Western analysts typically view China's non-governmental think tanks more favourably than government-sponsored ones. For example, they used to argue that the latter were established to shape the legitimacy of political authority.Footnote ³ Today, government-sponsored think tanks, such as the Chinese Academy of Social Sciences (CASS), have been accorded increased independence; however, Western researchers still tend to regard them as being politically steered by the government with the implementation of multiple administrative and financial measures.Footnote ⁴ By contrast, the emergence of non-governmental think tanks in China was highly lauded by Western observers; they regarded the competition between government- and non-government-affiliated think tanks in the policy domain as an indication of the Chinese government's relaxation of control over the ideological domain and the political development of civil society.Footnote ⁵

Nevertheless, many conflicting cases complicate the otherwise simple analytical frameworks of government-sponsored versus non-governmental think tanks. People customarily assume that non-governmental think tanks publicly discuss policy issues and criticize the government, or what is referred to as promoting democratization. Looking at the activities of China's think tanks, however, distinguishing the roles think tanks play between types proves to be a difficult exercise. For example, many non-governmental think tanks are staunch supporters of government policy or frequently conduct government research projects,Footnote ⁶ whereas high-level government-sponsored think tanks, on occasion, publicly criticize government policy.Footnote ⁷ In reality, every think tank in China, whether government-sponsored or non-governmental, decides on its individual roles in the policy process depending on internal and external factors. It is more appropriate, therefore, to evaluate think tanks by studying their specific activities. This article focuses on the roles of China's think tanks rather than their organizational structures, which is what previous studies have tended to concentrate on, and asks: what are the determinants that influence each think tank to choose its roles in the Chinese policy process?

Regional variations further complicate the analysis. In most previous studies, researchers have evaluated think tanks in Beijing or major cities such as Shanghai and have not examined regional variations in different institutional contexts throughout the country. Considering China is characterized by uneven development, a nationwide survey is necessary to understand fully the roles of think tanks.

This article presents two independent sets of survey data to explain the roles China's think tanks play in different regions. A nationwide survey was conducted by mailing questionnaires to think tank heads from September to November 2004 (China Think Tank Survey [CTTS] 2004). The questionnaires were sent to 1,124 qualified institutes using the 2003 registered official information from the Division of Soft Science at the Ministry of Science and Technology of China. A total of 301 responses was received from 25 provinces (municipalities and autonomous regions), or a response rate of 26.78 per cent. The selected targets covered various policy fields not limited to science and technology.Footnote ⁸ In this survey, think tank leaders were questioned on many variables related to their think tank's activities and characteristics. The second set of data used in this article is the 1998 regional Integrated Knowledge Development Index (IKDI) in China. Based on the concept of “knowledge” defined by the World Bank,Footnote ⁹ the IKDI 1998 was designed by Angang Hu and Yizhi Xiong 熊义志 in 2003 to measure the level of knowledge capacity in each province in China.Footnote ¹⁰ In this study, the IKDI is adopted as the key indicator of regional contextual variations.

China's Think Tanks

China's policy research institutes (or think tanks) can be defined as stable and autonomous organizations that research and consult on policy issues to influence the policy process.Footnote ¹¹ They have had a long process of development since the founding of the People's Republic of China. In the 1950s and 1960s, several state-owned research institutes based on the Soviet model were established. Meanwhile, intellectuals outside government experienced difficulties because of the capricious and ambiguous attitude of the Chinese political leadership and continuous political movements against intellectuals. After the reform and opening-up policy was initiated in 1978, many government-sponsored semi-official think tanks were established under the central and local governments. However, because of the Tiananmen Incident at the end of the 1980s, there was a two-year hiatus in China's think tank history.Footnote ¹² After Deng Xiaoping's 1992 South China tour speeches (Nanxun jianghua 南巡讲话), which ushered in a new era of reform, various types of think tank mushroomed, especially on-campus research institutes and purely private organizations. In 2007, the Party Report explicitly mentioned “the roles of think tanks” at the 17th Chinese Communist Party (CCP) Congress for the first time. This recognition and support from the authorities signified the anticipated rapid development of think tanks in China as well as their expanding influence. Overall, China's think tanks have become important and active policy actors. They have been regarded as some of the most important windows to a changing China by scholars who are interested in the Chinese policy-making system.Footnote ¹³

