Skip to main content

Central State vs. Local Levels of Government: Understanding News Media Censorship in China

Abstract

This study investigates the most frequently censored news in China. Existing studies show that the Chinese propaganda authorities are more likely to censor news considered harmful to the legitimate rule of the central state, yet allow news with negative consequences for local levels of government to be published. A content analysis of official propaganda notices (n = 728) reveals that the propaganda authorities, indeed, engage in selective news censorship. The selective censorship practice also reveals a structural difference: The central propaganda authority focuses more on news related to national guidelines and policies and the image of the central state and leadership, whereas local propaganda authorities tend to focus more on news that is harmful to social stability and the image of local government. The central authority is found to be more tolerant than its local counterparts, as it allows news media to report a considerable amount of news that have negative consequences for local levels of government—news which, on the other hand, is heavily censored by the local propaganda authorities.

Introduction

All political institutions want the news media to work as favorably for them as possible. In authoritarian states, with “the press being a servant for the state responsible for much of its content to the power figures in charge of government at any given moment” (Siebert et al. 1963), the authorities are able to censor whatever they want. Nevertheless, they do not censor everything. In authoritarian countries with more vibrant media, the rulers censor only very negative news and allow modestly negative news to be published (Shadmehr and Bernhardt 2012).

When compared to many other authoritarian states, the Chinese news media are not particularly free. A theory developed by Egorov et al. (2009), stating that dictators in oil resource-poor countries usually allow freer media, fails to explain the case of China: an oil resource-poor country which, nevertheless, controls its media tightly. Does this suggest that all negative news about the state are censored in China? There have been no studies looking directly at the news censorship program in China. Yet, research on China’s Internet censorship program shows that Internet censors do not censor all negative content, but rather only news which represents, reinforces, or spurs social mobilization (see, e.g., King et al. 2013). For example, censors usually do not delete Internet content criticizing central leadership (ibid). However, findings from research on the Internet censorship may not be directly applicable to news media censorship. Control over news media is considered to be tighter due to the stronger and wider influence of press information compared to the Internet posts. Moreover, research has shown that an increase in uncontrollable information on the Internet may lead to reduced media freedom (Lorentzen 2014).

While it is well known that the Chinese propaganda authorities censor media, there have been no studies investigating which types of news they are most likely to censor. This article will investigate this very question.

Media Censorship in China

The centralized state control of the news media is an aspect of authoritarianism (Friedrich and Brzezinski 1956). The basic objective of an authoritarian state’s media censorship is to serve its ruling capacity. There have been many examples, indicating that the downfall of authoritarian regimes is associated with the state’s loss of control over the media (Lawson 2002; Nye 2004; McMillan and Zoido 2004). Despite their capacity to do so, however, rulers do not censor every piece of news negative to the state and its leaders. In a political economy perspective, rulers stop censoring news when the direct costs of censorship are higher than the gains as far as the reduced probability of citizen revolt is concerned (Shadmehr and Bernhardt 2012). Moreover, when rulers are uncertain about the possibility of news leading to revolt, they stop censoring marginally bad news to prevent citizens from making inferences that the news could have been far worse than it is presented in the media (ibid).

This is especially the case as the news media become more commercialized. For example, in China, most news organizations became commercialized in the 1980s during the state-sponsored marketization reform. The reform allowed the existing party news organizations to establish market-targeting sub-outlets, i.e., nonparty outlets, to subsidize their operations. These operations were fully supported by the state finance before. As more nonparty news organizations began to expand the reporting boundaries by reporting on negative news about the government to meet market demand (Kuang 2017), it became impossible for the Chinese propaganda authorities to censor every piece of modestly sensitive news. In addition, to censor every piece of sensitive news in this expanding market would lead to vocal opposition from Chinese citizens. Internet news more specifically can disseminate at a viral rate and make its impact before censors can take action. When censoring every piece of negative or sensitive news is neither necessary nor possible, a pressing question becomes: Which types of “bad” news are censored by an authoritarian regime and which are not?

Studies on the broader topic of media and communication control strategies in China (see, e.g., Brady 2008; Du 2010; Liu 2013) provide some indication. They suggest that all media content harmful to the legitimate ruling of the communist state is censored. To protect its rule, according to Weatherley (2006), the Chinese Communist Party (CCP) has identified a few priorities including the sustaining of economic growth, nationalism, social stability, and (to a less extent) rational legal authority and electoral legitimacy. This is also what sustains the current political regime.

Thus, the news most likely to be censored is news that the propaganda authorities believe will have a negative effect on the legitimate rule of the communist party state. All Chinese news media must contribute to the enhancement of party ruling; if the priorities of the party are updated, media control strategies are adjusted accordingly (Du 2010). Indeed, what was prioritized in governance, including economic growth, nationalism, and social stability, was reflected in the media control strategies (Zhong 1996; Brady 2008; King et al. 2013). In return, these strategies have bolstered the ruling capacity of the regime (Stockmann and Gallagher 2011).

