2.1. What are the effects of coercive mobilization?

States seek to control the lives of citizens to ensure they pay taxes and cooperate with the regime’s policy objectives. There is enormous variation in the ability of states to achieve this (Lee, Reference Lee2019), partly due to the presence of non-state actors who compete with the state for political and social authority and seek to undermine state capacity when the state’s goals do not align with theirs (Suryanarayan and White, Reference Suryanarayan and White2021). While critical junctures such as wars and conquests are thought to offer opportunities to disrupt local networks of authority and increase state capacity (Dincecco and Prado, Reference Dincecco and Prado2012), states may prefer to create such junctures themselves by breaking entrenched local networks through intensive coercion. Often, such campaigns involve the mass resettlement of citizens, which tends to disrupt local networks and place the relocated citizens under more intensive regime control (Silwal, Reference Silwal2015; Lupu and Peisakhin, Reference Lupu and Peisakhin2017; Rozenas et al., Reference Rozenas, Schutte and Zhukov2017).Footnote ²

What is the effect of these programs on the resettled? Lupu and Peisakhin (Reference Lupu and Peisakhin2017) find that the violence and social dislocation inseparable from mass resettlement have led the resettled (in this case, Crimean Tatars) and their descendants to be more hostile toward the government that dispatched them. This finding builds on the large body of literature on the effects of repression and violence, which are widely thought to have profound effects on the victims’ psychology, overall social structure, and patterns of political participation. On this last point, the debate has generally been between those who emphasize the demobilizing effects of repression (Zhukov and Talibova, Reference Zhukov and Talibova2018; Komisarchik et al., Reference Komisarchik, Sen and Velez2022) and those emphasizing its tendency to encourage grievances against the regime (Balcells, Reference Balcells2012; Wang, Reference Wang2021). Both mechanisms are fairly intuitive: those who suffer at the regime’s hands will harbor anger against it and be wary of challenging it in the future—in fact, they may be wary of the efficacy of a wide range of activities after such a profound experience of personal powerlessness (Rozenas and Zhukov, Reference Rozenas and Zhukov2019).

2.2. Coercive mobilization and social control

However, coercive mobilization is usually a bundled treatment. While individuals are being coerced, they are also usually being exposed to more intensive monitoring and indoctrination. Such mobilization can also change the cognitive framework of individuals—how they view themselves and the world. To the extent that the regime has control of these changes, it can create citizens sympathetic to its goals. In many cases, such mass ideological refashioning was a major goal of the coercive state mobilization in the first place, and older accounts of “totalitarian” regimes claimed that this mechanism tended to dominate (Arendt, Reference Arendt1951).

Coercive state mobilization might lead to higher levels of support for the regime by influencing whom citizens trust. Typically, repression is thought to reduce trust in the regime (Desposato et al., Reference Desposato, Wang and Wu2021). Indeed, in authoritarian contexts as varied as the post-Stalinist USSR (Hosking, Reference Hosking2013) and the medieval Middle East (Greif, Reference Greif1989), citizens have relied on interpersonal networks of friends and family with high levels of trust, while adopting a suspicious attitude toward outsiders. In the language of Tilly (Reference Tilly2005, 9), while “networks of trust” exist, they are not “integrated into public politics.” Families, religious groups, and friend networks can all serve as alternative focuses of loyalty to the state in this way.

Banfield (Reference Banfield1958) conjectures that exclusive trust in family is an obstacle to development because it leads to distrust in outside actors, including the state. When individuals are suspicious of public institutions, they are capable only of caring about their private interests. A negative correlation between family ties and participation in political institutions indeed exists across countries (Alesina and Giuliano, Reference Alesina and Giuliano2011).

However, if the state could isolate citizens from private networks or discredit private networks, citizens would become incapable of trusting their friends and families as fully as they otherwise would, and might turn to the state as an alternative, even if an imperfect one. The state, in this conception, gains trust not by becoming more trustworthy, but by eliminating all alternative objects of trust and relying on the human need to rely on some other individuals for support and information. A regime capable of destroying all alternative focuses for social loyalty would have achieved Arendt’s (Reference Arendt1951) “total domination.”

