Cooptation without opposition: government response in the authoritarian legislature in China

ABSTRACT Cooptation is widely applied to study authoritarian legislatures, but it is often used to identify and placate the opposition groups. For one-party authoritarian regimes in which the hardcore opposition groups are not represented in the legislature, do autocrats still display dynamics of cooptation when responding to the delegates? Based on a unique dataset of government responses to delegate proposals in China, this article finds that the Chinese regime indeed responds to People’s Congress delegates based on the need for cooptation. Delegates who are members of the ruling Communist Party of China (CCP), who are more tightly controlled through the party institution, receive less favourable state responses than non-CCP delegates. However, the disadvantages of CCP delegates disappear when the risk of internal defection increases. The findings further show that even in the absence of strong opposition parties, authoritarian regimes still adopt cooptation strategies in legislatures.


Introduction
One of the most prominent theories regarding legislatures in authoritarian regimes is cooptation theory, which argues that autocrats use legislatures to make compromises vis-à-vis the public and to distribute spoils among ruling elites. 1 Existing studies show that delegates in authoritarian legislatures are active in making policy requests. 2Delegates from both the regime and the public can form policy coalitions to pursue certain policy outcomes. 3However, under the cooptation view, autocrats' strategy is to identify and placate key members of opposition groups. 4What happens, however, in oneparty authoritarian regimes such as China's, in which delegates are not necessarily selected from opposition groups?Does one-party authoritarian regime's response to delegates still reflect the logic of cooptation even when delegates are not affiliated with any organized opposition groups?Delegates make different demands, and the government certainly cannot satisfy all demands under resource constraints.Does the autocrat decide which demands to accommodate and which to reject based on the cooptation need?This is a critical gap in the literature on one-party authoritarian legislatures: scholars know quite well why and how delegates make policy requests in these legislatures, but relatively little evidence has been revealed on how the government responds to such requests.Many theories posit that public participation in legislative activities can affect policy outcomes, many of which are based on the logic of cooptation. 5owever, little evidence at the micro level that examines the validity of the cooptation logic in these legislatures is available.
Unfortunately, the details of government responses, which is necessary to answer the theoretical and substantive questions above, largely remains a black box.Only selective summaries of government response to delegates' proposals are available; the detailed texts are almost always confidential.Therefore, most studies only rely on limited qualitative evidence, in many cases even delegates' personal estimation, to evaluate government responses. 6One exception in existing study argues that the government response depends on the types of benefits requested under the proposal. 7owever, it does not explain whether delegates of a certain type are more advantaged in obtaining responses because they should be coopted, but rather argues that delegates of different types are better at obtaining favourable responses to different types of proposals.This study, by relying on a unique dataset that contains the full text of government responses from local legislatures in China, aims to provide empirical evidence on these theoretical and substantive questions.To my knowledge, this is one of the first studies that provides quantitative evidence on the presence of the cooptation logic based on comprehensive data on government responses in the one-party authoritarian legislature in China.
This article shows that cooptation is still one of the effective logics determining government responses in the one-party authoritarian legislature even when the opposition force is relatively weak.It finds that when the Chinese regime's control of a group is tighter and the regime's concerns about cooptation of this group are lower, the responses received by the delegates who represent this group are less favourable.Specifically, this article finds that despite the dominant role of the Communist Party of China (CCP) in the People's Congress system, CCP member delegates receive less favourable responses than non-CCP delegates.The fact that China is a one-party regime without organized opposition parties reinforces this logic because regime elites are less likely to defect to the opposition even if they receive less favourable treatment.This seems to be a counterintuitive finding because numerous studies have argued that authoritarian regimes must satisfy the needs of the members of the ruling coalition, which is usually the ruling party. 8However, the disadvantages of CCP members in obtaining responses disappear when more anticorruption investigations are conducted in the constituencies that they represent.After ruling out several alternative explanations, this article argues that the Chinese regime can largely ignore CCP delegates (and the group that they represent) because the cooptation need of CCP delegates is low given that the party institution maintains tight control over party members.Therefore, the government focuses on coopting social elites by responding favourably to their requests.Anticorruption investigations, however, reduce the access to spoils that officials and CCP members originally had and therefore increase their likelihood of noncooperation and defection. 9Facing this heightened crisis of internal cohesion, Chinese governments have to offer more favourable responses to CCP delegates to appease them, which makes the disadvantages that CCP delegates face disappear.
This study makes three major contributions.First, it provides quantitative evidence on an authoritarian government's selective responses to delegate proposals in a legislative system with one-party dominance.In such conditions, the regime's strategic considerations may not be about survival but more about broader cooperation from various classes within the regime since the legislature may not incorporate highstakes opposition groups critical for regime survival.This article finds that even in such a low-stakes consultative legislature, cooptation dynamics still appear in the regime's strategy for handling delegates who do not represent opposition forces.Since similar low-stakes consultative forums are critical in the literature on public policy influence in countries such as China and Vietnam, this finding indicates that the cooptation logic can still explain why some delegates obtain more favourable policy compromises than others.Second, this study provides the counterintuitive finding that members of the ruling party may be disadvantaged in legislature because they are a "low-risk group" with low need for cooptation.This article uses cooptation theory to explain how the Chinese government balances the demands from elites inside and outside the institution.Authoritarian regimes must satisfy the needs of both regime elites and outsiders, but these groups sometimes make competing requests. 10This article therefore, explains how autocrats strike balances between conflicting demands.Finally, this study may also reveal one unintentional consequence of China's anticorruption campaign.With the increasing evidence that officials are disgruntled because of the anticorruption measures, one may wonder how the CCP can maintain the loyalty of its members and officials while depriving them of their opportunities for spoils. 11This study provides one answer to the loyalty question from the perspective of legislative responses in the context of China's anticorruption campaign.

