Authoritarian Ruling Parties' Recruitment Dilemma: Evidence from China

Abstract In autocracies, party membership offers benefits to citizens who join the ruling party. The recruitment process consists of (i) citizens' applying to become party members, followed by (ii) ruling parties' selection among applicants. Hence, I propose that ruling parties can face a “recruitment dilemma” when the citizens who apply for party membership with an eye on its benefits do not overlap with the ruling party's targeted population. Previous research assumes that the Chinese Communist Party's (CCP) interest in co-opting white-collar workers is matched by those workers’ interest in becoming party members. However, it is their emergence as an essential social group that changed the CCP membership's pattern, leading it to adapt its co-optation strategy to solve the recruitment dilemma. Using surveys across multiple waves between 2005 and 2017, I show (i) changes in application patterns, (ii) the CCP's recruitment dilemma when they receive applications from more laborers than white-collar workers, and (iii) the CCP solution of rejecting laborers in favor of white-collar workers.

As the Chinese Communist Party (CCP) presides over a changing China, it must use party membership to accommodate those who are emerging as a new essential social group for Chinese development (Tsai 2007).Recent literature argues that CCP membership recruitment strategies have changed, co-opting more white-collar workers (Dickson and Rublee 2000;Gore 2015;Ji and Jiang 2020;Sato and Eto 2008).However, we have yet to define whether this change is due to an increase in applications from white-collar workers to become CCP members or whether the CCP has adapted its strategy to co-opt an increasingly vital targeted population for its stability, despite white-collar workers' decreasing interest in party membership.To answer this question, I disaggregate it into two sub-questions.Who is more likely to apply for party membership?And, as a result, what is the decision that the ruling party makes?Party membership is a signature characteristic of autocracies dominated by a ruling party (Angiolillo 2023;Koss 2018), which allows us to best observe citizens' political engagement with ruling parties and the parties' resulting membership recruitment strategies.Previous literature on autocracies shows that the regime attempts to co-opt social elites to strengthen their support and downplay liberal attitudes (Chen and Lu 2011;Rosenfeld 2020).As expected, adapting to the changing national conditions, the CCP's recruitment strategy has converged towards the emerging wealthier social group of white-collar workers (Gore 2015;Sato and Eto 2008).
Previous literature also assumes an overlap between the appeal that CCP membership holds for social elites and the CCP recruitment strategy in co-opting them (Nathan 2016).Although the observed relationship between party membership and professional status reflects the ruling party's co-optation priorities (Dickson 2014;Koss 2018), one of the most problematic assumptions this approach carries is that the CCP co-opts social elites more frequently because they are progressively more attracted by membership in the ruling party, and they seek to maximize their benefits by becoming party members.
However, citizens' interests in becoming CCP members and the CCP's strategic decisions about member selection must be treated consecutively rather than simultaneously.On the one hand, citizens strategically apply to become party members in order to benefit from party membership (Dickson and Rublee 2000), signaling a desire for the benefit of "rent" redistribution in exchange for legitimacy (Wintrobe 2000).On the other hand, a ruling party's recruitment strategy aims at co-opting a targeted population that is vital to maintaining a grip on the political system and stabilizing the party's legitimacy with the population.As a result, ruling parties do not necessarily aim at co-opting citizens belonging to social groups associated with the majority of applicants.Only in the best-case scenario does the majority of citizens applying to become members of the ruling party overlap with the ruling party's targeted population.
I propose that when the party's targeted population (i.e., membership supply) is outbalanced by citizens with a greater desire for party membership's potential rent redistribution (i.e., membership demand), the ruling party faces a selection dilemma.The stability of the regime could be challenged if the ruling party is not able to adapt its co-optation towards the growing vital social group.Hence, the party invests recruitment energies in co-opting its desired targeted population.This is to the detriment of citizens who may show greater interest in ruling party membership in favor a population who might: (1) secure coherence with the national structure, (2) increase the homogeneity within ruling party's ranks, and (3) strengthen the ruling party's appeal among social elites.
Empirically, I leverage on applicants to the ruling party as key to explore the mechanism of the CCP recruitment of its new members in the twenty-first century.Studying party membership as determined by both application and recruitment, I complement previous findings and present a more complex process that includes application patterns prior to recruitment outcomes.
Using rich survey data across multiple waves, and leveraging on citizens' application to the CCP membership, I find that urban laborers 1 are more likely to apply for membership than white-collar workers but are systematically less likely to be accepted.From the applicants' side, urban laborers' higher application rate than that of wealthier citizens reflects the higher value these citizens place on CCP membership.However, in facing a recruitment dilemma, the CCP reacts by adapting its recruitment strategy in favor of its targeted population by discarding other citizens who are more interested in joining the party but who fall outside the targeted population.
These findings depart from previous arguments on the implicit stable mechanisms between applicants and ruling parties' recruitment patterns (e.g., Dickson andRublee 2000; Wintrobe 2000).To further strengthen these results, I use the role of parental CCP membership and educational levels in determining applications, and I implement the Heckman selection approach to show that the observed prevalence of whitecollar workers in the CCP is a result of different admission rates by the ruling party, not an increase in application rates by white-collar workers.
The article contributes to the literature on authoritarian comparative political institutions in three primary ways.First, I focus on recruitment strategies as gateways to explore the party-society relationship in its dual dimension, bottom-up and topdown, where citizens apply for party membership and subsequently undergo a review process before being selected (or refused).Second, broadening the research on ruling parties' recruitment processes to applicants allows for further insights into who is more likely to apply and their socioeconomic background, and to report whether the ruling party adapts to applicants' demand for membership or maintains a firm recruitment policy to co-opt a targeted population.Third, in studying this process, I shed light on the differences between applicants and recruited, highlighting the adaptation strategy as key for the CCP to maintain its core role in the political system and avoid splitting political and economic capital between two different social groups.
The rest of the article is structured as follows.In the next section, I present a new theory on the recruitment dilemma ruling parties face when there is a divergence between supply and demand in party membership cards.I then introduce the data and research design, followed by an explanation of the main results and robustness checks.The last section concludes and presents possible new research paths.

