Explanations for Mass Provincial Protest in China


The occurrence of mass incidents such as protests, strikes, and social disorders has increased dramatically in China since 1989; however, these mass incidents are very different from the national demonstrations that occurred in 1976, 1978, 1989, and, most famously, in Tiananmen Square in 1989. Post-1989 protesters no longer challenge the legitimacy of the Central Communist government nor do they seek democratization but, instead, they try to gain the attention of the central government in order to force local governments to redress their specific local grievances. Therefore, I suggest that the spatial distribution of these mass incidents is shaped by the policies of local governments that give rise to these grievances, and, that it is China’s rapid industrialization and urbanization that are the sources of local grievances. Land disputes and labor disputes rise out of and escalate from government control over land and industry (state-owned enterprises, or SOEs). Government policies intended to foster industrial expansion increase the rate of urbanization. Local governments expropriate land from farmers and from residents for the purpose of industrial expansion, which leads to higher incidence of protests. I use a negative binominal regression model to estimate the impact of certain elements of China’s industrialization and urbanization on the incidence of protests across 31 provinces over three time periods. I posit a relationship between the frequency of mass incidents and several variables that are associated with land expropriation and the job security of recent migrants to urban areas who are not legal residents of those urban areas under the terms of China’s Household Registration System. First, the more large cities there are in a province, the higher the incidence of protests will be. Second, the higher the ratio of urban land to rural land, the higher the incidence of protests will be in a province. Third, the larger the number of state-owned enterprises relative to foreign owned enterprises in a province, the lower the frequency of mass incidents will be. In testing the relationships between these concepts, I find strong support for these hypotheses.

Table of Contents: 


    Since the Tiananmen Square Protest of 1989, there have been no national demonstrations in China. Nonetheless, there have been thousands of small, local protests throughout China. According to Western press reports, there were almost four times the number of mass incidents in the late 2000s as compared to the  early 1990s (Wedeman, 2009). These protests differ from the 1989 movement in several ways. Post-1989 protesters no longer seek either the democratization of the PRC’s political system or a change in the political structure of Chinese Communist Party dominance. The number of mass incidents, however, has not been reduced. Most of the mass incidents have involved violence, such as damaging public properties, self immolation, and physical conflicts between villagers and officials (Wines & Ansfeld, 2010). In addition, in the post-1989 period, protesters changed their tactics from directly complaining to the central government to intensifying their protests to gain the attention of the central government. In this manner, they believed that the central government would force local officials to address their grievances if the protests gain their attention.

    The distribution of mass incidents across provinces in China varies considerably. Some provinces experience much higher frequencies of mass incidents than others. In addition, the policies of local governments, which give rise to local grievances, are shaped by different social and economic conditions. Therefore, my main concern in conducting this research is with the impact of several provincial socioeconomic conditions on the distribution of mass incidents in China. The current seminal literature has attempted to link the occurrence of mass incidents to the socioeconomic situation in China. For example, Zweig and Ho argue that land expropriation by local governments leads to grievances among rural residents, and that it is China’s rapid urbanization that leads to these land disputes (Zweig, 2010; Ho, 1999). Wang argues that Household Registration-based discrimination contributes to mass incidents by generating grievances among the unregistered residents living in urban areas (Wong, 2007). Mason also argues that the mass unemployment which was caused by surplus rural labor increased the motivation of protesters in 1989 (Mason, 1994). Although these findings have made significant progress on understanding conflict in China, the current seminal literature focuses only on protests in single localities, micro-phenomena, or certain time periods, which may have overlooked significant relationships between various socioeconomic phenomena and mass incidents in all of the provinces in China over an expanded temporal domain. In order to explain the distribution of mass incidents in China, it is necessary to analyze variations in the socioeconomic conditions across provinces. Variations in a number of social, economic, and demographic conditions may contribute to variations in the number of mass incidents among provinces. Moreover, rapid urbanization affects local governments’ policies and the local economy in present day China. Its impact may be more strongly associated with variations in the occurrence of mass incidents than other variables. Therefore, this research focuses on the relationship between the number of mass incidents and the level of urbanization in the provinces of China.

    In the next section, I will review the existing seminal literature on mass protest in China, focusing especially on the wave of local protests that has occurred since the 1989 Tiananmen Square incident. In the paper’s third section, I present a theory to explain how certain policies of local bureaucrats increase grievances among certain segments of the local population and lead to large scale mass incidents. Then I will discuss the impact of urbanization on two different kinds of protests: land disputes and labor disputes. I then test hypotheses derived from this theory with two sets of provincial data. First, I use data from China’s provincial yearbook to construct variables that measure the extent of land expropriation and the effect of the Household Registration system. I also incorporate several control variables which do not support my theory but might be influential for mass incidents, such as income inequality, unemployment, and communication. Second, I utilize data from Wedeman on the number of mass incidents across China’s 31 provinces over three time periods (Wedeman, 2009). Finally, I estimate three models:  total disputes, land disputes, and labor disputes. Since the dependent variable in each of the models is a count variable, I employ a negative binominal regression model to determine if the number of mass incidents is affected by variation in the indicators of grievances generated by urban expansion.

    Literature Review

    Social protests (mass incidents) have increased dramatically in China since the late 1990s. The characteristics of the recent protests are significantly different from the democratic demonstrations prominent in the 1980s, most notably, the Tiananmen Square protest of 1989. Most scholars agree with Li and O’Brien that the protests reflected a form of “policy-based resistance” instead of an inherently democratic movement (O‘Brien, 2005). The goal of the current protests is not to threaten the regime of the Chinese Communist Party, but rather to seek the redress of their grievances by the national government. If the government fails to issue a response to the requests of the protesters or refuses to negotiate with them, the protestors will likely escalate the level of confrontation and persist for a long period. Nevertheless, there is little consensus in the debate over the causes and nature of these protests. The discussion covers not only government policies, but also the social and economic background in different areas of China. Previous seminal literature concerning conflict in China can be divided into two analytical categories: studies searching for the patterns, changes, and scale of resistance in certain regions, and studies seeking to explain the causes of the increase in mass incidents, particularly the rationales of the protesters, the central government and/or the local government behind the conflicts. In the seminal work on protests, scholars commonly link urbanization, economic performance, ethnic inequality, and political institutions as the main stimuli for protests in the various provinces in China.


    Work on urbanization and civil conflicts has centered on the expansion of urban areas and the growing population of urban immigrants. Here, two related policies of the Chinese government should be identified: land expropriation and the Household Registration System (Hukou). Regarding land expropriation, the current seminal literature including such scholars as Peter Ho, David Zweig, and Jeremy Wallace, states that land disputes have become the most important factor in regard to conflicts between the government and residents as the shortage of urban land contributes to the demand for suburban land and its soaring price (Ho, 1999; Wallace, 2007; Zweig, 1999). In this process, local governments tend to expropriate the land from suburban areas and sell it to land developers in order to increase the income of government and expand urban areas. However, the lack of compensation and negotiation for the residents who lose their land leads to an increasing number of disputes. Ho emphasizes that the ambiguity of collective land ownership gives way to illegal land expropriation (Ho, 1999). After the implementation of the Household Responsibility System in the 1980s, most land remained collectively owned by villagers and very little land was permitted to be privately owned by individuals (Nyberg & Rozelle, 1999). As a consequence, when land is ultimately expropriated, it becomes difficult to determine who legitimately holds the rights to it as land trade only requires the agreement of village leaders rather than all villagers. Moreover, some villagers, who do not have formal rights to the land, have lived on the land for a long time. Thus, they have customary claims over the ownership and control of land. Nevertheless, most of the local bureaucrats refuse or fail to recognize the informal rights of the peasants (Ho, 1999), and directly possess the authority to expropriate land from villagers. This is why land disputes often occur in rural areas.

    Without well-established clarification of collective land rights, it is easier for local officials to expropriate land from groups of villagers without their agreement or adequate negotiation. Even though villagers are not satisfied with the government policy regarding compensation and housing rearrangements, they cannot legally stop forced land evictions by local governments and developers. In Zweig’s case study, local officials sold villagers’ land to developers who promised to arrange a new home for the villagers in suburban areas. Since the village land merged with the urban areas, the rural registrations of the villagers were also replaced by urban registration. Unfortunately, villagers found that their new homes had not been built yet and the developer just arranged a dilapidated building for their temporary residence. At the same time, the local government who sold the land refused to help these villagers and provide welfare benefits for them because these people were no longer registered as rural residents and township governments do not need to be responsible for urban residents. As a consequence, the villagers resort to civil disobedience or violent protests, leading to large scale conflicts between local governments and residents (Zweig, 1999). Still, Ho and Wallace argued that privatization of land may not solve the land disputes (Ho, 1999; Wallace, 2007). Rapid privatization may destabilize the urban society and lead to more conflicts since the agricultural land functions as a social safety net for people who were born in the countryside (Economist, 2007). In addition, it is possible that many peasants tend to immediately sell their land and move to urban areas in order to work in cities after the privatization of their land. If factories are closed in these cities and that consequently results in an excess of labor, the workers cannot then be dispersed back to the farms in rural areas and the large number of unemployed or underemployed workers may then increase instability in urban areas. Furthermore, the absence of land rights reform suggests that the central government is less worried about the rural land conflicts than the urban instability caused by urbanization (Wallace, 2007).