The idea that the state mainly censors news that possibly questions the legitimate ruling of the state also means that not all negative news issues are being censored. Some negative news items about local officials, for example, may not be censored, as they are not considered harmful to state ruling. In fact, such news may, indeed, boost the image of the central state rather than damage it. Research has demonstrated that Chinese citizens have high political trust in the central government, but local levels of government do not enjoy the same degree of trust (Li 2004).

Negative news about local officials is more often seen in national than local news outlets, the latter being controlled by the local officials who have allegedly committed wrongdoings. This indicates that the focus of news censorship at the central state level is different from the local levels of government, including the provincial, municipal, district/county, and township levels.

Thus, news censorship in China cannot be fully understood if we ignore the structural difference of media control practices in China. There are news media at all of the politico-administrative levels, starting at the township level and upwards to the central level. Moreover, propaganda authorities control news media located at the same and at lower levels of government. They do so by sending the media and lower level propaganda authorities propaganda notices (Fig. 1). Yet, news that is considered extremely sensitive to the municipal leadership may not be seen to be so by the higher level authorities. Lower level propaganda authorities can submit an application to higher level authorities for propaganda notices that prohibit the spread of sensitive news about the local levels of government in higher level news media.

Fig. 1
figure1

China’s news censorship program at the central state and local levels. Is based on interviews with anonymous propaganda officers and news editors on the working of the news censorship program in China

The puzzles then become: 1. Which type(s) of news do the propaganda authority at the central level of government censor more frequently (Research Question 1) and 2. Which type(s) of news do the propaganda authorities at the local levels of government including Provincial, Municipal, County/District, and Township levels censor more frequently (Research Question 2)?

News Censorship at the Central State Level

As discussed earlier, the focus of news censorship at the central state level of government is very likely some news considered harmful to the legitimate ruling of the state (Weatherley 2006). To maintain its rule, the central leadership has a number of priorities in news censorship. In practice, the more frequently censored news issues could be those related to social stability, e.g., protest, revolt, and violence, and others which are not in line with national guidelines and policies on economic growth, nationalism, rational legal authority, electoral legitimacy, etc.

First of all, censoring news related to social stability is a top priority (Dowell 2006; King et al. 2013). Since the 1989 crisis, successive leaders have had Deng Xiaoping’s dictum of “stressing stability above all else” as their mantra (Brady 2008, 95). The state has increased spending on Internal security, which indicates that the central state is nervous about escalating public unrest (Hook 2011). This nervousness is also reflected in news censorship as the state has promoted the rhetoric of stability as its predominant theme since the 1990s (Zhang and Cameron 2003). The Central Propaganda Department has been strictly instructing editors to control or suppress any stories possibly contributing to instability in China (Brady 2008, 97).

Besides social stability, central leaders have also instructed propaganda authorities to keep issues that undermine the enforcement of the party’s guidelines and policies off the news agenda (Jiang 1989). Such topics include socio-economic issues, such as economic growth, the Chinese income gap, inflation and price increases, and political issues, such as those related to nationalism, e.g., Taiwan/Tibet independence, anti-Japanese movement, and others including opinions of political dissidents and international opposition, defiance of the one-child policy (Brady 2008, 96–104), and criticisms on the legal and electoral systems.

While Internet censorship in China mainly focused on issues spurring collective action and left out those criticizing the government and central leaders (King et al. 2013), studies on news censorship (e.g., Zhang and Cameron 2003) found that news that may be detrimental to the image of the central state and leadership on both the national and international levels is also suppressed. This is because information revealed in the news is usually considered authoritative and official as all news organizations in China are state owned. As Kotler (2001) defined image as “the set of beliefs, ideas, and impression that a person holds regarding an object”, we consider that any negative beliefs, ideas, and impression of the Chinese public holds regarding the central government and leadership in the news media would be devastating to the image of the Communist Party.

Another priority for the central state to maintain its rule is to monitor the local cadres. Lorentzen (2014) has found that the central leadership permits some investigative reports on lower level officialdom to help them improve governance. The central leadership benefits from an active watchdog that keeps them informed of wrongdoings among local officials (Zhao 2000; Shirk 2011; Liebman 2011). This allows central leaders to reduce public discontent that could possibly contribute to unrest despite the absence of an opposition party (Lorentzen 2014). Like the leaders in many authoritarian states, the central CCP leaders face the “Gorbachev dilemma”: while they have to reform the bureaucratic system to be perceived as legitimate rulers, reform is impossible without a freer media to monitor the process, which in return could undermine the foundations of the party dictatorship (Egorov et al. 2009). Part of the reform is to ensure that local government corruption is kept to a minimum while at the same time providing an incentive system for low-tier officials to move upwards to become qualified successors. Merit-based promotion is one of the developments of the CCP in this direction.