Ruling parties' recruitment dilemma
In party-based autocracies, the division between the ruling party and the state is narrow or virtually nonexistent, which promotes consistency in policy-making on the redistribution of economic and political resources.The state's social structure can influence how citizens engage with the ruling party, generating different patterns of political engagement based on individuals' socioeconomic position.Similarly, the party-society relationship, in general, and the role of party membership, in particular, are substantially different from those of any democratic regime, generating possible stability challenges peculiar to ruling parties in autocracies.
The vast majority of ruling parties allow citizens to apply to become party members (Angiolillo 2023).Though their role is severely limited in terms of policy-making influence, party members serve as useful tools for elites in at least three ways.First, elites can select new associates and appoint them as low-ranking officials who can benefit the elites' careers (De Mesquita et al. 2005;Gueorguiev and Schuler 2016;Jiang 2018).Second, rank-and-file members can help oversee, report, and implement central elites' policies on the ground (Koss 2018;Mattingly 2019;Truex 2016).Third, mass members are a vital source of legitimacy to the ruling party, especially in contexts where elections are not held, and their co-optation reinforces the ruling party's stability, also mobilizing in favor of the regime (Ekiert, Perry, and Yan 2020;Hellmeier and Weidmann 2020;Wintrobe 2000).
Within this framework, recent literature shows that rulers can co-opt depending on their needs for strengthening their legitimacy (Ekiert, Perry, and Yan 2020;Rosenfeld 2020).Hence, party membership recruitment strategies become a vital channel in helping a ruling party to weather social change and maintain legitimacy.This is possible primarily because the ruling party holds the reins of the application review process, allowing it to follow a strategic selection process that strengthens the party's role within the political system.In this context, the emergence of a new social group leads the ruling party to weigh whether to shift the socioeconomic profile of the party by co-opting citizens from an increasingly essential social group or to redistribute its resources among citizens belonging to social groups who have a stronger desire for party membership.Wintrobe (2000) presents a theory of redistribution of political capital in autocracies, in which the most "supportive" group is rewarded by the autocrat through redistribution of rents (i.e., formal and informal benefits of support).Nonetheless, the ruling party also needs to maintain social-group homogeneity within its ranks to avoid possible tensions and frictions in policy preferences.Many party-based autocracies have followed the Leninist party organization strategy, where social-group homogeneity refers to the overwhelming presence of one dominant socioeconomic group within the ruling party and a limited representation of other socioeconomic forces.On the one hand, this approach ensures legitimacy for citizens belonging to the predominant social group, whose representation within the party is predominant as the "vanguard" group, and grants them higher returns for their party membership (Lenin 1969(Lenin [1904]]).On the other hand, the limited presence of other social forces was necessary to penetrate and control critical sectors of society.A classic example of this effort is the CCP policy of the "Three Represents" during Jiang Zemin leadership, which was aimed at opening up their recruitment to owners of private businesses (Chen and Dickson 2010).
As the ruling party and the state virtually overlap in party-based autocracies, it is also in the best interest of the ruling party to maintain socioeconomic coherence between its rank-and-file and the national structure.In other words, it is possible to expect that a substantial national social change might be followed by the ruling party's adaptation to ensure a homogeneous redistribution of political and socioeconomic rents.
In the best case for the ruling party, citizens benefitting from social and economic development would result in higher levels of political engagement in support of the ruling party.In this scenario, party coherence and recruitment patterns develop naturally, and the ruling party would not need to intervene to ensure its adaptation to social change.This approach is implicitly followed when research observes only the outcome of the recruitment process-the changes to the rank-and-file of the ruling parties' social groups-and assumes there is growth on the applicants' side.But this may not always be the case.The "demand" side of party membership can decrease as the party increasingly recruits wealthier citizens who are less dependent on political redistribution.As a result, these wealthier citizens may grow uninterested in politically engaging with the ruling party.Further, wealthier citizens interested in joining the ruling party are often passive, frequently with the primary motive of grasping benefits such as otherwise unavailable employment opportunities such as permission to operate private businesses (Nathan 2016).In contemporary China, wealthier citizens are not universally dependent upon carrying a ruling party's membership card and earlier literature shows that the value of membership may be limited in important aspects of their lives such as their career perspectives (Pan and Zhang 2022), rendering it optional.At the extreme, these changes in the appeal of party membership could develop antagonistic sentiments towards the ruling party.The "modernization" theory depicts a worst case scenario for autocrats-that economic growth will lead to a wealthier society that will demand democracy (Chen and Lu 2011;Lipset 1959).
When the applicants' side does not follow the ruling party's expectations, the ruling party must intervene and adjust the recruitment strategy in its favor to maintain party membership coherence.I propose that ruling parties have active agency in adjusting their recruitment to ensure coherence with national changes, and that they adapt their recruitment strategies when the demand and supply of party membership are not aligned.This raises a dilemma for the ruling party: who to recruit?
Relaxing the assumption that citizens' demand for party membership is always overlapping with the ruling party's best expectations in strategizing recruitment, I propose that when the social structure changes, the "rent" is redistributed to citizens belonging to the ruling party's targeted population even if these citizens are the minority among the overall applicants' pool.This is a fundamental intervention ruling parties can enforce through the recruitment process to secure coherence between changes at the national socioeconomic structure and its rank-and-file demographics.On the one hand, the ruling party might risk a decrease in support from citizens who may be more dependent on the benefits of supporting the party (Truex 2022).These are citizens who, if recruited by the party, would be more likely to unconditionally support the party's legitimacy, as their potential dependence on the political benefits of membership is much more extensive than that of wealthier citizens.On the other hand, recruiting wealthier citizens who form a central social group would secure coherence with the national structure, increase the homogeneity within its ranks, and strengthen its appeal.In actively recruiting these citizens, the ruling party's strategy is to pursue a harder task of appealing to a more independent section of society, in order secure the party's legitimacy by avoiding possible detachments between this social group and the ruling party.
The case of China is a useful example of this selection dilemma.The recruitment process to become a CCP member takes between two and three years, during which citizens first become candidates and, subsequently, after a review process, can be promoted to full members. 2Figure 1 shows that, since 1978, there has been a sharp decrease in the laborers' share of the Chinese urban workforce.At the same time, a cohort of white-collar workers have emerged in the workforce as a result of economic development.A similar story affected the social groups' share within the CCP ranks.In 1978, laborers and white-collar workers had similar shares within the CCP ranks, around 20 percent for each (Gore 2015).
The primary reason for this is the overwhelming presence of peasants within the CCP (46 percent) inherited from the Maoist era (Figure 2 in Appendix A).However, proceeding through the marketization of China, Figure 1 shows how urban laborers' decrease in membership share mirrors white-collar worker's growth within the CCP ranks.By 2018, white-collar workers' presence in the CCP was almost 35 percent, while only 8 percent were urban laborers.At the same time, while in 1978 laborers accounted for 60 percent of the Chinese workforce, white-collar were less than 25 percent.During the following 40 years, the 40 percent difference between the two groups in the workforce virtually disappeared, showing a shrinking laborer force and a rise in white-collar workers, whose importance within the socioeconomic system is now undeniable.

