Regulatory capture of the Chinese social credit system: Bureaucratic self-interests in project implementation

Abstract

The construction of China’s social credit system (SCS) involves numerous bureaucratic agents who develop mechanisms aimed at contributing to the SCS’s overall objective of building mutual trust in society. This article traces the development of centrally designed censorship-related SCS mechanisms and examines them in a broader regulatory context. In doing so, we examine the effects that single SCS solutions produce for particular areas of governance, and the impact of bureaucratic interests on the proposed mechanisms and the entire SCS. The findings reveal that agents design mechanisms which support the underlying logic of their governance areas but which often diverge from the core SCS assumption of strengthening trustworthiness in society by raising the cost of violating legal provisions. This article also argues that, despite the trend towards centralization and controlling institutions, the ability of the Chinese party-state to develop a novel, complex, and coherent project may already be compromised at the central level due to the interplay of bureaucratic self-interests. Even though constructing the SCS has involved testing and introducing innovative mechanisms, these often fail to serve the project’s core assumption. As a result, long-existing mechanisms that now serve the SCS punitive regime may be more potent than the novel solutions in raising the cost of violating legal provisions.

Western observers initially regarded China’s social credit system (SCS) as an overwhelming tool that would strengthen Beijing’s ability to control citizens through the introduction of vast data collection and analysis, as well as quantitative scoring. ¹ This view has since changed among China watchers who acknowledge that, despite various local pilot mechanisms, the system is primarily focused on enforcing pre-existing laws. ² The SCS label encompasses various mechanisms that share a common goal: strengthening mutual trust in society by raising the trustworthiness of individuals, business entities, and other institutions. ³

The literature has focused on the technicalities of the SCS, the effects that its mechanisms produce, and public attitudes toward the system. It has devoted less attention to two interrelated aspects: the effects of individual SCS mechanisms on particular areas of governance; ⁴ and how bureaucratic agents mould mechanisms and translate the SCS framework to serve their self-interests. Despite scholars often mentioning power centralization and growing control over institutions under Xi, the Chinese party-state is not a monolith with consistent goals, such as the expansion of control. ⁵ Therefore, given the wide range of organs engaged in constructing and operating the SCS, looking at it through the prism of institutions is important for a better understanding of its work and impact.

This article also aims to contribute to the discussion on China’s bureaucracy. Scholarship on bureaucratic interests in the People’s Republic of China (PRC) often focuses on how local agents translate ambiguous central directives. ⁶ This article foregrounds central agents’ conduct in SCS implementation. Moreover, recent research indicates that central guidelines are, in fact, ‘a mixture of ambiguous and clear commands’. ⁷ The SCS mechanisms discussed here are responses to both clear and vague central directives. In sum, the research question navigating this article is: in what ways do bureaucratic self-interests affect the construction of the SCS? Bureaucratic self-interests are understood here as factors that influence particular actions of relevant commissions and ministries. These factors may include the agents’ need to pursue their ‘particular missions’, voice ‘the views of their units’, and push forward their own expertise. ⁸ Related to this point is the question of what the case of SCS construction reveals about the ability of the Chinese party-state to develop a complex and coherent project. SCS censorship-related mechanisms will serve as a case study. These cover both online and offline censorship – from crackdowns on illegal Internet content to dealing with teachers who spread inappropriate content in class. This choice was dictated by several factors. First, while belonging to the same category – efforts to control content – censorship-related mechanisms are a collection of SCS elements from different governance areas, and they produce varied results. Second, the SCS as a whole should not be considered to be a tool targeted primarily at individuals to achieve greater political and social control. However, the SCS remains a cog in an authoritarian machine: supporting law enforcement in the PRC naturally amplifies the authoritarian system under which the laws and regulations have been created. Censorship-related mechanisms imply that these laws include provisions that limit freedoms, for example, freedom of speech. Discussions on the SCS fail to explore content control in depth. Third, these mechanisms exemplify how agents may distort project implementation even at the central level.

The findings show that central bureaucratic agents mould their SCS mechanisms to serve their purposes and support the existing logic of content control activities in their areas. Often, the mechanisms hardly contribute to the development of a coherent project as envisioned by the organs leading the construction of the SCS. The two blacklists discussed in the article do not serve the core assumption of the SCS, that is, raising the level of trustworthiness in society through engaging multiple agents to develop and administer mechanisms that operate and share data under the SCS umbrella, hence raising the cost of violating legal provisions (with joint and reputational sanctions).