Consequently, the central leadership must be careful when using the news media to cleanse the political system. Too much negative reporting about local officials may lead to the assumption that corruption is widespread in the political system, including the central level. The central state must, therefore, balance the reporting of news media with the level of social tension. By maintaining a level of discontent that is constant, the regime can monitor and control the public’s willingness to challenge its authority (Lorentzen 2014).

News Censorship at Local Levels of Government

The fact that China is a decentralized authoritarian state has implications for news censorship at the local politico-administrative levels. Decentralization in the Chinese economic reform has significantly reshaped the power distribution and political relations within the party-state political system (Wu, 2000, 47). Today, local levels of government have much more autonomy than earlier in the policy-making process at their own level. This seems to have somewhat broken political homogeneity and loosened central dominance (Goodman 1992; Hao and Lin 1994; Naughton 1995). Yet, the central state has the instruments to control the local cadres by strengthening and institutionalizing mechanisms of administration and organization (Landry 2008, 12). The CCP personnel management system is believed to be an important explanation as to why China remains a highly authoritarian state despite the country’s profound decentralization reforms. The promotion mechanisms for local cadres have allowed the CCP to decentralize local development without weakening the party’s political control in that it is up to the CCP to reward officials for developing their localities (ibid, 16). As in most Leninist systems, the Chinese central party leadership has a monopoly on appointing officials at all levels of the bureaucracy (Harding 1981). In fact, the capacity of the central state to monitor and control lower level agents by promoting successful local leaders to hold concurrent positions at higher levels has been increasing (Edin 2003).

The merit-based promotion incentive program can be expected to lead local leaders to tighten control on issues that may undermine their performance. They seem to have complete control over the news media within their “territory” of governance. This is because the media system has been substantially decentralized alongside the decentralization of the economy. The news media have become platforms for provincial and local authorities to express their own views (Wu 2000, 47). Decisions concerning news media editors at the provincial and local levels are in fact made by party and state leaders who appoint the media managers at their own levels of government (Esarey 2006). Dowell (2006, 115) points out that a driving factor for local leaders in their attempts to control the media is their personal survival in the political system. Xiong (2008) also finds that local officials use the news censorship program at the local level to keep negative news unveiled, so that the image of the local government and their performance are not affected. For instance, during the early stages of the Severe Acute Respiratory Syndrome (SARS) epidemic, local officials hoped to cover up the seriousness of the outbreak to protect their careers (Zhang and Flemming 2005; Dowell 2006).

With the promotion program, another concern of the provincial and local cadres is to engage in a “court politics”; that is, to seek support from the central leaders for a promising political career by accepting their orders and pandering to their preferences (Teiwes 1995). The news censorship program at the provincial and local levels is also expected to reflect this concern, meaning that most news issues sensitive to the central leadership would also be censored by the local leadership. Among all of the issues considered essential to central state legitimacy, including those related to social stability, national guidelines and policies, and the image of central state and leadership, the type of issue that is also expected to be highly sensitive to local government performance is social stability. This is because most social unrest originates from the struggles of the citizens against the local governments and/or officials.

Hypotheses

The review of extant research indicates that the different politico-administrative levels in China diverge in their news censorship focus, i.e., the central state focuses more on news related to social stability, national guidelines and policies, and the image of the central state and leadership as such, while the local levels of government focus more on the image of the local government. Three hypotheses corresponding to the first research question (RQ1) looking into the types of news most likely censored by the central propaganda authority are formulated as follows.

H1a

The central propaganda authority censors more news issues related to social stability more frequently than the local propaganda authority;

H1b

The central propaganda authority censors more news issues that contradict national guidelines and policies than the local propaganda authority.

H1c

The central propaganda authority censors more news issues that are harmful to the image of the central state and leadership than the local propaganda authority.

As extant studies also indicate that the central state provides some autonomy to the news media to reveal wrong doings at local levels of government, we assume that the central state propaganda authority is more tolerant of news that is detrimental to the image of the local government than are the local propaganda authorities.

Meanwhile, one hypothesis corresponding to the second research question (RQ2) focusing on the type(s) of news issues the local propaganda authorities censor frequently is formulated as follows.

H2. The local propaganda authorities censor more news detrimental to the image of the local government than the central propaganda authority.

Methods and Data

The data consist of official documents called “propaganda notices” (PNs). These 728 documents in total comprise a complete list of the news media censorship orders issued by the Propaganda Department of Province Q to all of the news organizations in the province between late May, 2011 and early July, 2012, 728 documents in total. These PNs are usually official bans for some specific news, requesting that the news organizations refrain from reporting on some issues or some aspects of an issue. There are also a small number of PNs requesting the news organizations to report positive issues. Due to the sensitivity of the data, the sources providing the documents are kept anonymous. To further conceal the identity of Province Q, these PNs are not quoted here. Instead, illustrative quotations are taken from PNs collected and released on the Internet by the Counter-Power Lab at the UC Berkeley School of Information (CPL-UCB).

