The Economic Origin of “Loyal Opposition”: Homeownership and Political Participation in China

Abstract

What shapes the ways in which citizens participate in politics? This article investigates the association between private homeownership and the forms of citizens’ political behaviors using a Chinese nationwide social survey. Exploiting the abolishment of the welfare housing system in the late 1990s as a quasi-natural experiment, I find that owning a home and experiencing home value appreciation increases citizens’ willingness for political engagement as well as participatory behaviors through formal channels, but reduces their confrontational behaviors towards government such as participation in protests. Further evidence on political attitudes suggests that homeowners are more critical of government performance, yet they report higher political trust in the state and a stronger preference for maintaining the status quo. These findings highlight the critical role of asset ownership in preventing conflict and promoting stability by shaping the political behaviors and beliefs of citizens.

Introduction

Balancing active civic engagement with stable political order poses an inevitable challenge for political systems. On the one hand, extensive participation in public affairs is vital for a well-functioning polity because it enhances government responsiveness, ensures accountability, and fosters a sense of citizenship (Verba and Nie 1987, Putnam 1992). However, on the other hand, excessive political engagement can provoke conflicts (Huntington 1968, Lohmann 1994). When participation becomes unrestrained and exceeds institutional limits, it can lead to violent confrontations, endangering both political stability and societal cohesion. Therefore, managing civic participation in a controlled and orderly manner becomes a critical concern for governing authorities.

What shapes the ways in which citizens participate in politics? There is a long-standing claim in political theory that links private property ownership with forms of civic engagement. Property owners are widely perceived as active political participants who engage in politics in a law-abiding manner out of self-interest motivation. ¹ In recent years, a growing body of political science literature has empirically examined the impact of housing on political behaviors. Studies have shown that homeowners exhibit a more active sense of citizenship, with higher participation rates in various political activities such as voting, making political contributions, and attending local meetings (McCabe 2013; Einstein, Palmer, and Glick 2019; Bueno, Nunes, and Zucco 2022; Hall and Yoder 2022). However, much of the available evidence stems from industrialized democracies, with limited attention given to housing politics in other parts of the world. Moreover, while existing studies have primarily focused on whether or not individuals participate in politics, there is less understanding of how housing influences the specific modes of political engagement.

Adding to this burgeoning literature, this article highlights the important role of housing in shaping the forms of citizen engagement. I argue that homeownership influences political participation through two mechanisms: The first mechanism involves the cost of exit. Homeowners, due to transaction costs and liquidity limitations, exhibit less mobility than non-homeowners, leading them to prefer “voice” over “exit” (Hirschman 1970) to preserve their property’s value. The second mechanism is linked interest with the existing order. When their interests are at stake, homeowners tend to engage in politics by using institutionalized channels rather than by inciting disruption. Hence, homeowners represent a “loyal opposition” that participates actively in public affairs to advocate for favorable policies, while maintaining a less confrontational approach toward the authorities.

I substantiate these claims in the context of China. One major challenge in studying property ownership is that the acquisition of property is an entirely endogenous issue (Brady, Verba, and Schlozman 1995). Variables as diverse as talents, social status, and economic well-being can influence property ownership and confound the causal effect. To address this challenge, I focus on China’s housing reform in the late 1990s, which abolished the distribution of free public housing to urban employees. This reform created a plausibly exogenous distribution of homeownership among different working cohorts. Using a nationwide social survey in China, I employed a generalized difference-in-differences (DID) design, leveraging the interaction between a dummy variable indicating whether individuals began working after the reform and a housing shortage variable measuring the difficulty of acquiring homeownership in the post-reform era. Employing the DID design, I initially estimated the decline in home ownership and further used it as an instrumental variable to predict political behaviors.

The results show that owning a home increases the willingness to engage in politics and participate in institutionalized forms of political involvement, such as engaging in community organizations and communicating with the government about community issues. In contrast, homeowners are less likely to engage in direct confrontation with the government, as evidenced by lower participation in protests. These findings remain consistent across various additional tests, including adjustments in the sample time span, alternative model specifications, and assessments of heterogeneous effects. Further analyses of housing prices reveal that the participatory impact of homeownership strengthens as home value increases. Additionally, the research suggests a mixed pattern of political attitudes among homeowners: while they exhibit greater scrutiny of government performance and heightened awareness of civil rights, they also express higher levels of political trust in the state and a stronger inclination to maintain the existing state of affairs.

This article directly contributes to the emerging literature on the politics of housing. The boom of the housing market in recent decades has brought it into the view of political scientists (Ansell 2019). While existing studies link homeownership to increased participation rate in public affairs (DiPasquale and Glaeser 1999; Einstein, Palmer, and Glick 2019; Yoder 2020; Kumar 2022), few explore how homeowners specifically participate in politics. This article examines the varying impacts of housing on both institutionalized and non-institutionalized forms of political behavior. It contributes empirical evidence to the long-standing argument linking property ownership with political participation within a regulatory way. Moreover, existing empirical research primarily originates from industrialized democracies with well-established electoral systems, where voting predominates as the main form of political participation. This has led these studies to mostly focus on the impact of homeownership on electoral behavior and policy preferences (Ansell 2014; Ansell, Hjorth et al. 2022). However, the article provides empirical evidence from non-Western contexts, where there are various non-institutionalized and disruptive forms of political participation. This enables us to concentrate on residents’ choices of different political participation approaches and further elucidates the crucial role of homeownership in balancing active engagement and political order.