This article also argues that, despite the trend towards centralization and controlling institutions, the ability of the Chinese party-state to develop a complex, coherent project may already be compromised at the central level because of bureaucratic self-interests. The construction of the SCS has involved testing numerous new governance mechanisms; yet the more innovative mechanisms created under the SCS label may fail to contribute in concrete terms to the project. It is the long-existing, solid mechanisms that are used in a new manner, for instance, in the imposition of the SCS’s characteristic joint punishments, that serve its core assumptions. Resorting to older, established solutions may have been necessary for setting up a standardized SCS since fewer conflicting interests are involved than in the establishment of new mechanisms.

The analysis was based on documents available on the websites of Credit China and relevant organs, including central directives issued by organs leading the construction of the SCS, SCS-related documents issued by agents responsible for developing and implementing particular SCS mechanisms, and other regulations that provide the context for agents’ actions in SCS construction. This allowed for requirements and plans to be juxtaposed together with delivered results and the operational logic of particular governance areas, enabling the identification of possible self-interests that impacted the discussed SCS mechanisms. Three cases were identified among other (not censorship-related) mechanisms, and were chosen for detailed scrutiny.

This article begins with an introduction to the SCS. To provide the frames and context for the analysis, the article briefly discusses Chinese bureaucracy and describes Chinese content control efforts. The empirical section explores the chosen SCS mechanisms, after which follow the concluding remarks and discussion.

The SCS – a complex network of mechanisms

The term ‘SCS’ is used in various contexts. It is necessary to differentiate between various credit-related schemes, as well as to provide a discussion of the state-led SCS. The SCS discussed here is a state-led project which is unrelated to commercial credit-scoring systems, such as the Sesame Credit. ⁹ The state-led SCS basically comprises two parts. First, the financial credit reporting system, which emerged earlier; ¹⁰ and second, the trustworthiness system, which focuses on strengthening social governance and moral attitudes. This article pays particular attention to the latter. A point to keep in mind is that the trustworthiness system is not being implemented simultaneously across the entire country – it is introduced within local administrative units, such as cities. The particular mechanisms of the system are either executed in a top–down manner, with local organs implementing mechanisms designed by central-level agents (the focus of this article), or developed locally, with localities proposing their own mechanisms. ¹¹

The state-led SCS is a complex network of mechanisms ¹² that cover individuals (who are central to this article), legal persons, and other entities. The mechanisms appear in the key areas of governance, including the environment, labour rights, health care, culture, education, and the Internet. In essence, the SCS rewards the trustworthy and punishes the untrustworthy. Among the types of punishments are sanctions that stem directly from relevant legal regulations and blacklisting. Scholars have emphasized that the rewards mostly include reputational and moral incentives. ¹³ Trustworthiness is mostly determined by subjects’ behaviour in relation to legal provisions. The SCS mainly uses information produced or gathered by administrative organs, including data on administrative licences, professional qualifications, administrative punishments, and blacklisting. ¹⁴ Overall, the SCS does not sanction new categories of behaviour; the system is based on pre-existing Chinese regulations. It may, however, raise the costs of breaching legal provisions through the implementation of joint punishments and through public information disclosure that triggers additional consequences.

With its data disclosure procedures, the SCS enhances the imposition of not only formal punishments by state organs but also informal sanctions by society members. Scholarship has devoted considerable attention to the ramifications of creating and publicly disclosing blacklists of trust-breakers. For instance, Marianne von Blomberg and Haixu Yu investigated how blacklisting introduces ‘a strategy for regulatory shaming’, which enhances informal punishments that lack a legal framework. ¹⁵ Alexander Trauth-Goik and Liu Chuncheng focused on the ‘relational punishment’ imposed on blacklisted trust-breakers by other members of society and found that it is manifested ‘at the subtler microlevel’. ¹⁶

Joint punishments are introduced via various memoranda of understanding signed by state organs and entities. Once a leading organ blacklists a subject who has violated legal provisions, other organs – in accordance with their competencies – may impose additional sanctions that would normally not be imposed for such a violation under existing laws. Blacklists, therefore, serve two basic roles. One is public shaming and informal punishments, for example ostracization, imposed by non-state entities and individuals, and the other is imposing excessive sanctions.