Official policy research institutes outside the definition of China's think tanks are excluded from this article. The General Principles of the Civil Law of the People's Republic of China (中华人民共和国民法通则) stipulates that official policy research institutes established inside the government are “government agencies” (jiguan 机关). The behavioural patterns of official policy research institutes differ from those of think tanks; the main responsibilities of the latter are drafting policy documents, releasing information and organising studies on policy issues. Although official policy research institutes also investigate policy issues, they cannot independently decide on research objectives and timetables. Ogden describes establishment researchers in Chinese official research institutes as those who serve as pens of the party-state (bi gan zi 笔杆子) and have no ideas of their own.Footnote ¹⁴ Official policy research institutes may be described as the “internal brains” of the government, whereas think tanks are frequently viewed as its “external brains.”

China's think tanks can be divided into two categories according to organizational identity: government-sponsored semi-official think tanks and non-governmental think tanks. Semi-official think tanks are public institutions (shiye danwei 事业单位) founded and sponsored by the government (as a “supervising unit,” or yewu zhuguan danwei 业务主管单位 or guakao danwei 挂靠单位). They have well-defined administrative connections to the government, are headed by government-nominated personnel and accept start-up capital from their supervising government agencies. They also receive a steady flow of administratively appropriated funds as fees for regular research tasks assigned by their supervisors; thus, their policy output is somewhat shaped by government directives. Non-governmental think tanks include policy research institutes that are registered as enterprises (qiye 企业) and civilian non-profit institutions (minban fei qiye danwei 民办非企业单位), or affiliated institutes under universities. Unlike government-nominated think tanks, they are less characterized by administrative affiliation. Although some have supervising units, a number of which are also government agencies, the relationship between them and their supervising units is very loose.Footnote ¹⁵ In addition, non-governmental think tanks acquire seed capital from diverse sources: from universities, enterprises, overseas funding, domestic funds and partnerships brought about by private capital.Footnote ¹⁶

Advisors, Academics, and Advocates

To build influence, think tanks need to use their expertise to create an impression both on decision-makers and other actors in the policy-making system. Generally speaking, they tend to serve simultaneously as advisors to the government, academics in universities and research institutes, and policy advocates in the public sphere.

The basic tasks of think tanks are to carry out research and participate in academic activities. Creating expertise on policy issues is an essential distinction between think tanks and other policy actors, such as special interest groups that also try to influence policies. Moreover, think tanks can influence academic and other policy actors by, for example, publishing papers in academic journals and attending domestic or international conferences and seminars. Some policy ideas are initially proposed through academia and eventually influence policy-making. For example, a scholarly paper discussed in an international conference in 2001 drew the attention of top state leaders and consequently played an important role in the decision-making process concerning the provision of financial and organizational support for the establishment of the New-Form Rural Co-operative Medical System in 2002.Footnote ¹⁷

To influence policy directly, think tanks also spend considerable effort serving as advisors to authorities. They not only apply for research projects sponsored by the Chinese government, but also seek opportunities to provide literal or non-literal consultancy services. They submit internal research reports to the government and personally advise government officials. For example, the General Office of the CASS periodically submits two versions of the internal CASS Important Report (Yaobao 要报) to the central government and other agencies: Information Special Report (Xinxi zhuanbao 信息专报) and Leader Reference (Lingdao canyue 领导参阅).Footnote ¹⁸ Think tanks are also frequently invited to government meetings as well as to deliver lectures and offer advice on policy issues in their respective fields. The most significant example is the Political Bureau of the CCP Central Committee, which organized 44 collective study events inviting many think tank experts to give lectures to top Chinese leaders between the 16th and the 17th National Congresses of the CCP.Footnote ¹⁹