The findings of this article also speak to the literature on political stability in China. In Chinese politics, a persistent puzzle revolves around how the state maintains political stability despite active citizen participation (Li 2019). While the existing explanations predominantly focus on governmental strategies and political institution (O’Brien 2008; He and Warren 2011; Lorentzen 2014; Chen 2017), this article emphasizes the role of homeownership in fostering a self-motivated drive to preserve political stability. In contrast to the conventional belief that urban property owners are driving forces for democratization, this article suggests that Chinese homeowners act as a “loyal opposition” to the rule of CCP, and that they influence policy-making through political participation rather than seeking to overthrow the state. Additionally, this article support the argument that China’s political legitimacy stems from performance (Yang and Zhao 2015). However, in contrast to the prevailing emphasis on economic growth or social welfare policies within existing research, this study highlights the political influence of the rapid escalation of housing prices and the sustained enhancement of household assets in urban China over the past two decades. Housing prosperity has supported urban homeowners to exhibit a behavior pattern of active participation while also abiding by rules, being critical of policies yet maintaining support for the government. This dual disposition not only underpins the foundational stability of the political order but also facilitates the expression of public viewpoints and advances in governmental efficacy.

Who engages in politics in an orderly manner?

The challenge for governing authorities lies in enabling citizens to participate in politics without jeopardizing political stability. Since the work of Huntington (1968), institutionalization has been regarded as a key factor influencing whether political participation has positive or negative effects. In less mature political systems, the rapid increase in political participation due to modernization often leads to a greater risk of regime overthrow. Conversely, highly institutionalized political systems, be they democratic or authoritarian, maintain stability by orderly regulation of political involvement. However, in recent years, social unrest has emerged in countries traditionally considered to have mature democratic systems (Berman 2021). Examples include the storming of the Capitol by Trump supporters in the US and prolonged street protests in France, posing enduring challenges for their respective governments (Bauhr and Charron 2023; Piazza 2024). Similar disruptions have also been observed within authoritarian regimes in East Europe and Central Asia, which have introduced quasi-democratic institutions such as elections, government transparency, and free media (Enikolopov, Makarin, and Petrova 2020; Kudaibergenova and Laruelle 2022). These democratic institutions have been found to unintendedly encourage disruptive behavior and decrease institutionalized political engagement (Hurst et al. 2014). This emergence of new phenomena underscores the need to reconsider factors beyond political institutionalization that influence how people engage in politics.

What shapes the way in which citizens participate in politics? Looking back to Ancient Greece, active political participation was primarily reserved for property-owning males. This system stemmed from the belief that property owners had a vested interest in the stability and prosperity of the state. In his Politics, Aristotle argued that the middle class, characterized by moderate property holdings, represented the most rational and moderate segment of society. He believed that this class was best suited for governance, as they were neither too wealthy to be arrogant nor too poor to be desperate. These ideas were further developed in political theories concerning private property rights. For instance, James Madison emphasized the importance of protecting property rights for ensuring property owners had a voice in government, leading to more moderate and deliberate rule. Edmund Burke, in his speech on the reform of representation in Parliament, argued that property owners, possessing both independence and a vested interest in the future of the nation, were reliable and prudent participants in public affairs. John Locke, in his “Two Treatises of Government,” asserted that individuals consented to government authority in exchange for the safeguarding of their property rights, thereby motivating them to engage in the political process to protect these rights. In brief, the role of property ownership in promoting both political activism and stability recurs throughout these theories.

When discussing property, political philosophers primarily discuss real estate assets. With urbanization and industrialization, housing has evolved into one of the most substantial forms of fixed property over the last century. For instance, in the UK, housing assets surged from 27 percent of national wealth in 1970 to 55 percent by 2010, as estimated by Piketty and Zucman (2014). This surge in the real estate market positioned homeowners as the predominant asset holders. Given the critical role of housing in individuals’ lives, there has been a growing interest in political behaviors associated with homeownership within political science literature. A longstanding claim in this field posits that homeowners exhibit higher levels of civic engagement (Cox 1982), and recent empirical evidence has further substantiated this claim. In developed democracies, homeownership shows a consistent positive correlation with multiple forms of political involvement, including higher voter turnout in local elections (André et al. 2018; Hall and Yoder 2022), engagement in city council meetings (Einstein, Palmer, and Glick 2019), contributions to politics (Yoder 2020), and participation in neighborhood groups and civic associations (McCabe 2013). However, the existing research on the political behaviors associated with housing lacks depth in two aspects. First, much of the empirical evidence stems from industrialized democratic countries, which raises concerns about the external validity of these findings. Second, despite previous studies identifying a stable positive relationship between homeownership and political participation rates, there is limited understanding of how housing specifically influences the ways in which individuals engage in political activities.