As Arendt’s critics pointed out (Benhabib, Reference Benhabib2003, p. 73), the overwhelming majority of dictatorships, including the ones she studied, were incapable of fully eliminating the private sphere, and were forced to share their citizens’ loyalty with private networks. To make fully effective “total domination,” a state would have to separate individuals from their families and friends, prohibit them from forming intimate relationships, provide them food and clothing, completely control their work and leisure, and keep them from all sources of information not produced by the state.

In the same way that they prioritize the regime over the private sphere, high-capacity regimes tend to strengthen loyalty to the central government over other levels of government. Many regimes, including democratic ones, seek to encourage citizens to identify with the nation rather than with subnational identities through policies such as universal education and military conscription (Weber, Reference Weber1976). Resettlement, by relocating citizens to unfamiliar regions, is often designed to foster loyalty to the nation rather than to the locality.

The relationship between attitudes toward the regime and political participation is complex. When extreme costs are imposed on those who overtly oppose the regime, only those with very extreme anti-regime beliefs will be tempted to do so (Kuran, Reference Kuran1991). Conversely, the regime may use both rewards and punishments to encourage participation in its own institutions and activities, meaning that even regime opponents may participate in some of them. While attitudes and participation are thus correlated in authoritarian regimes, the relationship is much less pronounced than in democracies. We therefore expect the relationship between resettlement and participation to be more complex and context-dependent than the relationship between mobilization and attitudes.

3.1. The sent-down movement

In 1966, Mao Zedong and his supporters, the “Gang of Four,” launched the Cultural Revolution, which was designed to mobilize the “revolutionary masses” (students, workers, and peasants) against a bureaucratic establishment perceived as insufficiently radical and overly independent of Mao. Student supporters of the Cultural Revolution were organized into Red Guards (Walder, Reference Walder2009), playing a key role in the period’s violence and instability. All schools were shut down between 1966 and 1968, while college entrance exams were canceled from 1966 to 1977. However, urban high school and university students were also seen as a privileged group compared to workers and peasants and potentially in need of revolutionary education to counteract incipient elitism.

The term “Sent-down Movement” is short for the “Up to the Mountains and Down to the Villages Movement” and a policy of forcibly relocating young, educated people to work in the countryside between 1968 to 1978. Before 1968, the program was limited and voluntary, descending from earlier Soviet and Chinese rural resettlement programs (Bernstein, Reference Bernstein1977; Zhou and Hou, Reference Zhou and Hou1999). However, the Cultural Revolution led to a vast expansion of the program’s scope and profile, with Mao proclaiming that “it is necessary for the educated youths to go to the countryside, and be re-educated by the poor peasants. We need to persuade cadres and others in urban areas to send their children who graduated from junior high, senior high, and college to rural areas.” From 1968 to 1978, more than 16 million urban youths, who were called zhiqing (educated youths), went to the countryside through the program. In theory, all junior high school graduates were eligible for the movement, but only one third of the youths were actually sent to the countryside, with the others serving in the military or an urban work unit.

In its expansive period, the Sent-down Movement used both persuasion and coercion to recruit zhiqing, the youths to be sent-down to villages. Due to the political ferment of the Cultural Revolution era, many young people were enthusiastic about the program’s goals, while others were anxious to demonstrate their political loyalty to Mao and the Communist Party. However, when there were insufficient volunteers, local governments conscripted eligible youths. Both in the initial conscription and in the subsequent administration of exemptions, alternatives and punishments, those with “bad” (anti-communist) family backgrounds were more likely to be sent down (Gee, Reference Gee2011; Rene, Reference Rene2013), as well those without a sibling already in the countryside (Zhang et al., Reference Zhang, Liu and Yung2007). Relative to other political movements of the period, the Sent-down Movement thus influenced a large subset of urban families with a wide variety of views on the regime (Li et al., Reference Hongbin, Rosenzweig and Zhang2010).