As advocates in the public sphere, think tanks have the opportunity to influence government decisions. The Chinese government established filter mechanisms that monitor, collect and digest information and policy ideas from mass media, the internet and other academic or non-academic publications. For example, most ministries and local governments have information centres or public sentiment analysis institutes. Moreover, many official news agencies have types of internal reference (neican 内参) systems for governments at each level, such as the “internal references” of People's Daily and “domestic status proofs” of Xinhua News Agency.Footnote ²⁰ Think tanks are therefore able to draw the attention of decision-makers by initiating or participating in public discussions or even by publicly criticising government policies. With the rapid development of the internet, many policy changes are triggered by online discussions amongst the public initiated by scholars.Footnote ²¹Table 1 summarizes the major activities that think tanks tend to adopt in the Chinese policy process.

Table 1:

Activities of China's Think Tanks

The activities of think tanks can be considered the way of mobilizing resources to influence policies most efficiently.Footnote ²² In the Chinese context, the administrative relationship between think tanks and the government is the most important resource to help think tanks exert direct policy influence. If a think tank's administrative connection is not sufficiently solid, it would have to choose other approaches to output ideas. Government-sponsored think tanks certainly have many advantages over their non-governmental counterparts in terms of administrative relationships. Non-governmental think tanks sometimes have nominal administrative ranks and official supervising units, but they rely more on personal social networks of scholars and their academic knowledge to influence policy.Footnote ²³ My previous empirical analysis also provides convincing evidence illustrating the influence of China's think tanks. The regression models in that report show that expert knowledge, administrative connection and personal ties are the factors that determine influence on the Chinese policy process. Thus, individual resources enable think tanks to exert more influence on policies.

Regional Variations and Hypotheses

Theoretically, the regions where think tanks are located are also key factors in addition to individual resources. For example, it could be assumed that think tanks located in Beijing have more direct influence on the government, particularly government-sponsored think tanks. However, in my previous empirical report, I found that a Beijing location is advantageous only to the advocacy role of non-governmental think tanks, whereas for semi-official think tanks, this location shows no appreciable difference after other independent variables were controlled. Nevertheless, using Beijing as a dummy variable to show regional variation is a simplified method. China is a large and transitional country; thus there are notable variations in the development of think tanks and a need to study whether they play different roles in different regions. Furthermore, what contextual characteristics can we extract to show the differences that have had an effect on these roles?

Some scholars argue that the transformation of the relationship between Chinese intellectuals and the party-state can be attributed to marketization and the expanded intellectual public sphere.Footnote ²⁴ If so, variations in the degree of marketization and the intellectual public sphere can be used to explain the roles that China's think tanks play in different regions. However, this argument has two limitations. First, to borrow Andrew Walder's argument on redistributive power, in the specific circumstances of China “the decline of redistributive power is not caused by the shift to markets, it is in fact the definition of the spread of market.”Footnote ²⁵ Similarly, to say that marketization has allowed a higher degree of autonomy from political interference amongst Chinese intellectuals is actually tautologically saying that the decline of political power has weakened political control over Chinese intellectuals, which does not make any sense. The second problem is how to measure empirically the ever-changing intellectual public sphere in China. Scholars observe changes in the public sphere by examining changes in the roles of intellectual organizations. For example, Edward Gu and Merle Goldman state that one of the most remarkable changes in intellectual–state relations has been the burgeoning of non-governmental intellectual organizations, which has led to a structural transformation of the intellectual public sphere.Footnote ²⁶ However, if we argue that the expanded intellectual public sphere has made it possible for Chinese non-governmental intellectual organizations to mushroom, the result then becomes the cause. Therefore, another contextual factor must be used to illustrate the divergent roles of China's think tanks, one that is relatively independent and easier to measure in the public sphere.

