Expertise and Responsiveness in People's Political Consultative Conferences

Abstract

What benefits do inclusive institutions offer authoritarian rulers? Previous research has studied delegate behaviour in authoritarian institutions but has been less well-equipped to assess government reactions to it. Analysing the case of one People's Political Consultative Conference in China, I argue that an overlooked key benefit of inclusive institutions is their provision of expertise. Drawing on novel data comprising more than 9,000 policy suggestions submitted by delegates, delegates' biographies and the corresponding government responses, I illustrate that the government generally values suggestions that signal expertise. While this is especially true for departments of a more technocratic nature, I also find that members of the institutional leadership are systematically favoured. These findings provide an important addition to our understanding of the role of authoritarian institutions in policymaking processes.

摘要

包容性机构对威权统治者有什么益处呢？先前的研究已经涉猎威权机构代表，然而针对政府对其反馈的研究则欠奉。借分析中国人民政治协商会议，本文提出包容性机构一个被忽视的益处－提供专业知识。本文利用新颖的原始数据，包括代表们提交的九千多项政策建议，他们的简历以及相应的政府回应等，说明政府总体上重视体现专业知识的建议。虽然对于技术官僚性质较强的部门尤其如此，但我也发现机构领导阶层受到系统性的青睐。这些研究结果为我们理解威权机构在政策制定过程中的的角色提供了重要的补充。

What benefits do inclusive institutions offer authoritarian rulers? Existing literature based on the China context and elsewhere has primarily responded to this question by extrapolating findings from delegates’ behaviour. Delegates have been found to represent citizens,¹ demand policy concessions,² form policy coalitions,³ engage in power-sharing⁴ and portray regime unity.⁵ Yet, although inclusive institutions supposedly facilitate policy bargaining, the governments’ responses to delegates have largely been underexplored.⁶

Consequently, relatively little is known about whether the government values information provided by delegates or to whom it is more likely to grant policy concessions. For instance, while information provision has been highlighted as a key benefit of legislatures, a regime like China's can collect much more information through direct (online) engagement with citizens⁷ and sophisticated surveillance.⁸ Relatedly, research has also shown how and when authoritarian regimes are responsive to citizens,⁹ including through institutionalized consultations¹⁰ and deliberative polls.¹¹ Yet, systematic accounts of elite-level consultations are rare and research has not been able to assess dictatorships’ responsiveness to delegates in institutions.

This paper contributes to this literature by analysing how government departments respond to suggestions originating from a provincial People's Political Consultative Conference in China (sheng zheng xie 省政协, PPCC hereafter). Assessing delegate–government interactions through policy proposals submitted by delegates in Hainan province and their corresponding government responses, I argue that an overlooked but crucial benefit of the PPCCs and inclusive institutions more broadly is their regularized provision of expertise that ultimately may help to formulate better policies. They provide a forum where delegates can use their expertise to draw attention to previously neglected issues. In contrast, delegate identity appears to be a minor factor in that more trusted elites are not systematically advantaged, except for members of the institutional leadership, which is presumably owing to their influence on agenda setting.

This argument builds upon and goes further than previous accounts that highlight the provision of information by authoritarian institutions, as those accounts primarily focus on delegates accurately transmitting citizens’ preferences and not expert opinions.¹² Original data, including a sample of more than 9,000 delegate suggestions from 2005 to 2019, their corresponding government responses and delegates’ biographical information allow this study not only to focus on the question of whom the government responds to but also to explain variation in how suggestions are received by government departments.

Against this background, I find that the regime responds more positively – i.e. by signalling the government's commitment to implement the delegates’ suggestions and assuring implementation more often – to proposals that demonstrate familiarity with the legal framework, provide evidence and display more complexity. In particular, government departments involved with more technical issues are more open to such proposals. On the other hand, I find very limited support for alternative arguments that see responsiveness as co-optation of non-Party members in the PPCCs or as primarily guided by the elite status of delegates.¹³ With regard to the proposers’ identities, I identify a bias towards members of the institutional leadership but not towards other, more trusted elites, such as government officials and Party members. Instead, private entrepreneurs and professionals are no less likely than government officials to see their suggestions implemented.

The contributions of this article to the study of authoritarian regimes and institutions are twofold. First, this study contributes to the argument that institutions are important for authoritarian regimes’ information-gathering capacities. Previous literature has highlighted public consultation mechanisms,¹⁴ analysed the function of institutions to accurately transmit citizen preferences,¹⁵ or highlighted sporadic contributions of expertise to policymaking processes by NGOs.¹⁶ This study links these arguments to show that inclusive institutions can provide regularized channels of expertise for governments. As such, it builds on prior case studies and is the first to show quantitatively that this expertise is generally valued by the government, with the limitation that one's position in the institution's hierarchy also matters.