Access to the data could only be obtained using a social network based on mutual trust between the author and the sources. Obviously, it is possible for many scholars in China to access these documents, but they are hesitant to study the news censorship program with sensitive data like PNs. Rui (2010) is among the few scholars in mainland China who have analyzed the PNs, yet those he analyzed were “directives on the reporting of surging prices”, and they are rather moderate compared with notices banning, for instance, politically sensitive news.

In a content analysis, two aspects of the PNs were examined. The first aspect was the main theme of the notices. Based on existing studies and a pre-overview of the data, a code list consisting of “social stability” (Dowell 2006; King et al. 2013), “the image of the central state” (Weatherley 2006; Brady 2008), and “the image of the local government” (Xiong 2008) was produced to code the main theme of the PNs. Social stability was made a code equivalent, because issues related to social stability have always been considered highly sensitive to both the central state and local authorities.

PNs dealing with topics related to any form of collective action, including protests, demonstrations, sit-ins, collective petitioning, and any form of individual protests, resistance and struggles against the authorities, were coded as “social stability”. PNs that were coded as “national guidelines and policies” were related to issues that mention state policies in any field, including (1) political (e.g., Taiwan/Tibet/Xinjiang independence, anti-Japanese movement, opinions of dissidents, international oppositions, and criticisms on the legal and electoral systems), (2) economic (e.g., economic growth, income gap, inflation and price surges), and (3) social and cultural policies. Those coded as harmful to “the image of the central state/leadership image” refer to (1) matters negative or positive to the performance of the central government (including ministries) and leadership; (2) issues may potentially cause negative or positive citizen evaluations of the performance of the central government and leadership. Last but not least, those coded as “the image of the local government” are PNs related to: (1) the performance of the local leadership and government (including government at the provincial, municipal, county, and township levels), (2) the negative effects of local policies, and (3) matters that may potentially cause negative and positive citizen evaluations of the performance of the local levels of government.

The second aspect examined is the original source of the PNs within the propaganda system, which indicates whether the news was considered sensitive to the central state, to the local levels of government or to both. As shown in Fig. 1, propaganda authorities at all levels can issue PNs. For example, if a county/district level government wants to ban the news within the entire province, it will first issue a PN and submit an application to the propaganda authority at the provincial government level, which will decide if they would transmit the PN to all the news media in the province. The propaganda authority at the provincial government level also decides if the PN should be submitted to the central propaganda authority. If it is considered necessary, the central propaganda authority will decide whether to transmit the PN to all other provincial propaganda authorities in the country. Each provincial authority will then send the PN to the provincial level news media, the lower level propaganda authorities, and so on.

As the PNs collected for this study were all issued by the propaganda authority of Province Q, the origins of the notices could be the local governments in Province Q (LGQ), including all prefectural-level, county-level, township-level government in Province Q, the provincial government of Q (PGQ) itself, the central government (CQ), the provincial governments outside Q (PGOQ), and the local governments outside Q (LGOQ), including all prefectural-level, county-level, and township-level government outside Province Q. All local governments, i.e., municipal, county-district, and township governments, must first submit an application to the provincial propaganda authority if they do not want the news media in the entire province to report on the issue. The provincial propaganda authority then decides if the propaganda note should be sent to all news organizations in the province and if it is necessary to submit an application to the central propaganda authority to ban the news nationwide.

Thus, as far as origins are concerned, PNs originally coming from the CG are those most sensitive to the central state, PNs originally coming from LGQ, PGQ, PGOQ, and LGOQ deal with matters sensitive to the local levels of government, and PNs originally coming from PGOQ and LGOQ deal with matters sensitive to both the central state and local levels of government. The most frequently appearing themes on the PN items with the origins of CG, PGOQ, and LGOQ indicate that these types of news are most likely censored by the central state, while the themes most often seen in the PNs originally from LGQ and PGQ are those purely censored by the local levels of government.

A codebook was produced to guide the coding of the PNs. One external coder was trained to do systematic coding. The percentage agreement between the two coders for each variable was calculated as: social stability (94%), national guidelines and policies (90%), central state/leadership image (83%), local state image (83%), and origin of PNs (75%).

SPSS was used to perform a series of Chi-square tests to verify the hypotheses. In the presentation of the results, both the p value indicating the statistical significance level and the Cramer’s V measuring effect size are reported. The statistical significance level was set at α = 0.05, while the effect size measures of 0.5, 0.3, and 0.1 were used to represent the cutting points for the “strong”, “medium”, and “weak” associations (Ellis 2010).

Results

In the first part of this section, an overview of the analysis is represented with a classification of PNs by origin. The second part examines whether news harmful to social stability, national guidelines and policies, and the image of the central state and leadership is more likely censored by the central state than its local counterparts. The third part will check if news negative to the image of the local authorities is more often censored by local levels of government, while the last part compares the news more frequently censored by the central state and the local levels of government to see if the central state is more tolerant of news that is detrimental to the image of the local levels of government.