Delegates and government responses in one-party authoritarian legislature
Scholars have developed various theories of authoritarian politics to explain legislative activities.Evidence on how delegates are selected, behave, make demands, and coordinate in authoritarian legislatures has been well documented. 12One group of studies emphasizes the information dilemma of autocracies. 13Authoritarian regimes set up legislatures and allow the political participation of delegates to solicit information, monitor local officials, and ensure governance quality. 14Many studies are thus based on the assumption that regimes allow political activities in exchange for information. 15hile the information argument provides valuable evidence to explain why authoritarian regimes allow delegate activities in legislatures, it is less informative on government responses.There is evidence that policies put to public consultation are more likely to move ahead in the legislative process. 16However, the logic of information extraction does not directly address the question of to whom the government responds.Although regimes need to provide some incentives so that delegates are willing to provide information, such incentives may not necessarily be favourable responses.Delegates may want to continuously provide information in exchange for reselection and material benefits rather than favourable responses. 17nother line of literature stems from cooptation and power-sharing theory, arguing that authoritarian legislatures are institutions for coopting public elites and sharing power with internal elites. 18Although these arguments are widely aired, direct tests of them in one-party authoritarian legislatures such as China's are less common.One reason may be that cooptation arguments emphasize regime responses such as policy concessions, but regime responses in China are difficult to obtain.Another reason is that the cooptation strategy involves identifying and placating key members of the opposition to prevent opposition groups from forming or becoming overwhelmingly threatening, but the functions of controlling the opposition are normally not performed by the legislature in one-party regimes.
Nevertheless, there is some evidence that delegates can receive some feedback by making relevant proposals. 19There is also evidence that online public petitions influence government welfare policy. 20If a similar logic applies to government responses, then there should be more favourable responses to delegates who represent groups with greater cooptation needs and less favourable responses to delegates from the low-risk group with less cooptation needs.Furthermore, the responses received should be more favourable if the group becomes riskier.This study, therefore, empirically tests this conjecture.
One notable previous study in this line of inquiry argues that universalistic proposals receive more favourable responses if they are made by delegates further away from the regime, while particularistic proposals receive more favourable responses if they are made by delegates closer to the regime. 21This study builds on data from the previous Liu piece and extends its arguments, but the two studies differ from two critical perspectives.
First, this study and the Liu piece are based on different theoretical perspectives.The fundamental logic in the latter one is that autocrats can only determine whether a proposal credibly reflects the true policy demand by looking at its content and its proposer and that it is efficient for the government to spend resources on satisfying proposals that credibly reflect true demands.Thus, the Liu piece argues that all delegates can obtain a favourable response if they make a proposal with policy demands deemed credible by the government.This study, however, focuses on the cooptation logic and explores whether some delegates can obtain better responses than other delegates, controlling for the content of their proposals, and demonstrates that CCP delegates, in general, receive less favourable responses than non-CCP delegates because the regime needs to prioritize non-CCP delegates.It provides evidence to support the cooptation logic in that the regime selectively responds to delegates that who it deems in greater need of cooptation even when those delegates do not represent an opposition group in the legislature.
Second, this study differs from the previous Liu piece in that it provides a dynamic process of government response to explain the evolution of the autocratic cooptation strategy.The Liu piece describes a static condition of government response because both the proposal type and delegates' identity are constant.This study argues that the delegates' cooptation needs change across situations and that government response changes accordingly.Utilizing anticorruption campaign data, this study reveals a dynamic process in the cooptation logic: the authoritarian regime favours non-CCP delegates more when regime insiders' morale is at no risk, but this favourable treatment of non-CCP delegates disappear when regime insiders become disgruntled.This article therefore, shows that even the same delegates with similar proposals will obtain different responses when the overall cooptation dynamics on both sides are different.This study, therefore, provides additional and unique insights into the scholarly understanding of government responses displaying the cooptation logic in a single-party authoritarian legislature.