Reasons to join the ruling party
In single-party regimes, the most accessible formal channel of individual political engagement with the ruling party is to become a party member.Official affiliation with the ruling party increased individual political capital in autocracies such as East Germany and the Soviet Union (Bourdieu 1998).Though rank-and-file have very limited influence over the ruling party's policy-making, carrying a party membership card can have different implications depending on citizens' socioeconomic status.Disempowered citizens may perceive joining the ruling party as a way to limit their socioeconomic disadvantages through the benefits of party membership.Previous literature has widely documented the intrinsic benefits that having a ruling party membership card can have for citizens (Dickson and Rublee 2000;Li et al. 2007), which implies that citizens in autocracies can see joining the party ranks as a gateway to increasing their social, political, and economic status.For these reasons, it is reasonable to expect that citizens exposed to social and economic hardship would value membership more than wealthier citizens.
Furthermore, joining a ruling party is an individual action rather than a collective effort, a fact with at least two implications.First, in many autocracies citizens must take the first step in applying to become party members.This points at the active motivation for party membership: applicants may be driven by ideology (Ji and Jiang 2020;Pan and Xu 2018) or they may simply seek the potential benefits of party membership (Dickson 2014;Dickson and Rublee 2000).Second, among citizens within similar social groups, those who find success in their application for party membership are in social, political, and economic privileged positions compared to their coworkers.These two individual dimensions taken together and applied to urban workers results in those affiliated with the ruling party substantially increasing their socioeconomic position, especially in autocracies with high levels of patronage (Tsai 2007).
Conversely, white-collar workers might not apply at the same rate as their economic power grows and their direct dependence on party membership decreases, limiting its effective benefits and appeal.This is not to say that white-collar workers are independent of the state.Previous research shows both that there is a strong relationship between white-collar workers and the state in China (Li 2010;Pearson 1997) and that jobs in the public sector are highly valued by professionals who perceive them as the "golden standard" (Nathan 2016).Nonetheless, the development of both the private sector and businesses provide Chinese professionals with alternative paths that decrease their dependency on the CCP membership to advance their careers and increase their wealth (Pan and Zhang 2022).Hence, the value of party membership differs across social groups, explaining possible differences in application rates, where economically disempowered citizens attempt to gain more political capital to improve their socioeconomic status while white-collar workers are increasingly less dependent on their association with the ruling party.
Some findings on China (Nee and Opper 2010; Nikolov, Wang, and Acker 2020), post-Soviet Russia (Gerber 2000), and Vietnam (Markussen and Ngo 2019) maintain that party membership is the cause of individual economic prosperity, overlooking the possible differences in members' socioeconomic background prior to being recruited (Dickson and Rublee 2000;Li et al. 2007).However, none of the scholars involved in this debate approach party recruitment as an interaction between the party and potential members, limiting the scope of research to the dichotomy members vs. non-members.In understanding whether the adaptation to social change is the result of a higher engagement of citizens who are increasingly at the center of Chinese development or of the ruling party's needs to actively adjust its membership base to them, applicants and those recruited must be analyzed together.By focusing on the socioeconomic differences between the applicants and those who are selected it is possible to determine whether a ruling party faces a recruitment dilemma.Through this approach we can unpack ruling parties' strategies in solving potential co-optation challenges.
In China, joining the CCP would increase urban laborers' political capital and limit their socioeconomic disadvantage.The decline in laborers' in the urban workforce leads to an increase in the appeal of party membership for this social group.Following the argument above on the differences in the value of membership across and within social groups, laborers affiliated with the CCP would have a significant comparative advantage to laborers without party membership.Laborer CCP members would gain job benefits in both the private and public sector.In working for a private company, they would increase their job security and safety nets by being affiliated with a party cell (Chen and Dickson 2010).As these workers are categorized as low skilled, ensuring a political safety net is highly important for both finding employment and avoiding possible layoffs-increasingly so since the marketization of China.Similarly, both accessing and keeping a job in a state-owned enterprise would be easier for laborers with CCP membership.
Furthermore, there is a long-lasting ideological relationship between the CCP and laborers.Though the ideological dimension in contemporary Chinese politics is decreasing in centrality, and it is mostly reduced to a mere formality, laborers' tendency to engage more than wealthier citizens may be influenced by the CCP formal identity recalling a workers-based party.Lastly, the extensive penetration of the CCP into economic production has also contributed to the growth of the political appeal of party membership for laborers (Andreas 2019; Heilmann and Perry 2011).Hence, my first hypothesis tests the appeal of CCP membership to laborers as opposed to wealthier citizens, more formally: Hypothesis 1: In China, economically disempowered citizens apply to the ruling party's membership more than do more economically advantaged individuals.