The Origin of PNs

More than half of the 728 PNs collected (55%) were sent from the central propaganda authority to the propaganda authority of Province Q. The PNs originate from “central government” (CG), “provincial governments outside Q” (PGOQ), and “local governments outside Q” (LGOQ) (cf. Table 1). Of these central propaganda authority PNs (55%), about half (27%) originally come from provincial and local governments outside Province Q. The remaining PNs (45%) come directly from Province Q, i.e., “provincial government of Q” (PGQ) and “local governments in Q” (LGQ). This indicates that the central propaganda authority does not automatically censor nationwide all news issues censored by the local levels of government; otherwise, the number of PNs with origin of PGOQ and LGOQ would have been much higher.

Table 1 Propaganda notices by origin

News More Likely Censored by the Central State

More than half of all of the PNs coded (58%) relate to “the image of the local government”. This comes as no surprise, since the vast majority of PNs, 526 of the total 728 PNs, originally come from the local levels of government, i.e., PGOQ, LGOQ, PGQ, and LGQ (Table 1). 30% of the censored themes relate to the “image of the central state/leadership”, 13 to “social stability”, and 11 to “national guidelines and policies” (Table 2).

Table 2 Propaganda notice themes by origin

A comparison of the themes of censored news purely from the central propaganda authority (PNs from CG, PGOQ, and LGOQ) and those from one local-level government (PGQ and LGQ) (see Table 3) shows that the central government censors more news issues that contradict “national guidelines and policies” than the local levels of government (18 vs. 2%). The extremely small p value of 0.000 (smaller than 0.004) indicates that the difference in percentage is statistically significant. Moreover, the Cramer’s V value, indicating the association between the levels of government, and this PN theme falls at 0.232, which is between the “weak” (0.1) and “medium” (0.3) cutting points. This suggests that H1b is confirmed.

Table 3 Comparison of PN themes by origin (central state level vs. local levels of government)

Similarly, the central government censors many more news issues related to “the image of central state and leadership” than the local government. While more than half of the PNs issued by the central propaganda authority (51%) are concerned with this theme, only 5% of those distributed by the local governments are also associated with this theme. The extremely small p value (0.000) and the large V (0.496) indicate a strong association, and thus H1c is also confirmed.

Surprisingly, H1a is rejected as the central government censors less news issues about “social stability” than the local government (6 vs. 22%). The difference is, however, significant with an extremely small p value (0.000) and the V value of 0.232 falls between the “weak” and “medium” cutting points. This can be attributed to the fact that local governments censor a substantial number of items related to “social stability”, because social unrest at the local level also damages the image of the local authorities. Moreover, since social stability is also of great concern to the central leadership, local authorities will suppress a news story before it spreads widely.

News More Likely Censored by Local Authorities

Besides news issues related to “social stability” as indicated above, those which are detrimental to “the image of the local government” are heavily censored by the local governments. As it is shown in Table 3, 77% of PNs with the theme of “the image of the local government” were distributed by the local propaganda authority. In comparison, the percentage of such PNs distributed by the central counterpart is smaller at 42%. The extremely small p value of 0.000 and the V of 0.357 (between the cutting points of “medium” and “strong”) also confirm H2.

Although the focus of censorship at the local levels of government is news harmful to social stability and the image of the local government, we cannot infer from this that the local levels of government do not censor news related to national guidelines and policies and the image of the central state and leadership. On the contrary, the local propaganda authority transmits most of the PNs handed down from the central propaganda authority to all of the news organizations in their control. The only exceptions are those (just a few) which are tolerated and distributed by the central propaganda authority but are negative to the local government in question (see more details in the next section).

News Tolerated by Central State but not Local Levels of Government

Another piece of evidence indicating that the central state and local levels of government do not always align in how they censor news is that sometimes news that is to the detriment of the local levels of government is tolerated by the central propaganda authority but not by their local counterparts.

While the central propaganda authority frequently censored news that was not in line with national guidelines and polices and those which are harmful to the image of the central state and leadership, it did not maintain a consistent level of censorship when news proved to be the detriment of local governments. As shown in Table 1, the PNs from the local levels of government outside Province Q and accepted by the central state propaganda authority for a ban nationwide only added up to 199 items (PGOQ + LGOQ). This is in sharp contrast to all of the PNs issued by the Province Q propaganda authority itself (PGQ + LGQ), which made up 327 items.

If all the PNs from all of the other 31 province-level states in China were accepted by the central propaganda authority, we would see a huge number of notices originating in PGOQ and LGOQ, which would also outnumber the combined number of PNs originating in PGQ and LGQ. Furthermore, if we only consider the PNs relating to the image of the local authority, the contrast between the central state and local propaganda authorities is even sharper, with only 8 originating from the Central Authority (“central government”), 160 from lower level governments outside Province Q (PGOQ + LGOQ), and a total of 253 from one single province, namely Province Q (PGQ + LGQ). These two comparisons show that the central propaganda authority tolerates some news that is not tolerated by the local levels of government. This news tends to portray local authorities in a negative light.