In this article, I posit that homeownership influences individuals’ political participation through two primary mechanisms. First, homeownership engenders linked interests with the status quo. Homeownership implies a substantial stake in the political and economic system (Fischel 2002). Fluctuations in housing values depend on factors such as economic growth, public services delivery, political stability, and law enforcement within a locality (Hsieh et al. 2011). These linked interests provide homeowners with a stronger motivation to engage in local public affairs to advocate for favorable policies. In democratic countries, homeowners tend to express their support or opposition to the incumbent government through economic voting. For instance, in Denmark, scholars have found that homeowners who have experienced an increase in property prices are more likely to support the incumbent government in their voting (Larsen et al. 2019). In an authoritarian context, the political participation patterns of homeowners are more complex. On one hand, homeowners need to protect their property rights from government infringement, which increases the frequency of their political participation. On the other hand, due to the fragility of authoritarian regimes, homeowners are less inclined to support the overthrow of the existing government, as this could lead to political unrest and property losses. For example, in Taiwan during the 1980s, property owners collectively pursued their interests but avoided direct confrontation with the authorities to prevent undermining the prevailing political-economic system (Hsiao 1990).

The second mechanism is that homeownership increases the cost of exit. As argued by Hirschman (1970), citizens facing a government policy that negatively impacts their well-being respond by utilizing either “voice” or “exit.” Voice refers to an individual seeking to change a dissatisfactory policy by “persuading” the government, while exit refers to an individual accepting a deleterious change but altering his or her behavior (Clark, Golder, and Golder 2017). Due to transaction costs and liquidity constraints, homeowners have lower mobility than renters, making them more inclined to attempt to influence and maintain their property’s value through “voice” rather than “exit.” In democracies, homeowners tend to be more actively involved in movements such as Not-In-My-Back-Yard (NIMBY), as renters can easily relocate, but property owners cannot (Enos 2016; Diamond and McQuade 2019). In authoritarian regimes, the higher exit costs suggest a stronger inclination toward political stability. For instance, in South Korea, one of the driving factors behind the democratization movement was the loss of political certainty under military dictator Chun Doo Hwan. Homeowners in Korea took to the streets to advocate for an orderly leadership transition, aiming to ensure stability and prevent upheaval during the movement (Jones 1998).

To briefly summarize, I argue that homeowners are more inclined to engage in politics to safeguard their private property. Moreover, their participation tends to be within institutional boundaries, as they avoid direct clashes with governments. Their active yet controlled involvement is shaped by their vested interests in the prevailing political and economic order and the high costs associated with relocation. In contrast, renters, being less invested in the system, exhibit less motivation to engage in political participation. When they do participate, they are less inclined to confine their actions within formal channels.

Hypothesis 1 (Willingness to participate): Compared to renters, homeowners exhibit a greater willingness to engage in politics to influence policies.

Hypothesis 2 (Ways of participation): Compared to renters, homeowners are more inclined to engage in politics through institutionalized channels and tend to avoid direct confrontations with the authorities.

The argument is extended to suggest that the impact of homeownership on political engagement intensifies amid rising housing prices. As home values appreciate, homeowners’ vested interests in the existing system strengthen, while the costs of exit increase. Furthermore, the growth in housing prices accentuates the divide between owners and renters (Ansell 2014; Dancygier and Wiedemann 2024). Homeowners experience additional wealth appreciation due to rising property values, fostering stronger incentives to engage in politics using institutionalized approaches. Meanwhile, renters face exacerbated housing affordability issues that make it challenging for them to transition into property ownership, further diminishing their ties to the existing system and dissuading political participation. This scenario hints at a more polarized participation trend in tandem with housing booms and growing wealth inequality.

Hypothesis 3 (Intensified effect of housing prices): As housing prices increase, the disparities in the willingness for and ways of political participation will widen between renters and homeowners.

Homeowners and political stability in China

I substantiate the claim using the case of China. The emergence of homeowners was a consequence of China’s market reform. During the era of the planned economy, the government took a substantial role in housing provision (Walder 1983). Urban residents predominantly lived in publicly owned housing distributed by state-owned work units at subsidized rental rates, with only a small minority residing on private property. This housing system, known as the “Welfare Housing Allocation System,” underwent changes with China’s market-oriented reforms beginning in 1978 (Wang and Murie 1996). A fundamental shift took place in 1998, when the State Council decided to abolish the allocation of public housing nationwide and replaced it with monetary subsidies. Employees in the urban sector were encouraged to purchase prior public housing at discounted prices. In the following two years, more than 60 percent of the publicly owned housing stock was sold to individuals. The housing privatization reform enabled many Chinese families to acquire their first property, leading to a surge in urban family-based homeownership rates from 24 percent in 1990 to 72 percent in 2000 (Adams 2009).