The SCS is in the process of standardization, which takes place via central documents, such as the Central Punishment List (hereafter List). This document – first issued in 2021 – sets out the list, typology, and scope of sanctions that may be used within the SCS and to which the local systems are obliged to adhere (although localities may propose original mechanisms in locally published lists). While the List includes some contentious mechanisms, such as blacklists, it is filled with punishments that are embedded in laws, many of which existed before the current SCS framework was put in place. According to the List, punitive measures for breaking trust may be imposed if a decision has been made, in accordance with regulations, to employ an administrative punishment against a subject. ¹⁷

Apart from reputational and joint punishments, the SCS comprises other elements. These include trial local scoring systems which ‘introduce a new way of deepening datafication in identifying and classifying people’. ¹⁸ These systems are separate from the private scoring initiatives and belong to the state-led SCS, but are not centrally commissioned or standardized. Beijing has limited their impact by ordering that individuals must not be punished for having a low credit score. ¹⁹ Currently, the local scores are ‘benefit-oriented’ and often voluntary, resembling loyalty schemes. ²⁰

The construction of the SCS involves the establishment of an information-sharing infrastructure. Such shared use of data could enhance cooperation and coordination between different state organs, yet this potential may be hampered by bureaucratic agents following their own interests. As a result, the network of SCS mechanisms does not always have clear boundaries, and the operability of some of the mechanisms under the SCS label remains unclear. Vincent Brussee has pointed out that such vagueness is caused by ambiguous SCS objectives and various interests, and that the problems of fragmentation and lack of coordination are ‘fundamentally rooted in China’s bureaucracy’. ²¹ In his discussion, he primarily pointed to the discretion of localities. ²² This article takes up this bureaucratic angle and elaborates on how this may be observed in the case of centrally designed mechanisms before local circumstances even come into the picture.

Chinese bureaucracy and bureaucratic self-interests

Scholars have discussed Chinese bureaucracy from different perspectives. ²³ For over 30 years, they have referred to the model of ‘fragmented authoritarianism’ to explain how central policies are reshaped, mostly vertically (tiao 条) or spatially (kuai 块). ²⁴ Researchers have studied relationships between bureaucratic entities that cause policy implementation to be fragmented, often foregrounding the principal–agent problem. ²⁵ The central feature of the problem is the agent’s information advantage over the principal. This advantage results in agency discretion, where agents ‘pursue different ends and strategies than desired by the principal’. ²⁶ Discussions on agency discretion in China have leaned towards central–local dynamics, because local agents are often presented with general goals but ‘are given a great deal of latitude in choosing the means and personnel for achieving these targets’. ²⁷

Discretionary power is one of the phenomena triggered by China’s government administration structure. Another phenomenon is the ‘erosion and segmentation of the authority of the state’, stemming from the potentially conflicting legal provisions issued by various organs. ²⁸ Yet another phenomenon is bureaucratic agents’ use of these legislative capabilities to serve their self-interests and increase their power. ²⁹ The agents’ ability to use their legislative power is central to this case study. Actors’ interests in SCS construction are mapped primarily by looking at the regulations they issue. The discretionary power of agents requires more discussion because it appears in a somewhat unusual context here.

Analysing SCS construction within the principal–agent model is not a straightforward task. The construction is not restricted to the tiao and kuai axes or to the central horizontal dynamics. Moreover, the multitude of bureaucratic agents and the interrelations between them lead to various authority chains. Therefore, it is challenging to determine the universal SCS principal in the first place. The State Council issued the major planning document for SCS construction, making it the natural candidate for the principal. However, it does not have authority over the Chinese Communist Party (CCP) organs responsible for some SCS mechanisms. The State Council has transferred its SCS coordinating role to the National Development and Reform Commission and the People’s Bank of China, who themselves fail to execute power over all the participant agents in SCS construction. Moreover, the development of detailed mechanisms mostly lies with particular departments. Therefore, different organs, also below the central level, may be commissioned to act as additional principals; this is a normal occurrence in China. ³⁰ Such principals may appear either on the tiao axis or within the spatial dynamics (kuai): a local agent can act either under a higher-level ministry or under its local government, depending on whether the SCS mechanism in question is a local or a top–down one. These organs, however, only act as principals in the construction of individual mechanisms. They are not principals in the entire SCS.

Without a universal principal, the agents’ informational advantage over the principal may not be the primary reason for discretion in the implementation of SCS mechanisms. Discretion at the central level – where no additional principals are involved – may stem from the lack of any principal, or from an unclear authority chain. This environment may enhance the central agents’ focus on serving their own interests rather than contributing to the construction of a cohesive SCS. But before mapping these interests, we must first discuss the nature of Chinese censorship.

Censorship and control in China

The literature points out that China does not exert full authoritarian control over its citizens. Its control is ‘more fragmented and heterogeneous’. ³¹ However, Beijing’s increasing focus on stability, security, and the expansion of its domestic security apparatus ³² might suggest that new, controversial governance tools, such as the SCS, would tighten control. Therefore, the functioning of the chosen SCS mechanisms should be juxtaposed against the regular logic of PRC censorship and control.