Cooptation without opposition
Authoritarian regimes need to both coopt the public and maintain the cohesion of the ruling coalition. 22As Gandhi suggests, the legislature is an appropriate venue for the cooptation of political elites and distribution of benefits. 23A key question in cases such as China is why cooptation is necessary in a one-party authoritarian legislature, in which delegates do not belong to high-stakes opposition groups.Hardcore regime opponents are heavily repressed by coercive forces, so one-party autocracies like China do not use the legislature to placate the opposition.All delegates in China's People's Congress, to some extent, support the regime or at least do not oppose the regime.Nevertheless, they still represent different social sectors that are key to the regime's functioning.For example, some delegates are more likely to represent local interests while others tend to work with local officials. 24Similarly, delegates with official backgrounds may advocate for their factions when elites are split on policy issues. 25Despite being a one-party regime, China still has various classes.They may all support the regime but still have different, and sometimes conflicting, demands.Balancing those demands from various classes is key for regime and can be a task for the People's Congress.The role of delegates, therefore, is designed by the Chinese regime to reflect the interests of the different sectors, constituencies, or classes that they represent.The Chinese regime has always deliberately selected delegates from different sectors so that it can have broad coverage of those classes and their demands. 26herefore, although China's legislature is not designed to control and handle opposition groups, it still needs to solicit cooperation from different sectors via the legislative channel.The Chinese regime can be responsive to avoid regime instability. 27A regime's stability and continuous functioning require the cooperation and loyalty of different social sectors to ensure state performance, enhance regime control, and generate economic gains.In a public policy process with finite resources, no government can satisfy all demands, especially since the demands from different social sectors are sometimes contradictory.Therefore, there must be some social groups whose demands are not well met, and they may in turn become dissatisfied with the regime.
Repression is certainly a solution to such dissatisfaction, but autocrats cannot govern by repression alone.Repression is costly and sometimes causes stronger backlash.Furthermore, even if repression can prevent elites from rebelling and defecting, political elites may still engage in passive resistance and noncooperation.That is why the literature has found that, even with a low level of threat, autocrats still need to make concessions in exchange for cooperation from social segments, and that such concessions are likely to be policy concessions. 28herefore, this article argues that even in a low-stake legislature without an opposition party, the need to solicit cooperation is strong enough for authoritarian regimes to respond to policy requests from delegates.This is consistent with the underlying assumption of many studies that apply the logic of cooptation to the Chinese case. 29he Chinese regime uses a combination of repression and cooptation to maintain regime stability.It may repress hardcore dissidents who strongly oppose the regime.However, for social elites who are neither insiders nor opponents of the regime, the Chinese regime may need to provide some compromises to coopt them in exchange for their loyalty and cooperation.If the Chinese regime is still willing to make concessions in some situations to coopt non-oppositional social groups, it seems more reasonable to expect cooptation of political and social elites who represent key social groups that are critical to the stable functioning of the regime.
Therefore, one-party authoritarian regimes like China still make policy concessions in the legislature, even though it is perhaps a low-stakes cooptation channel not designed for controlling high-stakes opposition.Government responses to delegates in the People's Congress serve as a credible compromise because the responses are regulated by law and carry some weight, 30 but the government can only make such promises selectively because it does not have resources to satisfy them all.Thus, I argue that the regime needs to strategically prioritize the groups that it has the highest need to coopt.When the Chinese government perceives a stronger need to solicit public cooperation and support, it tends to make policy compromises to delegates who represent the public.This is consistent with findings on other authoritarian legislatures. 31When the Chinese government is more worried about the cohesion and loyalty of internal elites in the ruling coalition, however, it needs to provide benefits and spoils to the delegates who represent the interests of political elites within the institution, as has been shown in different authoritarian settings. 32

Hypotheses
If the Chinese government selectively responds to delegates based on cooptation needs, how does the Chinese government identify whom it has a high or low need to coopt?One way is to differentiate the delegates who represent groups that the state deems "low risk" and those who represent groups with higher risks and with which the state needs to compromise.If the government can satisfy only some demands, it seems reasonable to satisfy more urgent needs from the high-risk group and ignore those of the low-risk group because ignoring the former is costly but ignoring the latter is affordable.Although the regime cannot ignore the latter forever, the regime may provide other exclusive benefits, such as promotion opportunities or access to bribes, to reward them.Thus, when the regime has limited resources to provide policy compromises to everyone in the legislature, the low-risk group seems to be the delegates who can be temporarily ignored.
This article argues that one dimension on which to differentiate the low-risk group from the group urgently in need of cooptation is political affiliation, i.e.CCP membership.Party institutions are crucial to China's authoritarian resilience. 33The strong organizational capacity embedded in the party institution guarantees the regime tight and strong control over party members.The well-organized party structure from the top to the grassroots level ensures that even if one does not work in the government or public organizations, one is still bound by party rule and closely connected to and monitored by party organizations.
Because of the tight control over party members, it is expected that CCP delegates receive less favourable government responses since they are in the low-risk group, about which the regime does not need to worry too much.Furthermore, unlike elites in competitive authoritarian regimes, CCP members in China do not have opposition groups to join and thus are less likely to defect.Furthermore, the regime can offer other exclusive perks (e.g.career promotions, access to corruption incomes, exclusive welfare, etc.) to maintain the loyalty of party members.This argument can also be understood from a different perspective.Failure to obtain the desired commitment may make delegates and the group that they represent angry, unsatisfied, and non-cooperative.However, this may not be a serious concern for CCP members because their support, cooperation, and loyalty are guaranteed by party controls and other perks associated with membership.Given that the government is unlikely to satisfy all policy requests, ignoring those from CCP delegates seems to be a safe option.
This argument does not mean that being a loyal party member does not pay off.As mentioned above, party members can obtain various perks and side payments outside the legislature.In an authoritarian regime like China, the legislature might not be the most important channel through which party members obtain spoils.However, if CCP members can obtain spoils from other (more important) channels that are exclusive for them, the government may not need to provide favourable policy compromises to them in the legislature because those compromises may be reserved for public delegates, who cannot be coopted through other channels.
To summarize the discussions above, because CCP members are both tightly controlled by party institutions and paid off via other channels, I offer the following hypothesis.
H1: Delegates who are CCP members receive less favorable responses to their policy proposals than delegates who are not CCP members.
One immediate alternative explanation is that CCP delegates obtain less favourable responses because their policy requests can be met through other channels.Since CCP delegates have more access to the government, this explanation is plausible.However, this study provides robustness checks and discussions that rule this theory out.Details can be found in the "Extensive analysis and robustness check" section.
One may also question whether all CCP members belong to a low-risk group that poses little need for cooptation.As the CCP is a party with more than 90 million members, it is certainly not the case that no CCP members require cooptation.However, People's Congress delegates are elites, and CCP delegates should represent the core CCP members, over which the regime and the ruling party institution have stronger control and influence.Therefore, even if not all CCP members follow the regime closely, CCP delegates should be generally more loyal to and strongly influenced by the regime than others, thus posing little need for cooptation.
I now try to theorize a dynamic process of cooptation strategy: under what conditions would the government change its strategies to respond to CCP delegates more favourably?As previously discussed, the Chinese regime can provide other benefits via other channels to CCP members, so the regime does not need to respond to them too favourably in the legislature, but what happens if CCP members' access to other benefits is restricted?Members of the ruling party can defect from the ruling coalition when they feel that their interests are no longer well protected. 34Although CCP members may not completely defect from the party, passive resistance by disgruntled members is still possible and not uncommon in the CCP's history.Therefore, the logic of cooptation implies that the government should not entirely ignore demands from CCP delegates, especially when CCP members are at higher risk and become disgruntled with the regime.When access to spoils via other channels is restricted, the regime may need to offer new extra benefits to internal members via the legislature and can therefore make fewer compromises to the public.Thus, the level of spoils is a moderator of the effect of CCP membership on government responses.H2: The negative effect of CCP membership on government responses is weaker when access to internal spoils is restricted.