The eligibility procedures for being forcibly sent down were complex and inconsistently enforced, yet one consistent criterion remained: students were not sent down until they had graduated from middle school, the point at which formal education typically ended during most of the Cultural Revolution. While no nationwide document or guideline explicitly defined the eligibility criteria for the Sent-down movement, local party committees issued their own guidelines, following standards communicated by the central committee through meetings and newspapers. For example, official documents from the local governments of Yueyang, Hunan (Yueyang Municipal Government, 2015) and Yulin, Shaanxi (Yulin Municipal Party History Research Office, 2022) both specify that eligibility was junior high school and high school graduates at the time.

Mao’s exhortation covered only these students, as those without this level of education were not considered “educated youths.” Bernstein (Reference Bernstein1977) wrote that “it is in the school that each graduating middle school student is assigned to a production unit, either industrial or agricultural, or to the People’s Liberation Army (PLA).” Middle school usually occurred around the age of 15, but the disruptions of the Cultural Revolution period meant there was some variation in the age of graduates. While in the countryside, the sent-down youths were generally unable to attend high school or university, except for a tiny number recommended as “worker-peasant-soldier college students.”

The Sent-down Movement declined gradually after Mao’s death in September 1976 and then ended abruptly following the political disgrace of the Gang of Four in October 1978. In 1978, the National Sent-down Movement Conference officially decided to end the program, sending existing zhiqing back to urban areas and arranging jobs for them. However, it took several years for all the zhiqing to be able to return home. Deng Xiaoping, a leading figure in the policy reversal, remarked, “The nation spent thirty trillion yuan to buy three unsatisfactory things: zhiqing are unsatisfied, parents are unsatisfied, and peasants are unsatisfied.” The Communist Party was thus both the force that sent students to the countryside and the force that brought them back. This ambivalence is evident in the disturbances in Yunnan in 1978, where zhiqing eager to return home rallied against the policies of the local government while proclaiming their support for Deng (Zhou, Reference Zhou2010), a phenomenon termed “rightful resistance” (O’Brien and Li, Reference O’Brien and Lianjiang2006).Footnote ³

3.2. Experiences in the countryside

During the Sent-down Movement, most of the youths being sent down were rusticated within their home provinces, while many students from the biggest cities, such as Beijing, Shanghai, and Hangzhou, were sent to border provinces like Inner Mongolia, Xinjiang, Yunnan, and Heilongjiang. Most of the zhiqing, coming from relatively sheltered urban backgrounds, faced difficulties adjusting to rural life and dealing with the local officials who controlled their work, leisure time, and the distribution of food and clothing. Many were shocked by the shortage of food and poor living conditions in the countryside, where vegetables and meats were scarce and heavy manual labor was required. One zhiqing interviewed in Heiming (2006) recalled that:

Another zhiqing interviewed by (Rene, Reference Rene2013, 139) remarked that:

Besides tough living conditions in the countryside, the sent-down students were under intense re-education. As one zhiqing recalled:

Their collective learning strengthened their beliefs in the blueprint that painted by the central leadership. Moreover, the sent-down students had vicarious contact with the national government at a distance. When Mao personally responded to a sent-down student’s letter and sent him money. Zhiqing were touched that the government understood them and tried to improve their conditions (Pan, Reference Pan2009, 140). Their convictions in the well-intentioned national policy in part explain why their later protests were respectful to the central leadership.

Zhiqing clashed with local officials, and were often contemptuous of the cadres who assigned them work, considering them corrupt and unsophisticated (Bernstein, Reference Bernstein1977). They observed malfeasance at the local level, including embezzlement of the resettlement fee and sexual violence against female zhiqing. When these problems broke out, it was the national government that performed top-down accountability by sending delegations and drafting orders to address the problems (Pan, Reference Pan2009, 140-145).