Censorship in China operates within a ‘strong but fragmented system’. ³³ Provisions on content and behaviour control in various governance areas are often diluted by non-censorship-related legal articles in regulations that serve broader purposes. In the area of education, for instance, the idea of controlling and maintaining ideological quality ³⁴ is found among other educational norms, including bans on bribery and corporal punishment in schools. Securing the correct ideological and political direction itself requires different measures (from overseeing teachers’ behaviour to designing curricula and textbooks), which are found in various documents. In the area of culture, provisions banning particular content appear in the same regulations that, for instance, protect intellectual property rights. Moreover, whatever the subject of regulation, be it print publications or movies, the lists of illegal content seem standardized. As for online censorship (also fragmented into various measures), the regulations concern not only banning information but also guiding news and public opinion in the desired direction. ³⁵ Importantly, regulations often order Internet content providers – instead of state organs – to fulfil censorship tasks themselves under the threat of penalties or closure. ³⁶

Information control in China is intensified and loosened at different times in response to ‘symbolic events and policy and leadership changes’ ³⁷ or in accordance with ‘the level of underlying social tensions’. ³⁸ The motivation for enforcing censorship goes beyond silencing criticism towards authorities; it includes silencing any collective actions because they may jeopardize regime stability. ³⁹ Overall, during targeted crackdowns, particular provisions that are not always enforced are executed. In other words, despite the legal institutionalization of censorship, enforcement depends on circumstances. Such balancing is characteristic of various types of repression in China, for example, in labour organizations. ⁴⁰

SCS censorship-related mechanisms: The case study

The mechanisms chosen for analysis in this article are two blacklists and a group of sanctions that execute pre-existing provisions. This section determines their operability and the extent to which they belong to the scope of the SCS. It also maps how bureaucratic self-interests influence the functioning of the mechanisms and discusses how they affect the PRC’s efforts to control content.

The Internet Credit Blacklist

The Internet Credit Blacklist is among 41 blacklists in the Central Punishment List. ⁴¹ This blacklist was announced by the State Council in 2014. It includes enterprises and individuals whose actions on the Internet are considered serious trust-breaking behaviour, one of which is starting and disseminating rumours. The blacklisted enterprises and individuals are subjected to public exposure and restrictions on online activity. The Cyberspace Administration of China, the Ministry of Public Security, and the Ministry of Industry and Information Technology are responsible for this blacklist. ⁴²

Despite the clear directive obliging organs to impose the sanctions from the Central Punishment List, attempts to find this Internet Credit Blacklist on the Credit China websites (central and local) and relevant departmental websites were to no avail. A search with phrases containing the words ‘Internet Credit Blacklist’ and ‘breaking trust on the Internet’ using the Baidu search engine produced similar results. At the same time, the Cyberspace Administration of China – a major actor in Internet regulations and censorship – already limits online activity for sharing improper content. People are targeted individually ⁴³ or during crackdowns. ⁴⁴ The Cyberspace Administration, however, likely has no major interest in meting out the punishment within the scope of the SCS.

The Cyberspace Administration chooses when to enforce online censorship provisions and when to publicly reveal its actions. These provisions are found in various regulations that prohibit certain types of online content. For instance, according to the ‘Management Regulations for Online Audio and Video Information Services’, websites, online platforms, and other channels cannot be used for actions that could endanger political security and social stability. The Regulations ban actions that are considered illegal by the Cybersecurity Law. ⁴⁵ These include using the Internet to jeopardize national security, incite subversion of state power, or overthrow the socialist system. ⁴⁶ Similar content-limiting provisions are found elsewhere, for example, in regulations related to the use of deep synthesis technology ⁴⁷ or governance of the ecosystem of online information content. ⁴⁸ Violating the provisions may lead to punishments by relevant departments in accordance with their duties, regulations, and laws. ⁴⁹ Under the SCS, the consequences for violating relevant content-related provisions that are stipulated in the various documents could include blacklisting.

Viewing these regulations, we see that numerous bureaucratic agents are involved. The Management Regulations were promulgated jointly by the Cyberspace Administration, the Ministry of Culture and Tourism, and the State Administration of Radio and Television. Regulations on deep synthesis technology were issued by the Cyberspace Administration, the Ministry of Public Security, and the Ministry of Industry and Information Technology. The provisions of the online information content ecosystem were compiled by the Cyberspace Administration. As the sole agent appearing in all of the documents, the Cyberspace Administration has the leading role. Incorporating the mechanism that limits online activity to the scope of the SCS could limit the discretion of the Cyberspace Administration in sanctioning those who violate the provisions. The Cyberspace Administration would share responsibilities with the Ministry of Industry and the Ministry of Public Security, ⁵⁰ which already participate in the issuance of content-related provisions, albeit not all of them. If the Internet Credit Blacklist was implemented, then the two ministries could potentially gain access to relevant online activity data and trigger the mechanism in accordance with their own interests and outside the Cyberspace Administration’s authority. However, the ministries may be unable to activate the blacklist mechanism because, usually, only one leading organ is responsible for creating a blacklist. Apart from authority issues, the blacklist could also generate coordination problems for the Cyberspace Administration since other already-mentioned actors participating in online content control are left out of the SCS mechanism.