Anhui: the empirical case
The Anhui data This study uses data from the Anhui Provincial People's Congress in China to test the hypotheses.Anhui is a province in eastern China, whose land area, population, and economic scale are all ranked at the medium level among all the provinces in China.When this study was conducted, Anhui was the only province to systematically release the full text of all government responses in some years to the public.This study collects all delegates' policy proposals and the corresponding government responses to all proposals from the fourth Annual Meeting of the 12th People's Congress in Anhui in 2014 for empirical testing and collects 2016 data for additional analysis.When this study was conducted, the Anhui People's Congress still had not released the response text of other years.While there are no theoretical reasons to believe that the data for 2014 and 2016 are particularly unique, issues of representativeness may still be a concern.I tend to believe that the slow progress in the release of the response text is because of administrative dilly-dallying or a shortage of labour. 35In the online appendix, further discussion is provided.
The advantages of the Anhui data are obvious.The full texts of each government response provide details to analyse trends in government responses.In addition, the full texts allow in-depth evaluation of the responses to provide more variation for measurement purposes.Since the Anhui data are perhaps the only publicly available dataset of government responses to delegate proposals in China, they are the only data that this study can use.Some shortcomings, however, are noteworthy.The dataset covers just one provincial People's Congress with a few years of data available.Whether Anhui is representative of the whole China remains questionable.I tend to believe that the Anhui data are still informative because there is no apparent reason why the Anhui government should act differently from other provincial governments or the central government.In addition, some qualitative evidence from other provinces is also used as supplemental evidence.I determine it to be worthwhile to work with full-text government responses even at some cost to generalizability, especially since the Anhui data seem to be the only detailed data available.Nevertheless, readers should be cautious in generalizing from the findings.

Delegates and government in the provincial People's Congress
Delegates' policy proposals usually make a clear policy request with suggestions.Proposals are usually sent to government agencies, and the agencies are required by law to provide responses with detailed explanations that indicate whether they accept the request.Almost all delegates have other full-time jobs, so it is not surprising to see some proposals with empty words and ambiguous topics.However, since delegates' chances of being reselected are partly determined by the proposals that they submit, they do have incentives to ensure the quality of their proposals. 36In fact, all proposals used in this study contain explicit policy topics and include some suggestions.
One may wonder whether the government responses to delegate proposals are merely symbolic or simply not credible.While this is possible, I argue that this issue is not a serious concern for two reasons.First, delegates have the right to evaluate government responses, and negative evaluations have impacts on the performance appraisal of government officials. 37Therefore, it would be problematic for the government if its responses were cheap talk or not credible.Second, even if some responses are symbolic, a proposal that receives a response that is favourable in tone is still more likely to be implemented than one that receives a negative response.Therefore, I believe that government responses to proposals should be closely associated with any policy compromise that the government provides to delegates.This is consistent with the conclusions of other studies in which scholars find that delegate proposals do have influence and that government officials take them seriously. 38In the online appendix, I provide a detailed discussion on this issue.
This study focuses on a local legislature in China.Cooptation theory was originally developed with reference to national legislatures and regime leaders, not necessarily with reference to provincial officials, who are the officials providing the responses in the Anhui data.However, I believe that provincial leaders are susceptible to a similar logic for two reasons.First, provincial officials also require cooperation from the local public and elites.Adequately satisfying their needs is critical for provincial leadership, and local People's Congresses seem to be an appropriate venue for such compromises.Second, the National People's Congress has a limited number of seats.Thus, most local political elites, who are important for the functioning of the regime at the local level but are not important enough to be selected as one of the approximately 3000 national delegates, will inevitably be coopted at the local level.Provincial congresses, therefore, play a role in coopting these elites.