On the other hand, the interactions of zhiqing with local peasants in the countryside shaped their political beliefs. Zhiqing became more knowledgeable about the rural part of the country and sympathetic to the peasants’ perspectives. This ideological transformation demonstrated the success of the reeducation goal of the Sent-down Movement, which was conducive to building trustworthiness and righteousness in the decision-making by the central leadership.

The zhiqing were cut off from their families and those friends who remained in the city or were displaced to different areas, whom they could visit only with difficulty. This isolation meant that their networks were smaller than those of their counterparts who stayed in the city, and they tended to marry much later (Wang and Zhou, Reference Wang and Zhou2017). In the villages, their means of communication were limited mostly to cadres, peasants, and fellow zhiqing.

3.3. Possible long-term effects of the sent-down experience

The Sent-down Movement had several effects on zhiqing that correspond to those discussed in Section Two. Firstly, the sent-down tend to trust their families relatively less. Zhiqing spent virtually no time with their parents and friends in cities during a crucial formative period of their lives. Many of them had no chance to spend holidays at home with their families until they were allowed to return to the cities after their service in the countryside. The most common way for them to express their homesickness to their family members was through letters.Footnote ⁴ At the same time, they spent their days in close proximity to previously unfamiliar zhiqing and local peasants, with whom they bonded through repeated collaboration on production work and over shared struggles against boredom and material scarcity. Research from developmental psychology shows that social trust depends on the environment. When adolescents are physically away from their parents, their parental trust goes down. In addition to direct physical separation, it is generally more stressful for adolescents to make important decisions when they could not consult with their parents (Smetana, Reference Smetana2010). Under such circumstances, the sent-down students were forced to turn to alternative authorities, such as their peers and the government, for guidance.

We should thus expect the zhiqing to have lower levels of trust in family relative to strangers than others. This was, in fact, the goal of much of the political mobilization to which the sent-down were exposed, emphasizing the virtues of the regime and the collective over the family and individual. This double shift created a broader focus of trust among the sent-down. Qualitative evidence indeed suggests that one of the political purposes of the Sent-down Movement was to replace family ties by ties between the individual and the state, compatible with Marxism (Rene, Reference Rene2013, 91).

The issue of the effect of being sent down on trust in the Communist Party is more complex. On the one hand, the Sent-down Movement tended to pit the zhiqing against local officials. Local officials oversaw the day-to-day implementation of this very unpopular policy: forcing individuals to leave the cities, feeding and clothing them in rural areas, and disciplining them if they protested. Given the flawed manner in which the policy was often executed, this made local officials the targets of zhiqing’s anger. Local officials were also, by definition, local, and outsiders relative to the urban students in a way that national and provincial officials were not. The national government encouraged this trend by condemning local abuses and stressing the need to hold “local emperors” accountable while praising zhiqing for their bravery, tenacity, and contributions.

Because the national government operated at a greater distance from the pubic, citizens relied on portrayals from selected media rather than direct experience to form their perceptions. Ordinary citizens were not possible to observe decision-making processes, policy mistakes, and personal imperfections at the national level. In contrast, citizens could assess policy implementation through their direct experiences at the local level. This proximity allowed citizens to observe political tensions and personal flaws of local officials and bureaucrats (Li, Reference Li2016; Wu and Wilkes, Reference Cary and Wilkes2018).

At the same time, despite being in the countryside as a result of a national policy, zhiqing were often reluctant to condemn the central government—in fact, even zhiqing’s demonstrations demanding to be sent home were extremely respectful toward the central government (Pan, Reference Pan2009). Students were naturally unwilling to acknowledge that their labor and conditions were meaningless, and thus would often adhere to the indoctrination of the period, which highlighted the party’s wisdom. If the party was the source of the zhiqing’s problems, it was also seen as the source of all solutions to those problems, and the entity upon which they depended for their material survival.