Moreover, data sharing via the National Credit Information Sharing Platform (the main platform for information sharing under the SCS) would lead to the shared use of information with a wider variety of ministries. This is the objective of the National Development and Reform Commission, which seeks to interconnect Internet credit information for the purpose of achieving online and offline multi-area shared use. ⁵¹ This would enable the introduction not only of joint sanctions but also reputational sanctions via the disclosure of the blacklist. Even with the Cyberspace Administration retaining its discretion over whom to include on the blacklist before sharing the data on untrustworthy individuals and enterprises via the Information Sharing Platform, this data sharing would limit its flexibility in deciding the intensity of sanctions. Raising the cost of violating relevant provisions through the imposition of additional sanctions by other actors could act against the logic of adjusting control and punishments to circumstances and it might prove to be poorly balanced. As the major Internet watchdog, the Cyberspace Administration could ultimately be blamed for poorer performance. Importantly, the Cyberspace Administration’s decision not to use the SCS mechanism does not necessarily mean that it is reluctant to raise trustworthiness on the Internet. Rather, it implies that the Administration believes that this goal can be better achieved without the SCS.

Despite its limited interest in implementing the Internet Credit Blacklist, the Cyberspace Administration has shown some devotion to SCS construction. Although various documents mention constructing a credit system ⁵² or establishing a credit management system that would determine access to online services, ⁵³ they fail to present details. Most importantly, in July 2019, the Cyberspace Administration published Draft Measures introducing a blacklist for seriously trust-breaking subjects in the area of Internet information service. ⁵⁴ One goal of the blacklist was to establish credit in the area of Internet information service, in accordance with, among others, the PRC Cybersecurity Law, ⁵⁵ which mentions notions of honesty and credit. Another goal was to record illegal behavior that violated the Cybersecurity Law in credit files. ⁵⁶ The Draft Measures were relatively specific about the types of untrustworthy behavior that ought to be punished; these included publishing or disseminating information that violates social morality, honesty, and credit, or activities that seriously disrupt order in cyberspace or harm the public interest and the lawful rights and interests of others, creating an abominable social influence. ⁵⁷ According to the Draft Measures, the blacklist would be valid for three years. ⁵⁸

Under the Draft Measures, the Cyberspace Administration would use data on untrustworthy netizens provided by Internet information service providers and verify whether the requirements for inclusion on the blacklist or focus list are met before the lists are compiled. ⁵⁹ The blacklist would either be determined and published by the central Cyberspace Administration or determined by a provincial cyberspace administration and published by the central organ. ⁶⁰ The document, therefore, overlooked the roles of the Ministry of Industry and the Ministry of Public Security. Retaining full authority over the blacklist mechanism diverged from the clear directive to operate the mechanism together with the two ministries. The document never came into effect, and drafting it was likely a move to superficially demonstrate efforts to fulfil orders from above.

The Cyberspace Administration’s lack of interest in executing the Internet Credit Blacklist has not been successfully tackled by the top authorities, possibly due to the lack of a proper authority chain. The basis for the blacklist was the 2014 Outline, which was issued by the State Council, and it is a policy basis of a general nature. The earlier-mentioned Internet regulations were fully overlooked. The State Council should, therefore, be the principal since it ordered the implementation. It could be supported by the National Development and Reform Commission, which reaffirmed the task in a document complementary to the Outline. ⁶¹ However, neither the State Council nor the Commission has authority over the Cyberspace Administration which is responsible before the Central Committee of the Communist Party of China.

In terms of censorship and control, the blacklist would mostly be a re-labelling of a mechanism that already exists. It could, however, raise the cost of breaching content-related provisions through blacklist disclosure and the imposition of joint punishments. This, in turn, would make it more challenging for the Cyberspace Administration to control the precise consequences of punishments. This could be against the underlying logic of PRC censorship actions, which are characterized by weighing who, when, and how to punish. While showing signs of commitment to the SCS project, the Cyberspace Administration avoids real implementation. This shows how bureaucratic self-interests might hamper the introduction of a project at the central level. The next blacklist in our discussion throws light on another interests-induced phenomenon.