From the 2000s onward, China experienced nearly two decades of booming real estate growth. The country’s rapid urbanization saw a large scale migration of people from rural areas and towns to cities in search of residence and employment, creating a substantial demand for housing. Additionally, following the housing privatization reforms in 1998, the commercial housing market became the primary means for urban residents to obtain housing (Yang and Chen 2014). These factors collectively led to a remarkable increase in urban housing prices in China. Nationwide, average housing prices surged by more than four times from 2000 to 2020, surpassing the growth rate of individual incomes. According to statistics from the People’s Bank of China, housing represented 60 percent of Chinese families’ total wealth, while financial assets comprised less than 20 percent in 2019. ² Housing emerged as the primary form of wealth for Chinese households, with homeowners becoming the predominant property owners in the country.

How do homeownership and housing prices influence Chinese politics? Many studies have explored the link between private homeownership and political engagement in urban areas. With the privatization of housing and the disbanding of once predominant work unit organizations, residential communities have taken on a central role in state–society interactions in urban China. Given the high stakes that homeowners have in their property, they are strongly motivated to defend their interests through local political involvement. Li and Wang (2012) discovered that homeowners are more inclined than non-owners to participate in both neighborhood residents’ committee elections and local legislative elections. Other studies also indicate that homeowners participate in demonstrations and activism (Kelly 2006; Zhu and Wang 2007; Read 2008). For instance, in movements such as NIMBY aimed at halting the construction of polluting industries or welfare housing projects, homeowners residing in the affected neighborhoods often emerge as the primary force (Sun 2015; Zhang, Xu, and Ju 2018).

Some observers of China’s grassroots democracy expect that Chinese homeowners might catalyze democratic forces challenging the established political order. However, the majority of homeowners assume a less confrontational position toward the existing order, opting for participation in lawful political processes (Cai 2005, Shi 2014). This differs from the participatory tendencies of farmers and workers, who more frequently employ nonlegal avenues, such as strikes, boycotts, and protests (Cai 2006). A major approach that homeowners leverage to voice their demands is the residents’ committee, which serves as a community-level political organization (Heberer and Göbel 2011). Homeowners also employ other methods, such as establishing self-governing homeowner committees, utilizing party connections to express their concerns, drafting suggestion letters, and attending community meetings (Read 2003; Tomba 2014; Xie and Xie 2019). It is certainly true that homeowners may engage in conflict when their rights are affected, however, this conflict dynamic often lacks a distinct political agenda and their grievances predominantly target real estate developers or property management companies rather than the government (Wang and Li 2005). Furthermore, homeowners often adopt a strategy of rightful resistance, utilizing laws, policies, and established norms to challenge local authorities and assert their property rights. This approach allows homeowners to express their allegiance while voicing concerns, minimizing their political risks (Xiong 2018).

Homeowners’ patterns of engagement also influence the governance mode of the Chinese government. Recent studies have pointed out that political participation in China helps to make the government more responsive to public opinion (Chen, Pan, and Xu 2016, Chen and Xu 2017). In urban areas, homeowners are active groups in civic engagement. Fu, Li and Yan (2023) found that housing prices serve as important clues for local governments to respond to public opinion, often favoring responses to the demands of residents in higher-priced communities. Research by Zeng, and colleagues (2023) aregue that local governments in China show considerable tolerance for the self-governance of homeowners. Since homeowner associations normally have fewer major disagreements with the state, local governments are even willing to encourage such homeowner’s association to effectively address intersocietal conflicts arising from economic modernization.

Research design

Data

To test the hypotheses regarding homeownership and political participation, I utilized data from the 2010 wave of the China General Social Survey (CGSS). The CGSS is one of the longest- running and most professionally managed social surveys in China. The contents of the questionnaire in the CGSS vary widely across waves, and I selected the survey wave in 2010, in which the variable settings most closely suit the needs of this article. ³ Focusing on homeowners and renters in urban China, I employed a sample of urban residents from the CGSS 2010 for analysis.

Scholars have diverse viewpoints on the definition of political participation in the context of China. Shi (1997) offers a widely adopted definition, defining political participation as “activities by private citizens aimed at influencing the actual results of government policy.” I follow Shi’s definition of political participation and use four variables to measure political participation among urban residents in China:

I first utilize willingness to participate in politics as a measure of urban residents’ inclination to engage in politics. This variable is captured through the survey question “To what extent do you agree with the statement: I think I can participate in politics.” It is coded as 5 when respondents answer “Totally agree,” and 1 when they answer “Totally disagree.” The variable exhibits an average score of 2.77, with 31.7% of respondents reporting either “Totally agree” or “Agree.”

To gauge the modes of political participation, I utilize three variables:

1. (1) participation in community organization is a binary variable measured by the survey question “Have you participated in the village committee, neighborhood committee, or residence committee in your community in the past year?” It is coded as 1 when the respondents answer “Yes.”

2. (2) reporting community issues to the government is captured by the question “Have you reported any community-related issues to the relevant government departments in the past year?” This variable is coded as 1 when the respondents answer “Yes.”

3. (3) participation in protests is determined by the question “Have you participated in any protests in your community in the past year?” It is coded as 1 when the respondents answer “Yes.”

