge, MA: Harvard University Press.

64.

Wallace

Weiss

(2015) The Political Geography of Nationalist Protest in China: Cities and the 2012 Anti-Japanese Protests. China Quarterly 222: 403–429.

65.

Wang

Sun

et al . (2013) Leadership in China’s Urban Middle Class Protest: The Movement to Protect Homeowners’ Rights in Beijing. China Quarterly 214: 411–431.

66.

Wong

Wan

Hsiao

H-HM

(2011) The Bases of Political Trust in Six Asian Societies: Institutional and Cultural Explanations Compared. International Political Science Review 32 (3): 263–281.

67.

Wooldridge

(2002) Econometric Analysis of Cross Section and Panel Data. Cambridge, MA: MIT Press.

68.

Wright

(1999) State Repression and Student Protest in Contemporary China. China Quarterly 157: 142–172.

69.

Dong

(2014) How Does Experience of Being Unfairly Treated Affect Protest Participation? Evidence from CGSS. Guangdong Shehui Kexue 6: 201–210.

70.

Xie

Zhou

(2014) Income Inequality in Today’s China. Proceedings of the National Academy of Sciences 111 (19): 6928–6933.

71.

Yang

(2005) Environmental NGOs and Institutional Dynamics in China. China Quarterly 181: 46–66.

72.

Yip

N-M

Jiang

(2011) Homeowners United: The Attempt to Create Lateral Networks of Homeowners’ Associations in Urban China. Journal of Contemporary China 20 (72): 735–750.

73.

(ed.) Zhongguo Shehui Yuqin Lanpishu [Annual Report on Public Opinion in China]. Beijing: Renmin Ribao Chubanshe, 2010.

74.

Zhang

Zhou

et al . (2014) Are Poverty Rates Underestimated in China? New Evidence from Four Recent Surveys. China Economic Review 31: 410–425.

75.

Zhao

(2001) The Power of Tiananmen. Chicago, IL: The University of Chicago Press.

76.

Zuo

Benford

(1995) Mobilization Processes and the 1989 Chinese Democracy Movement. The Sociological Quarterly 36 (1): 131–156.

What Drives People to Protest in an Authoritarian Country? Resources and Rewards vs Risks of Protests in Urban and Rural China

Abstract

Keywords

Introduction

Theoretical Framework: What Drives People in Authoritarian Countries to Protest?

Urban vs Rural Protests

Research Design, Data Source, and Methods

Data Source

Dependent Variable

Independent Variables 6 and Hypotheses

Personal Resources

Rewards vs Risks of Action

Urban Middle Class

Control Variables

Descriptive Statistics

Analysis and Results

Protest Participation in Urban vs Rural China

Conclusion

Footnotes

Appendix 1

Appendix 2

Appendix 3

Acknowledgements

Funding

Notes

Author biographies

References

Independent Variables⁶ and Hypotheses