Second, this study provides the largest analysis of PPCC proposals to date and, in contrast to previous research, examines not only the delegates’ proposals but also the government departments’ responses. This angle is a necessary step forward in the debate on whether information collected through institutions, and PPCCs, is valuable for autocracies or whether they merely resemble “echo chambers.”¹⁷ The findings presented here suggest that the nature of the information that is collected must be considered, not merely its accumulation.

This paper will first provide some background on the PPCCs before introducing its overall argument, with the corresponding hypotheses and then the empirical strategy to test those hypotheses.

The People's Political Consultative Conferences (PPCCs)

The PPCCs are advisory bodies that exist at all administrative levels above the county/district level. The history of the national PPCC dates back to 1946 when its first session was attended by the Chinese Communist Party (CCP), the Kuomintang and other political parties of the time. While it was the main legislative institution in the immediate aftermath of the Communist victory in the civil war, the National People's Congress later took over that function.¹⁸ As a result, the PPCCs lost their grip on state power and were reduced to consultative entities. Despite not surviving the Cultural Revolution, the political elite during the reform and opening-up era recognized the merits of these institutions and revived them.¹⁹

The PPCCs are among the most inclusive political institutions in China and are a key element in the party-state's system of co-opting important sectors of society.²⁰ In the founding days of the People's Republic of China they were primarily concerned with co-opting the minor political parties.²¹ More recently, however, private entrepreneurs have become the focus of co-optation work.²² In consequence, the membership of the PPCCs is more inclusive than, for instance, that of the People's Congresses. It comprises not only government officials but also private entrepreneurs, academics, professionals, state-owned enterprise executives, government-affiliated organization representatives and military cadres. The first three groups are the largest and account for around 70 per cent of all delegates.²³ Overall, around 60 per cent of delegates are not CCP members.

The mechanisms for selecting individual PPCC delegates are tightly controlled by the party-state to ensure that political loyalty remains a key prerequisite for serving as a delegate.²⁴ All candidates undergo an internal recommendation and review process, which is led by the United Front work department.²⁵ Among the politically suitable candidates, members’ occupations appear to be crucial for nomination, with the aim of projecting diversity and expertise.²⁶ Nevertheless, membership turnover in the PPCCs appears to be generally high after every five-year session. Data from the Hainan PPPCC show that only about 40 per cent of delegates retain their seats, with slightly higher retention rates among non-government officials.²⁷ This suggests that most delegates do not remain in the institution beyond a single term, which makes it difficult for them to build lasting relationships with the government departments that may respond to their proposals.

As consultative institutions, the main task of the PPCCs is to provide advice and recommendations to the government, which ultimately may or may not act upon them. The primary way for delegates to serve the institution's advisory role is to submit formal policy proposals (ti'an 提案). While these proposals can be submitted throughout the year, most are put forward during the annual sessions. They are the main channel through which delegates can express their opinions, suggestions and criticisms, and provide the party-state with a mechanism for receiving information from societal elites. The PPCCs on all administrative levels have seen a dramatic increase in activity in recent decades, with the number of proposals that delegates submit annually skyrocketing.²⁸ However, this number of submitted proposals can also be very uneven, as most delegates do not submit any proposals in any given year.²⁹

In comparison to Peoples’ Congress delegates, PPCC delegates have more room to manoeuvre regarding the issues they write about, as their proposals are not always designed to serve as templates for new bills.³⁰ In contrast to representatives of the People's Congresses, who generally need 30 co-sponsors to introduce a bill (yi'an 议案), PPCC delegates can author proposals individually or with fewer co-authors. PPCC delegates, therefore, have fewer administrative hurdles which makes their proposal writing more flexible.