Moreover, the very act of traveling to a distant province might tend to strengthen the “national” identification of students, since they were manifestly not local to the areas they were sent to and yet were very far from their areas of origin. Finally, the national government at times intervened on behalf of the zhiqing, providing reimbursements for resettlement expenses, accepting petitions against local officials, and ultimately permitting all the zhiqing to return home after the program’s conclusion.

We thus expect the sent-down to have divergent attitudes towards the local and national governments; the sent-down experience should lead to higher levels of regime approval, but lower levels of approval for local officials. China’s political system, including its power structure and social control, facilitates the persistence of distrust in local governments. The system fosters identification at the national level, shifts blame to lower level governments, and exposes the malfeasance of local officials. From the public’s perspective, the national government represents political commitment, whereas the local governments is seen as responsible for policy implementation. Such divergence is still common in China, where the national government is much more popular than local government (Chen, Reference Chen2004; Li, Reference Li2016; Wu and Wilkes, Reference Cary and Wilkes2018).

The relationship between being sent down and political participation is more ambiguous. Intuitively, we might expect citizens who approve of the regime to be more likely to engage in political activities favored by the regime (the Communist Party, official local elections, etc.) and less likely to become involved in activities disfavored by the regime, such as protests and unofficial community groups. However, this relationship is complicated by official policy. The party rewards those who participate in officially favored activities with material benefits and sometimes sanctions those who engage in unofficial activities. As a result, many people with only mild support for the regime are party members, and even individuals with strong anti-regime views may refrain from protesting. Therefore, we should thus expect the relationship between being sent-down and pro-regime political participation, while positive, to be weaker than that for political attitudes.

Supplementary Material, Appendix Section J provides additional discussion of the historiography of the Sent Down movement and more recent scholarship

4.1. The regression discontinuity design

Since selection into being sent down was non-random, a naive estimate of the influence of being sent down on subsequent political participation and attitudes is likely to be biased. In particular, since qualitative evidence suggests that those politically active and hostile to the regime were more likely to be sent down than others, we might expect estimates of the effect of the experience on participation and regime hostility to be biased upwards.Footnote ⁵

An alternative approach would be to consider eligibility for the sent-down program, rather than participation itself, by comparing those schooling cohorts whose members were eligible to be sent down with those who were too young or too old. However, this method would not only yield an attenuated estimate of the effect of being sent down (since many non-zhiqing would be included in the “treatment” group), but it would also be biased by various other policy changes during the Cultural Revolution that differentially affected specific age cohorts in this politically tumultuous period. The most significant of these was the shutdown of the high school and university system during the Cultural Revolution, which began in 1966 and lasted until 1976. Therefore, the sent-down cohorts were, in general, less educated than the age cohorts before and after them, even among those who were not resettled. However, the last two age cohorts eligible for the program, those who graduated from junior high school in 1975 and 1976, could (if they stayed in the cities) complete their high school education and enter university through the newly restarted university examination system, just as subsequent age cohorts did.

To address the selection problem, we use a fuzzy regression discontinuity (RD) design based on age cohort. Since only middle school graduates were eligible to be sent down, middle school graduation year, our running variable, is strongly associated with whether students were eligible. We take advantage of the sudden end of the movement determined by the Communist Party in October, 1978, which created a discontinuous drop (to nearly zero) in probability of being sent down. We are thus comparing individuals who differ in age by only a few years—those who were “barely” eligible to be sent down and those who were barely ineligible. Both the treatment and control groups would have the same memories of the Cultural Revolution and experience the same subsequent life courses, with only minor differences in the age at which they experienced these events. Recall that the gradual winding down of the policy meant that even the last sent-down students spent lengthy periods in the countryside: Supplementary Material Figure A.1 shows that these individuals, on average, spent 3.9 years in the village.