    On the other hand, Wallace (2007), Wang (2010), and Solinger (1995) link the Household Registration System to the control of urbanization. In order to decrease the population in cities, the Chinese government tried to restrict the migration from rural to urban areas through the Hukou system. According to Wang’s description, all Chinese citizens have to be registered with the Hukou authority from birth. According to the rules of the Hukou system, citizens are allowed to inherit only the mother’s Hukou. More importantly, the Hukou designation determines whether the person can have legal permanent residency, community-based rights, opportunities, and social benefits in certain regions. Therefore, in this manner, local governments can limit access to resources to selected groups of the population (Wang, 1999). For example, a peasant may migrate to urban areas for employment for many years, but he/she still cannot obtain permanent urban registry and will not have the same status and opportunities as urban residents. Wang also demonstrates the disparity created by the Hukou system in an example regarding health insurance. When approximately 79 percent of registered residents had health insurance in urban areas, there were only approximately 14 percent of unregistered residents who had health insurance in their work position (Wong, 2007). Therefore, this system imposes economic costs on migrants and reduces the incentives for the migration of rural residents to urban areas.

    Since urban stability has the highest priority in China‘s economic plan, the central government continues to differentiate its citizens and discriminate against people designated as rural Hukou (Wallace, 2007). In addition, Wang states that the economic inequality between rural and urban status also creates a legal double standard in judicial and administrative systems (Wang, 2006). For example, the family of a victim with a rural Hukou generally receives much less accident compensation than those with urban Hukou. Moreover, the temporary residential status of people with rural Hukou also makes it easier for the urban government to legally repatriate them from cities to the rural areas. The forced expulsion usually triggers anger in those affected, who may already harbor grievances due to the discrimination they experience while living in cities. In short, all these situations reflect a form of Hukou-based discrimination in society, leading to various forms of inequality and, consequently, an increase of conflicts within China.

    Nevertheless, Dorothy Solinger argues that while rural migrants have become second-class citizens, they can still enjoy a certain level of citizenship in cities. She also argued that peasants can earn as much as twenty times what they could doing agricultural work (Solinger, 1995). Since the money they bring back into the rural economies improves the economic status of the rural population, most of them hesitate to participate in protests that may lead to their forced repatriation from the cities so most of them assent to the control of Hukou system. They would rather purchase Hukou, or even pay bribes to obtain entry to work in cities. Moreover, Solinger suggests that the Hukou system makes migrant workers more disorganized. Without spare resources and permanent residential status, workers are less likely to incite protests and instead resort to working overtime, saving money for families in the countryside and returning to rural areas if there is a need (Solinger, 1995). Without reform, the Hukou system will continue to maintain the urban stability, but at the same time solidify the social and economic inequalities between the rural and urban population.

    Ethnic conflicts

    The study of Chinese ethnic conflicts has largely focused on two ethnic minorities: Tibetan and Uyghurs. The most problematic of the minority areas are those with high populations of Tibetans, such as Sichuan, eastern Gansu, and Qinghai provinces, which cover almost one-fifth of the area of China. In order to maintain stability in these areas, the central government established a Tibetan autonomous region and allowed the Tibetans a limited degree of autonomy and preferential policies, including religion, customs, using their own language in government, and an exemption from the one child policy (Sautman & Eng, 2001). The Chinese national government also invested in infrastructure projects and implemented policies to encourage economic development and modernization in the Tibetan region. Nevertheless, the Chinese authorities refuse to tolerate any forms of separatism or any religious activities with the potential to promote it (Mackerras, 1993). In response to any form of protest concerning separatism, authorities will not hesitate to suppress them brutally through the Chinese military.

    Although the central government claimed that the exercise of mobilization will reduce regional inequality and promote ethnic integration between the Han and Tibetans, some scholars argue that most Tibetans have not benefited from government policies and the expansion of the Chinese economy. Hillman stated that Han immigrants are aggressive in opening businesses and have managed their businesses well; however, most Tibetans do not admire and learn the Han mode of business (Hillman, 2008). Wang also found that the Tibetans are poorly equipped to take advantage of government policy which promotes the economy of West China, and most Tibetans are less competitive than the Han when trying to find employment (Mackerras, 1993). Hillman (2008) suggests that poor education is one of the most crucial factors explaining why Tibetans are more vulnerable to unemployment. The standard of education varies by province and only a few educated people are willing to teach in Tibet. As a result, the education level of most Tibetans is very low, leading to their inability to secure employment. On the other hand, the Uyghurs, a predominantly Islamic, Turkic people, experience a similar problem as the Tibetans, and complain that they are marginalized regarding education and employment in comparison to the Han immigrants. At the same time, Smith mentioned that some Han people, who behave like colonial masters, regarded minorities such as the Uyghurs as an inferior stratum. This “great chauvinist attitude” of the Han contributed to their hostile relationship with the Uyghurs (Smith, 2002). After September 11, 2001, terrorist attacks also deteriorated the already tense relationship between the Han and Uyghurs by increasing the mutual fear and mistrust among their populations. Ethnic inequalities and discrimination also foster instability in the autonomous regions. Although it did not effectively challenge the legitimacy of the Chinese government, a range of terrorist attacks and violent protests that targeted the Han people occurred in Xinjiang during the Beijing Olympics of 2008. The purpose of the resistance of minorities is completely different from the policy resistance in the other provinces of China. In comparison to other conflicts, this conflict demonstrated an overall higher level of violence, a stronger ethnic identity, and grievances stemming from injustice and the hegemony of the central government rather than a restriction on religious freedom by local governments.

    Economic factors

    The study of the relationship between the Chinese economy and protests often focuses on unemployment and the agricultural tax. Mason argues that the scale of student demonstrations was small and not nationwide at the early stages of the Tiananmen Square protest. Students eventually formed a coalition; however, with a number of unemployed workers who shared the belief with students that democracy can be a solution for their economic difficulties (Mason, 1994). In the late 1990s, labor protests continued as workers did not have enough income for their subsistence while economic growth during the 1990s brought huge profits for private or state-owned enterprises. Therefore, the motivation for protests increased among workers because they believed their economic difficulties were exacerbated by managerial corruption (Chen, 2000). On the other hand, Cai found that economic welfare drives the grievances for resistance while the majority of the rural resistance is directed against local governments at the city and lower levels (Cai, 2008a). Without adequate funding from the central or provincial government, local governments were forced to rely on informal fees, taxes, fines, and apportionments in order to fund public goods, such as building infrastructure, education and medical services. If peasants refused or were unable to pay the taxes and fees, some cadres hired local thugs to threaten the peasants through brutal methods, leading to villager resentment.

    Political structure

    Research on political institutions has focused largely on village elections and the bureaucratic system. After the economic reform of the 1980s, peasants sought to gain a certain level of property rights, such as land use rights, transfer of lease rights, and residual income as well as more rights to participate in the policy making process (Shi, 1999). He also mentioned that the Chinese Communist Party used elections to mobilize peasants to carry out “predetermined party goals, to socialize them into official norms, and to legitimate the rule of the CCP” (Shi, 1999). Therefore, local elections were permitted in rural villages. However, Zweig argued that the elected village representatives only have rights to manage village finances and do not possess the power to supervise the performance of local officials and resolve disputes, such as land expropriation and inadequate compensation. In Zweig’s case study, the elected village representatives also embarked on a frustrating effort to meet with local officials who possess the authority to solve the problems (Zweig, 1999). If local officials refuse to make any responses to their requests, the representatives are unable to make any adjustments in policies or resolve the disputes by negotiation.

    On the other hand, the Chinese bureaucratic system affects how local officials respond to the resistance of citizens. Since the officials are accountable to upper level authorities rather than the citizens, they are only concerned whether the resistance threatens their performance in reaching economic plans and social stability. In order to prevent the protests from gaining the attention of top level officials, local officials rarely tolerate any form of resistance (Cai, 2008b). They tend to suppress the protests and selectively arrest the leaders of the protests, especially the village representatives (Bernstein, 2004). As a result, protesters resort to escalating the level of resistance until they can gain attention from the central government.