Specifically, the first two forms of participation are institutionalized, as they involve expressing political demands through official channels provided by the state. In contrast, participation in protests is considered non-institutionalized, as it typically occurs outside of formal political structures. Within the research sample, the actual engagement in urban political activities is relatively low compared to the expressed willingness to participate. Specifically, 6% of respondents had participated in community affairs over the last year, 4% reported community issues to the government, and only 1% engaged in protest actions. ⁴

In terms of the independent variable, I measure homeownership through a binary variable denoting whether the respondent has full ownership of a private house. In the sample, 47% of respondents own their home property, while an additional 25% do not possess the property themselves but have family members (such as parents, spouses, or children) who own property. The remaining 28% neither own property nor reside in their own homes but instead inhabit rented accommodations. Detailed statistics on these variables are available in Table A1 in the appendix.

Identification strategy

In studies exploring the relationship between wealth and politics, a significant challenge involves mitigating the influence of other socioeconomic factors, such as income, occupation, and education. For instance, obtaining property is closely intertwined with income levels, which makes it difficult to isolate the distinct impact of property ownership. To address this endogeneity issue, I exploit an exogenous distribution of housing ownership during the Chinese housing reform in 1998.

In China’s housing reform, two crucial variations in homeownership emerged. The first distinction lies among urban employees based on the year they began employment. Employees who had begun working before the housing reform were provided the opportunity to obtain public homes and purchase them at lower costs during the reform, facilitating their path to homeownership. Conversely, those who began employment after the reform no longer received housing allocations from the state, had to purchase homes in a costly housing market, which elevated the hurdle to homeownership (Zhiyuan 2024). As the Chinese housing market continued to grow, affordability concerns intensified, making attaining private homeownership increasingly challenging. Grouping the research sample by the first year of employment in Figure A1 in the Appendix, the data around 1998 display consistent or stable trends in age, gender, education, income, and occupation. Notably, employees employed before 1998 maintained a steady homeownership rate of approximately 55 percent. However, a noticeable decline in homeownership rates was observed among those employed after 1998.

Another variation concerns the housing shortage stemming from differences in the construction of public housing. Following China’s housing reform, the housing market was largely supplied by privatized former public housing and commercial properties developed by real estate firms. Generally, the latter tends to be more expensive than the former. In cities with a larger stock of public housing constructed during the planned economy period, the privatization of these properties typically led to lower housing prices in the post-reform era. Using data from China’s 2000 census, which tracks the urban population living in public housing, a trend is evident (see Figure A2 in the Appendix). It shows that in regions where the per capita supply of public housing was higher during the planned economy, housing prices after the reform tended to be lower, as these privatized units entered the market at more affordable prices. As a result, cities with a larger share of public housing historically tended to experience less severe housing affordability issues in the post-reform era, while those with fewer such units faced higher housing prices, making homeownership less accessible for urban residents.

China’s housing reform in the 1990s provides an opportunity to apply a difference-in-differences design to predict homeownership. Following the specification of Duflo (2001), Function (1) is presented as follows:

(1) $$ {Homeownership}_{ijk}={\alpha}_k+{\beta}_j+\gamma {I}_k^{After}\times {Housing\ shortage}_j+\sum {I}_k{\varGamma}_j+{\varPi}_i+{\varepsilon}_{ijk} $$

Specifically, $ {I}_k^{After} $ estimates the differences between cohorts of work using a five-year work cohort bandwidth. Specifically, respondents who began to work between 1993 and 1997 compose the control group. $ {I}_k^{After} $ is coded as 0, indicating that they are under the welfare housing system and not affected by housing reform. Respondents who began to work between 2000 and 2004 are the treatment group. $ {I}_k^{After} $ is coded as 1, suggesting that these work cohorts are affected by the housing reform and are no longer distributed of public housing. ⁵ $ {Housing\ shortage}_j $ is the reciprocal of the number of prior public home sales amount in the $ {city}_j $ and indicates the difficulty of buying homes for the treated cohorts. γ is the co-efficient of interest and is expected to be negative, which suggests that respondents who began employment after the reform and were exposed to a greater housing shortage will report a lower probability of owning a home. $ {\alpha}_k $ and $ {\beta}_j $ denote cohort-of-work and city-of-work fixed effects, respectively, and $ \sum {I}_k{\varGamma}_j $ denotes that cohort-of-work fixed effect interacts with the following city-level covariates: GDP per capita, GDP growth rate, the proportion of industry on GDP and fiscal expenditure per capita, which may affect the amount of public home sales. $ {\varPi}_i $ denotes the individual-level control variables.

The findings in Table 1 reveal consistent and significant negative DID estimators across columns 1 to 4, indicating that the post-reform group facing housing shortages is less likely to own a private home. This signifies a reduced probability of homeownership among individuals who entered the workforce after the housing reform, particularly in the face of increased housing shortages. ⁶ Additionally, the examination in Figure A3 confirms the parallel trends assumption of the DID design. When regressing homeownership among older work cohorts against the current DID estimator within a fully controlled model, the results show the significance of the DID estimator in the context of current changes but indicate no correlation with past homeownership trends. This underscores the exclusion of observable factors that could have confounded variations in work cohorts related to homeownership.

Table 1. Difference-in-difference effect on homeownership

Note: Models 1 to 4 are linear estimations with fixed effects. Robust standard errors are clustered at city level and reported in parentheses.