Blacklisting employees of off-campus tutoring institutions ⁶²

In 2021, the Ministry of Education and the Ministry of Human Resources and Social Security issued trial measures for the management of employees of off-campus tutoring institutions. ⁶³ These measures stipulate that among the requirements to which employees have to conform are a comprehensive implementation of the CCP’s education policy, a very good ideological and moral character, and professional ethics. ⁶⁴ Such an ideological focus is regularly found in education-related documents.

What distinguishes the Trial Measures from other documents is the Blacklist of Employees of Off-Campus Tutoring Institutions (hereafter Blacklist of Employees), which bars individuals from being employed in an off-campus tutoring institution. ⁶⁵ The Trial Measures state explicitly that statements and actions harming the CCP’s authority or violating the party’s line, guidelines, or policies can lead to blacklisting. Publishing and transmitting erroneous viewpoints can result in the same punishment. The prohibitions also include conduct unrelated to censorship, such as meting out corporal punishment on students or demanding improper benefits from students or parents. ⁶⁶

There is no doubt about the execution of this blacklist. China Daily’s coverage confirmed that, following three-month-long investigations at the turn of 2021–22 carried out nationwide by local education departments and covering nearly 770,000 employees, 64 individuals were identified and included on the locally created blacklists. ⁶⁷ However, whether the Blacklist of Employees falls within the scope of the SCS remains unclear because it seems to operate within two separate projects. Employees are not blacklisted as a direct result of breaching provisions but in the aftermath of an investigation, which itself is not limited to the supervision of their work. For instance, investigations also include the verification of the content of teaching materials. ⁶⁸ Such examinations are part of a policy aimed at limiting the workload put on students and a step towards the implementation of the ‘Double Reduction Policy’. This policy, implemented in 2021, aims to reduce the pressure on the youth by curtailing the amount of homework and limiting the work of off-campus tutoring institutions.

This case reveals how bureaucratic agents created one mechanism that works for two separate projects. ⁶⁹ That the blacklist exists within the Double Reduction Policy is certain; the simultaneous affiliation of this blacklist with the SCS is less explicit, but three facts speak for it. First, blacklists have generally become characteristic elements of the SCS. Second, there is a similar blacklist (called the Blacklist of Off-Campus Tutoring Institutions – institutions, not persons), which is on all editions of the Central Punishment List. Finally, the Trial Measures are found among SCS regulations on a Credit China website, and Credit China reported the introduction of these measures. ⁷⁰ This indicates that the Ministry of Education and the Ministry of Human Resources likely reported them as their input into the SCS framework. In this manner, the ministries can carry out two separate tasks in one go and, possibly, obtain double funding for the implementation of one product – from the Double Reduction Policy and SCS budgets.

However, one question is why the National Development and Reform Commission included the Blacklist of Off-Campus Tutoring Institutions in the Central Punishment List but failed to do the same with the employees’ blacklist. There might be several reasons for the omission. First, introducing measures targeted at individuals may be politically risky. Blacklisting individuals and disclosing the lists could raise privacy concerns and shaming-related outcries, especially amidst emerging privacy regulations, such as the Personal Information Protection Law. ⁷¹ Therefore, since the Central List is an easily available document, the Commission may find it beneficial to exclude mechanisms targeted at individuals from this list. Should this blacklist be contested, it would remain the responsibility of the Ministry of Education and the Ministry of Human Resources. At the local level, the local agents are to blame, so the National Commission avoids responsibility. Second, the introduction of the blacklist may clash with a central directive to limit the number of blacklists. Third, the Trial Measures are found among thousands of SCS documents, and the absence of the blacklist in question from the Central Punishment List may be a result of simple omission or oversight. Also, this is a trial document, which may well simply disqualify it from inclusion on the Central List.

The exclusion of the said blacklist from the Central Punishment List, and the overlapping of projects, give rise to two problems. The first is the ambiguous scope of the SCS; the second is the inability to determine the principal and the corresponding project when assessing the operations of the mechanism. In other words, while the SCS (a means for dealing with trust-breakers) and the Double Reduction Policy (which is actually more of a policy goal aimed at limiting the work of off-campus tutoring institutions as well as the workload put on students) are not mutually exclusive and may be complementary, it is unclear whether the said blacklist should mainly improve the trustworthiness of employees or limit the pressure on students.