    In conclusion, ethnic conflicts comprise a large portion of the aggregated number of mass incidents in China over the past 20 years (Wedeman, 2009). However, most of these ethnic conflicts only occur in the peripheral regions, and do not spread over the majority of China. The protests of some ethnic minorities, such as the Uyghurs, differ greatly from the other protests in that they seek autonomy and independent power in their respective regions, which triggers most of the central government’s fear of separatism. Most Tibetan and Uyghur citizens, outside of the fringe terrorist organizations, are also resentful of the terrorist attacks by separatists. Consequentially, this type of mass incident is relatively less influential and limited to only a few autonomous regions. In addition, some scholars such as Fischer (2008), and Sautman and Eng (2001) argue that most ethnic conflicts are more strongly related to urbanization. They believe that the misallocation of economic resources leads to conflicts between rural and urban residents, and thus deteriorate the relationship between the Han and the various ethnic minorities. The Han can obtain more opportunities and economic benefits primarily because most of them live in urban areas of Tibet; however, most Tibetans and Uyghurs are excluded from the benefits of urbanization and economic improvement (Sautman and Eng, 2001; Fischer, 2008).

    In the end, a strictly economic perspective cannot completely explain the source of the conflicts in the various provinces in China. Although unemployment is not high in some major cities, such as Beijing, Shanghai, and Tianjin, this does not mean there are fewer mass incidents in urban areas than in rural areas. Moreover, the main source of high unemployment is strongly related to urbanization and the Hukou system. A large quantity of surplus rural labor cannot be absorbed by cities because peasants, the major economic demographic of China, are restricted by local governments in their decision to migrate. Additionally, the level of urbanization has grown rapidly since 2000, leading to a larger income gap between urban and rural areas. Therefore, the study of urbanization might provide a better explanation about the main source of conflict in recent years.

    The bureaucratic system might also be one of the main sources of conflict between local governments and residents. The institutional perspective explains why local governments have to suppress the protesters and why protesters have to use violent resistance and escalate the scale of protests in order to get the attention of upper-level authorities. This political structure is consistent among all provincial and lower level governments. Therefore, this alone cannot explain what characteristics of a province would make it more likely to generate contentious interactions between local officials and residents and a higher occurrence of protests in certain regions. As a result, in order to understand the source of conflicts and the spatial distribution of mass incidents, a study of the bureaucratic political structure must be combined with urbanization in China. In the face of urbanization, the central and local governments exercise different policies in certain regions which may affect the lives and economic benefits of local residents. If local officials cannot provide adequate responses to public grievances, these regions will unavoidably become a hotbed of mass incidents.


    In 1981, the Chinese Communist Government abandoned the state socialism model of economic development and implemented a range of market-based economic reforms referred to collectively as the Four Modernizations. In the agricultural sector, control over state-owned land resources were transferred to township and village enterprises (TVEs) which stimulated competition among these enterprises. The reform involved a certain level of decentralization of political and economic power from the central government to the regions and localities (Harvey, 2007). During the reform process, local officials obtained more administrative autonomy in the management of local economic resources (Iredale, Bilik, Su, Guo, & Hoy, 2001). After the dismantling of the People’s Commune system, peasants could lease land from the local government, sell their surplus crops in open markets, and no longer needed to be allocated economic resources by the state. Instead, they were granted more property rights, including residual income rights, transfer of lease rights, and land use rights. They could accumulate surplus crops and sell them in markets. After they gained more economic freedom, they requested more opportunities to participate in the process of policy-making (Shi, 1999). When the world was convinced that the economic reform might eventually lead to political democratization in China, the calamity in Tiananmen Square protest interrupted the progress and deterred people from the pursuit of democracy.

    After 1989, problems arose whereby local bureaucrats, who obtained the power of administration and legislation in localities, did not need to be accountable to the local population for their policies. They could make policies in favor of their own interests. First, they gained more opportunities to receive bribes from the land transactions they oversaw. Second, they took an official interest in producing growth that would reflect well on their performance in the eyes of their superiors in Beijing. Local officials also have their interpretation in the laws stipulated by the central government (Wong, 2007). Indeed, the central government is less able to supervise how local governments implement their policies. Some of the senior officials are also interested largely in performance indicators rather than the strategy of local governments for producing the outcome. In 1987, the National People’s Congress adopted the Organic Law of Village Committees, allowing the chairman, vice chairmen, and members of village committees to be directly elected by villagers, and enabling some local bureaucrats, who were responsible for holding the elections, to employ their authority to disrupt the elections during the process of implementation (O’Brien, 2005). Although elections were conducted under the Organic Law of Village Committees, the party secretaries are only elected by party members and peasants do not possess the right to nominate candidates for party secretary who supervises the village committees and has more influence in the process of policy-making than the elected village committees (Shi, 1999). Without the rights of nomination, the local elections are still not universal for all village members. Thus, the elected leaders cannot effectively bring the voice of villagers to the national congress and affect the governance of local governments.

    On the other hand, there are frequent conflicts between the elected village representatives and the unelected party cadres and local officials. Since the elected village leaders have to be accountable for the economic plights and demands of their constituents, they cannot just obey the orders from above (Bernstein, 2004). On the contrary, the evaluation of local officials by the national party and government is subject to their capacity to maintain social stability and achieve development goals, particularly GDP growth. In order to gain promotion in the bureaucratic hierarchy, officials must be more concerned about reaching clearly discernible standards, such GDP growth and the increase in urban areas, rather than placating the local public’s grievances. As a result, the policies of local party officials, especially land policies, may be strongly opposed by local elected leaders, who are accountable (through elections) to local constituencies. The conflicts between local officials and elected leaders would contribute to wide scale peasant protests in suburban areas.

    The bureaucratic system also influences how local officials “mediate” disputes that arise at the local level. The presence of petitions may harm the image of a government official and signal local officials’ failure of administration to higher-level authorities. If the officials tolerate the protest, it may threaten the evaluation of their performance in maintaining social order and following the development plan of the central government (Cai, 2008a). If it is unnecessary for them to be accountable to citizens, the choice with the lowest cost for them is to suppress and conceal the occurrence of petitions at the early stage. Accordingly, unlike the Tiananmen Square movement, the goal of protests is to force local governments to redress local grievances rather than to threaten the regime of the communist party. However, most local officials overreact to the level of the protests and suppress the protesters with coercive means instead of negotiations.

    On the other hand, the Party-dominated political structure contributes to the unprecedented power of local officials and the weakness of the judicial system. The laws, which are made by the bureaucracy, are generally designed to serve the development priorities of the central party. If local bureaucrats who are in charge of implementing policy are also responsible for monitoring the implementation of those policies, there are no incentives for officials to give adequate attention to the grievances of citizens. In addition, courts have little authority to protect the rights of citizens by overturning the decisions of local government (Zweig, 1999). Therefore, the judicial protection of citizens is restricted to a considerable extent. The courts are not an effective means for citizens to seek redress of grievances that arise as a result of development policies adopted and implemented by local officials in the service of national party/state development priorities. Local officials can selectively enforce the laws and legally arrest a limited number of petitioners, such as the elected village leaders, in order to reduce the occurrence of resistance (Cai, 2008b). Consequently, citizens lose their confidence in the judicial system and in negotiations with local officials as means of seeking redress of their grievances. Radical resistance becomes the only remaining option for them to force officials to take account of their grievances.

    It is interesting that citizens gradually learn the rules of the game in this political structure, meaning they recognize how to struggle for their interests in the bureaucratic system. In order to resolve the problems, they believe that they only need to inform the top levels of the hierarchy of their plight. If their problems cannot be resolved by a certain rank of officials, they will seek help from a higher level official until they receive a satisfying response from the government. During this process, they believe there must be a response from the central government as long as they protest on a large enough scale to gain attention.

    In short, the decentralization of political-economic power and the limited function of local elections make it more difficult for the central government to constrain the power of local bureaucrats in China. Local bureaucratic agencies captured more autonomy and control over local affairs. Without universal elections, it is not necessary for the local officials, the policy makers, to put a priority on public interests. Some of their policies may only favor the vested interest groups and not gain the agreements of most local citizens, leading to anger and grievances. On the other hand, since the bureaucratic system fails to respond to the demands of citizens and hampers the function of courts, the only solution for dissatisfied citizens is to inform the central government of their plight through large scale and violent protests.