* p < 0. 1,**p < 0.05, *** p < 0.01 (two-tailed test)

Following Duflo (2001), I further exploit the Function (1) as the first stage to predict the value of $ {\hat{Homeownership}}_{ijk} $ , and use it as the main explanatory variable in the second-stage regression to predict political outcomes. In Function (2), $ {Y}_{ijt} $ is the dependent variable regarding to politics; $ {\alpha}_k $ and $ {\beta}_j $ denote cohort-of-work and city-of-work fixed effects, respectively; and $ \sum {I}_k{\varGamma}_j $ denotes the interaction of cohort-of-work fixed effect with the following city-level covariates, and $ {\varPi}_i $ is the individual-level variables.

$$ {Y}_{ijk}={\alpha}_k+{\beta}_j+\eta {\hat{Homeownership}}_{ijk}+\sum {I}_k{\varGamma}_j+{\varPi}_i+{\varepsilon}_{ijk}\;(2) $$

Figure 1 tests the instrumental variable’s exclusive restriction assumption. It reveals the impact of the DID estimator on homeownership and other socioeconomic factors. When altering the work cohort bandwidths, shifting from 1 year (work cohort bandwidth from 1997 to 2000) to 10 years (work cohort bandwidth from 1989 to 2009), significant patterns are observed in predicting homeownership. However, this DID specification does not predict other essential socioeconomic variables, such as age, gender, education, income, or working organization. These findings indicate that the DID design can specifically explain political behaviors through homeownership, affirming the robustness and reliability of the identification design.

Figure 1. DID Results for homeownership and socioeconomic variables with different time spans.

Note: This figure visualizes the DID estimator with 95 percent confidence intervals.

Baseline results

Homeownership and political participation

Table 2 presents the effect of homeownership on urban Chinese citizens’ willingness for and modes of political participation. I begin with a difference-in-differences model in Panel A, which controls for various individual-level and city-level variables, including age, gender, education, income, employment status, work organization, CCP membership and city economic factors, as well as work cohort and city fixed effects.

Table 2. Homeownership and political participation

Note: Panel A shows linear estimations with fixed effects. Panel B shows two stage least square estimations. Robust standard errors are clustered at the city level and reported in parentheses.

* p < 0. 1,**p < 0.05, *** p < 0.01 (two-tailed test)

Column 1 uses willingness to participate in politics as the dependent variable and the coefficient of DID estimator is negative and significant at 99 percent confidence intervals. It suggests that respondents who started working after the housing reform and faced greater housing shortages exhibit reduced willingness to participate in politics. Columns 2–4 use respondents’ actual participation in community organization, government reporting, and protest as dependent variables. Specifically, participation in community organizations and reporting community issues to the government signifies institutionalized modes, while protest involvement represents an un-institutionalized form of participation.

The results denote that respondents who are less likely to own a home have lower involvement in community affairs and government reporting, but instead, display increased engagement in protests. This suggests lower enthusiasm for public affairs and a higher likelihood of adopting confrontational approaches in cases of involvement in politics among the non-homeowners. This suggests that China’s housing reform, by decreasing homeownership rates among the younger generation, impacts both individuals’ willingness to participate politically and their adoption of institutionalized forms of participation.

Moving to Panel B, I further use the instrumental variable approach with the same set of controls as Panel A. The results of panel B display significant and consistent effects of homeownership on political participation. Column 1 shows that homeowners exhibit greater propensity for participation than renters. Columns 2 and 3 further emphasize homeowners’ heightened participation in formal channels, including community organization and communication with local governments. Finally, Column 4 highlights that owning a home decreases the likelihood of engaging in mass protests against the government. These results corroborate with the hypothesis that homeowners are more active but self-regulated participants in the political landscape.

Robustness

Table 3 extends the analysis to ensure the robustness of the initial findings. In Panel A, I test the sensitivity of baseline results to different bandwidth settings. Narrowing the work cohort bandwidth from five years to three years before and after the housing reform, the results show that the IV estimation remains consistent with the primary findings across these different bandwidth settings.

Table 3. Robustness

Note: Panel A, C, D and E shows two stage least square estimations. Panel B shows logistic model estimations. Robust standard errors are clustered at the city level and reported in parentheses.

* p < 0. 1, ** p < 0.05, *** p < 0.01 (two-tailed test)

Panel B examines whether baseline findings are affected by different model specifications. With the employment of ordered logistic and logistic models for the ordinal and dummy dependent variables, the DID estimation continues to mirror the baseline findings from Table 2. This reaffirms the robustness of the participatory effect of homeownership, irrespective of the model settings.

Additionally, Panel C explores the heterogeneous effects of homeownership on state-owned enterprise (SOE) workers’ political participation in urban China. As a major part of China’s transition to the market economy, SOEs underwent reconstruction and reduced their employee welfare commitments, such as job security and public housing provision. As anticipated, the hypothesis aligns with the consistent findings from the baseline results, suggesting that SOE workers who own homes are more active in political engagement.

Panel D focuses on placebo tests conducted on a rural resident sample. During the 1998 housing reform, urban employees were allowed to purchase public housing at lower prices, while rural residents did not have the same opportunity. Therefore, if the causal identification strategy holds, there should be no significant effects in the sample of rural residents. The results are as anticipated, providing additional support for the causal identification strategy.