Solving the recruitment dilemma
Parties are at the core of many successful authoritarian regimes, as one of the regime's institutional pillars (e.g., Pepinsky 2014).Authoritarian parties have different structural roles than political parties in democratic regimes, where the state, party, and civil society are independent.As the distinction between state and party is narrow or virtually non-existent in authoritarian regimes, ruling parties may need to adapt their recruitment strategy over time to maintain this close relationship to the state; this is especially important during substantial national development.I use party membership as a precious connector between ruling party and society, so that a change in the socioeconomic structure leads to changes in the ruling party's recruitment strategy to maintain the stability of the relationship with society and to co-opt a targeted population that is central for the regime.
The ruling party seeks legitimacy from its population; but if the party recruitment process becomes a mere formality rather than a selection process, it risks losing citizen support.This happens when there is either a devaluation of membership (Wightman and Brown 1975) with subsequent disenfranchisement, or an increase in heterogeneous rank-and-file membership that pulls policy-making practices in different directions, stretching the ideological identity of the ruling party.
The appeal of becoming a party member can be volatile as well (Schnytzer and Susterisic 1998).Hence, recruitment strategies that redistribute resources to a targeted population, ensuring similar political and economic redistribution of resources to similar citizens, are crucial.When this adaptation does not happen, the ruling party's accountability to its members is seriously challenged (Roeder 1989) and the necessity of forcefully removing party members increases (Wightman and Brown 1975).
For these reasons, it is not possible to define a static recruitment strategy; and when the socioeconomic structure changes, the ruling party's selection strategy must adapt to prevent a growing separation between its masses and the new emerging social group.When Jiang Zemin legally opened the CCP to white-collar workers through his policy of the "Three Represents," the Chinese ruling party was in need of adapting its rank-and-file structure to its fast-changing socioeconomic country by engaging with emerging private businesses (Pearson, Rithmire, and Tsai 2021).
Finally, recent literature shows that ruling parties may stabilize their regimes by sponsoring regime-complicit civil social movements or mass protests (Ekiert, Perry, and Yan 2020;Hellmeier and Weidmann 2020;Mattingly 2019).In this context, the mobilization of party members, especially during crisis periods or shock events (e.g., the mass mobilization during the Sichuan earthquake in 2008 or the Covid-19 pandemic in 2020-2022 and the socialist rhetoric of "frontline workers"), is essential for channeling political participation to overcome severe challenges.Nonetheless, to effectively channel participation in the ruling party's favor, there should be an alignment between the ruling party's policy-making and the mobilization of citizens to enforce it-and the primary link is once again party membership.
Faulty recruitment strategies are recurrent in party-based autocracies (Angiolillo 2023), and these present challenges to the ruling party.One of the most important roots of faulty recruitment strategies is the inability to solve the dilemma between focusing on citizens who have a high level of desire for party membership and those in the targeted population, who may be more detached from the party, but who are critical for securing adaptation and stability.In the event of a discrepancy in application rates between these two groups, a ruling party co-opting its targeted population chooses to deprioritize citizens who have a strong desire for party membership and who may be more compliant with the regime.
In the case of China, previous literature on the state-society relationship shows the increasingly central economic role of white-collar workers and how the state's institutions have been able to adapt in order to allow their further growth, in the hope of a deeper economic development of the country (Heilmann and Perry 2011;Pearson, Rithmire, and Tsai 2021;Tsai 2007).The consolidation of white-collar workers is of vital importance for the CCP political system's stability, and this results in the CCP's willingness to recruit white-collar workers regardless of their application rate, leaning towards recruiting its targeted population at all costs.
The economic role of white-collar workers is complemented by their growing social and political capital, which puts additional pressure on the CCP to favor them over the majoritarian laborer applicants who place a higher value on party membership.The new emergent generation of young professionals is generally highly educated and culturally active, and some even have professional experience abroad (Ji and Jiang 2020;Truex 2022).The CCP is motivated to pull this cohort towards supporting the regime rather than criticizing it, as has frequently happened online, resulting in the CCP feeling compelled to censor online content (Chen and Xu 2017;King, Pan, and Roberts 2013).
Hence, in solving the selection dilemma, the CCP would take agency in recruiting its targeted population of citizens central to national interests to the detriment of citizens who fall outside that population and have a greater inclination to join the ruling party' ranks.This leads to the second hypothesis: Hypothesis 2: Given a fixed number of absolute admissions to the ruling party per year, the ruling party is more likely to accept professionals than laborers as professionals' socioeconomic national relevance increases.This hypothesis assumes that the percentage growth of recruitment per year remains constant.When it is possible to define a clear social group selected by the ruling party, distinct from the majority of applicants, this indicates that the party is adapting to changing social structure.Between 2000 and 2015, the CCP registered a 2.17 percent average membership growth with a peak in 2012 (3.06%) and a lowest level in 2015 (1.1%), positioning it as one of the most constant acceptance rates among ruling parties.Complementary to this result, the acceptance rate on total population during the 2010s has been highly selective as well (Figure 1 in Appendix A).
As a useful comparison to the CCP membership growth stability, the Lao People's Revolutionary Party (LPRP) membership growth during the same period ranged between 5 percent and 10 percent and the Communist Party of Vietnam (CPV) between 8 percent and −3 percent (Angiolillo 2023).As Figure 2 shows, the absolute admissions are constant at around two million recruited per year.However, what changes is the percentage of relative acceptance rate, swinging between 15 percent to less than 9 percent, supporting the opening puzzle on who is more likely to apply.The change in acceptance rate is simply the mirror of a substantial change in the number of applicants, increasing the relative refuse rate precisely because of the fixed number of absolute admissions to the CCP.