    The next question should be which local policies lead to the grievances of most citizens and, thus, increases in the occurrence of mass incidents in certain provinces. Since the implementation of local policies is usually triggered by certain social and economic changes, the answer is certainly related to the most dramatic social and economic phenomenon in China – urbanization. In the face of urbanization, local governments employ different policies in their regions in order to maintain rapid economic growth and social stability. Therefore, certain discernible patterns of spatial distribution of mass incidents can be identified with the understandings of urbanization.

    Urbanization, Land Disputes, and Protests

    In the late 1970s, the implementation of the Household Responsibility System not only increased the productivity of agriculture, but also created a large amount of surplus labor in rural areas. Thus, rural workers migrated to urban areas and engaged in non-agricultural production (Zhang, 2003). This large flow of internal migration created the rapid process of urbanization and imposed a huge burden on the resources available in big cities, such as food, water, jobs, and infrastructure (Goldstein, 1990). Considering that major cities such as Beijing, Shanghai, and Tianjin have limited capacities to absorb more than 200 million surplus laborers, the central government adopted a Household Registration system (Hukou) to restrict migration to large cities and encourage citizens to migrate to middle or small towns and cities around the large cities (Goldstein, 1990, p.3). On the other hand, in the suburban areas, officials tended to promote their localities to cities or urban towns in order to reap more economic benefits and freedom from the central government (ESCAP, 1990). Therefore, more villages and towns were approved as urban rosters through annexation to existing urban areas or through reclassification, expanding the urban areas and urban population. Nevertheless, most of their land area is still farm land (Goldstein, 1990).

    However, in order to promote the localities to urban areas, local officials must develop the land and increase state-owned enterprises to absorb surplus labor from large cities and the countryside. In this process, they have to increase the quantity of land available for industrial and commercial expansion. When incorporating this land policy, local governments commonly expropriate farm land and residential neighborhoods from villagers who live in suburban areas and the newly classified urban towns. However, the policy of land expropriation commonly leads to disputes. According to my case study, some local governments expropriate and sell the land to developers without the agreement of residents (NY Times, 2010). Because of the soaring land prices in suburban areas, local governments can earn huge amounts of revenue from selling land. Still, the local governments are not willing to devote an adequate portion of the land revenue for compensation to the residents. Instead, they leave the developers to negotiate with the residents over compensation and housing relocation. When the residents complain to the developers about the inadequate compensation and the issue of forced eviction, the local officials stay away from these problems and refuse to respond to their grievances. As a result, affected villagers form resistance protests against the land policies. Since most local officials tend to suppress the resistance, the conflicts between the local officials and villagers are more likely to occur in these areas.

    In other words, land disputes are encouraged by the urbanization in China. If a certain province has a higher level of urbanization, meaning a rapid increase in the size of urban land subject to land expropriation, it will also experience an increase in the occurrence of mass incidents. In addition, local governments only expropriate land from the regions surrounding urban areas. Thus the provinces which are located closer to major cities, such as Beijing, Shanghai, and Tianjin should have higher frequency of mass incidents compared to other cities and provinces.

    In order to slow down the growth of population in big cities, the central government has tried to control the flow of internal migration through the implementation of the Household Registration System (Hukou) that was adopted in the 1950s.Most citizens are registered either as rural Hukou or urban Hukou. The registration of a citizen’s location determines whether he/she is entitled to economic welfare benefits in the area. For example, if citizens registered in rural villages migrate to cities, they cannot obtain the same social benefits as registered citizens of those cities, including medical benefits, social security, job opportunities, education, and housing subsidies (Wong et al., 2007). They may be assigned temporary residential registration to work in urban areas, mostly in state-owned enterprises. However, they may earn a lower salary and be prohibited from becoming permanent residents of cities, even though they have stayed there for periods extending from a few days to several years or more (Goldstein, 1990).

    Temporary residential status also makes it easier for local governments to repatriate the rural residents to the countryside if urban economic growth slows and demand for surplus labor from rural areas declines (Solinger, 1995). Therefore, temporary residents, or the “floating” population, are allowed to live in cities only when there is a high demand for workers in physically demanding jobs such as manual labor, construction, toy factories, electronics, and service industries, as those are also the jobs that most urban residents detest (Roberts, 2000). The migrant workers are also subjected to a great deal of exploitation. Many factories delay paying wages in order to prevent migrant workers from leaving freely (Dittmer & Liu, 2006). In urban areas, workers with rural Hukou suffer from various kinds of Hukou-based discrimination, leading to their anger and grievances with the local governments.

    There is no doubt that labor disputes stemming from Hukou-based discrimination contribute to the occurrence of mass incidents in cities. The administrative power of the bureaucracy, however, and the effect of a planned economy also affect the spatial distribution of labor protests in the various provinces in China. The implementation of the Household Registration system and the large share of the work force that is still employed by state-owned enterprises increase the opportunity costs of resistance for migrants, and thus function to restrain the protests against local governments. Nevertheless, the control of the Hukou system may be diminished by the presence of foreign enterprises, causing higher occurrence of mass incidents in certain provinces.

    First, a large proportion of residents, especially unregistered residents, work in state-owned enterprises in most of the provinces in China, except Guangdong, Fujian, and Zhejiang (China Statistical Press, 2010a). Employment and internal migration are not fully market-driven, meaning most of the labor allocation has to be bureaucratically arranged by local officials (Solinger, 1995). Migration is not subject to the incentives of rural residents and the demand of labor in urban areas. Local officials can control the ability of rural residents to move into urban areas in order to work in state-owned enterprises. Moreover, local officials devote much effort to regulating, taxing, and fining the private firms that belong to the temporary residents because unregistered residents are prohibited from obtaining licenses for doing business in urban areas (Solinger, 1995, p.135). Thus, temporary residents have fewer opportunities to work outside the state-owned enterprises.

    Local officials possess the power to send unregistered workers back to the rural areas as a punishment for any type of social disruption in urban areas. Furthermore, the income gap between urban and rural areas has become extremely large since the economic reform and openness of 1980s (Wang, 2006). Despite the fact that migrant workers suffer from Hukou-based discrimination, they can earn much higher salaries and have more job opportunities as unregistered urban workers than they can earn in the rural communities where they are registered. As a result, the opportunity costs of protests are high for unregistered workers in cities (rural Hukou). Most of them are hesitant to participate in any kind of resistance, such as strikes, protests, and petitions, which may lead to their expulsion from urban areas and the loss of their jobs in state-owned enterprises.

    Some registered residents may stage protests against outside laborers because they are worried that the temporary workers may undermine their own job security, social benefits, and wage levels (Solinger, 1995). However, the proportion of unregistered residents is still relatively small in some provinces and they are excluded from social welfare benefits provided through the urban government. Moreover, the central government is not willing to allow local managers of state-owned enterprises to fire workers in order to maintain urban stability (Wallace, 2007). As a result, the presence of state-owned enterprises and a rural-urban income divide decrease the likelihood that migrant workers and urban workers will resort to protests for better working conditions in urban areas.

    Second, there are significant numbers of workers employed by foreign enterprises in the urban areas of some provinces, particularly Guangdong and Zhejiang which are located close to Taiwan and Hong Kong. Since these workers do not work in state-owned enterprises, local officials are less able to control them through the Hukou system because they cannot dismiss them from jobs in foreign-owned enterprises. Local governments have more difficulty deterring temporary workers from engaging in protest behavior because the managers of foreign enterprises do not possess the same power as local officials and, therefore, cannot punish workers who engage in political or social protests through forced expulsion from urban areas.

    In some provinces, the majority of the urban populations are residents designated as rural Hukou rather than urban Hukou. For example, there are approximately 8,760,000 permanent residents in Shenzhen, one of the large cities in Guangdong. However, the size of the unregistered population is 6,440,000 which is approximately 73.5 percent of the total population (China Statistical Press, 2010b). Thus, most workers suffer from Hukou-based discrimination and exploitation by employers. According to Frohlich’s and Oppenheimer’s theory, “…as group size increases, the size of the contribution required from any one individual approaches zero while the probability of the good’s provision approaches one” (Frohlich and Oppenheimer, 1970; Mason, 1994, p. 406). Accordingly, the increase in group size will diminish the effect of the free rider problem and enhance the rational actor’s incentives to join the group of resistance (Mason, 1994).