The contextual factor extracted in this research to explain regional variations in China's think tanks is regional knowledge capacity. The 1999 World Bank Report states that knowledge capacity in developing countries includes abilities to acquire, absorb and communicate knowledge.Footnote ²⁷ Acquiring knowledge involves tapping and adapting knowledge available elsewhere in the world, creating knowledge locally through research and development, and building on indigenous knowledge. Absorbing knowledge involves ensuring universal basic education and extending education, creating opportunities for lifelong learning and supporting tertiary education. Communicating knowledge involves taking advantage of new information and communication technology to ensure access to knowledge. On the basis of this report, Angang Hu and Yizhi Xiong designed an integrated index of knowledge development and collected all relevant data in all provinces (municipalities and autonomous regions) in China (IKDI 1998).Footnote ²⁸ In their research on regional knowledge development, they argued that there are “four different regions” (sige shijie 四个世界) within China because of the gap in regional knowledge development. Table 2 shows the 1998 IKDI in each region, as surveyed and calculated by Hu and Xiong.

Table 2:

The Integrated Knowledge Development Index in 1998

Source: Hu and Xiong's survey and calculation in 2000 and 2003.

A straightforward logical connection between think tank influence and contextual characteristics is knowledge capacity of the environment surrounding think tanks. Whether an expert idea that a think tank produces is accepted by other policy actors and eventually gets on the policy agenda depends not only on the think tank's capabilities and quality of expertise, but also on whether all potential audiences have the ability to acquire, absorb and communicate the policy idea. Therefore, synthetically speaking, a factor that determines the ability of think tanks to build influence is the knowledge capacity of the entire region where they are located, because all the residents may be regarded as potential participants who might influence eventual policy decision-making. In brief, I hypothesize that regional knowledge development is the variable that influences think tanks on a regional level. In this article, Hu–Xiong's regional IKDI is employed to show whether and to what extent regional knowledge capacity affects the roles think tanks play in the policy process in China.

Table 3 shows hypotheses on the roles of China's think tanks in regional contexts. Hypothetically, semi-official think tanks have nothing to do with the level of regional knowledge development because they rely mainly on administrative networks to influence policy. By contrast, regional knowledge development strongly affects non-governmental think tank activities. Moreover, the latter tend to appeal to public opinion first before eventually influencing decision-makers if they have no administrative connection with the government. Consequently, they more actively serve as advocates in the public sphere in regions where regional knowledge development is high, because there are more people who are willing to listen to and understand what they are advocating.

Table 3:

Hypotheses on the Roles of China's Think Tanks in Regional Contexts

Notes:

 All other related variables are controlled. The higher the number of “ + ” the more significantly the factor contributes to think tank influence. NS means no significance.

A think tank's connection with the government is the second key variable. Culturally and traditionally, China's think tanks prefer to influence decision-makers directly as advisors rather than through indirect channels.Footnote ²⁹ As mentioned above, semi-official think tanks rely mainly on administrative connections to influence the government; therefore, the stronger a think tank's administrative relationships, the more actively it can serve as an advisor for the authorities. Non-governmental think tanks, on the other hand, can only build connections with the government through their personal social activities. If non-governmental think tanks have already developed personal social networks with decision-makers they have the advantage of being able to advise the government directly, and can spend less energy promoting policy ideas to the public, assuming locality does not confer an advantage to another think tank.

Variable and Index Design

In the CTTS 2004, I designed an indicator system of the roles of China's think tanks from which typical and effectively influential activities amongst the three roles of advisor, academic and advocate were selected. Two indexes in each role – “literal” and “non-literal” activities – were likewise selected. Table 4 provides the 3 × 2 indicator matrix of the activities of China's think tanks from the CTTS 2004.

Table 4:

Indexes of the Roles of China's Think Tanks

Two indexes are especially mentioned in Table 4. First, the number of research reports submitted to the government by a think tank does not denote “literal” decision-maker influence, so we need to look for evidence showing that a report has been noticed and recognized by decision-makers. The “commentary” (pishi 批示), or written comment and remark placed by government leaders on a report, indicates that they consider it important and it should be paid attention to, circulated, adopted and so on. Therefore, the number of commentaries on a research report submitted by a think tank can be designed as an indicator of that think tank's “literal” activities that effectively influence decision-makers. In addition, when a research document or media report contributed by a think tank scholar is given a commentary, related government sectors or news agencies inform and congratulate the author. Considering that receiving a commentary is very rare (fewer than five times annually for almost all think tanks in the CTTS 2004) and is a significant event for the scholar's institute, a think tank leader will remember it quite clearly. Consequently, this enables a relatively accurate measurement of the number of commentaries even if the only method used is the questionnaire.