Consequently, PPCC delegates can address very diverse issues in their proposals, ranging from fixing a particular street to reforming the entire provincial social security system, although both extremes are rare in practice. Some delegates are inspired by the inspection tours organized by the PPCCs and minor parties and write about lessons they have learned from these trips. Most delegates, however, raise issues that are not new to them, but which are linked to their professional identity. For local government officials, for instance, this may mean that they appeal to provincial government departments to request additional resources. Academics and professionals, on the other hand, often highlight issues related to their expertise.³¹

Although some observers doubt the usefulness and influence of PPCC proposals,³² more recent research shows that the PPCCs can be arenas for policymaking conflicts³³ and that delegates often attempt to advance their own interests through proposals.³⁴

While in the past, many delegates were pleased to receive any official response to their proposals at all, it is now mandatory for government departments to respond.³⁵ After receiving proposals from delegates, the PPCC administration will assign and forward them to the responsible government department(s), which then has to respond within five months. Generally, government responses are extensive and provide many details concerning the current legal situation, the government's challenges in implementing the proposal, what the government has done so far to put the proposal into practice, or what else it will do in response to the proposal. The detailed responses convey a feeling of being taken seriously to the different members of the PPCC, even though ultimately their suggestions may not be implemented for various reasons.³⁶ In other words, procedural responsiveness is an important factor in the handling of proposals in itself and contributes to the legitimacy of the institution.³⁷

Upon receipt of the government response, the delegates have an opportunity to express their satisfaction and are provided feedback forms. These forms show that the government is interested in learning about how the implementation of the proposal was advanced, how the delegates assess the government's attitude towards the proposal, and finally, how satisfied they are with the response of the government and the further actions that will be taken. The relevant government departments are instructed to provide opportunities for further discussion and negotiations if delegates are not satisfied with the government's response. These evaluations form part of government officials’ performance evaluations and are therefore taken seriously by the handling officials.³⁸

While the actual policy implementation goes beyond the scope of this paper, there are reasons to believe that government responses are more than perfunctory.³⁹ Some responses provide details as to when and how exactly the government will implement delegates’ suggestions, which indicates that governments are taking suggestions seriously. Several cases indicate this. In 2017, for instance, delegate Chen Xuejun 陈学军 submitted a proposal to the Guangzhou municipal PPCC to include disabled persons in family-based pension services, a proposal that was well received by the municipal Party committee and government. The Guangzhou PPCC then coordinated with the municipal disabled persons’ federation, the municipal civil affairs bureau and the municipal finance bureau to fully negotiate the details and the solutions for their inclusion. The civil affairs bureau took the lead and drafted the “Work plan for the severely disabled persons in Guangzhou,” which proposed that disabled persons be included in the pension system. The municipal disabled persons’ federation, the municipal finance bureau and the district governments were all consulted over the implementation of the plan and approved it to the satisfaction of the initial proposer. Appendix 1 describes another case in which a proposal led to a change in policy.⁴⁰

While the delegates’ side has received much scholarly attention, the side of the government has remained unexplored, largely owing to limitations in data availability. Below, I present my empirical expectations, namely that government departments will be more inclined to respond positively to delegate suggestions that seem to be higher quality.

Quality over Quantity: Empirical Expectations

The Chinese leadership has long been concerned with the quality of proposals submitted to the PPCCs at all levels. Official remarks emphasize a narrative of policy rationalization and signal that the regime may particularly value high-quality proposals. In 2020, Xi Jinping 习近平 expressed the hope that delegates would continue to serve as “good advisers, good aides and good colleagues” and held that the Chinese system of multi-party cooperation can prevent the flaws of oversight in one-party regimes.⁴¹ These remarks align with official documents, such as the “Regulations of the National Committee of the Chinese People's Political Consultative Conference” (2005) and the “Opinions of the CPPCC National Committee Proposal Committee on improving the quality of proposals” (2018), which encourage delegates to improve the quality of their proposals.⁴²

The 2018 “Opinions” underline that PPCC delegates “must emphasize quality rather than quantity … continuously improve the ability to propose high-quality proposals, and encourage members to concentrate on submitting 1–2 high-quality proposals every year.”⁴³ To judge these, PPCCs across the country have established quality-evaluation guidelines that, among other things, highlight the necessity of in-depth investigations to collect comprehensive data, familiarity with the relevant laws, regulations and policy documents, and specific recommendations. While governments and PPCC administrations had to regularly issue calls to stop embarrassing proposals in the past, delegates submitting high-quality proposals are now publicly praised, offering further incentive to perform well.

In fact, proposals submitted to the PPCCs can be remarkably informative and enrich policymakers’ perspectives with practical experience. In some of the most developed proposals, delegates even conduct their own studies to direct attention to previously neglected issues. For instance, in 2017 the Provincial Committee of the Democratic League distributed a survey among teaching staff in private colleges (“Suggestions on the development of teachers in private colleges and universities”). The results of this survey highlight that most teachers share the belief that their employment conditions are generally very unstable and unsatisfactory. In addition, in 2013, an individual delegate addressed the issue of community health services (“Some suggestions on promoting the development of community health services”). Relying on a variety of statistics, the proposer revealed the lack of construction land and shortage of qualified staff available for community health service institutions. Studies such as these can provide an additional source of information for government officials, which can further support evidence-based policymaking.