    The income difference between state-owned enterprises and foreign enterprises is also a crucial factor in the cause of mass incidents. In some provinces, such as Guangdong and Zhejiang, the incomes of workers are higher in state-owned enterprises than in foreign enterprises (China Statistical Press, 2010a). Since the salary of workers in state-owned enterprises becomes the expected income for some migrant workers, they often feel they are exploited by their foreign managers more so than do those who work in state-owned enterprises. Most migrant workers resent the fact that they cannot enjoy the full benefits of economic growth although the policy of openness boosted the economic growth in some provinces. In fact, their income grows at a much slower rate than the growth of GDP and inflation, generating a dangerous level of income inequality in urban areas (Wang, 2006). In this manner, income inequality contributes to their grievances and thus the occurrence of mass incidents.

    The central government may also be less capable of resolving the labor disputes between foreign enterprises and workers. Not only do the foreign enterprises serve as the main competitors for the state-owned enterprises which are supported by local governments, but also the central government and local governments cannot effectively force foreign enterprises to pay minimum wages to the workers through state laws (Wines & Ansfield, 2010). Since the 1980s, the central government has tried to regulate the labor relations and wage levels in foreign enterprises through a series of regulations and laws (Tan, 2000). Workers were also granted the right to establish and join trade unions; however, many foreign enterprises now discourage or even restrict the establishment of trade unions within their enterprises (Wong et al., 2007). Some local officials also avoid forcing foreign investors to establish trade unions because those officials may have an ownership stake in the foreign enterprises (Wines & Ansfield, 2010). Without the support of local governments and trade unions, workers have fewer opportunities to negotiate with the foreign enterprises unless they resort to large scale resistance.

    On the other hand, sometimes local governments are less aggressive in resolving labor disputes for foreign enterprises because they try to narrow the wage gap between SOEs and foreign enterprises. This makes it more difficult for the central government to reform state-owned enterprises and reduce the financial burden of providing public goods to employees of SOEs in urban areas. If the wages of workers are still much higher in state-owned enterprises, they are not willing to leave the SOE for work in the privately-owned or foreign-owned enterprises. Local governments cannot fire workers in SOEs or reduce the number of state-owned enterprises without harming urban stability. Similarly, the central government cannot reform the communist public good regime, referred to as the “iron rice bowl” of state workers, if the income is still not enough for workers to maintain their living standard. To make matters worse, the central government and local governments lack enough tax revenue to support or extend the supply of public goods. Since the 1980s, the central government has offered preferential tax policies to attract foreign investors, including generous tax holidays, reduced tax rates, and tax refunds (Wang, 2006). Some local governments even provide extra tax treatment to foreign enterprises in order to compete with other provinces. However, in the process of attracting foreign enterprises, the central government and local governments have difficulty providing public goods for citizens because the tax burden on foreign enterprises is almost half of what it is for domestic enterprises, leading to a large loss of tax revenue (Wang, 2006). They may not have enough funding to support public goods for urban citizens, let alone the residents with rural Hukou. Nevertheless, most of them avoid putting too much pressure on foreign enterprises in order to maintain economic growth. As a result, if the foreign enterprises are compelled to increase the wages of workers under the pressure of workers’ resistance, such as protests and strikes, it can improve the living standards of citizens, accelerate the reform of SOEs, and reduce urban citizens’ reliance on social welfare.

    In summary, the proportion of workers, especially migrant workers, who work in foreign enterprises, is much larger in some provinces. Many of the foreign enterprises also pay a much lower salary than state-owned enterprises. At the same time, the effect of the Hukou system is less significant for workers in foreign owned enterprises. With less suppression by government, they are more likely to form large scale resistance for their rights and interests, such as the wage rates and the establishment of independent trade unions.

    In order to understand how the political structure and rapid urbanization lead to the occurrence of mass incidents, it is necessary to analyze the effect of land policy and migration policy in each province. For the land policy, the major indicator should be the level of land expropriation, which can be determined by the expansion rate of urban areas and the number of large cities in each province. In regard to migration policy, it would be better to focus on the impact of the presence of SOEs and foreign enterprises on the Household Registration system among provinces, and the social and economic conditions which are influenced by these enterprises, including wage rate, number of employed workers, and the unregistered population. Based on my theory, all of these factors are significant to the differential occurrences of mass incidents among provinces in China.

    Research Design

    Research Design

    This research will use data from 22 regular provinces, 5 autonomous regions, and 4 municipalities in China over three five-year time frames from 1995-2008 to determine if certain features of the political, economic, and social structure affect the frequency of mass incidents in certain provinces when China experiences rapid urbanization. Officially, the autonomous regions and municipalities are not classified as provinces (China Statistical Press, 2010b). All of them, however, have their own government. In the statistical yearbooks, the China Statistical Bureau puts these regions into the same categories with the remaining provinces and has the same period of observation. Accordingly, this research categorizes the autonomous regions and municipalities with the other 22 provinces, totaling 31 provinces. Furthermore, in order to comply with my dependent variable of mass incidents, this study focuses on three 5-year time periods. Therefore, the unit of analysis for this study is 31 provinces over three 5-year time periods, which generated 93 observations.

    Dependent Variable

    In order to measure the dependent variable, the number of mass incidents (massincidents), I utilized data compiled by Andrew Wedeman (2009). In his study, he defined mass incidents as protests, demonstrations, strikes, and riots, and then he counted all available news stories about mass incidents reported by the Western press between 1995 and 2008 to analyze the sources and significance of social unrests in 31 provinces. In regard to the time series, Wedeman divided fifteen years into 3 groups of five-year increments. He did not, however, provide the number of mass incidents in certain provinces and certain five-year increments. This does not mean that there were no occurrences of mass incidents in the province during that period, but the scale of resistance might not have been large enough to gain attention from the Western press. Considering that these cases mostly did not reflect the major source of mass incidents in China, I will set the number of their occurrences as zero in the dataset. Among 31 provinces, the report showed that Tibet, Xizang, and other peripheral regions had the highest level of mass unrest in the 1990s. However, in the 2000s, the provinces in the coast region, such as Guangdong and Zhejiang, assumed the top place. Moreover, the occurrence of mass incidents in Sichuan dropped in the early 2000s, but increased dramatically in the late 2000s. This indicates that the occurrence of mass incidents declined because of certain phenomena. Therefore, urbanization, which occurred in coastal China and parts of western China, may serve as a predictor for the spatial distribution of mass incidents.

    Independent Variables

    To measure the effects of urbanization, I used data regarding local governments’ land policies and the effect of the implementation of the Household Registration system. This data was compiled from the China Statistical Yearbooks. For the effect of land policy, land expropriation is a common strategy for local governments to expand urban areas (Jacobs, 2010). Since land expropriation would also lead to the eviction of peasants from their cultivated land in each province, I calculated the ratio of urban land to cultivated land per province (ratioland). This variable indicates the level of urbanization in each province. Thus, the province with larger ratios of urban land to cultivated land will experience more mass incidents.

    The number of big cities (numberofbigcities) also affects the level of urbanization in provinces. Urban growth is more likely to occur around big cities where the size of the population is more than two million. Therefore, if a province has more big cities than the others, it is expected that the local governments will exercise more frequent land expropriation to increase the land available for urban development. This should be associated with higher frequency of mass incidents.

    To test the effect of the implementation of the Household Registration system (Hukou system), I use the ratio between registered population (totalrp) and unregistered population (totalup) in each province. The size of the unregistered population, however, is close to zero in a few provinces, leading to the presence of an infinite number in dataset. Thus, in order to determine the size of the population affected by the Hukou system, I will utilize the data regarding the ratio of registered population to total population. This ratio should have an inverse relationship with the frequency of mass incidents. The smaller the proportion of registered population in a province, the more people who live under Hukou-based discrimination in the urban area of a province. Although it would provide clearer pictures of the impact of the Hukou system if I were to analyze the population at the county level, this must be left to future research due to time and data limitations.

    The number of staff and workers employed in the state-owned enterprises determines if the local government is able to discourage people from resistance. If a large proportion of residents are employed in state-owned enterprises, there should be fewer protests because SOE employees may hesitate to join protests. On the other hand, the foreign enterprises, including the enterprises funded by the investors of foreign countries, Hong Kong, Macau, and Taiwan, do not possess bureaucratic power to send migrants back to the countryside as punishment for protesting. Therefore, the proportion of the workforce employed in foreign-owned firms increases relative to the size of the workforce employed in SOEs should be associated with higher levels of mass incidents. Another variable representing the ratio of the wages of workers in SOEs to the wages of workers in foreign enterprises is also important because foreign-funded enterprises are known to hire large proportions of unregistered residents. Therefore, if such an enterprise is present in a certain province and pays a lower salary to workers, resistance is expected to increase, due to a lower opportunity cost of participation in protests. To analyze this combined effect of foreign-funded enterprises and state-owned enterprises, I included in the analysis the ratios of SOEs workers to workers in foreign funded enterprises (ratiosf) and the wage of those two categories of workers (ratiowsf). If the values of these two variables are high, state-owned enterprises have more influence on the work force than foreign funded enterprises, and thus serve well to decrease the frequency of labor disputes.