In terms of control, the blacklist under discussion institutionalizes a seemingly new punitive mechanism to reprimand teachers acting against the CCP line. Notwithstanding this, sharing inappropriate content in class already triggers actions outside the SCS, not only in the off-campus tutoring institutions but also in the public education sector. ⁷² The blacklist may be a new way of identifying untrustworthy teachers, but it fails to reinforce the SCS’s characteristic features. The Trial Measures do impose a ‘single’ punishment (a ban on teaching), but they do not entail additional consequences – joint punishments or public data disclosure. The blacklist is not available on the websites of Credit China or the two responsible ministries. The lack of transparency is against central SCS rules that advocate the public disclosure of information related to instances of trust-breaking, an example of which is blacklisting. ⁷³ On the other hand, disclosing credit information on individuals requires a law, regulation, State Council decision, or consent from the individual. ⁷⁴ Thus, it is possible that the lists of individuals are not publicly exposed because of the lack of relevant provisions. All this questions the reason for the existence of the blacklist under the SCS, besides the already-discussed interests of the Ministry of Education and the Ministry of Human Resources. Finally, the blacklist makes no significant alterations to the underlying logic of the governance area of education, because it seems to guard various types of educational norms at the same time – from proper ideological attitude to the ban on obtaining illegal benefits. In other words, the blacklist does not foreground censorship since it also punishes those who violate different education-related norms – albeit not in the manner characteristic of the SCS.

The Blacklist of Employees illustrates how agents deliver a mechanism that superficially carries out a task commissioned in the 2014 SCS Outline to raise the trustworthiness of teachers. The mechanism, however, operates under two labels, and it diverges from the already-mentioned core SCS assumption because it does not raise the cost of violating legal provisions with data disclosure or joint punishments. The last section discusses pre-existing punishments that are now used within the scope of the SCS.

Pre-existing punishments within the SCS scope

The Central Punishment List delineates the sanctions that may be lawfully imposed under the SCS umbrella. Most of the 14 punishment categories on the List include pre-existing administrative punishments. The sanctions were analysed to track censorship-related provisions applicable to individuals. This section covers the punishments for wrongful behaviour in the field of culture. According to the general guidelines, one of the responsibilities of the SCS is to regulate the culture and entertainment market and manage various cultural products. ⁷⁵

The Central Punishment List includes sanctions that restrict individuals from working as senior managers and legally authorized representatives in particular cultural sectors if they violate content-related provisions. ⁷⁶ Untrustworthy individuals with individually owned companies may also be barred from engaging in business activities in a particular sector. ⁷⁷ The length of the ban varies, covering periods of 5 ⁷⁸ or 10 years. ⁷⁹ These sanctions, however, are not triggered immediately by content-related violations. They follow a business entity license withdrawal, which itself is an administrative punishment for serious violations of legal provisions that prohibit specific types of behaviour. Among these provisions are lists of prohibited content; these are almost identical for the movie industry, the publishing industry, audiovisual products, and venues providing Internet access services. The prohibitions include content violating the basic principles stipulated in the Constitution (which include the leadership of the CCP); jeopardizing national unity and security; jeopardizing the national interest and public morality; and disrupting social order and stability. ⁸⁰

These sanctions were introduced by laws and regulations that do not mention and are not related to the SCS. Long-established, with a strong legal basis, these punishments do not generate problems such as overlapping or unknown authority. Even though including them in the scope of the SCS may at first glance seem purposeless in terms of introducing novelty to the PRC’s governance, these punishments, in fact, serve the core assumption of the SCS. Their imposition may now be a part of the SCS joint punishments scheme, ⁸¹ which aims to raise the costs of violating the law.

Compared to the two blacklists, the mechanisms that use pre-existing punishments trigger fewer interests-driven problems. The agents include the ministry responsible for the movie industry (in charge of individuals’ self-registered companies in that industry), the Ministry of Culture and Tourism (responsible for venues providing Internet access services), and the State Administration for Market Regulation (which deals with senior managers and representatives in the following sectors: movies, audiovisual products, and publishing). These agents carry out their tasks set out in the general guidelines, but they do so in the most secure manner possible and without controversy because the mechanisms are embedded in laws. Such a strategy for participating in the development of the SCS may mirror the agents’ unwillingness to create central mechanisms that would go far beyond basic law enforcement. Should novel centrally designed mechanisms be poorly received, there would be no local agent to blame. Therefore, attempts to introduce bolder mechanisms may be left to localities. With this attitude, central agents may also avoid public controversy because the subjects against whom SCS punishments – based on pre-existing sanctions – are used may, in fact, not even be aware that they are being sanctioned under the SCS. In brief, the organs responsible for these sanctions carry out their tasks without generating risk or workload for their agencies. The State Council and National Development and Reform Commission may accept them as relatively simple law-based mechanisms, which take into consideration the characteristics of the SCS and do not cause major challenges for the central government.