Finally, Panel E address concerns about an unspecified shock during the housing reform. In 1998, the state-owned company reform occurred. Many workers lost their jobs during the reform, and there was a systematic difference between those who began to work before and after 1998. I attempt to address this concern by excluding the sample from the Northeast provinces (i.e., Heilongjiang, Jilin, Liaoning) in Panel E, which suffered most from the reform and faced the most severe unemployment problem. The results are consistent with the baseline findings, suggesting that the results are not driven by other major events.

Home value and political participation

Apart from homeownership, how does the fluctuation in housing prices influence political behaviors? In recent years, the boom in the global housing market and its widespread socio-economic impact have prompted scholars to pay increasing attention to the effects of changes in housing values on political participation, especially voting behaviors. For example, Adler and Ansell (2020) found that rising housing prices make homeowners less likely to vote for populist parties. However, although existing studies have explored how fluctuations in housing prices influence the overall level of political participation, fewer studies have delved into the specific ways in which changes in housing prices affect how residents engage in politics.

I investigate this relationship in Table 4. Specifically, I estimate the respondent’s housing value by calculating the product of the housing size and the city-level urban housing price. Housing size is measured from CGSS survey data. The city-level average housing price is taken from the statistical yearbooks of each city in China and then matched with the survey data. Similar to the approach for predicting homeownership, I initially employed a difference-in-differences design to assess the influence of home value. Table A3 in the Appendix reveals an overall negative correlation between the DID estimator and home value. This suggests that the housing reform in the late 1990s not only reduced the property ownership rate of a younger generation but also widened the disparity in property ownership.

Table 4. Home value and political participation

Note: Robust standard errors are clustered at the city level and reported in parentheses.

* p < 0. 1, ** p < 0.05, *** p < 0.01 (two-tailed test)

Table 4 presents the results of a second-stage regression following the instrumental variable approach. In column 1, it is evident that individuals with higher housing values are more inclined to participate in public affairs. Further examining the ways of participation, columns 2 to 4 indicate that an increase in home value motivates engagement in community public affairs and interactions with local governments. However, it decreases participation in protests. These findings provide support for Hypothesis 3, suggesting that rising home values intensify homeowners’ linked interests and exit cost and thereby strengthen their incentives for political participation. This implies that the participatory effect of housing stems from homeowners’ self-interest in sustaining wealth growth.

I also conduct robust tests on the effect of housing prices. Following the approaches of Ashraf et al. (2020), Table A4 examine the effects of housing prices based on the difference-in-differences-in-differences (DDD) design. The triple interaction term “Work after reform $ \times $ Housing shortage $ \times $ $ \varDelta $ Housing prices” is negative in columns 1 to 3, and positive in column 4. This suggests that for the respondents who are in disadvantaged conditions of obtaining homeownership, the housing affordability is aggravated, which decreases their interest in participating in public affairs. This evidence further strengthens confidence in the main argument of this article.

Mechanism

I performed several additional analyses to verify the proposed mechanisms.

1. Linked Interest with the Existing Order. Homeownership generates linked interests with the current political and economic systems. To evaluate this, I used the variable Political Stability is Important . The respondents ranked four political issues in order of priority: political stability, civil rights, free speech, and flat prices. I coded their choices by assigning a value of 2 if “political stability” was the highest priority and a value of 1 if it was the second highest priority, while the other choices were attributed the value 0. Figure 2 illustrates the IV estimator on the respondents’ preference for stability. The positive and significant coefficient indicates that homeowners exhibit a stronger preference for political stability than for other political values.

Figure 2. IV results on mechanisms.

Note: This figure visualizes the instrumental variable estimator of homeownership with 95% confidence intervals.

I also employed the variable Current Social Distribution is Fair , which measures respondents’ perceptions of the fairness of social distribution. This five-point scale variable ranges from “very fair” (coded as 5) to “not fair at all” (coded as 1). Figure 2 portrays the results, showcasing a positive and significant effects on feeling that the current distribution system is fair among homeowners. In conclusion, the evidence suggests that homeowners tend to prioritize maintaining the existing order and exhibit a greater aversion to disorder.

2. Exit and voice. Another mechanism is the increased cost of exit among homeowners. According to Hirschman (1970), individuals tend to either voice their concerns or exit when faced with adverse changes. A higher cost of exit typically leads individuals to voice their grievances in an attempt to change policies. To evaluate this mechanism, I utilized the variable Doing Nothing When Government Harms Interests to directly measure the willingness to exit. Specifically, the survey question is “To what extent do you agree that, if the government harms my interests, I have to endure.” It is a five-point ordinal variable, with a code of 1 indicating “totally disagree” and 5 indicating “totally agree.” Figure 2 displays a negative and significant coefficient for homeownership, indicating that homeowners are more inclined to choose voice rather than exit when encountering conflicts with the government.