    China’s three decades of openness to foreign investment have contributed to that nation’s remarkable growth but has also contributed to income inequality between urban and rural residents. Therefore, the variables representing the ratio of rural income and urban income (ratiourincome) are also used to determine if the income gap is related to the occurrence of mass incidents among provinces. The higher the level the income inequality in a province, the more mass incidents that province will experience. In short, the theory generates the following six hypotheses regarding the explanatory variables for land disputes and labor disputes.

    H1: The higher the level of urban growth, the higher the number of mass incidents in a province.

    H2: The more large cities (i.e., cities with a population greater than 2 million) in a province, the more mass incidents that province will experience.

    H3: The higher the proportion of unregistered residents relative to total residents in a province, the more mass incidents a province will experience.

    H4: The larger the number of workers in state-owned enterprises relative to workers in foreign enterprises, the fewer mass incidents a province will experience.

    H5: The lower the income of workers in foreign enterprises relative to that of workers in state-owned enterprises, the more mass incidents a province will experience.

    H6: The higher the income of urban residents relative to rural residents, the more mass incidents a province will experience.

    I also included several control variables in the models. First, I collected data on unemployment (uemployee) among provinces. I incorporated this variable to determine if the occurrence of mass incidents is related to the number of unemployed registered residents. Second, the total numbers of post and telecommunication offices (comm.) is also included in the control variables. Some provinces with large proportions of underdeveloped rural areas may have fewer resources for interprovincial communication. This means residents have more difficulty communicating with people outside of their province, especially with the top level officials who can intervene in local disputes. Therefore, these residents may have to resort to large scale local protests to gain attention from the mass media or top officials outside. Third, the ratio between rural population (ruralpp) and urban population (urbanpp) also reflects how many residents are affected by the land policy. If the size of the rural population does not change when the cultivated land decreases dramatically, it may affect their livelihood and thus their motivation for resistance. Fourth, the ratio of GDP per capita to rural income (ratiogdprural) was added to the model to test if the GDP grows at a much faster rate than rural income. Even though rural residents gain more income during the growth of the economy, they still suffer from inflation if the growth of their income cannot keep up with the GDP growth, especially for the unregistered residents who are not entitled to subsidies and welfare.

    In analyzing the causes of mass incidents in the various provinces of China, this study faces several limitations regarding data collection. First, numerous observations regarding the size of urban areas per province are missing from 1995 to 2000. Therefore, the lack of observations for this variable may affect the statistical significance of this variable. Second, since the Chinese government fails to provide data regarding the number of mass incidents within the various Chinese provinces, the data used for my dependent variable of mass incidents collected by Wedeman contains only the mass incidents reported by the Western press and the official Chinese news agency, Xinhua (Wedeman, 2009). Reports by the mass media, however, are commonly limited to the mass incidents involving violence. Therefore, the dependent variable for this study may be underestimated in this study.


    In this analysis, I argue that there is a relationship between the number of mass incidents and urbanization in the form of land expropriation and Hukou-based discrimination in Chinese provinces. The dependent variable, the number of mass incidents, measures how often this type of event occurs over a time interval (19 years). It is generated from random variation by social conditions. Using OLS regression with this even count variable may lead to inefficient, inconsistent, and biased estimates (Long, 1997). Moreover, the probability that a protest occurs during each five year increment is independent of the occurrence of other protests during that interval, but the probability of the occurrences is understood to be related to the independent variables. As a result, I will use a Negative Binomial Regression Model in order to evaluate the relationship.


    This analysis focuses on the causal relationship between the number of mass incidents and various features of urbanization and the effect of the Household Registration System. Using a negative binominal regression model, the results support the principal contentions of the theory. Briefly, there are several initial significant results from the test using a negative binomial regression. First, as expected, increases in the ratio of urban land to rural land leads to increases in the number of mass incidents. The expansion of urban land comes at the expense of rural land, and we expect that rural residents displaced by this process to react by engaging in protests. Second, the ratio of the number of workers in state-owned enterprises (SOEs) to those in foreign funded enterprises is inversely related to the occurrence of mass incidents. SOE employees are entitled to a set of benefits and job security (referred to as the “iron rice bowl”) that employees of foreign-owned firms do not enjoy. Therefore, we would expect the number of mass incidents to be higher where there is a larger proportion of the workforce employed in foreign-owned firms.

    I test three different models using a negative binomial regression in this analysis. The first one includes all explanatory variables and control variables, with the total number of mass incidents as the dependent variable. The second model focuses on land disputes only and includes the control variables and the explanatory variables associated with land disputes in the theory. The third model focuses on labor disputes only and includes the control variables and the explanatory variables associated with labor disputes in the theory. I describe the results of each model in more detail below.

    The results from Model 1 (Table 1) show that increases in the ratio of urban land to rural land (meaning expanding urban areas at the expense of cultivated land) reflect the rate of urban growth in provinces. The number of mass incidents in a province increases with higher levels of urban growth, lending support for Hypothesis 1. Table 2 shows that an increase of one standard deviation in the rate of urban growth translates into six additional episodes of mass incidents. Similarly, the number of big cities with a population greater than two million also has a significant effect on the total number of mass incidents. The more large cities there are in a province, the more government units are making claims on rural land. Therefore, we would expect more mass incidents to occur. Table 3 shows that a province that has five big cities, the maximum number of big cities in every province, experiences thirteen more mass incidents than a province that has no cities with population greater than  two million, all else being equal. These results indicate that the number of big cities in a province and the ratio of urban land to rural land (representing the level of land expropriation during a five-year period) have strong significant effects on the number of mass incidents. In this manner, local governments increase the quantity of available land around the big cities, leading to increases in the rate of urban growth. Similarly, a province with a large number of big cities has more city governments engaging in land expropriation to relieve the population pressure in their dominions and, therefore, we would expect more clashes between urban governments and rural populations who are affected by the land expropriation.

    Model 2 differs from model 1 that it excludes explanatory variables associated with labor disputes. Unlike total incidents, there appears to be no connection between the number of land disputes and the number of big cities in a province. Therefore, further research will be needed to determine why the number of big cities in a province affects the total number of mass incidents but not labor disputes or land .On the other hand, urbanization is one of the most influential social and economic factors in present-day China. In model 2, the ratio of urban land to cultivated land which represents the rate of urban growth continues to be positively associated with the occurrence of land disputes and its magnitude is similar to that in model 1, demonstrating support for Hypothesis 1.

    Model 3 shows the results for the relationship between mass incidents and variables representing significant features leading to labor disputes. The ratio between the number of workers in state-owned enterprises versus those in foreign-funded enterprises and the ratio of the wages of these two categories of workers are significantly related to the number of labor disputes in a province. The results provide evidence in support of Hypothesis 4. As the number of workers in SOEs increases relative to the number of foreign-funded enterprises, all else being equal, the province will experience fewer labor-based mass incidents. An increase of one standard deviation in the proportion of workers in state-owned enterprises is expected to result in a decrease of five labor-based mass incidents (Table 2). As noted earlier, workers in foreign owned enterprises have less job security and fewer benefits than those in SOEs. Therefore, we would expect fewer labor disputes where the ratio of SOE employees to foreign-owned enterprise employees is higher. Similarly, the ratio of wages of workers in state-owned enterprises to foreign funded enterprises also has a negative relationship with the occurrence of labor-based mass incidents, lending support for Hypothesis 5. With a one standard deviation increase in the wage ratio, the number of mass incidents decreases by eight (Table 2). These results in model 3 indicate that the presence of state-owned enterprises serves to maintain urban stability. Compared to foreign-funded enterprises, state-owned enterprises have more authority to punish disobedient workers, particularly by sending unregistered residents back to the countryside. If most of the urban residents work for state-owned enterprises and their wages are higher than those who work in foreign-funded enterprises, they are more likely to tolerate the conditions and curb their grievances. On the contrary, the rising number of foreign-funded enterprises in China also leads to a decrease in the ratio between the number of workers in SOEs and those employed by foreign-funded enterprises. The number of mass incidents will increase because the managers of foreign-funded enterprises lack authority to suppress the resistance of workers.