Data and research design
To study the politics of CCP party membership, I use data from three surveys published by the Chinese General Social Survey (CGSS) in 2010, 2015, and 2018, covering the years from 2005 to 2017.The surveys implement a multi-stage stratified sampling design and are nationally representative, with a sample size of 11,783 in 2010, 10,968 in 2015, and 12,582 in 2018.Although CGSS published waves prior to 2010, I selected the surveys that explicitly asked questions on the year of application and (if available) recruitment of respondents.In previous waves, the CGSS asked simply whether citizens had applied, which does not allow refining by year of application and acceptance.Furthermore, CGSS 2018 is the most recent survey available to date, which allows a substantive comparison with previous waves.

Measuring political behavior: applying and recruiting
The two dependent variables capture applications to the CCP and recruitment into the CCP.The first focuses on applicants, with the following question: "When did you submit the application to join the Chinese Communist Party for the first time?"This question allows us to determine whether an individual has applied for party membership during the period under consideration.I subset the sample to include citizens applying to the CCP within five years prior to each survey release to gain precision in my estimates (e.g., for CGSS 2010, I selected individuals applying between 2005 and 2009), dropping those applying six years or more from the survey's year.This conservative approach to trimming the data is necessary to limit the possibility that individuals' social group does not change during the time under consideration.The broader the time bandwidth, the higher the chances of changes in social group.Setting the time to only five years provides confidence that very few, if any, individuals surveyed by the CGSS could have changed their macro socioeconomic group.The dependent variable for applicants to the CCP takes the value of 1 if the individual applied, and 0 otherwise.The second dependent variable is constructed from a separate question which asks, "When did you become a member of the Chinese Communist Party?"This question makes it possible to identify successful applicants who became party members within a three-year review process and those rejected by the CCP.I apply a trimming procedure to this second variable as well.I limit the data to citizens who applied within five years prior to each survey to capture the three-year review process, which allows citizens applying in the two years before the truncated data to be observed in the data under consideration.Hence, the dependent variable for those recruited by the CCP takes the value of 1 if the individual is accepted and 0 if rejected.Hence, this second variable's sample is smaller than the first one as it only considers those who applied to CCP membership while the first variable is constructed to contrast those who applied and those who did not.

Independent variables
My main explanatory variable is a dummy variable that captures the main urban jobs reported by the Central Organization Department of the CCP (ODCCP): laborers and white-collar workers.In coding this variable, I cluster together two questions on the individual's professional status, capturing individuals' work experience and current employment.As a result, I create a new variable divided into seven categories (i.e., casual work, farmers, freelance, self-employed, labor worker, clerk, business owner, manager).I drop farmers (农民) from the main models, which focus on urban areas, and create a dummy where the reference category is white-collar workers (i.e., self-employed, clerk, private business owner, and manager), while laborers take value 1.In the robustness checks, I further assess this key independent variable with different measurements and add farmers.I do not use other variables present in CGSS surveys describing working hours, salary, and whether their work in the public or private sector, because the missingness exceeds 60 percent of the entire sample, making them hardly usable for this study and likely to introduce sample biases.
I also control for individual characteristics that might be associated with professional status.Until the end of the twentieth century, higher education was sufficient to be selected by the CCP; however, it became necessary but not sufficient in the twenty-first century due to the widespread availability of higher education.To assess how age and education influence the decision to apply for membership and be accepted by the CCP, I follow a similar procedure.I measure education using eleven category levels, from primary school to doctoral degrees.This would also account for younger citizens' decision to apply for the CCP and their subsequent acceptance rate.I add individuals' age using a five-category variable from under 30 to over 80 with cutoff points at each decade.Figure 3 in Appendix A shows the descriptive relationship between age and application/acceptance rate.Third, I use a variable that measures whether at least one parent has graduated from college (1 if affirmative, 0 otherwise).I also include control variables for gender (1 is male, 0 female), and ethnic background (1 if Han and 0 otherwise).
To account for geographic determinants of the CCP recruitment, I include a dummy for the household registration system (hukou) (1 for urban areas, 0 for rural) and provinces by CCP membership density.I build on Koss's (2018) three-level provinces categorization by CCP density (i.e., red, light-red, and pink) to create the province variable, which approximates the geographic distribution of CCP members better than using only provinces' fixed effects.This approach is helpful for at least three reasons.First, I can detect whether citizens living in "red" provinces are more likely to engage with the CCP than those in less politicized provinces.Second, I can create an interaction between the key independent variable and this geographic variable, which would unveil whether there are differences in the application rates by citizens belonging to different social groups depending on CCP density at the provincial level.Third, using the CCP density across three survey waves allows me to see whether and how the CCP changes its recruitment strategy at different times.Hence, this specification provides a glimpse of possible changes in the political geography strategy of the CCP between 2005 and 2017.