Data
I collected information on all delegates at the 12th Anhui Provincial People's Congress, including 708 delegates from 16 municipal constituencies. 39For approximately half of the constituencies, the city government published a nomination list for candidates in local newspapers before the delegate election, and information on the elected delegates can be found in the list.For the delegates from the other half of the constituencies, an internet search was used to find their information.For some delegates, there is almost no public information on their background, so missing data are inevitable; therefore, more discussion and robustness checks related to the missing data are provided later. 40 then collected all responses for each proposal in 2014 and 2016 from the official website of the Anhui People's Congress. 41It is notable that all government responses had been copied and pasted into a single Word file at the time of data collection, which indicates a haphazard work style and a shortage of labour for processing government responses.Two research assistants read each government response and rated each response separately on a four-scale index following guidelines provided, and I made the final decisions on items on which the two coders disagreed.The four-scale index represents a direct and firm rejection (1), an indirect and soft rejection (2), an abstract acceptance without explicit implementation plans (3), and a firm and direct acceptance (4).The responses are also coded binarily as rejections (0) and acceptances (1).Delegates usually make one request per proposal, so I do not have any proposals for which both positive and negative responses exist.I provide two examples of proposals and responses (one is an acceptance, and one is a rejection), both of which are about a wastewater processing fee in a rural village.Details on the response coding and examples are in the appendix (Section I).
I also collected data on whether each proposal was cosponsored by multiple delegates.

Variables and measurement
The unit of analysis is the delegate since the need to be coopted lies in each delegate.I match proposals with delegates based on the proposer's name (or the lead proposer's name if the proposal was cosponsored).
The dependent variable is a response favorability index for each delegate, calculated as the average of the response index for all proposals that the delegate submitted.Therefore, the dependent variable takes values from 1 (all proposals received firm rejections) to 4 (all received firm acceptances).I also calculate the average of the binary response for each delegate, which reflects the acceptance ratio of proposals.
The independent variable is party affiliation.Many delegates are members of the eight democratic parties in China.Members of the democratic parties are not CCP members, but they might still be influenced by the CCP institution and are more incorporated into the regime than unaffiliated public delegates.Therefore, this study treats party affiliation as a continuous variable with three values: public delegates (0), democratic party delegates (1), and CCP delegates (2).When the value is larger, a delegate is closer to the CCP, and thus, he or she is more tightly controlled by the regime.To further check the validity of the findings, I also code party affiliation as a nominal variable with three categories, and the results are similar.
The spoils and benefits associated with CCP membership are more difficult to measure directly because of their secretive nature, so I proxy them by measuring the level of anticorruption efforts under the rationale that the stronger the anticorruption campaign, the less access to spoils and benefits delegates have.I use the number of corruption investigations of each city-constituency in the anticorruption campaign, as generated from the dataset collected by Wang and Dickson, to measure the level of anticorruption in each city-constituency. 42This dataset includes all investigations at the county level and above and is perhaps one of the most comprehensive datasets on anticorruption investigations.The number of investigations reflects the intensity of the anticorruption measures in a region.Highly intensified anticorruption investigations create disgruntled officials who may be dissatisfied with the regime because their welfare and opportunities to obtain private spoils are reduced, as several existing studies show. 43In these studies, the number of investigations is used as the proxy for anticorruption investigation levels, reflecting the spoils and benefits that regime insiders can obtain.Following these measures, this study also argues that the number of corruption investigations can reflect the level of anticorruption, which determines officials' access to internal spoils and benefits.Since all the investigations recorded in the dataset are at the county level or above, these investigations should be very influential at the city-constituency level.Thus, when a city has more investigations, the level of spoils and privileges received by party members is likely to be lower among the ruling coalition elites in that city, whose requests, represented by CCP delegates from that city, might be more urgently processed to preserve their cooperation.To account for differences in city size and number of officials, the number of investigations is calculated as the number per 10000 people in each city, with the absolute number used for additional checks. 44In the appendix, more justifications of anticorruption investigations as a measurement are provided.
The following variables are also included as control variables: gender, ethnic minority status, age, education, total number of proposals submitted by each delegate, and percentage of cosponsored proposals.A constituency fixed effect is included.

Models
Two models are used.First, since not all delegates submitted proposals in 2014, the Heckman selection model is used to address the fact that delegates are self-selected and different in terms of their proposal submission.In the selection equation, whether a delegate submitted proposals is used as the dependent variable, and all personal characteristics of delegates (gender, ethnicity, political affiliation, age, and education) are used as the independent variables.I then include the estimation of selection into the main regression to analyse the effect of party membership.Second, ordinary least squares (OLS) regression is used as an additional check and an analysis only on those delegates who are active in proposal submission.For both models, the error term is clustered at the constituency level.More details on the model selections are provided in the appendix (Section IV).

Results
This study uses data from the 2014 annual meeting to explore how the government responded to delegates.Table A1 in the appendix presents summary statistics for the main variables.Briefly, the political affiliations of 637 delegates are confirmed, with 449 (70.49%) being CCP delegates, 60 (9.42%) being democratic party members, and 128 (20.09%) being unaffiliated public delegates.A total of 375 delegates submitted proposals in the 2014 session and have a corresponding response favorability index. 45The mean favorability index value is 2.81.The average number of proposals per delegate (excluding delegates who did not submit any) is 2.26, and approximately 92% of the delegates submitted fewer than four proposals.