    Model 3 also demonstrates that the ratio of registered population to total population is negatively associated with mass incidents, showing support for Hypothesis 3. A one standard deviation decrease in the proportion of registered population to total population translates to an increase of six more expected mass incidents per province (Table 2). In other words, the decrease in the ratio of registered residents to total population implies that more unregistered residents are migrating to cities and suffering from the Hukou-based discrimination. The finding regarding the relationship between the structure of the population and the number of mass incidents also supports Mason’s theory: if the population of unregistered workers becomes large enough to diminish the free rider problem, they tend to join the resistance (Mason, 1994).

    The ratio of urban income to rural income also has a positive relationship with the occurrence of mass incidents, lending support for Hypothesis 6. The province with the highest ratio of urban income to rural income experiences thirty-one more labor incidents than the province with the lowest ratio. Considering that the ratios of wages and number of workers between state-owned enterprises and foreign-funded enterprises are also significant, these results supports Wong’s theory that the openness policy for foreign-funded enterprises contributes to a large income gap between urban residents and rural residents (Wang, 2006), and eventually leads to increases in mass incidents. Moreover, it is interesting that the province with the highest ratio of urban income to rural income was Tibet (Xizang) from 2000 to 2004. This result supports Sautman and Eng (2001), who argue that ethnic conflicts are strongly associated with urbanization leading to the misallocation of economic resources between urban residents and rural residents in Tibet (Sautman and Eng, 2001). Since there is a large rise in the Han population in the cities and the Han are more aggressive in opening and maintaining businesses (Hillman, 2008), the Han have many more economic opportunities and benefits than the Tibetans, leading to a higher occurrence of ethnic conflicts in the peripheral regions such as Tibet.

    Finally, the number of post and telecommunication offices, one of my control variables, is significant in all three models but the coefficients for the occurrence of mass incidents are positive when the opposite effect is expected. Tong’s study (1998) found that this variable demonstrated a moderate relationship with mobilization in China in the late 1980s. However, communication is related to the occurrence of mass incidents from 1995 to 2008 in every model. A one standard deviation increase in the total number of post and telecommunication offices translates into eight additional mass incidents per province (Table 2). The result suggests that it will be easier for protesters to mobilize people to join a protest if they live in a province with a developed system of communication. On the other hand, Mason found a positive relationship between unemployment and the protests in the 1980s, especially the Tiananmen Square Protest of 1989 (Mason, 1994). Therefore, I included unemployment as a control variable in the models. This variable is not significant in all models, and the coefficient performs in the opposite direction of the expected relationship with the number of mass incidents. This finding suggests that unemployment is not a significant explanatory variable for the occurrence of mass incidents from 1995 to 2008. Despite this finding, the study shows that land disputes and labor disputes play significant roles in explaining the distribution of mass incidents among provinces.


    While China’s economy expanded more than 200 percent over the last ten years, as measured by the yearly change in GDP (China Statistical Press, 2010a), it is shocking that the number of mass incidents increased by 110 percent over roughly the same time period (Wedemen, 2009). Scholars offer several explanations for the increase in mass incidents, but quantitative analyses of these trends are a small minority of the studies on this topic in the field of Chinese politics (Dittmer & Liu, 2006), particularly regarding civil conflict between the Chinese government and its population. Most of these studies are based on interviews and data collected from a single locality over limited time periods. At the same time, there is a considerable variation in the occurrence of mass incidents across provinces, and this variance needs to be analyzed with quantitative methods. Most of the mass incidents are caused by local grievances. Without a cross-provincial comparison of the mass incidents in China, it is difficult to determine the main source of the grievances, and consequently predict the spatial distribution and the number of mass incidents across provinces in China. As a result, this study has attempted to quantitatively test the relationship between mass incidents and various socioeconomic factors in 31 provinces over three 5-year time periods: 1995-1999, 2000-2004, and 2005-2009.

    The analysis of the impact of urbanization on the occurrence of mass incidents in China highlights several possible implications. First, this analysis shows significant support for the ratio of urban land to rural land which is positively associated with the occurrence of mass incidents. This indicates support for my theory that the land expropriation policies by local officials accelerate the growth of urban areas, but also lead to the grievances of peasants who have lived in suburban areas for an extended period. Second, this analysis finds support for the contention that the ratio of SOEs workers to workers in foreign-funded enterprises and the relative wage of these two categories of workers have an inverse relationship with the number of mass incidents within a province. Specifically, a province with more residents working and earning higher salaries in the SOEs may experience fewer occurrences of mass incidents than a province with relatively more workers in foreign-owned enterprises earning lower wages than workers in SOEs. This also illustrates the effect of the Hukou system on maintaining urban stability: to the extent that larger proportions of the work force in a province are unregistered migrants from rural areas working in foreign-funded enterprises instead of state-owned enterprises, the number of mass incidents should increase.

    Third, this analysis finds support for the contention that as the ratio of registered population to total population increases, the number of mass incidents decreases. In other words, if the size of unregistered population increases, the province will experience more mass incidents than other provinces. Fourth, this study finds strong support for the relationship between the development of communication networks and the occurrence of mass incidents, which is different from Tong’s findings of the Tiananmen Square Protest (Tong, 1998). It indicates that the development of communications networks decreases the risk of free rider problems and makes it easier for the protest leaders to mobilize protesters.

    This study, however, faces several limitations in regard to data collection. First, the missing variable regarding the size of urban areas decreases the number of observations and affects the significance of the variable. Second, the variable regarding unemployment includes only the registered population rather than total population. Thus, the level of unemployment is underestimated in the analysis. Third, the Chinese government does not publish data on some variables in Tibet and Qinghai, such as unemployment and number of workers in foreign-funded enterprises, thus the estimate of any casual relationship between unemployment and mass incidents could be biased by these data problems. Fourth, the number of mass incidents is collected by Wedeman instead of the China’s official statistical bureau. Since his data collection is based on reports in the Western press, this dependent variable may be slightly different from the actual number of mass incidents.

    Future research in this area can focus on a diverse array of issues. First, although most ethnic conflicts occur only in peripheral regions and do not reflect the grievance of the Han Chinese population, the occurrence of ethnic conflicts was still high in China during the 1990s. It would, thus, be advantageous to segregate ethnic conflicts from the category of mass incidents in order to provide clear findings of its characteristics and specific causes. Second, this analysis focuses only on the distribution of mass incidents among provinces; however, some features that may contribute to mass incidents will be more significant in an analysis conducted at a country level. For example, the proportion of unregistered population is approximately 80% in Shenzhen, one of the major cities in Guangdong. If we analyze this variable at the provincial level, the proportion of unregistered population falls to approximately 15%, leading to a decrease in its significance. Therefore, future endeavors may find more significant findings if the research is based on a country-level analysis of mass incidents in China. Third, the explanatory variables used in this research such as urban growth, the ratio of workers in SOEs to workers in foreign-funded enterprises, and the ratio of registered population to total population may be effective to predict the distribution and frequency of mass incidents in China. We could use these explanatory variables as indices to predict the probability of increases in the number of mass incidents in provinces. Furthermore, if one of the variables, such as the rate of urban growth, increases too rapidly in a province, it may also indicate certain policies the central government will implement to redress or compensate public grievances in order to prevent large scale mass incidents and protect the image of the Chinese Communist Party.

    The analysis has found positive support for the hypothesis that land disputes and labor disputes influence most of the mass incidents in China. It provides a new picture of mass incidents in China since the 1980s. The purpose of protests has been transformed from challenging the legitimacy of the Chinese Communist Party to informing upper-level party members about the blunders of local governments and local grievances. Thus, this theory suggests that the socioeconomic impact of urbanization encouraged the deeper contradiction between the public interests and local officials’ interest, leading to the different frequency of mass incidents among provinces.