Of course, one may argue that provincial levels of government lack the incentive to submit news ban applications to the central propaganda authority every single time when they notice a negative news item about them. It is also costly for the local leadership to do so. However, if the central state favors the censorship of all negative news, we would expect the central propaganda authority to ask province-level propaganda authorities to submit an application whenever negative events occur at the local level, but this is actually not the case.

Furthermore, there is another piece of evidence, indicating that the central state does not censor everything that is negative from the perspective of the local governments: Local propaganda authorities sometimes send PNs to the news media under their control, requesting them not to report news that is reported by the central level news media. A PN from the Guangdong Province Propaganda Department and released by CPL-UCB serves as a good example:

Do Not Reprint Related News Articles from Xinhua News Agency

Propaganda Department of Guangdong Province: In regards to the event about Yuexiu District Party Committee Member and Head of Armed Forces Department beating an air hostess, strictly execute the requests of our department’s previous notice. Do not reprint related news articles from the Xinhua News Agency.

Thus, this piece of negative news about a high ranking official in a local government in the Guangdong Province was actually tolerated by the central propaganda authority, since central level news organizations, such as the Xinhua News Agency, were allowed to report it. Yet, the local propaganda authority did not tolerate the news and, therefore, issued a PN to the news media in the province to ensure that they would not reprint the news.

We expect the same pattern as far as the provincial authorities and municipal authorities or the latter and the district/county level authorities are concerned, although we do not have a complete list of PNs from a municipal propaganda authority to investigate this matter. Yet, the following example from CPL-UCB provides some evidence:

Zheng Beiquan Discipline Violation Case

Propaganda Department of Guangdong Province: According to the notice of the Commission for Discipline Inspection of Qingyuan City (CDI-QY), Guangdong, Zheng Beiquan, former vice mayor and police chief of Yingde Municipal Government, has been file-inspected by the CDI-QY and is now being investigated by the organization. The case can be reported normally.

One point that our data set does not allow us to investigate is why the central state chose to censor some news to the detriment of the local levels of government but not others. One possible explanation is that some news is seen as potentially harmful to the legitimacy of the state once it was reported by the news media nationwide. Other central propaganda authority PNs may be the result of local leaders’ personal network within the central government. Thus, factions or patron–client relations within the Chinese political system may be part of the explanation.

Discussion

Every state authority wants the news media to work to their advantage, but their capacity to do so is constrained by either costs or the political system. In an authoritarian state like China, the rulers are believed to have the capacity to control all news organizations (Siebert et al. 1963). This study did not intend to examine whether the Chinese state has the capacity to fully control news reporting but instead examine the content of news media censorship and answer a question that yet to be addressed: What kind of news do the Chinese state authorities censor? The study has assumed that even if the state authorities had the full capacity to censor every item of negative news, they do not censor everything. Extant studies on media control in China indicate that news involving social stability (Dowell 2006; King et al. 2013), national guidelines and policies, and the image of the central state and leadership (Brady 2008; Zhang and Cameron 2003), which represent the priorities of the state in maintaining legitimate ruling, is mostly likely censored. Yet, the fact that China is a decentralized state implies that a structural difference exists between the central state and local levels of government when it comes to news censorship. On the one hand, the central state wants to use the news media to control local leaders by permitting some negative reports about the local levels of government (Zhao 2000; Lorentzen 2014). On the other hand, local leaders, who control the local propaganda authorities, do their best to suppress bad news about themselves to protect their reputation for future career promotions (Zhang and Flemming 2005; Dowell 2006; Xiong 2008).

By analyzing PNs from a province-level propaganda authority in China, our study shows that both the central and local authorities are engaged in selective censorship practices. The focus of the central state’s news censorship is news considered not in line with the national guidelines and policies, and is harmful to the image of the central state and leadership. The majority of the PNs that the central propaganda authority sent to the province-level propaganda authorities were related to these two themes. This finding is in accordance with those from most existing studies. Although the central propaganda authority did censor some news which was to the detriment of the image of the local authority, it did so mainly, because such news is also found to be harmful to the image of the central state. In fact, the data suggest that the central state allows news media to reveal part of the negative news about local levels of government, whose propaganda authorities on the other hand strive to suppress such news. One explanation why the central propaganda authority tolerates negative news about local levels of government is that the central leadership controls the decentralized states and local cadres through the career promotion system (Landry 2008).

When it comes to censorship, the local levels of government, on the other hand, focus on social stability and the image of the local government. In particular, the Province Q propaganda authority issued many more PNs related to the image of the local government than related to the image of the central state and leadership as such. As noticed by Dowell (2006) and Xiong (2008), the main function of the local censorship program is to serve the local leadership, whose career prospects for advancement through the political system would be dim without a well-maintained image of the local state that they govern. Furthermore, local propaganda authorities do request that local news organizations not reproduce news reports on the local government that are, otherwise, allowed by the central state authority. Moreover, though it is surprising to find that the local governments censor more news related to social stability than the central government, it could be easily inferred that there are two reasons behind. The first is that news about social unrests happening in their area would also damage the image of the local government if the problems are not resolved properly. The second is that social stability is also important to the central government and the political life of the local cadres will be at risk if the social unrests originated from their area spread.