Note that since the fuzzy RD design focuses on “intention to treat” determined by age, the control group to which we compare the sent-down are those who were barely ineligible to be sent down, not those of the same age who managed to avoid being sent-down. To the extent that the barely ineligible were exposed to a significant amount of ideological indoctrination and regimentation in their urban childhoods during the Cultural Revolution, our estimate of the effects of the sent-down policy is likely much smaller than the overall effect of state mobilization during the Maoist period.

Figure 1 shows the proportion of urban youth within a given graduating year who were sent-down, taken from the 2010 China Family Panel Survey (CFPS). The percentage of junior high school graduates being sent-down dropped significantly after the National Sent-down Movement Conference in 1979, from mostly above 20% to nearly 0. Individuals who graduated from junior high school in 1978 were thus the last urban youths to be sent-down in any numbers. Seven individuals reported being sent-down after 1978, almost certainly incorrectly. We include these misreported observations because the deletion of a non-random subset of observations, conditional on the endogenous regressor, might lead to bias. However, excluding these seven observations has virtually no effect on the reported results and, in fact, improves the precision of the second stage estimates.

Note: This figure illustrates the likelihood of being sent down based on individuals’ graduation year. It shows that junior high school graduates were no longer sent down after 1978. Note that there are a few data points showing individuals being sent down after the 1980s, which are likely errors in data collection. To preserve the integrity of our research, we keep these points in our figures and regressions.

Figure 1. Probability of being sent-down by graduation-year.

The fuzzy regression discontinuity design uses a two-stage least squares (2SLS) design (Hahn et al., Reference Hahn, Todd and Van der Klaauw2001), with the first stage being the influence of the cutoff on treatment (in our case, the effect of age cohort on being sent down) and the second stage regresses the outcomes on the predicted treatment values from the first stage. The model can be written as:

\begin{equation*} Pr(SentDown_i = \textbf{1})= \begin{cases} p_0 (c_i) & c_i \ge c_0 \\ p_1(c_i) & c_i \lt c_0 \end{cases}, \end{equation*}

where $c_0$ is the cut-off of graduating year, which is 1978. $c_i$ is the running variable, which is graduating year. Since the movement ended following the National Sent-down Conference in 1978, it must be $p_0 (c_i) \gt p_1 (c_i)$, which represents the sudden drop in the probability of an individual being sent down.

The first stage regression, which uses a triangular kernel function, is:

\begin{equation*} SentDown_i= \alpha_1 + \beta_1 Eligibility_i + \gamma _1 (c_i - c_0) + \theta_1 Eligibility_i (c_i-c_0) + u_i, \end{equation*}

where $Eligibility_i = \textbf{1} (c_i \lt c_0) $ represents whether the individual was eligible to be sent-down. The reduced form RD regression is:

\begin{equation*} Y_i= \alpha_2 + \beta_2 Eligibility_i + \gamma _2 (c_i - c_0) + \theta_2 Eligibility_i (c_i-c_0) + \epsilon _i. \end{equation*}

The estimated coefficient is $\beta_{RD} = \beta_1/\beta_2$. Standard errors are clustered by junior high school graduating year. In most models, we used optimal bandwidths calculated using the procedure described in Calonico et al. (Reference Calonico, Cattaneo and Titiunik2014). We report results without covariates adjustment in the main paper, and the regression results with covariates adjustment are reported in Supplementary Material Appendix. Given the young age at which people were sent-down, the number of plausible pretreatment covariates is small, but we include gender, membership in a minority ethnic group, and self-reported family “class background.”

Note are estimating the local treatment effect of being sent down by comparing those sent down immediately before the program’s end in the 1970s to those who came of age immediately after them. This local treatment within the bandwidths might be different from the effect of being sent down on all zhiqing. However, we think it likely that the LTE in this case is biased towards zero. In the last years of the program, the original ideological enthusiasm had become muted and those who were sent down spent less time in the country, while the barely untreated group had received extensive indoctrination during the cultural revolution. By contrast comparing early zhiqing to those who came of age after China’s economic reforms (using the whole sample) would represent a much larger difference in exposure to forcible resettlement and reeducation (though also, of course, to unobserved confounders).