Main findings
First, the effect of party affiliation is presented.Table 1 demonstrates the regression results of the Heckman selection model (OLS results are reported in Table A1 and Figure A6 in the appendix).Models S1-S3 use the response favorability index as the dependent variable, while Models S4-S6 use the acceptance ratio as the dependent variable.I present the results of the basic model that controls for the delegates' personal characteristics with no missing data and then expand my models with additional control variables (proposal number, percentage of cosponsored proposals, age, and education).Both the OLS and Heckman models demonstrate similar results.
The results demonstrate a consistent and significant pattern.CCP membership has a clear and negative effect on obtaining favourable responses from the government when all control variables are included.This finding supports H1.To further demonstrate the effect of CCP membership, I then treat the party affiliation variable as a nominal variable with three categories.The results are shown in Figure 1 (Table A7 in the appendix).CCP delegates receive significantly more unfavourable responses.
This finding is consistent with my argument.With the strong party institution, CCP delegates are closely embedded in the regime, which makes compromise with them less necessary.However, public and democratic party delegates should represent the key social elite sectors that the regime needs to coopt.In fact, they have always been important targets in the united front work in China.It is therefore reasonable that non-CCP delegates respond more positively than CCP delegates because the regime needs to provide extra benefits to coopt those non-CCP delegates.To provide some direct examples of how my theory works, I briefly compare the two sample proposals on the water processing fee that I include in the appendix.The comparison shows that, even on a similar issue, the proposal by CCP delegates was rejected while the proposal by a public delegate was accepted.
Next, this article presents evidence for H2 and explores how the effect of CCP membership on the response changes when the level of risk to the regime changes.The number of corruption investigations is used in interaction with political affiliation.The marginal effect of party affiliation is presented in Figure 2 (Table A8 in the appendix).The negative effect of being a CCP member on favourable responses is strongest and significant only when CCP delegates come from a city with relatively few investigations.When the city-constituency has a relatively high number of corruption investigations, CCP delegates are not significantly disadvantaged in obtaining favourable responses.This pattern shows that anticorruption investigations change the need for  cooptation of public and CCP delegates, with CCP delegates needing more cooptation and policy compromises if they represent a group of disgruntled officials whose access to spoils is more restricted.To check the robustness of the findings, I replace the number of investigations per 10000 people with the absolute number of investigations in each city and present the results in Figure A1.The results remain the same.
In the online appendix (Section IV), I provide further discussion on why proposal responses can serve as compensation for dissatisfaction over anticorruption.First, many of the delegate proposals made substantial requests that brought enormous benefits, such as infrastructure construction, salary increases, and extra bureaucratic welfare.Accepting these requests, therefore, should provide substantial benefits and rewards to the local officials and local government that the delegates represent.Second, the effect of the investigation number is also substantive.As I show in the appendix, I find that 20-25% more proposals from CCP members are accepted when the anticorruption level in their constituency is high.This is certainly a significant increase.Therefore, it is plausible to say that more favourable responses in cities with more intense anticorruption investigations can be regarded as extra benefits for cooptation.
Finally, this article presents the marginal effect of democratic party affiliation and CCP affiliation compared to the effect of the public delegates.The results in Figure 3 are consistent (Table A9 in the appendix).When a city-constituency has relatively few investigations, democratic party delegates from that city show no difference in the responses obtained in comparison with those of public delegates, while CCP delegates have a significant disadvantage.However, when a city has a relatively high number of investigations, both democratic party delegates and CCP delegates obtain more favourable responses than public delegates.Since CCP members and democratic party members are more likely to work inside the regime and, therefore, affected by the anticorruption campaign more than the public, it is reasonable to expect these two groups to be more disgruntled when the level of anticorruption activity is high.
In general, the empirical evidence based on the 2014 session of the Anhui Provincial People's Congress supports both H1 and H2.CCP delegates receive less favourable responses from the government than other delegates because there is less need to ensure the loyalty of the groups that they represent.However, when there is a highlevel anticorruption campaign that reduces spoils and creates the potential for disloyalty among CCP and democratic party members, delegates from these two groups receive equally favourable responses.Altogether, the evidence is convincing that the Chinese government provides responses to delegates' policy proposals based on the need for cooptation.When there are imminent risks of noncooperation, the responses are more favourable, and more policy compromises are offered.CCP delegates usually represent a low-risk group, so the responses that they receive are usually less favourable.However, when their spoils are reduced and their dissatisfaction is strong, the government is willing to offer them more compromises and more favourable responses.
Despite the response strategies revealed in this article, more evidence is still needed to fully demonstrate how the government can treat CCP delegates as a low-risk group.The dynamic pattern shown with the anticorruption data provides some support to show that the change in response strategy is based on the level of risk and need for cooptation, but it is still not entirely clear how CCP delegates are regulated to pose low risk.More qualitative evidence is, therefore, encouraged.
Exploratory findings I now discuss several exploratory findings regarding delegates' proposal submissions and the responses received.
First, most personal characteristics other than party membership do not have significant effects.This finding may be interesting to scholars of Chinese politics since conventional wisdom usually indicates that delegates who are public, intellectuals, ethnic minorities, or women should be advantaged because of their importance for cooptation.My finding confirms the advantages of being a public delegate but no advantage for the other three types.There are some exploratory explanations.First, the advantages of being an intellectual, from an ethnic minority or a woman may be significant for delegate selection but not for proposal responses.Second, there may not be an explicit need for cooptation of delegates with these characteristics.Female and better-educated delegates may not be riskier.Although delegates from ethnic minorities in general may be riskier than delegates of Han Chinese ethnicity, this may not be the case in Anhui since ethnic minorities account for only 0.66% of the Anhui population and none of the minority delegates belong to risky ethnic groups (e.g.Tibetans, Uyghurs).Nevertheless, since the scholarly theorization of proposal response behaviour is still limited, these are only exploratory explanations.I encourage further study on this point.
Second, the selection model reveals that delegates who are CCP members, male, older, or better educated are less likely to submit proposals.I suspect that this finding may coincide with the abovementioned argument that these characters are advantageous in delegate selection, as delegates with these advantaged identities may dare to submit fewer proposals because they know that they will still be reselected as delegates anyway.I encourage further study on this issue.