The first hypothesis therefore investigates whether government departments respond more positively to proposals that signal expertise.

H1: Proposals that can display expertise in the subject matter receive more positive responses from the government.

In addition to the proposals’ characteristics, expertise may also be conveyed through the identity of the delegates. For instance, academics and professionals often submit proposals that are directly linked to their research or work experience and they may therefore be regarded as experts on the subject irrespective of specific proposal characteristics.⁴⁴ Similar to the participation of NGOs in policymaking processes, academics’ knowledge has also been shown to be valued by the government.⁴⁵ However, whether this is limited to a few cases or reflects the regime's general approach remains unknown. Thus, the second hypothesis analyses whether suggestions submitted by academics and professionals are valued more highly by the government than those submitted by other delegates, in particular, those submitted by the reference category of government officials, which forms the largest group of delegates. “Professionals” here mainly refers to medical professionals and lawyers, but the category also includes a smaller number of accountants, engineers and teachers.

H2: Proposals submitted by academics and professionals receive more positive responses from the government.

In contrast to the information and expertise argument above, an alternative mechanism may stress the importance of the elite status of the proposer instead of the content of the proposal. Given that the projection of party unity has been highlighted as the main task for authoritarian assemblies, especially in one-party regimes,⁴⁶ the government may be inclined to respond positively only to submissions that adhere to the party line and which are submitted by more trusted elites. In consequence, some authoritarian institutions have been dismissed as “echo chambers” that do not consider policy alternatives based on their merit.⁴⁷ Thus, the extent to which delegates are more trusted elites in the political system is tested by the third hypothesis.⁴⁸

H3: Proposals submitted by trusted elites receive more positive responses from the government.

Empirical Strategy

To test the abovementioned hypotheses, this paper utilizes an original dataset of proposals, and their corresponding government responses, which were submitted to Hainan province's PPCC between 2005 and 2019. While several provinces and municipalities make (some) proposals publicly available, government responses are not accessible at the national and provincial levels, except for in Hainan.⁴⁹ Thus, the primary reason for choosing Hainan's PPCC is data availability. This leads to two potential concerns with data and results. First, it is unclear why Hainan was the only province to release governmental responses. The reasons behind these decisions are impossible to know.⁵⁰ Although it cannot be completely dismissed, the data presented below do not suggest that the provincial government has made substantial efforts to misrepresent its (public) track record of responses, as there is also a relatively large number of non-positive responses. Dongshu Liu also suggests that it is not in the interest of government departments to manipulate records of their responses as they may be (partially) held accountable for false promises.⁵¹

The second concern relates to the representativeness of the case of Hainan and its PPCC. Hainan is China's only island province and may be peripheral in elite-level politics, but it falls in line with national averages in key statistics, such as GDP per capita and urbanization, and is not a substantial outlier in terms of Party membership and “high” corruption. Processes within the PPCCs have by now become fairly standardized across the country, but it cannot be fully established if government officials in Hainan respond to proposals in the same way as officials do elsewhere. Although I offer indirect evidence that this may be the case in the online Appendix, limits to generalizability remain inherent, as is typical in studies of this kind.⁵² A longer discussion on these two concerns is provided in the Appendix (Section 2).

For the purposes of this study, it is also important to distinguish between proposals (ti'an) and suggestions (jianyi 建议). Proposals can consist of one particular suggestion, such as fixing a specific road section, or several suggestions that aim towards reaching a certain goal, such as improving rural education (see Appendix Section 3 for an illustration of the common proposal structure). Since the government can respond differently to suggestions, the unit of analysis for this study is the suggestion, rather than the proposal. From a total of 5,446 proposals submitted during this time, a random sample of 2,473 proposals (45 per cent), amounting to 9,589 suggestions, was taken.⁵³

The dependent variable for the analyses is the government's response to each suggestion. Research studying responsiveness towards citizens has focused on the question of whether the government responds to queries at all.⁵⁴ In the case of PPCCs, however, government agencies must respond to every proposal and thus the key question is how they respond. For classification purposes, this study has developed a 5-point scale coding scheme with the primary goal of capturing whether the government will act in response to the proposal and implement the suggestions. In many cases, this aspect is different from linguistic or textual characteristics, such as the responses’ sentiment for instance, since the government may be sympathetic to suggestions but may not have the necessary budget to implement them. Thus, for this analysis, a non-automated approach for coding the responses is deemed more suitable.