4.2. Data and variables

We use two datasets in our research: The China Family Panel Study (CFPS) from 2010-2016 (Institute of Social Science Survey, Peking University, 2015) and the 2008 China Survey.Footnote ⁶ The CFPS, a nationally representative survey initiated in 2010 by the Institute of Social Science Survey (ISSS) of Peking University, collects individual, family, and community-level longitudinal data in contemporary China. Our attitudinal measures primarily consist of thermometer scores evaluating feelings toward particular institutions (on a scale from 1 to 5) and the severity of specific social problems (on a scale from 0 to 10). Participation measures are mostly binary, indicating whether the respondent has recently participated in a certain type of activity. Since our independent variable of interest (being sent down) does not change over time, we estimate only cross-sectional models. Given that both attitude and participation variables are measured across all four survey waves, we use their four-year average in our models to minimize the impact of year-specific variations in the attitudinal variables. A detailed description of the variables is available in Section G of the Supplementary Material Appendix.

The disadvantage of the CFPS is that it features a limited number of questions on political issues and lacks measures assessing opinions about the national government or non-state political participation. For this reason, we augment our analysis with the 2008 China Survey, conducted by the Research Center for Contemporary China at Peking University. The scope of binary participation measuresFootnote ⁷ and thermometer-based attitudinal measures is significantly wider in the China Survey than in the CFPS. Since the China Survey does not measure whether an individual was ever sent-down, or when they graduated from middle school. We are forced to estimate a reduced form, single stage model where being a member of an age cohort that should have spent the normal amount of time in school would have been eligible to be sent-down, rather than being sent-down itself. We believe that this, combined with a much smaller sample, will result in an attenuation of our estimates.

Since the sent-down program applied only to urban youths, we restrict our sample to those who lived in urban areas as children and graduated from middle school. Our sample includes all urban graduates born between 1949 and 1972, though most of the RD models are estimated within much narrower bandwidths. Supplementary Material Table A.1 reveals that our CFPS samples consist of 2,110 individual observations, with 408 of them having participated in the Sent-down Movement. Among the individuals who were sent down, 208 were females, and 200 were males. Their sent-down experiences ranged from 2 to 9 years, and Supplementary Material Figure A.1 illustrates that later cohorts spent shorter periods in the countryside compared to earlier cohorts. The majority of individuals in our samples completed middle school or high school, with the younger generations not affected by the movement receiving slightly higher education on average.

6.1. Trust

In Section 3, we suggested that when two institutions compete for trust, families’ loss is the state’s gain. The zhiqing are more trusting of those outside their immediate circle than their age-matched peers, making them more likely to hold pro-regime attitudes and participate in pro-regime activities. Table 3 provides evidence for this mechanism. Sent-down respondents are more likely than others to trust strangers and cadres, but they are less likely than others to trust their parents. The evidence supports the hypothesis that when the state is able to isolate citizens with their families, they would turn to the state as an alternative, and our evidence is consistent with findings from Zhang et al. (Reference Zhang, Zhang and Zhang2023) that adults with the sent-down experience provide less financial support to their elderly parents relative to their non-sent-down siblings. In the Supplementary Material Appendix, we show a consistent and positive relationship between trust in strangers and cadres and attitudes, and a consistent and negative relationship between trust in parents and attitudes toward the regime (Supplementary Material Table B.2–B.3).

Table 3. Effects of sent-down on trust

Note: Standard errors in parentheses are clustered at the graduation-year level. All RD estimations use local linear regressions, triangular kernels, and MSE-optimal bandwidth on both sides of the cutoff. *p < 0.1, **p < 0.05, ***p < 0.01.

The results from the 2008 Survey are consistent, with the circle of trust of the sent-down cohorts being larger than their slightly younger contemporaries. Supplementary Material Table B.1 shows that the sent-down cohorts are more likely to trust people who they do not personally know and equally likely to trust people who they personally know. Trust in unknown people is indeed positively associated with pro-regime attitudes (Supplementary Material Table B.8).