Difference across proposals: an endogeneity problem
One may be concerned that the current analysis does not include the nature of proposals, but the nature of proposals can be a strategic choice of the delegates and determine the response that the delegates receive.The core of this concern is that CCP delegates may systematically submit less favourable proposals so that they receive less favourable responses.In the appendix (Section III), I address this concern in several ways.
First, to address the potential nature of the proposals, I add additional control variables that the existing literature identifies, including 1) whether the proposal demanded parochial benefits for local constituencies or encompassed broad benefits to society, 2) whether the proposal directly requested money, 3) whether the proposal referred to issues that require central approval or coordination with other provinces, and 4) whether the proposal was about legislative bills.I find, after controlling for what types of proposals delegates wrote on average, that CCP delegates still receive less favourable responses than non-CCP delegates.Hence, my findings remain valid after I consider the nature of proposals.
Second, I discuss why some nonquantifiable proposal characteristics that cannot be included in the statistical model (e.g. proposal topic) are unlikely to bias my findings.In sum, I show both theoretically and empirically that CCP delegates do not systematically submit less favourable proposals, so the difference in responses received cannot be explained by proposal types.
Theoretically, CCP delegates usually have a closer relationship with the government, so they should better understand government preferences.They also have better access to party guidelines and policy priorities than non-CCP members because many such priorities and guidelines are discussed within the routine activities of local party organizations.Additionally, many CCP delegates are full-time government professionals, while many non-CCP delegates are part-time amateur delegates with other jobs.Therefore, if CCP delegates were submitting systematically different proposals, they should be more likely to submit proposals relevant to government priorities or with better-quality policy suggestions.This should boost their likelihood of receiving more favourable responses rather than less favourable ones.There seems to be no reason why CCP delegates would be particularly motivated to submit proposals that would be rejected.
Empirically, I conduct a textual analysis of proposal topics that CCP and non-CCP delegates submit.It shows that CCP and non-CCP delegates do not significantly differ in the proposal topics submitted.In other words, CCP delegates and non-CCP delegates submit similar types of proposals, but they receive, on average, different responses.Thus, the types of proposals should not explain why they receive different responses.
In general, I believe that the nature of the proposals does not bias my main findings.Nevertheless, I cannot control for all potential proposal characteristics because of data limitations, and I encourage further studies on this question.

Other channels for CCP members
One alternative explanation is that CCP delegates, because of their closer ties with the government, can have their demands met through other channels and therefore do not need to be placated by favourable responses in the People's Congress.If this were the case, delegates who take a leadership role in government or other important quasiofficial organizations would be more disadvantaged in obtaining favourable responses because their leadership position provides them with more and stronger alternative channels for having their requests satisfied, such that there is no need to placate them in the congress.However, if it is the party membership argument rather than the "alternative channel" argument that explains the government response, there might be no difference between leadership and non-leadership delegates as both groups have CCP members.
To test this alternative explanation, I generated a new independent variable indicating whether one delegate has a leadership role in regime institutions.The results with this new variable are found in the appendix (Table A2).As the results show, being the leader of a regime institution does not affect the favorability index or the acceptance rate, although leader-delegates are much less likely to submit proposals.Leader-delegates indeed have other channels for making requests (thus, they are less likely to submit proposals in congress) but having other channels does not lead to less favourable responses to their proposals.Thus, the disadvantages of CCP delegates in obtaining favourable responses, as I observed, cannot be explained by this alternative argument of having access to other channels.

The 2016 data
Data from the 2016 session provide an opportunity to further test the hypotheses of this article.One important difference between the 2014 and 2016 settings is the intensity of the anticorruption campaign.Compared with 2013, the 2013-2015 period had 11 times more investigations in all 16 city-constituencies.This would inevitably create enormous dissatisfaction among CCP members in these regions, and the regime was likely to be under considerable pressure to appease them.Therefore, the negative effect of CCP membership should be weaker or even disappear for the 2016 session.The results based on the 2016 data are presented in the online appendix.In brief, they align with expectations and further support my argument (appendix, Section III).

Missing data
The next concern relates to missing data for party affiliation, age, and education of some delegates.This subsection provides additional discussions and robustness checks on how these missing data may affect the results.