Model specification
I specify the two main models according to the following general equation: where i indexes each individual and p the provinces by CCP density.The dependent variable y captures each of the two outcomes of interest.First, Applied is a dummy variable for those who applied for the CCP membership or not, and Recruited represents the successful recruitment by the CCP or refused.The key independent variable is individuals' Job, which I already defined previously.The π p are province dummies by CCP members density and X is a set of covariates include age, gender, parents' higher education, household registration system (hukou), dummies for each education level, and Han ethnic background.I primarily implement probit regressions to estimate the main models.I replicate the main results using OLS models in the robustness tests and find that all of my main results are qualitatively unchanged.

Main results
Table 1 presents the main results for applicants, summarized in Figure 3 and I highlight three primary takeaways.First, there is an evident growth in the application rate by laborers compared to white-collar professionals.At the beginning of the twenty-first century, laborers were applying significantly less compared to whitecollar workers, aligning with previous literature on the blossoming relationship between white-collar workers and the CCP.One of the primary reasons for the whitecollar workers' growing interest is explained by the CCP Secretary General Jiang Zemin's "Three Represents" declaration in 2001, allowing wealthier social groups to join the CCP.Hence, Model 1 in Table 1 captures this immediate influence on white-collar workers' desire to apply, outnumbering laborers.Nonetheless, Models 2-3 in Table 1 show that white-collar workers' interest in the CCP faded quickly and is still significantly less prominent than that of the laborers.During 2010-2015, laborers were 30 percent more likely to apply for CCP membership than white-collar workers.Although the strength of this difference has decreased in recent times, laborers are still applying systematically more than white-collar workers.
The findings support hypothesis one: in more recent years, Chinese laborers have turned more consistently than white-collar workers towards the CCP to join the party ranks.However, these results also shed light on an understudied factor.There might be a progressive political disinterest from the white-collar workers towards the CCP.One explanation can be that their rising economic power can be linked to a decrease in their political needs once required to operate in the country.As shown in Figure 1, white-collar workers have sharply taken the center of the country's socioeconomic structure, and nowadays they are at the core of the Chinese economic growth.Hence, this social group is necessary to the CCP, but the growing number of whitecollar workers has allowed many to enrich themselves without official party membership affiliation.
These results are somewhat puzzling because common expectations are that, the CCP's progressive co-opting of professionals within its ranks is naturally associated with the rising rate of white-collar workers' applications.However, Models 2-3 in Table 1 report a different story.Since 2010, laborers have been applying to become CCP members significantly more than white-collar workers.This shows how the increasing number of professionals within the CCP rank-and-file is not defined by white-collar workers' increasing interest in the CCP.
Second, I test the stability of the model using a conservative interaction between job and province of residence accounting for CCP member density.Following Brambor, Clark, and Golder (2006) checklist, I contextualize the relationship of interest by political geography to define CCP recruitment strategies by using unconditional and interaction probit models.Koss (2018) defines "light-red" provinces those with higher density than pink but lower than red provinces, and these  1 and 2, showing a progressive increase of laborers' application to the CCP followed by a progressive rejection rate by the CCP in favor of white-collar workers.experience a higher application rate than the most densely populated by CCP members.In more recent years, the CCP received from 12.81 percent (e 0.115 ) in 2010 to 24.11 percent (e 0.216 ) in 2018 more applicants in light-red than in red provinces.In pink provinces this result increased even more, moving from a −0.7 percent (e 0.070 ) applicants in 2010 to a positive 18.4 percent more applicants in 2018 (e 0.169 ).The similarities between unconditional and interaction models not only shows the relationship's stability, but it also presents stronger results when conditioning on CCP membership density.
Third, if we do not define CCP members by profession, young and highly educated citizens are the highest share.For the sake of space, the education and age values in Table 1 can be fully consulted in Tables A1 and A2 in Appendix B. Results are unsurprising, however, as higher education and younger age are both associated with a greater application rate.To become party members, citizens have to undergo a long and selective process that takes around two to three years.The benefits of party membership for college students also imply more job opportunities after graduation; and becoming a CCP member could help secure jobs both in private and public sectors.Figure 3 in Appendix C shows that most citizens applying to become CCP members are between 18 and 24 years of age, which falls within the years of higher education.
Table 2 reports the CCP response to applicants, 3 and Figure 3 summarizes the main results.Although some of the results are expected, I am interested in the comparison between applicants' behaviors and the CCP response to the changing application by socioeconomic background rather than only looking at the CCP final decision.This allows us to answer questions of whether the CCP has been adapting to the applicants' socioeconomic background or if the CCP has been implementing a recruitment policy against citizens more dependent upon its membership benefits.From the CCP's standpoint, previous literature posits that accommodating more dependent citizens would grant higher legitimacy through lower levels of redistributions (Wintrobe 2000).
Nonetheless, if those more reliant on benefits and the ruling party's targeted population do not overlap, as we saw in Table 1, what is the CCP response?In this vein, Table 2 does not retrace previous findings, but shows the recruitment dilemma the CCP has faced in the first two decades of the twenty-first century.In this way, I contextualize previous correct findings pointing at the growing share of white-collar workers within the CCP ranks.Together with Table 1, I explain these results as one of the decisions taken by the CCP's officials to refuse economically disadvantaged citizens showing higher desire for party membership to the advantage of a targeted population that is harder to co-opt.
These findings are complementary to previous literature in at least two ways.First, previous research points at the progressive increase of white-collar workers within the CCP (Gore 2015;Sato and Eto 2008).Other studies have presented the CCP adaptation to the emergence of the white-collar workers, accommodating their entrance into the party, surveilling them in the workplaces through party cells (Chen and Dickson 2010), and adapting to their economic rise, especially through informal institutions (Tsai 2007).Nonetheless, Tables 1 and 2 show that the CCP is implementing an active recruitment strategy attempting to address the lower interest shown by its targeted population.In other words, I show the CCP's political stretch and internal policy-making decision to fulfill its adaptation strategies in a changing socioeconomic structure.Furthermore, Tables 1 and 2 also provide a broader picture of the political decision the CCP takes in co-opting white-collar workers, which is not an automatic and natural process arising from a higher application rate of white-collar workers.Second, white-collar workers are vital to the CCP, making it stretch its recruitment process to accommodate them.Although it is not a new finding that white-collar workers are recruited at a higher rate, it is puzzling that white-collar workers are applying in significantly smaller numbers than laborers but are recruited in significantly higher numbers.This highlights a possible challenge the CCP faces with whitecollar workers' political engagement.Especially in more recent years, it seems that white-collar workers are less interested in becoming party members than their laborer counterparts (Models 5-6 Table 1).This can be explained by their economically advantaged position in society that does not require them to hold a CCP membership card to secure their socioeconomic status.Nonetheless, despite their low application rate, the CCP's systematic recruitment of white-collar workers resonates with a strategy of adaptation to the changing socioeconomic national structure, as shown in Figure 1.
Furthermore, these results glimpse the progressive geographic changes the CCP might be undertaking.With very limited access to micro-data on the CCP presence in different provinces and its transformation at different times, results in Tables 1 and 2 can be helpful in providing insights into this hardly accessible dimension.I build on previous research on the slow development of the CCP in co-opting citizens in provinces with fewer CCP members (Koss 2018).Similarly, Table 2 does not seem to show the CCP's interest in expanding in provinces with fewer CCP members.One possible explanation can be the threat of overstretching the ruling party.Expanding in a country of these dimensions can expose the elite to threats in strongholds and more secure areas as the CCP focus might be diverted in other directions.
Results in Tables 1 and 2 also seem to rule out the possibility of intense CCP propaganda in areas with fewer CCP members.This might have been the case only looking at Table 1, as the significant positive application rates from areas with fewer party members can suggest heavy propaganda to politicize the population in politically marginal provinces.However, Table 2 reveals the negative and significant acceptance rate in provinces with fewer CCP members to the advantage of the core red provinces.The most extreme value is Model 6 in Table 2, where applicants in "pink" provinces have 77 percent fewer (e 0.574 ) chances of being recruited than those applying in "red" provinces.