Second, The China Core Journal (Zhongguo hexin qikan 中国核心期刊) is an authoritative academic journal index widely used in China. The academic journals listed in it have to meet a comprehensive indexed standard based on seven indicators: the overall indexes count, overall digests count, overall citations, overall indirect citations, digested frequency, impact factor, and national awards or inclusion in large search engines in China or abroad.Footnote ³⁰ The number of published papers listed in The China Core Journal can therefore be regarded as another index of the “literal” academic influence of a think tank. This number for 2002–03 was coded by searching the database “China national knowledge infrastructure.”Footnote ³¹ Other indexes of think tank activities are all clearly defined in Table 4.

Other independent variables as well as their correspondent indexes and coded values are shown in Table 5. Many independent variables were designed by asking think tank leaders about the sizes of their personal networks (decision-maker networks, social elite networks and media networks), statuses of administrative connections (administrative ranks of organizations, xingzheng jibie 行政级别), level of expert knowledge (educational background including degrees, overseas education and work experience), history (when the institute was established), organizational identities (government-sponsored or non-governmental) and locations, budgets, and personnel sizes. In addition, the regional IKDI was added to show contextual characteristics of regions where think tanks are located.

Table 5:

Other Variables, Indexes, and Roles of Coding

Note: * The principal component analysis was employed to merge all dimensions into one independent variable, the Standardized factor of the Decision-maker Networks and Standardized factor of the Social Elite Networks.

Empirical Findings

China's think tanks are developing unevenly in different regions. From the CTTS 2004, we obtained 301 think tanks samples in 25 provinces. Table 6 shows the distribution of these samples. Each institute's average personnel size and average expenditure on policy research were also calculated. Table 6 shows that Beijing, Guangdong, Hubei, Jiangsu, Shandong, Shanghai, Tianjin and Zhejiang are relatively well developed in terms of the number of established institutes. Beijing has the highest number of think tanks in China, similar to the high density in Washington DC. Of all the provinces with at least three evaluated think tanks, Beijing, Guangdong, Jiangxi, Liaoning and Shaanxi have larger institutes in terms of organizational personnel size than those in other regions. Think tanks in Beijing, Guangdong, Hunan, Henan, Shanghai and Sichuan have more research funding resources than those in other regions where think tanks are well developed, not only in terms of institutional research expenditure but also research expenditure per person.

Table 6:

Distribution of Think Tank Samples and Average Personnel Size and Funds (from CTTS 2004)

Table 7 presents the descriptive statistics of the activities of China's think tanks without considering regional variations; factor analyses were then performed on the activities in each role. Similarities exist between government-sponsored and non-governmental think tanks in terms of their activities. It is not true that government-sponsored think tanks rely only on direct channels to influence policies and non-governmental think tanks can only explore indirect strategies. Both have a similar range of activities from advisors to advocates. The data in Table 7 show that both attempt to produce expertise and exert policy influence directly and indirectly with various influential activities. On the other hand, the differences between them exist statistically in the active extent of each role. Table 7 also shows that government-sponsored think tanks are more influential as advisors, whereas non-governmental think tanks have advantages when it comes to academic activities. In addition, non-governmental think tanks are nearly as active as government-sponsored think tanks in serving as advocates.

Table 7:

Descriptive Statistics and Factor Analyses of Activities in Each Role of China's Think Tanks (CTTS 2004)

Note: *The principal component analysis was employed for the dependent variable of Role.