    • Bernstein, T.P. (2004, August 1). Unrest in rural China: A 2003 assessment.eScholarship, University of California, Irvine: Center for the Study of Democracy. Retrieved from http://escholarship.org/uc/item/1318d3rx
    • Cai, Y. (2008a). Local governments and the suppression of popular resistance in China. The China Quarterly, 193 (March), 24–42.
    • Cai, Y. (2008b). Power structure and regime resilience: Contentious politics in China. British Journal of Political Science38, 411-432.
    • Chen, F. (2000). Subsistence Crises, managerial corruption and labour protests in China. The China Journal44 (July), 41-63.
    • China Statistic Press. (2010a). China Statistical Yearbook [1996-2010]. (Adobe Digital Editions version), Retrieved from http://chinadataonline.com/member/yearbooksp/default.asp? KeyTitle=&StartYear=1981&EndYear=2009&KeyType=0& Source=1&Region=All&ybcode=CHINAST&page=2
    • China Statistic Press. (2010b). Guangdong statistical yearbook 2009 [Chapter 4]. (Adobe Digital Editions version), Retrieved from http://chinadataonline.com/member/yearbook/default.asp?StartYear=1981&EndYear=2009&ybcode=GUANGDONG#
    • Dittmer, L., & Liu, G. (2006). China’s deep reform. Lanham, MD: Rowman & Littlefield Publishers.
    • Economist (2007, July 26). Urbanisation in China: China’s Chicago. The Economist. Retrieved from http://en.vigerjiang.com/documents/pdf/21-china’s%20chicago.pdf
    • Fishcher, A.M. (2008). What is the Tibetan population? How is it distributed? In A.M. Blondeau & K. Buffetrille. (Eds.), Authenticating Tibet, Answers to China’s 100 questions (pp. 151-152) Berkeley, CA: University of California Press.
    • Frohlich, N., & Oppenheimer, J.A. (1970). I get by with a little help from my friends. World Politics, 23(1), 107.
    • Goldstein, S. (1990). Urbanization in China, 1982-87: Effects of migration and reclassification. Population and Development Review16(4): 673-701.
    • Harvey, D. (2007). Environment and planning. New York: Oxford University Press.
    • Hillman, B. (2008). Money can’t buy Tibetans’ love. Far Eastern Economic Review,171(3), 8-16.
    • Ho, P. (1999). Contesting rural spaces. In E.J. Perry & M. Selden (Eds.), Chinese society: Change, conflict and resistance (pp. 93-112). New York: Routledge.
    • Iredale, R.R., Bilik, N., Su, W., Guo, F., & Hoy, C. (2001). Contemporary minority migration, education, and ethnicity in China. Northampton, MA: Edward Elgar Publishing, Inc.
    • Jacobs, A. (2010, July 12). Chinese factories now compete to woo laborers. NY Times, Retrieved from http://www.nytimes.com/2010/07/13/world/asia/13factory.html?_r=1&scp=7&sq=labor%20protest%20china&st=cse
    • Long, J., Leon, Joseph, Brown, Wayne, Sr., Sr., Ruch, Libby, Sr., Sr., Johnson, Thomas, Sr., Sr., Leon, Joseph, Brown, Wayne, Ruch, Libby, & Itamura, John. (2005). Telephone survey methods and practices. Streamline Surveys Inc.
    • Mackerras, C. (1993). Tibetans, Uyghurs, and multinational “China”. New York: Routledge.
    • Mason, T. D. (1994). Modernization and its discontents revisited. Journal of Politics,56, 400-424.
    • Nyberg A., & Rozelle, S. (1999). Accelerating China’s rural transformation. Washington, DC: World Bank.
    • O’Brien, K. J., & Li, L. (2005). Popular contention and its impact in rural China. Comparative Political Studies, 38, 235-259.
    • Sautman, Barry, & Eng, Irene. (2001). Tibet: Development for whom? China Information: A Journal on Contemporary China Studies15(2), 21.
    • Shi, T. (1999). Village committee elections in China: Institutionalist Tactics for Democracy. World Politics50, 385-412.
    • Smith, J. N. (2002). Making culture matter: Symbolic, spatial and social boundaries between Uyghurs and Han Chinese. Asian Ethnicity3(2), 172-174.
    • Solinger, D. (1995). China’s urban transients in the transition from socialism and the collapse of the Communist urban public goods regime. Comparative Politics27(2), 127-146.
    • Tan, S. (2000). The relationship between foreign enterprises, local governments, and women migrant workers in the Pearl River Delta. In L.A. West & Y. Zhao, Y. (Eds.), Rural Labor Flows in China (pp. 293-309). Berkeley, CA: University of California Institute of East Asian Studies.
    • Tong, J. (1998). The 1989 democracy movement in China: A spatial analysis of city participation. Asian Survey, 38(3), 310-327.
    • Wallace, J. (2007). Managing urbanization and measuring unrest in China. Paper presented at the 103rd Annual Meeting of the American Political science Association, Chicago, IL.
    • Wang, F. (1999). Conflict, resistance, and the transformation of the Hukou system. In E.J. Perry & M. Selden (Eds.), Chinese society: Change, conflict and resistance(81-99). New York: Routledge.
    • Wang, S. (2006). Openness and inequality: The case of China. Lanham, MD: Rowman & Littlefield Publishers.
    • Wedeman, A. (2009). Enemies of the State: Mass Incidents and Subversion in China. Paper presented at the 105th  Annual Meeting of the American Political science Association, Toronto, Ontario.
    • Wines, M., & Ansfield, J. (2010, May 26). Trampled in a land rush, Chinese resist. The New York Times, 1-2.
    • Wong, D.F.K., Li, C.Y., & Song, H.X. (2007). Rural migrant workers in Urban China: Living a marginalized life. International Journal of Social Welfare, 16(1), 32-40.
    • Zweig, D. (1999). To the Courts or to the Barricades. In E.J. Perry & M. Selden, M. (Eds.), Chinese society: Change, conflict and resistance (113-135). New York: Routledge.
    • Zhang, Honglin. (2003). Rural–urban migration and urbanization in china: evidence from time-series and cross-section analyses. China Economic Review, 14(4), Retrieved from http://www.sciencedirect.com/science?_ob=MImg&_imagekey=B6W46-4B505V6-77&_cdi=6534&_user=10&_pii=S1043951X03000695&_ orig=na&_coverDate=12%2F31 %2F2003&_sk=999859995&view= c&wchp=dGLzVzz-zSkWA&md5=3680a5aa3158ae351d1c8 02fa1dae088&ie=/sdarticle.pdf

    Table 1: Negative Binomial Regression Model

      Negative Binomial Regression Model: Dependent Variable: Number of mass incidents within 5 years increments
    Model 1   (Total disputes) Model 2  (Land disputes) Model 3  (Labor disputes)
    Number of big cities 0.2047*  (0.1145) 0.1302  (0.1153) 0.0829  (0.1150)
    Total number of post offices and telecommunication 0.0003***  (0.00008) 0.0003****  (0.0001) 0.0003****  (0.0001)
    Size of cultivated land 0.00003  (0.0005) 0.00003  (0.00004)  
    Unemployment of registered residents -0.0026  (0.0110) -0.5783  (0.0098) -0.0023  (0.0106)
    Ratio between urban land and rural land 0.8957***  (0.2961) 0.5783****  (0.1687)  
    Ratio between urban population and rural population -0.0695  (0.7036) -0.8587  (0.6410) -0.1390  (0.7310)
    Ratio between wage of workers in state-owned enterprises and foreign enterprises -0.3675  (0.6221   -1.0536*  (0.5871)
    Ratio between number of workers in state-owned enterprises and foreign enterprises -0.0049  (0.0057)   -0.0112**  (0.0047)
    Ratio between urban residents’ income and rural residents 0.4376*  (0.2621)   0.5407**  (0.2558)
    Proportion of registered population relative to total population 2.3615  (2.4957) 0.1531  (0.1287) -3.4770**  (1.7476)
    Ratio between GDP per capita and rural income 0.0164  (0.1344)   0.1815  (0.1363)
    *        p <  .10

    **        p <  .05

    ***        p <  .01

    ****        p <  .001

    N=60  Pseudo R2=0.1231 N=63  Pseudo R2=0.1076 N=62  Pseudo R2=0.0969

    Table 2: The Marginal Effects of the Number of Mass Incidents - Part A

    Independent Variables Total incidents Land disputes Labor disputes
    Number of big cities 1    
    Telecommunication 5 8 8
    Ratio of urban land to rural land 6 6  
    Ratio of urban income to rural income 6    
    Income ratio of workers in SOEs to Foreign enterprises     -5
    Ratio of number of workers in SOEs to foreign enterprises     -8
    Ratio of registered population to total population     -6
    Ratio of urban income to rural income     3
    ***Value determined by calculating the difference between the predicted number of mass incidents for one standard deviation above the mean and one standard deviation below the mean for the independent variable

    Table 3: The Marginal Effects of the Number of Mass Incidents - Part B

    Independent Variables Total incidents Land disputes Labor disputes
    Number of big cities 13    
    Telecommunication 36 52 51
    Ratio of urban land to rural land 582 89  
    Ratio of urban income to rural income 24    
    Income ratio of workers in SOEs to Foreign enterprises     -7
    Ratio of number of workers in SOEs to foreign enterprises     -8
    Ratio of registered population to total population     -25
    Ratio of urban income to rural income     31
    ***Value determined by calculating the difference between maximum and minimum