Robustness checks
Heckman selection models I implement a Heckman selection model to account for possible sample selection bias (Certo et al. 2016;Sartori 2003).To identify the Heckman model, I use two crucial variables: the parents' CCP membership and education level.This requires a substantive assumption that parents' membership and higher levels of education no longer influence CCP recruitment.A wide range of evidence supports this assumption.
Previous literature suggests that a father's CCP membership has decreased in centrality in contemporary CCP rank-and-file politics.One of the reasons for this decrease comes from the CCP membership recruitment changes over decades.During the Maoist era, membership recruitment was not maintained at a stable rate (Angiolillo 2023), the masses were mobilized and purged with high frequency and the risk of "counter revolutionary elements" was limited by recruiting primarily CCP family members.A drastic change initiated by Deng Xiaoping stabilized CCP recruitment members, following a strict strategy of low and constant yearly recruitment growth.This "ceiling" limitation in absolute numbers provoked a decrease in the efficiency of parents' membership for new applicants to be recruited by the CCP.The value of membership increased and the competition between CCP families naturally increased.For more quantitative evidence, in their historical analysis of CCP membership recruitment between 1949 to mid-1990s, Walder, Li, and Treiman (2000) show consistent evidence that a father's party membership was central to access the party during the Maoist period but decreased during the 1980s and disappeared at the beginning of the 1990s.Later studies have confirmed this primary finding, extending it also to the early twenty-first century (Walder 2004).More recently, Dickson (2014) maintains that father's CCP membership has come back as a driver to determine CCP membership recruitment as a loyalty assurance for younger citizens, though cautioning that this practice is "not officially acknowledged in CCP reports" (p.48).
I present a similar argument for higher levels of education.Many scholars have recently pointed out that the most important source of CCP recruitment is students in higher education (Guo 2007).The primary sources for these correct conclusions are official CCP reports, especially those released by the ODCPP, which insistently show the rise of students within the CCP ranks.Over the last three decades, the Chinese higher education system experienced a surge of college students.Recent studies show that, already in the mid-2000s, China shifted from an elitist to a mass higher education system (Mok 2016), also raising some concern about the possible negative results of this sharp increase-first and foremost the increase of unemployed with a college degree (Mok 2016).The oversupply of citizens with higher education has decreased the absolute value of this variable as crucial to selection.As Figure 3 in Appendix C shows, during college is the period when Chinese citizens are most likely to apply to the CCP, which happens in massive numbers.Hence their oversupply to the CCP should raise some concerns about education being one of the selection criteria rather than an application criteria.
Analyzing members vs. non-members does not take into account that an overwhelming majority of applicants with parents in the CCP or higher education would likely result in a greater number of CCP members with CCP parents or college degree because the selection is done over the applicants rather than the entire population.Moreover, CCP recruitment per year does not exceed two million individuals and as the size of the CCP rank-and-file exceeds 90 million, there is an over-supply of CCP parents among applicants as well.Greater competition among CCP families nullifies the family political premium to similar extents to which higher levels of education nullify the educational premium.
Although the assumptions of lack of influence of parents' CCP membership and education on CCP recruitment are not testable, I can present evidence that they are defensible.If parents' CCP membership still mattered, we would expect to find that this variable predicts admission in the CCP.To test this expectation, I incorporate it in the main recruited models presented in the main findings.Tables A3 in Appendix C shows that parents' CCP membership does not influence the chances of applicants to be recruited more than refused.As I used education in the main models, I implement a different approach.Instead of adding the variable to the main model, I drop education from the recruitment models, which should not modify our key independent variable.As a result, Table A3 in Appendix C also shows similar results for the relationship between individual's job and the chances of being recruited by the CCP, especially in more recent years as the number of graduated increased.Hence, the empirical evidence presented here, together with previous literature and historical data on the CCP recruitment policy adaptation, provides strong evidence that parents' CCP membership and education are excludable instruments in the Heckman selection model.
Tables A4-A6 in Appendix C report the results, highlighting how results for CCP selection remain unchanged compared to the main models, implying that there is no selection bias driving results for the main models.In the first stage, parents' CCP membership and education do strongly influence the decision of individuals to apply (Certo et al. 2016).The Heckman selection models present strong evidence that the CCP recruitment adaptation strategy adjusts to changes in social groups applying to become party members.The ruling party draws a decisive line on who must join its ranks, and systematically applies its recruitment adaptation strategy to a targeted population.