Although the focus of news censorship at the local levels of government is to preserve and maintain their image, the local propaganda authorities must transmit all of the PNs handed down from the central propaganda authority to the newsrooms. Consequently, the local propaganda authorities become more repressive than the central propaganda authority in that they censor the news that is to the detriment of both state legitimacy and the image of the local government. The central state, on the other hand, allows some degree of freedom for the news organizations to reveal wrongdoings among the local leadership as this helps the central leadership monitor lower level cadres and maintains the legitimate ruling of the central state. This may partly explain why the general public in China, as some researchers point out, e.g., Li (2004), has more political trust in the central government than in local levels of government.

Similarly, a closer look at some PNs indicates that the provincial propaganda authorities are also more tolerant than their counterparts at the lower levels of government, including the municipality, district/county, and township levels. This explains the fact that “bad news” is already seldom seen in district/county level news media.

The study also shows that the central propaganda authority does censor some news that is to the detriment of local levels of government but not harmful to the central state and leadership. One possible explanation is that censors fear that the news may potentially undermine central state legitimacy once it is spread widely. Another is that faction leaders in the central government interfere in the censorship of news that can be harmful to the local levels of government.

This research is the first of its kind to investigate the focus of propaganda authorities on both the central state level and the local levels of government. Despite its uniqueness, the study only examined PNs distributed by one provincial government. Though we can see almost all of the PNs issued by the central propaganda authority as propaganda authorities at lower levels are requested to distribute all the PNs handed down from the central level, we do not see all the PNs distributed by other provincial and lower level propaganda authorities. Having examined only PNs from one provincial government, the conclusions from this exploratory research would require tests in future studies in which more local governments should be included in the comparisons.

References

  1. Brandy, Anne-Marie. 2008. Marketing dictatorship: Propaganda and thought work in contemporary China. Lanham: Rowman and Littlefield Publishers Inc.

    Google Scholar 

  2. Dowell, William Thatcher. 2006. The Internet, censorship, and China. Georgetown Journal of International Affairs 2006: 111–119.

    Google Scholar 

  3. Du, Huizhen. 2010. The Changing Strategies of Media Control in China’s Reform Era, PhD Thesis, University of Leeds (Institute of Communications Studies), 2010.

  4. Edin, Maria. 2003. State capacity and local agent control in China: CCP cadre management from a township perspective. The China Quarterly 173: 35–52.

    Article  Google Scholar 

  5. Ellis, Paul D. 2010. The Essential guide to effect sizes: statistical power, meta-analysis, and the interpretation of research results. Cambridge: Cambridge University Press.

    Book  Google Scholar 

  6. Egorov, Georgy, Sergei Guriev, and Konstantin Sonin. 2009. Why resource-poor dictators allow freer media: a theory and evidence from panel data. American Political Science Review 103 (4): 645–668.

    Article  Google Scholar 

  7. Esarey, Ashley. 2006. Speak no evil: Mass media control in contemporary China. Freedom at Issue: A Freedom House Special Report (February 2006).

  8. Friedrich, Carl, and Zbigniew Brzezinski. 1956. Totalitarian dictatorship and autocracy. Cambridge: Harvard University Press.

    Google Scholar 

  9. Goodman, David. 1992. Provinces confronting the state? In China Review, 3rd ed, ed. H.C. Kuan, 2–19. Hong Kong: CUHK Press.

    Google Scholar 

  10. Hao, Jia, and Zhimin Lin. 1994. Changing Central-Local Relations in China: Reform and State Capacity. Boulder Co: Westview.

    Google Scholar 

  11. Harding, Harry. 1981. Organizing China: the problem of Bureaucracy, 1949-1976. Stanford: Stanford University Press.

    Google Scholar 

  12. Hook, Leslie. 2011. Beijing raises spending on internal security. Financial Times online, accessed on June 30, 2017. https://www.ft.com/content/f70936b0-4811-11e0-b323-00144feab49a.

  13. Jiang, Zemin. 1989. Several Questions on the Party’s Press Work (In Chinese: Guanyu dang de xinwen gongzuo de jige wenti). News Front (Xinwen Zhanxian) 1990 (3): 2–7.

    Google Scholar 

  14. King, Gary, Jennifer Pan, and Molly Roberts. 2013. How Censorship in China Allows Government Criticism but Silences Collective Expression. American Political Science Review 107 (2): 1–18.

    Article  Google Scholar 

  15. Kotler, Philip. 2001. Kotler on marketing: how to create, win and dominate the market. New York: The Free Press.

    Google Scholar 

  16. Kuang, Xianwen. 2017. Negotiating the boundaries of news reporting: journalists’ strategies to access and report political information in China. MedieKultur: Journal of Media and Communication Research 62: 35–51.