Party affiliation
The party affiliation of each delegate was recorded in very conservative ways.A delegate was recorded as a CCP member, democratic party member, or unaffiliated public delegate only if there was explicit evidence.If no information on a delegate's affiliation could be found, then the party affiliation was recorded as missing (rather than "public").Under this coding rule, missing values are much more likely to correspond to unaffiliated public delegates than to the other two types because it is relatively easy to find evidence of party affiliation, but it is difficult to find evidence that confirms non-affiliation. 46However, observations of the latter type would be treated as missing rather than coded as public delegates because of the conservative coding rule applied.This article adopted the following three options to address missing values on party affiliation: 1) treating all delegates with missing affiliation values as public delegates; 2) treating all delegates with missing affiliation values as democratic party members; and 3) treating all delegates with missing affiliation values as CCP delegates.Intuitively, option 1 is the most likely scenario, while option 3 is the least likely one.
Figure 4 presents the results based on these three options.In general, the results with the recoded variables are very similar to the results of the original analysis (Table A3 in the appendix).

Age and education
This article uses the multiple imputation method for missing age and education data based on multivariate normal imputation.Other personal characteristics without missing data are used for imputation.
After I generate the imputed values for the missing age and education data, all delegates can be included.Party affiliation is still recoded under the three options described above.The results are presented in Table A3 in the appendix.They again show that the effect of CCP membership is negative across all models, while it becomes slightly nonsignificant when I recode missing party affiliation values as CCP.Therefore, only under the least likely scenario (option 3) do the disadvantages of CCP delegates become less significant.

Discussion and conclusion
This study provides empirical evidence on the Chinese government response to delegates' policy requests in the legislature.This study finds that the Chinese regime usually responds less favourably to delegates who are CCP members because the robust party institution is strong enough to control and ensure their loyalty.Instead, the public and democratic party delegates, who are important to coopt, receive more favourable responses.However, when the cooperation of CCP members is in question, the disadvantages of CCP delegates in obtaining favourable responses disappear because the regime also needs to offer compromises to placate them.This finding contributes to the literature that overwhelmingly focuses on who participates in China's legislature but that has relatively little evidence on the government response to this participation.
This finding for the legislature in China also offers evidence of the cooptation logic, consistent with findings on competitive authoritarian regimes. 47It shows that even in a legislature whose main purpose is not to placate opposition groups or prevent the formation of them, the cooptation function is still salient.It further shows that autocrats do not necessarily favour regime elites in the legislature even though there is little pressure from opposition forces.This article, together with other existing studies, reveals that authoritarian regimes use their legislature as a main venue for both resource distribution and cooptation. 48However, although regime insiders can still obtain some good responses in the legislature, this study shows that the authoritarian legislature is still more like a venue for processing cooptation requests from nonregime members even when those non-regime members are not necessarily from opposition groups.Public demands can still be met even without an organized opposition party in the legislature.Nevertheless, this article also finds that the regime can quickly respond to changes in the ruling coalition's stability and loyalty, indicating that its responsive strategies within the legislature are indeed affected by the changing nature of political dynamics outside the legislature as well.
This study also echoes the growing stream of literature regarding authoritarian legislatures and public participation. 49It indicates that all else equal, public delegates have advantages in obtaining policy compromises.Public delegates have been found to affect policy processes outside the legislature. 50This finding may suggest that even though public delegates are tightly controlled within the legislative process, their participation does influence policy outcomes.It may advance the understanding of these delegates' role and influence in the policymaking process.
The finding of this study appears contradictory to some existing literature since I show that winning coalitions (CCP members) can sometimes be disadvantaged rather than treated favourably in authoritarian legislatures. 51However, I do not argue that winning coalitions are disadvantaged overall; they can obtain much more spoils outside the legislature.The key takeaway from this article is that the winning coalition or most loyal regime supporters are not always favoured and sometimes will be sacrificed to coopt social elites in some low-stake venues, such as one-party authoritarian legislatures.
Regarding the study of Chinese politics, the findings may further reveal the parochial logic accompanying the anticorruption campaign in China.The anticorruption campaign has created instability and dissatisfaction among many party members in China.I find that CCP members who are suffering from anticorruption measures seem to receive compensation from other perspectives.Although CCP members may have multiple channels for making requests, delegate proposals still seem to be one of the few effective methods that they can use to reach regime agencies, especially those agencies with which they do not have a close connection. 52Thus, the People's Congress seems to be suitable for taking care of political needs when internal loyalty is in question.
This finding raises some long-term questions on the consultative authoritarianism model in China and other autocracies. 53In China, the continuous development of the market economy has removed many longstanding privileges of CCP members in areas such as food, housing, and medical welfare.The new trend involves calls for stronger accountability of and tightened control over CCP members in terms of both party rules and rule-of-law requirements. 54There are similar trends in many one-party autocracies in which political elites lost some of their old privileges when the regime no longer monopolized the economy and society.If these trends continue, then the government may need to compromise more with political elites to retain loyal members, and the legislature may become more like a parochial venue for sharing spoils with regime insiders.However, tightening social control, together with a possible slowdown in economic growth (as has happened in China in recent years), may also make it necessary to expand the policy compromises offered to public delegates.How a balance is struck between the two sides, therefore, will be a long-term question for China and other authoritarian legislatures.

Figure 1 .
Figure 1.Party affiliation of delegates and response received.

Figure 2 .
Figure 2. Marginal effect of party affiliation moderated by investigation intensity.

Figure 3 .
Figure 3. Marginal effect of democratic party and CCP membership at each investigation level.

Figure 4 .
Figure 4. Response level of the three party affiliations (missing data recoded).

Table 1 .
Heckman selection model: regression of party affiliation on responses.