The response categories are as follows. Category 5 responses are the most positive in that they have either been implemented already (since the submission of the proposal), or else they will be implemented as proposed, and details are provided of how and when the government will do so. The following is an example of a Category 5 response:

The “Suggestions on setting up traffic lights at XXX” … has been received. Thank you for your concern and support for our city's traffic management work … Before the Spring Festival in 2006, our city set up traffic lights at this intersection and perfected the installation of traffic signs and markings and other supporting traffic facilities.⁵⁵

Category 4 responses generally offer support for the suggestion and pledge that it will be implemented. However, these responses use relatively vague terms, such as “strengthen,” “develop,” “explore” and “pay attention,” and do not offer clear plans of action. Category 3 responses are “neutral,” which indicates that the relevant government department has not concluded its work and will conduct more research before giving a substantive response. Category 2 describes the abovementioned case, in that a department may support the suggestion, but owing to a lack of financial resources, the government is currently unable to implement it. An example of a Category 2 response is:

Owing to the heavy task of the reconstruction of rural highways and the construction of the “Access project” in our province this year, the funding gap is relatively large. Therefore, there is no funding source for your proposal to build the highway from … However, our department will consider it in the Eleventh Five-Year Rural Road Plan and implement it as soon as possible, depending on the funding situation

The lowest category, Category 1, includes responses that reject suggestions outright on a substantive basis because the problem described in the proposal is not viewed as an issue or because the suggestion is not considered a solution to a problem. This is illustrated by an example from 2005 in which a proposer suggested the government regulate teacher training in private schools:

Private schools are responsible for the training of teachers in private schools … Education administrative departments should not force teachers at private schools to participate in training.

The complete coding scheme is provided in the Appendix (Section 4), with more details and full examples (Section 5) for each category.⁵⁶

Three coders classified the responses according to this categorization and four steps were taken to safeguard the intercoder reliability. The first was the design of the different categories. By focusing on the question of whether the government will take any action, the study provides mutually exclusive and relatively easily distinguishable categories. Second, all coders received identical extensive individual training before coding. Third, the entire coding was monitored and a quality improvement feedback mechanism established for continuous communication. Finally, a subsample of proposals (248 proposals/10.0 per cent/1,000 suggestions/10.4 per cent) was coded by all coders to directly observe and measure the intercoder reliability. The results of the intercoder reliability tests are provided in the Appendix (Section 6) and show a high degree of reliability. The coding decisions for the entire sample also do not diverge significantly from each other, as shown in the Appendix (Section 7).

Owing to the vastly different topics of the proposals, it is difficult to find common measurements for the expertise they demonstrate. Nevertheless, this study utilizes three different factors to identify proposals that signal competence and which are comparable across proposals: presentation of relevant statistical evidence, knowledge of relevant existing laws and complexity of the language used in the proposal. The first two are also recognized by the government as important factors when evaluating the quality of proposals.⁵⁷

Providing relevant evidence is crucial to convincing decision makers of the severity of the problem at hand, and most delegates recognize this.⁵⁸ Statistical evidence can convey credibility when raising an issue and highlights that the proposer knows the issue well. Often the process of gathering evidence itself entails substantive work and is taken as a signal of expertise by many government officials.⁵⁹ In addition, I consider the mention of laws as an important indicator of expertise. It signals the proposers’ familiarity with a certain issue if they can refer to the current state of legislation.⁶⁰ Proposers can pinpoint how existing legislation relates to the issue at hand and illustrate, for instance, whether authorities have failed to follow national laws and guidelines. Expertise in the field can also help delegates to highlight how issues are not covered by existing legislation. Thus, this measurement does not only apply to issues where the government has already promulgated laws. Finally, the linguistic complexity of the proposals serves as an indicator of expertise. Prior research has shown that linguistic complexity does not reflect personal writing style but shows that writers are more familiar with the topic.⁶¹ Thus, the complexity of language used in proposals suggests that delegates use more technical jargon and can signal their expertise in this way. It can also indicate that the proposer has more knowledge of the topic and thus can use a variety of words to explain a phenomenon.⁶²

First, a binary variable is used to document whether the proposals cite any statistical evidence to support their claims. This variable is coded as 1 if the proposal contains any empirical data to indicate to what extent the issue at hand is a problem that should be addressed or how the suggestion will help to resolve it. For example, in a proposal related to the health condition thalassemia, delegates cited statistics on how many people were affected by it in the province. In another example, a delegate advocated the promotion of vocational education in ethnic minority areas, but to convey the current lack of attention paid to this issue, he referred to statistics on current enrolment rates and funding of vocational schools across the province.⁶³ When such evidence is absent, this variable is coded as 0.