    Google Scholar 

  17. Landry, Pierre F. 2008. Decentralized Authoritarianism in China: The Communist Party’s Control of Local Elites in the Post-Mao Era. New York: Cambridge University Press.

    Book  Google Scholar 

  18. Lawson, Chappell. 2002. Building the Fourth Estate: Democratization and the Rise of a Free Press in Mexico. Berkeley: University of California Press.

    Google Scholar 

  19. Liebman, Benjamin L. 2011. The Media and the Courts: towards Competitive Super-vision? The China Quarterly 208: 833–850.

    Article  Google Scholar 

  20. Li, Lianjiang. 2004. Political trust in rural China. Modern China 30 (2): 228–258.

    Article  Google Scholar 

  21. Liu, Jun. 2013. Mobilized by Mobile Media: How Chinese People use mobile phones to change politics and democracy, PhD Thesis, University of Copenhagen (Department of Media, Cognition and Communication), 2013.

  22. Lorentzen, Peter. 2014. China’s Strategic Censorship. American Journal of Political Science 58: 402–414.

    Article  Google Scholar 

  23. McMillan, John, and Pablo Zoido. 2004. How to subvert democracy: Montesinos in Peru. Journal of Economic Perspectives 18 (4): 69–92.

    Article  Google Scholar 

  24. Naughton, Barry. 1995. Growing out of the Plan: Chinese Economic Reform 1978-1993. New York: Cambridge University Press.

    Book  Google Scholar 

  25. Nye Jr., Joseph S. 2004. Soft power: the means to success in world politics. New York: Public Affairs.

    Google Scholar 

  26. Rui, Bifeng. 2010. News production, and the reproduction of news productive relations – Citing the ‘propaganda notices’ and their implementation (In Chinese: Xinwen shengchan yu xinwen shengchan guanxi de zaishengchan – yi ‘xuanchuan tongzhi’ jiqi zhixing qingkuang weili). Journalism Quarterly (Xinwen Daxue) 103 (1): 86–92.

    Google Scholar 

  27. Shadmehr, Mehdi, and Dan Bernhardt. 2012. A theory of state censorship. APSA 2012 Annual Meeting Paper. Available at SSRN: http://ssrn.com/abstract=2105407.

  28. Shirk, Susan. 2011. Changing Media, Changing China. In Susan Shirk, ed. Changing Media, and Changing China, 1–37. New York: Oxford University Press.

    Google Scholar 

  29. Siebert, F.S., T. Peterson, and W. Schramm. 1963. Four Theories of the Press. Urbana and Chicago: University of Illinois Press.

    Google Scholar 

  30. Stockmann, Daniela, and Mary E. Gallagher. 2011. Remote Control: how the Media Sustain Authoritarian Rule in China. Comparative Political Studies 44 (4): 436–467.

    Article  Google Scholar 

  31. Teiwes, Frederick C. 1995. The paradoxical post-Mao transition: from obeying the leader to ‘normal politics’. The China Journal 34: 55–94.

    Article  Google Scholar 

  32. Weatherley, Robert. 2006. Politics in China since 1949: legitimizing authoritarian rule. London: Routledge.

    Google Scholar 

  33. Wu, Guoguang. 2000. One head, many mouth: Diversifying press structure in reform China. In Power, money, and media: communication patterns and bureaucratic control in cultural China, ed. Chin-Chuan Lee, 45–66. Evanston: Northwestern University Press.

    Google Scholar 

  34. Xiong, Lei. 2008. The relations between news media and state image (In Chinese: Xinwen meiti yu guojia xingxiang de guanxi). Youth Journalist (Qingnian Jizhe) 23: 53–54.

    Google Scholar 

  35. Zhang, Ernest, and Kenneth Fleming. 2005. Examination of characteristics of news media under censorship: a content analysis of selected Chinese newspapers’ SARS coverage. Asian Journal of Communication 15 (3): 319–339.

    Article  Google Scholar 

  36. Zhang, Juyan, and Glen T. Cameron. 2003. The structural transformation of China’s propaganda: an Ellulian perspective. Journal of Communication Management 8 (3): 307–321.

    Article  Google Scholar 

  37. Zhao, Yuezhi. 2000. Watchdogs on party leashes? Contexts and limitations of investigative reporting in post-Deng China. Journalism Studies 1 (4): 577–597.

    Article  Google Scholar 

  38. Zhong, Yang. 1996. Legitimacy crisis and legitimation in China. Journal of Contemporary Asia 26 (2): 201–220.

    Article  Google Scholar 

Download references

Acknowledgements

This work was supported by the Danish Research Council under Grant 94-556-85013.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Xianwen Kuang.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Kuang, X. Central State vs. Local Levels of Government: Understanding News Media Censorship in China. Chin. Polit. Sci. Rev. 3, 154–171 (2018). https://doi.org/10.1007/s41111-018-0091-5

Download citation

Keywords

  • China
  • News censorship
  • Propaganda
  • Central state
  • Local levels of government