The second dependent variable addressing exit and voice is derived from the question “To what extent do you agree that you cannot affect politics.” Similarly, I Cannot Affect Politics is a five- point scale variable ranging from “totally disagree” to “totally agree,” coded as 1 to 5. Figure 2 presents evidence suggesting that homeowners perceive themselves as having a greater capacity to participate in public affairs. This further supports the notion that homeowners are more willing and capable of “voicing” their concerns rather than “exiting” from political situations.

Are homeowners supporters or critical citizens in China’s Political System?

Finally, I explore the impact of homeownership on citizens’ political beliefs. Initially, the focus is on political support, gauged through three dependent variables: political trust in the police , political trust in the army , and political trust in the central government . These are measured as five-point ordinal variables, where “totally trust” is coded as 5 and “totally not trust” is coded as 1.

Figure 3 illustrates the results of the IV estimation of the effect of homeownership on political support, employing the 2SLS specification. The findings reveal that homeowners exhibit higher levels of trust in various political institutions, including the central government, police, and the army—core institutions within a state. This suggests that compared with renters, homeowners are more inclined to support the rule of the CCP compared to renters.

Figure 3. IV results on homeownership and political attitudes.

Note: This figure visualizes the instrumental variable estimator of homeownership with 95% confidence intervals.

However, the study also reveals a critical aspect of Chinese homeowners. I measure citizens’ attitudes toward government performance through three variables: government must consider citizens’ views in policy-making , taxpayers have the right to discuss public spending , and citizens should be allowed to criticize the government . These variables employ ordinal scales, ranging from 1 to 5, representing “totally disagree” to “totally agree.” In Figure 3, the 2SLS results presented demonstrate a positive and significant effect of homeownership on these political attitudes.

The findings show a complex pattern of homeowners’ political attitudes. Homeowners are among the supporters of the one-party state and report greater trust in various political institutions. However, homeowners also exhibit a heightened awareness of political rights. They express more skepticism toward government performance and showcase a greater inclination to hold the authorities accountable. These dual inclinations among homeowners define their role as the “loyal opposition” in political participation. On the one hand, they display confidence in the government, aiming to preserve political order and social stability. On the other hand, homeowners are also keen on safeguarding their private property from potential misuse by public powers.

Concluding Remarks

The key finding of this study concerns how homeownership shapes both the inclination and modes of political engagement. By leveraging variations in housing ownership distribution among different work cohorts, I observe that homeowners show a stronger willingness to participate in politics, especially via institutionalized activities, such as community organizations and interactions with local governments. In contrast, homeowners exhibit a lower tendency to engage in direct confrontations, such as protests. These conclusions remain robust to various tests, including alternative model specifications and sample variations. When considering housing price appreciation, I find that homeownership has a stronger impact on political involvement when home values increase. Homeowners present a nuanced stance: while being critical of government performance, they also display greater political trust and a preference for maintaining the status quo.

This study provides new evidence for understanding the complex political characteristics of homeowners. On one hand, homeowners are motivated to protect their wealth and guard against the abuse of public power, leading to active participation, higher political consciousness, and criticism of government misconduct. On the other hand, the alignment of homeowners’ interests with the existing order curbs more destructive forms of engagement, making homeowners more likely to play the role of a “loyal opposition.” Although the empirical evidence in this paper is primarily drawn from China, I believe the findings contribute to understanding the increase in violent struggles worldwide. Both developed industrialized countries, and emerging economies have experienced continuous housing price increases over the past decade. The worsening housing affordability may lead to more forms of participation beyond institutional boundaries, which still requires further exploration in research.

The findings of this study also offer a new explanation for understanding how China maintains long-term political stability. Citizen expression and engagement provide the necessary information for the government to adjust policies and monitor subordinate officials (Distelhorst and Hou 2017, Jiang, Meng et al. 2019). However, the Chinese government has always been concerned about political participation spinning out of control, which could threaten the stability of its rule (Weiss 2013). Compared to existing research that primarily focuses on the exogenous institutional and strategic approaches governments use to maintain political stability, this paper discusses more about the endogenous motivations of citizens in spontaneously upholding political stability. The findings of this study imply that, in the context of China, the relationship between homeowners and the government is not simply one of confrontation or support. Rather, it entails both overall support and criticism of specific policies, as well as active participation and self-regulation of involvement. In other words, private property rights, represented by homeownership, do not necessarily position urban residents against the government. Instead, they shape their characteristics as a “loyal opposition.” Moreover, given that homeownership represents a considerable proportion of urban residents, this suggests that the role of homeownership in shaping political stability in China is quite significant.

However, the Chinese version of the “loyal opposition” currently faces challenges from two fronts: First, the housing-to-income ratio in major Chinese cities is among the highest globally, meaning that many young people entering these cities for work are experiencing decreased housing affordability. This could lead to a lower interest in political participation among the younger generation and a reduction in institutionalized civic involvement, indicating potential apathy and a trend toward confrontation. Additionally, a more significant challenge arises from the recent downturn in the Chinese real estate market. With widespread declines in housing prices, urban residents are facing shrinking family assets. Consequently, the homeowners who played a crucial role in supporting China’s political stability during the real estate boom, may become a source of political risk during the downturn. This deserves further empirical research and discussion in the future.