Second, all mentions in the proposal of any existing laws promulgated by the State Council, provincial and local governments, as well as national and provincial government departments, are recorded. Finally, the proposals’ lexical diversity is calculated with the R package quanteda. The lexical richness here is measured by the type-token ratio that denotes the variation in words used in each proposal. More variation often highlights that a text is more difficult to understand.⁶⁴

In addition to proposal characteristics, I also created a biographical database of Hainan province's PPCC delegates. The name lists of the delegates came from the PPCC website and further biographical information was collected from Baidu and university and company websites. Importantly, I collected their occupations, which allowed me to identify academics and professionals for the purposes of H2. In addition, I recorded whether delegates are CCP members or not, and whether they are members of the institutional leadership.⁶⁵ These variables form the basis of investigation for H3. Trusted elites here are coded as those that belong to the institutional leadership, are Party members or are government officials. To see whether government officials (or other occupations) are systematically favoured by government responses, the following analysis compares all occupations to the reference category of government officials.

Additional personal details, such as gender, ethnicity, birthplace, experience in the PPCC, and the number of proposals submitted to the PPCC each year, were also collected and included as control variables.⁶⁶ It could be argued, for instance, that more experienced deputies are better able to tailor a proposal so that it is more likely to be read and acted upon by a government ministry. Similarly, it may be expected that local delegates may receive more favourable responses owing to their potentially closer networks with government departments.

However, especially at the provincial level, these additional personal details were not available for every delegate. Thus, the problem of missing data has been addressed with a data imputation approach.⁶⁷ The results presented here are based on the mean values of 35 imputation runs on the original dataset. More details on how many observations are missing and were imputed are provided in the Appendix (Table 8).

The analysis also includes delegates’ displays of loyalty to the regime in the proposals as a control variable. The most straightforward way to display loyalty to leaders and the CCP is to cite them or their ideologies in proposals.⁶⁸ I particularly focus on whether delegates cite the Party secretary, the National Congress of the CCP as well as provincial Party secretaries and governors.⁶⁹ The latter two are included because they are the key decision makers at the provincial level. All references to these leaders and institutions were recorded for every proposal. The National Party Congress, which is the main event during which the country's guiding ideology and national development strategies are aligned and confirmed, only takes place every five years. References to the preceding National Party Congress will carry more or less weight depending on the time that has passed since the last one and may be particularly important in the earlier phases of the five-year plans and less so once the next National Party Congress is on the horizon again. Therefore, in the statistical models, a binary variable of whether a proposal makes reference to the National Party Congress is interacted with the number of years that have passed since the last one took place.

Finally, the models include further control variables regarding the proposals, such as the number of departments that are required to respond to the proposal, whether the proposal focuses on a local (i.e. city/county/village-specific) or more general issue,⁷⁰ and the length of the proposal as measured by the number of characters. Descriptive statistics of both the imputed and non-imputed data are available in the Appendix (Tables 9 and 10).

Results

Figure 1 highlights the distribution of government responses across the five categories in the sample. The histogram illustrates that there is significant variation in government responses to delegates. While most responses are indeed positive (Category 4), a much smaller number of responses commits the government to observable actions (Category 5). The next highest number is of Category 2 responses, while only a few responses fall into Category 1 and Category 3.

Figure 1. Histogram Distribution of Responses

The empirical analysis, shown in Table 1 with robust standard errors, was conducted in two steps. First, ordered logistic regressions were utilized to analyse government responses according to the ordinal 5-point coding scheme. The results of these analyses are illustrated in Models 1 and 2. Model 1 analyses how proposal characteristics are related to the government's response and, as such, also includes proposals submitted by groups such as the minor political parties. Model 2 complements this with additional variables controlling for delegates’ personal characteristics.

Table 1. Baseline Results

Notes: † p < 0.1, * p < 0.05, ** p < 0.01, *** p < 0.001. Robust standard errors are in parentheses. These are ordered logistic regression results (Models 1–2) and logistic regression results (Models 3–4).

Model 1 shows that when delegates include statistical evidence to support their cause and reference existing laws, they are more likely to receive positive responses. The type-token ratio, however, is not associated with the government's response in Model 1. Variables capturing displays of loyalty are largely not significant.