6.2. The local-national gap

Why do the sent-down approve of most aspects of the current political order, but disapprove of local officials? Section 3 proposed that zhiqing are more likely to dislike local government officials than national ones, for two reasons. Firstly, since they often had negative experiences with local officials while being sent-down, they should be more likely to dislike them. Secondly, the nationwide process of being resettled strengthened their national identity.

The direct experience mechanism is difficult to test directly, since we have no information on the details of particular sent-down experiences. However, one indirect test provides some imperfect evidence for this mechanism. A particular group of sent-down individuals, likely harboring grievances against local officials, includes those who remained in the countryside for years beyond the policy’s official conclusion due to bureaucratic delays, often related to the complex process of residential registration (Bonnin, Reference Bonnin2013). Supplementary Material Table B.5 shows that among the sent-down, those who left the country after 1980 but before 1990 are more likely to rate local officials poorly than others.

Supplementary Material Table B.6 provides suggestive evidence that the sent-down cohorts have a stronger national identity. Model 1 shows that the sent-down cohorts are more likely to feel proud of the country, and Model 2 shows that they are more likely to believe that China is a better country compared to other countries. sent-down cohorts are more likely to demand a powerful government to regulate the economy (Model 3). Supplementary Material Table B.5 indicates that among the sent-down, those who left the countryside after 1980 but before 1990 tend to evaluate local officials more negatively than their counterparts.

6.3. Alternative explanations: education, biased response, and baselines

Education: We can rule out several plausible alternative explanations for the difference in political attitudes between the sent-down and not sent-down. Most obviously, the sent-down might have lower levels of education and income than others, due to the fact that they spent several years of their youth outside of the educational system and performing unskilled labor with little transferability to other tasks. However, even youth who remained in urban areas during the Cultural Revolution had poorer educational and occupational prospects than subsequent generations of Chinese. Our results in Supplementary Material Table C.1 show that the negative effect of being sent-down on education is substantial (one level of education), but poorly estimated and statistically insignificant. Intuitively, while it was very difficult for the sent down to become educated, education was provided at very low levels in the China of the mid-1970s even in urban areas.

However, the presence of a small educated group among the not sent-down does not influence the results. In Supplementary Material Appendix Table D.12–D.15, we show that the results are substantively similar among educated and high income individuals.

Biased Response: Another alternative explanation is that the sent-down, perhaps because of their close experience with regime coercion, are more likely to give insincere responses to surveys for fear of punishment, leading them to give artificially pro-regime responses. Supplementary Material Table C.2 provides suggestive evidence that our study does not suffer from this type of political or social desirability bias. According to the assessments of the interviewers, the sent-down respondents appeared to be less concerned about their responses and more reliable when they were answering questions than others. This finding is also consistent with the higher levels of trust in strangers found in the sent-down. Supplementary Material Table C.3 shows that the sent-down are often more likely to respond to survey questions. Similarly, if the sent-down fear coercion more than others, they might be more likely to participate in officially encouraged activities even if they dislike the regime. However, Model 3 of Supplementary Material Table C.2 reports that sent-down respondents are no more likely than others to report that they were forced to vote in local elections, indicating that their higher levels of participation are not a result of coercion.

Cognitive Baselines: A final alternative explanation is that the zhiqing have a different, and lower, cognitive baseline compared to those who remained in the cities. After several years of experiencing rural poverty and deprivation, they might view conditions in urban China as more favorable compared to those who have never experienced rural hardships, potentially leading them to regard the improvements in living standards over the past three decades as a significant accomplishment of the Communist Party. However, the evidence suggests that the zhiqing tend to be more pessimistic and backward-looking. Model 1 in Supplementary Material Table C.4 indicates that sent-down respondents actually exhibit lower confidence in the future compared to others. Model 2 reveals that sent-down respondents are, on average, less happy, though the difference is not statistically significant.