Figure 1 shows the dramatic variations in the roles of think tanks in regions where there are more than three samples in the CTTS 2004. The average value of three standardized factors of the roles of think tanks in each province was calculated in order to observe the overall status of the roles think tanks play within each region. The degree of influence and activities in different regions is uneven. In some provinces, such as Fujian, Hubei, Hunan, Shanxi and Sichuan, all three of the standardized factors are below the average of all provinces. Thus, think tanks in these regions are very inactive in the policy process. In regions such as Hebei, Henan, Jiangxi, Shandong, Xingjiang and Zhejing, think tanks tend to serve mainly as an academic research institute and adoption of the other two roles is much less likely. Think tanks in Chongqing, Guangdong, Liaoning and Tianjin tend to serve more as advisors than advocates in the Chinese policy process, whereas those in Jiangsu, Shaanxi, Shanghai and Zhejing serve more as advocates than advisors. Finally, of all the regions, only Beijing and Shanghai have think tanks which are active in all three roles in the policy process.

Figure 1:

Regional Variations of the Roles China's Think Tanks Play (CTTS 2004)

The following empirical analysis explains the individual role selection of think tanks in the Chinese policy process, focusing specifically on the relationships between the contextual characteristics and the regional variations of the roles that think tanks play. Table 8 provides statistical estimates of all three roles in the overall samples, and Table 9 compares different roles between government-sponsored and non-governmental think tanks. Table 8 shows that regional knowledge capacity has a significant effect only on the role of advocates. This finding is similar to my previous empirical research in which only the dummy variable, Beijing, was incorporated into the regression models. As calculated by Hu and Xiong, Beijing has the highest IKDI value of all the provinces in China. Nevertheless, as Table 9 shows, if the two types of think tank are compared, we find that only non-governmental ones are affected by the knowledge capacity of the regions where they are located. Moreover, the significance of regional knowledge capacity on the roles of non-governmental think tanks gradually increases as roles vary from advisor to advocate.

Table 8:

The Roles of China's Think Tanks (CTTS 2004)

Notes: Each column is derived from ordinary least squares (OLS) analysis. Data in table are non-standardized regression coefficients (B), with standardized regression coefficients (β) in parentheses. ⁺, *, **, and *** indicate significance at 10, 5, 1, and 0.1% levels, respectively (one-tailed).

Table 9:

Comparison between Government-Sponsored and Non-Governmental Think Tanks (CTTS 2004)

Notes:

  Each equation is derived from OLS analysis. Data in table are non-standardized regression coefficients (B), with standardized regression coefficients (β) in parentheses.

 Other items not shown in the underlying equations include expenditure in policy research, integrative personnel size, history of think tanks, social elite networks, media networks, Ph.D. and M.A. education overseas, and years of employment of think tank leaders.

*, **, *** indicate significance at 5, 1, and 0.1% levels, respectively (one-tailed).

The contributions of connections to the government on the roles think tanks play differ between government-sponsored and non-governmental think tanks. Table 9 shows that only government-sponsored think tanks benefit from their administrative ranks. Moreover, the significance of administrative ranks gradually decreases as roles vary from advisor to advocate. Conversely, non-governmental think tank roles have no correlation with their administrative ranks. Instead, the personal decision-maker networks of non-governmental think tank leaders strongly influence their role as advisors. By contrast, it is interesting (and expected) that the decision-maker networks of government-sponsored think tanks have nothing to do with their roles.

In summary, there are two differing forces that determine the individual choice of roles of Chinese think tanks as either advisors or advocates in the policy process. One is connections with the government: this sways think tanks to select inside approaches to influence policies as advisors for the authorities. The other is knowledge capacity in regions where think tanks are located: this influences them to select indirect approaches through public opinion to influence policies as advocates in the public sphere. Owing to their well-defined superior-subordinate relationships and long-term administrative association with the government, government-sponsored think tanks mainly rely on the former. On the other hand, because of loose administrative connections between the government and non-governmental think tanks, the personal decision-maker networks of experts outside the government become the force that influences non-governmental think tanks to access the establishment. When experts outside the government have strong personal networks with decision-makers, they tend to be reluctant to promote their policy ideas to the public, even though they have non-governmental identities. On the other hand, knowledge capacity in regions where think tanks are located exerts a strong influence on the roles of non-governmental think tanks. When the level of regional knowledge capacity becomes higher, civilian think tank experts are more likely to serve as advocates in the public sphere.