In Model 2, the characteristics of the proposals are complemented by those of the proposing delegates. The results illustrate that references to existing laws remain a significant predictor of positive government responses, while there is no evidence for the importance of statistical evidence. Concerning the occupational backgrounds of delegates, I find that it is only academics who are significantly disadvantaged when compared to the reference category of government officials. Compared to ordinary delegates, members of the institutional leadership generally tend to get more positive responses, but Party members’ suggestions are not treated favourably.

Models 3 and 4 utilize the same independent variables, but instead of using the 5-point coding scheme, the dependent variable for these models is the binary variable of whether the government replies with a Category 5 response or not. As mentioned above, Category 4 responses often only offer rather vague promises that the delegate's recommendation will be acted upon. Category 5 responses, however, highlight how the suggestions have either been implemented already since the submission of the proposal or else provide assurances, with specific timelines, on how the government will execute them. Models 3 and 4, therefore, arguably, offer a more direct analysis of policy implementation. These models are estimated using logistic regressions.

Model 3 highlights similar important proposal characteristics to those in Model 1. First, variables related to expertise are also significant here. Providing statistical evidence as well as demonstrating a familiarity with existing laws significantly increases a proposal's likelihood of success. In addition to Model 1, there is also a significant positive coefficient for the type-token ratio. In other words, the greater the variety of words in a proposal, the higher its chances of implementation.

Model 4 again combines proposal and delegate characteristics. Importantly, regarding expertise, references to existing laws and statistical evidence remain significant while the type-token ratio marginally misses statistical significance. Regarding delegate occupation, I find a significant negative coefficient for academics but no significant effects for entrepreneurs and professionals when compared to the reference category of government officials. In addition, the coefficients of the institutional leadership remain significantly positive, whereas the effect for Party members is indistinguishable from zero.

Overall, I find considerable evidence supporting H1 – proposals displaying expertise are well received by government departments – but no support for H2. This may be a limitation of the case of Hainan where higher education is relatively less developed and resourced than in other parts of the country. A look at the descriptive statistics reveals, for instance, that less than half of academics’ proposals cite evidence and only one-third reference existing laws. The evidence for H3 is limited to the relatively small group of members of the institutional leadership that seems to receive systematically more positive responses, presumably owing to their agenda-setting power and ability to set annual key topics and schedule special investigations. However, this does not apply to Party members at large or the category of government officials. There is also no evidence that longer-serving PPCC members, those with a local background or those displaying loyalty have systematic advantages in seeing suggestions implemented.

Robustness Tests

I conducted a number of robustness tests to understand the validity as well as the scope conditions of the results. First, the sample of the analyses was modified. While the models for the baseline results above utilized the coding of Coder 1 for the intercoder reliability subsample, the same analyses were run with the coding of Coders 2 and 3 as well (Appendix, Tables 11 and 12). The results of those models are largely in line with those presented in Table 1 in terms of effect magnitude and significance. Moreover, as mentioned above, all models including the personal characteristics of delegates were analysed with imputed data. Thus, for the robustness tests, Models 2 and 4 were replicated with the non-imputed data. Despite the much lower sample (3,540), the results appear to be largely robust here as well (Appendix, Table 13). The coefficients for the reference to existing laws and the higher responsiveness to members of the institutional leadership remain significant.

For an additional test, I also noted the locations of delegates’ workplaces and excluded any delegates not based in Hainan, such as those from Hong Kong and Macau (Appendix, Table 14). The results provide further evidence of the robustness of the baseline models. Moreover, the results also remain largely robust when excluding responses from Category 3 (neutral) and analysing replies on a 4-point scale instead (Appendix, Table 15).

Second, I modified the estimated models. Proposals may receive different responses depending on the issues they address and, relatedly, which government department responds. To discount this possibility, I ran additional analyses, including the corresponding responsible government units for each proposal and the topics of the proposals. Topics were determined using the government's classification of proposals into five categories, namely economic, political, cultural, social and environmental development. Fixed-effects models with both variables (Appendix, Tables 16 and 17), as well as a model with standard errors clustered by government department (Appendix, Table 18), speak to the robustness of several important variables, especially given that the number of different government departments is very high (N = 85) and that their duties in responding are distributed very unevenly.⁷¹ The results also remain robust when estimating Models 1 and 2 with an OLS regression as opposed to an ordered logistics one (Appendix, Table 19).

Third, significant differences may also be expected across the different administrations of Hu Jintao 胡锦涛 and Xi Jinping. While expertise did not seem to be valued before 2013, most indicators of it are positive and significant in the period after 2012. Also, members of the institutional leadership are only systematically advantaged in the Xi Jinping era, while academics are not generally disadvantaged anymore during the same period. Full results of the analyses on the Hu and Xi eras are available in the Appendix (Tables 20 and 21). Furthermore, I provide an alternative way of retrieving the proposals that are most likely to be implemented (Appendix Section 22). While the expertise of these proposals cannot be verified, this approach also shows that different occupational groups see their proposals implemented and thus provides further evidence for the quantitative analyses above.