In terms of censorship, these mechanisms follow the previously established governance logic in the cultural sector. At the same time, they are supposed to be used ‘collectively’ under the SCS; that is, as joint punishments, and to sanction more severely those who seriously violate censorship-related provisions. This shows that, by strengthening the enforcement of legal measures that are, for instance, targeted at potentially dissident activities, the SCS will naturally amplify the authoritarian nature of the Chinese regime.

Conclusion and discussion

Bureaucratic agents responsible for policy and project implementation in China follow their own interests, which determine their actions. Scholarship has acknowledged this in various contexts, including censorship efforts ⁸² or even SCS implementation. ⁸³ Researchers tend to focus on the local manifestations of interests shaping bureaucratic agents’ actions and policy outcomes. This article has looked at bureaucratic self-interests in the development of SCS content control mechanisms, and found instances whereby it was the discretion of central – not local – agents at work.

It is not possible to identify all existing interests; especially officials’ personal motives, which affect their actions. However, by tracing the evolution or by examining the nature of the chosen SCS mechanisms and placing them in a broader context, we can set out the basic rationales behind the divergence of the mechanisms from the core assumption of the SCS. In the examined cases, the various interests induce three phenomena affecting the construction of a coherent SCS.

The first is the lack of intention to implement a mechanism despite a clear order to do so. In the case of the Internet Credit Blacklist, this was likely due to the reluctance of the Cyberspace Administration of China to lose control over the consequences of punishments imposed by additional actors. The failure to motivate the agent to implement the task is attributable to the absence of a direct line of authority between the SCS coordinators and the Cyberspace Administration. The second phenomenon is the introduction of one mechanism under two separate, overlapping labels. The Blacklist of Employees seemingly operates under both the SCS and the Double Reduction Policy, enabling the agents to achieve two tasks at once and, perhaps, obtain double funding. These labels can technically coexist in parallel. However, this coexistence seems to be inefficient for serving the core assumption of the SCS. Even though blacklisting itself seems characteristic of the SCS, the mechanism does not lead to data disclosure or – relatedly – informal punishments, nor does it trigger joint punishments. ⁸⁴ The overlapping, or coexistence, may also pose difficulties in determining who oversees the operations or who is authorized to order alterations. The third phenomenon involves central agents choosing the most secure and least labor-intensive manner of contributing to the construction of the SCS by incorporating pre-existing sanctions into its scope. These are the simplest mechanisms that do not reveal conflicting interests between the implementing agents and SCS coordinators, and diverge the least from the core assumption of the project.

These interests-driven phenomena have a common denominator: the outcomes of the agents’ actions support the underlying logic of their governance areas. In the case of online censorship, the Cyberspace Administration retains its flexibility in adjusting punishments. The Blacklist of Employees still introduces the censorship provisions among a broader set of education-related rules. The pre-existing punishments that may be used under the SCS do not reshape censorship-related provisions. They may be used, however, to intensify punishment for violating long-established provisions. Importantly, as previous literature has discussed, the SCS is about law enforcement, ⁸⁵ and the mechanisms discussed here do serve this goal. ⁸⁶ However, looking at how the enforcement takes place reveals that the mechanisms do not always serve the entire SCS network.

This case study is a modest example of a considerable challenge that the Chinese leadership faces in developing a complex, coherent project: participating agents must simultaneously fulfil their general functions and meet the goals assigned to them. However, the agents’ main objective is not necessarily the development of a coherent SCS that would foster a trustworthy society as envisioned by the State Council or the National Development and Reform Commission. Sometimes, however, a mechanism meets the requirements of SCS designers and does not clash with agents’ interests. For example, the use of pre-existing punishments within the scope of the SCS does not dramatically alter agents’ operations. This appeals to those who aspire to introduce law-based mechanisms and constitutes a solid base for SCS joint punishments.

The effect is, however, that an extensive project such as the SCS, which involves numerous and often innovative pilot mechanisms, has become, in its most functional and standardized parts, only a moderately novel instrument. Where an innovation or new trend appears, the mechanisms can turn out to be less controllable and not necessarily beneficial for the entire project, as the two blacklists illustrate. This article has also exemplified how the (rather universal) principal–agent model ⁸⁷ may not always reflect various phenomena within policy implementation processes. The discretionary power of the agents who introduced the two discussed blacklists is primarily a result of the lack of clarity over who the principal is and which project is the main reference for the assessment of the delivered results.

Finally, the biggest limitation of this study and its findings is the narrow focus on a small case of censorship measures. Similar analyses are needed in more governance areas where the SCS is present. Other interests-driven phenomena may occur beyond the domains of the Internet, education, and culture.