Additional tests
I test the main findings presented in Tables 1 and 2 using a wide array of robustness checks, possible to consult in Appendix D. First, I test the main models using OLS regressions.Tables A7 and A8 show similar results to the main models.
A second set of tests assesses the robustness of our key independent variable's measurement on individuals' jobs.First, in Tables A9 and A10, I add "self-employed" in the laborers category, while in the main model this is a job featured in the whitecollar workers.The reason is that "self-employed" might refer to either laborers or white-collar workers, as the survey provides no further breakdown.Results are, however, unchanged from the main model.Second, in Tables A11 and A12, I add "rural workers" (or peasants) to the main probit regressions.On the one hand, laborers' disadvantage results compared to the white-collar workers are similar.In more recent years, laborers apply significantly more but are refused significantly more by the CCP.On the other, peasants have progressively increased their application to become CCP members compared to their white-collar workers counterparts, especially in more recent years.However, they are not as disadvantaged in their applications for membership as laborers are, due to peasants' fundamental centrality within the Chinese socioeconomic system.Figure 2 in Appendix A shows a graphical representation of the divergence between laborers and rural workers similar to the one portrayed in Figure 1.
Lastly, some concerns may arise for the data truncation in creating the measurements for the two dependent variables, applicants and recruited.Hence, I replicate the probit models presented in the main results using different years.Tables A13  and A14 broadens from five to six years, while Tables A15 and A16 shrinks it further to four years.Each of these further tests replicate similar results to those presented in the main models.

Conclusion
This article contributes to the comparative authoritarian literature by focusing on the recruitment dilemma faced by ruling parties when there is a change in applicants' socioeconomic background.I first study applicants and then move to ruling party selection strategies, unfolding the dilemma that might arise while selecting new party members.
I take the CCP as a case study and show that over the last few decades it has implemented a recruitment strategy that does not reward citizens who value party membership for the benefits it offers, but instead favor a smaller minority belonging to a targeted population of white-collar workers.White-collar workers' initial interest in CCP membership has decreased, and the party receives applications from laborers in significantly higher numbers.Previous theories suggest that this would lead the ruling party to recruit more laborers, prizing their higher interest in becoming party members.Nonetheless, the CCP systematically recruits white-collar workers over any other social group, as they represent an essential social group who are at the core of Chinese national interests.I explain this changing strategy in the context of the changing socioeconomic national structure in China at the beginning of the twenty-first century.The CCP rank-and-file resembles the national socioeconomic structure, thereby avoiding policy incoherence.Furthermore, the laborers are systematically rejected in order to avoid increasing socioeconomic background heterogeneity within the CCP, which can create ideological challenges and frictions.
Further research on autocracies governed by a ruling party might build on these findings to explore the role of the impact of party membership variation on ruling party survival, the role between elites and rank-and-file recruitment strategies, and the changing behavior of individuals in participating in the ruling party organization structures.We have yet to identify the main recruitment strategies implemented by ruling parties, their differences and similarities among different party-based autocracies, and how and why these strategies change over time.
2. This is a standard practice most ruling parties have.The review process lasts at least one year but can be extended to several years (e.g., the Iraqi Ba'ath Party used to have seven years of review process).3. Figure 1 in Appendix A explores the small number of observations.The four density plots represent the absolute CCP recruited between 2005-2015, the CGSS 2010, CGSS 2015, and CGSS 2018.The similarity rules out the possible under-representation of CCP recruited in the analysis (on average 0.29% in CGSS 2010, 0.21% in CGSS 2015, and 0.19% in CGSS 2018) vis-à-vis the CCP national recruitment (on average 0.20%).

Figure 1 .
Figure 1.White-collar and Laborers Share in Urban Workforce and CCP membership, 1978-2018.Note: This figure presents the development of CCP membership (x-axis) and urban workforce ( y-axis) by social group between 1978 and 2018.Each dot represents each social group's workforce and CCP membership shares during the year associated with it.The two arrows indicate the direction of development for white-collar workers, on the right, and laborers, on the left.It is possible to notice that laborers decreased in workforce and within the CCP, while white-collar workers increased their workforce and CCP shares.Figure 2 in the Appendix shows the same plot featuring also the peasants.Data source: CCP Central Organization Department, 2020; National Bureau of Statistics, 2020.

Figure 2 .
Figure 2. CCP Applicants and Recruited, 2000-2015.Note: This figure shows the growth of applicants to the CCP in the twenty-first century and the stable absolute recruitment at around 2 million party members per year by the CCP.Data source: CCP Central Organization Department, 2020.

Figure 3 .
Figure 3. CCP Applicants and Recruited for CGSS 2010, 2015, 2018.Note: This figure summarizes Tables1 and 2, showing a progressive increase of laborers' application to the CCP followed by a progressive rejection rate by the CCP in favor of white-collar workers.