Finally, I also tested two alternative explanations. The first one accounts for the possibility that the government responds more positively to proposals that highlight grievances which may potentially trigger collective action. The results can be found in the Appendix (Section 23) and indicate that the collective action and grievance argument can hardly be sustained in this case. In addition, I highlight that cross-locality and/or cross-sectoral collaborations are not favoured by the provincial government either (Appendix Section 24).⁷²

Since the PPCCs are primarily used to co-opt high-status individuals from different sectors,⁷³ it may be the case that government departments respond more positively towards individuals of a high social status. I tested this potential mechanism (results shown in Appendix Section 25) but did not find empirical evidence that, within occupational groups, individuals of higher status receive more positive responses from government departments.

Technocratic Pattern

Department-fixed effects models and interaction terms reveal a technocratic pattern. That is, most departments that are open to suggestions, especially when relevant evidence is provided, are of a technical rather than political nature. Table 2 highlights a selection of departments that show significant interaction effects with occupational groups and empirical evidence. These include, for instance, the water affairs bureau, the departments of finance, environmental protection, and agriculture. Generally, the public security bureau, the supervision department and the poverty alleviation office are less open to delegates’ suggestions, irrespective of the occupational background of delegates and whether evidence is provided or not.

Table 2. Positive Interaction Effects Occupational Groups/Evidence + Departments

A closer look at the individuals who receive more positive responses from the more technocratic departments also reveals that these individuals are not always the same delegates. Owing to the relatively high turnover rate, it is extremely difficult for delegates to build such close connections with departments that their proposals are more likely to be accepted. Even the delegate with the most proposals submitted to the water affairs bureau, for instance, only sent a total of five proposals there in the entire universe of proposals between 2005 and 2019.

Conclusion

This study adds to our understanding of inclusive institutions in authoritarian regimes through an investigation into delegates’ proposals and the corresponding government responses in a PPCC in China. More specifically, this analysis goes beyond prior research, which focuses only on delegates, and provides an analysis of government–delegate interactions, which enriches our knowledge of policymaking in China and the role of inclusive institutions in these processes. This paper reveals that inclusive institutions can serve the important purpose of regularized provision of expertise, which may ultimately lead to better policymaking. On the other hand, the analysis also reveals that the government does not systematically respond to the status of the elites, with the exception of members of the institutional leadership, to accommodate either more trusted elites or co-opted delegates. In other words, not only have authoritarian regimes long recruited technocratic elites into institutions, but this research highlights that such regimes can also be responsive to their suggestions and open to incorporating their expertise into policymaking. This marks an important contribution to previous literature that largely focused on responsiveness towards citizens but not expert opinions.⁷⁴

As such, the findings also relate to the current debate in the literature on whether information gathered through authoritarian institutions is, in fact, helpful for regimes, and China in particular.⁷⁵ It appears that claims that policymaking is being improved with the help of institutions and consultants may deserve more credit than they are currently given.⁷⁶ Policy reports rich with evidence such as statistical data provide a different kind of information than that gleaned from direct links with citizens, for instance, through petitions. Although a systematic link between governments’ written responses and policy outcomes and changes cannot be guaranteed and verified here, as the analysis shows, this kind of information may be particularly valuable for government departments of a more technical nature.

Finally, this study also draws attention to the particular role of PPCCs in Chinese politics. Although the institutions have been largely neglected, my research highlights that they are not mere “echo chambers” but instead can be promising research subjects, especially for researchers interested in policymaking processes. Findings from this study also suggest that it may be better to think of different institutions as complementing each other by collecting different categories of information. While some may be more useful in gauging citizens’ grievances and pre-empting collective action, other institutions may be more helpful in receiving expert opinions, as is the case in this study.⁷⁷

Although this study provides a rare opportunity to assess the government responses to different proposals and delegates in-depth, it should also be acknowledged that generalizations beyond the case in question should be made with caution, as is usual with studies of this kind.⁷⁸ While the internal processes of PPCCs have become fairly standardized across China, it cannot be fully known if government officials respond to proposals similarly in other provinces. The further analysis provided in the Appendix does not indicate that this could be the case; however, future research would be well advised to study the rationale of government officials handling the proposals and the resources of delegates to produce high-quality proposals in more detail qualitatively.