Administrative Power and Academic Resources Allocation in China ----An Empirical Study Based on the National Excellent

Paper for APPAM 2013 "Collaboration among Government, Market, and Society" International Conference

Administrative Power and Academic Resources Allocation in China ----An Empirical Study based on the National Excellent

Doctoral Dissertations

FAN Yongmao, Associate Professor

School of Public Administration and Policy, Renmin University of

China

Tel:86-18612537703

E-mail:[email protected]

LUO Danglun, Associate Professor

Lingnan (university) College, Sun Yat-sen University

YING Qianwei, Associate Professor

School of Management, Sichuan University

Abstract: Scholars are aware that the higher education sector in China is highly affected by its administrative system, but the questions of how and to what extent the administrative power impacts academic resources allocation have yet to be answered. By examining the empirical data from 2003 to 2010 of China’s National Excellent Doctoral Dissertations, this article finds that the administrative power of doctoral advisors who hold a university or college post will have important influence on academic resources allocation. For example, even though a doctoral student's academic ability to win the award matters, the advisor’s administrative power does play an important role in helping the student to get the award. This paper’s further finding is that the greater an advisor’s administrative power, the lower the standards of evaluating his or her students’ academic ability to win the award. This is especially true for those highly bureaucratized disciplines or those with a high “degree of officialdom”. This is to say, to some extent, the platform of academic resources allocation has been transformed into an arena of power competition.

Key words: administrative power; higher education; China; advisors; academic resources; bureaucratization

Introduction For many modern countries, higher education reform, coupled with many other social changes, took place along with decentralization and marketisation which are seen as the best strategies to reform the government and the economy. However, an inevitable conundrum that higher education reform has to deal with is the problem of the state-university relationship. The term “bureaucratization”—has been popularly used to analyze and criticize the close state-university relationship. Dating back to over one century ago, the notable German Sociologist Max Weber published a number of articles that voiced concerns over the direction being taken by the German academic profession at that time. Later published collectively as Max Weber on Universities (Weber, 1974), these articles exhibit a strong disappointment at what he considered German universities’ passive and subservient approach to the state. Shils, in his introduction to the collection (Weber, 1974, p.2), states that the key concern for Weber was the “complaisance of the German academic profession in its eager subservience to the authority of the state and the erosion of its moral rectitude”. In one of the articles, “The Alleged ‘Academic Freedom’ of the German Universities”, Weber argued that university faculty concessions to government and market imperatives

“take their revenge in the ultimate weakening of the moral authority of the faculties” (1974, p.6). This concern over the extent of government interference with academic freedom (Samier, 2002, p.33) was part of a more extensive exploration into the increasing bureaucratization of everyday life (Murphy, 2009). The academic life, Weber perceived, was unable to escape the fate of the bureaucratic iron cage of societal rationalization – a fate that, in his famous words, produced “specialists without spirit” and “sensualists without heart” (Weber, 1958, p.182). The relentless march of bureaucratic rationality in capitalist societies was sweeping everything before it, leaving organizations and cultural activities increasingly devoid of freedom and empty of meaning (Murphy, 2009). Unfortunately, the recurrence of Weber’s concern over bureaucratization has become much more obvious in today’s Chinese universities, where the undue bureaucratization and unnecessary administrative intervention have become a major factor constraining the development and reform of higher education. There is growing concern and consensus within China that its higher education reform has been unsuccessful; and, as for the general public, universities are no less corrupt than other sectors. To reform-minded people, the higher education sector is the “last fortress of the planned economy.” Essentially, China’s higher education sector mirrors the larger system and, thus, its bureaucratization can be shown in two aspects. On one hand, all Chinese public universities are assigned an administrative rank which is used to judge or evaluate how good or prestigious a university is. On the other hand, bureaucratization determines how Chinese universities are governed. Externally, university presidents and Party secretaries are controlled by the Party’s Organization Department. They are appointed, promoted, and even evaluated as bureaucrats. Internally, in terms of institutional layout and personnel system, China’s universities are similar to government departments (this will be shown in a later section). As a result of “bureaucratization”, professionalism is lacking in Chinese universities and academic professionals often get the short end of the stick when competing with their administrative colleagues and are forced to take administrative posts in order to get 3 better pay and greater say on resources allocation. Although it should go without saying that the Chinese higher education sector is run by the state whose governing system administers universities, there have been criticisms, as perceived by many from both inside and outside of academia alike, that the bureaucratization and administrative power in Chinese universities have gone far beyond their spectra and started influencing academic affairs through resources allocation. Institutions such as universities cannot maintain their unique value orientation in the face of increasing rationalization. As Weber put it, the “ultimate, most sublime values have withdrawn from public life, either into the transcendental realm of mystical life or into the brotherhood of immediate personal relationships between individuals” (2004, p.30). But here we have a question: to what extent can bureaucratization impact the allocation of academic resources in China’s higher education sector? There is a further, trickier question: in the causal relationship between administrative power and resources allocation, which determines which? Although our intuition tells us that the administrative power could have an impact on the allocation of academic resources, we still need solid research evidence. As a quantitative study, this paper tries to quantify and recognize how bureaucratization (specifically, the administrative power) influences academic resources allocation by selecting the National Excellent Doctoral Dissertation Award (NEDD hereafter) as research samples, controlling the variables of the academic ability of advisors and students, and taking into account other endogeneity issues. This research is expected not only to show how serious the problem of bureaucratization in China’s higher education sector is for theoretical summary, but also to put forward the possible ways of reform for practical exploration.

Administrative Structure and Power in Chinese Universities As a part of the public sector, Chinese higher education is a state system without a significant degree-granting private sector. Its institutions are of various types including general universities (natural and social sciences/humanities), technical universities, specialized institutions (e.g. medicine, agriculture, and foreign

4 languages) and teacher-training colleges. Despite reforms in recent decades, the Chinese system has remained highly centralized (Yang et al., 2007). The administration of higher education institutions follows the vertical and horizontal patterns of general public administration in China (Cheng, 1998). The vertical system is controlled by the central (national) government and the horizontal system is controlled by local authorities, mainly provincial governments. This differentiation of universities is a major factor influencing the level of autonomy for universities and academics (Yang et al., 2007). Within a university, its organizational structure is composed of the Chinese Communist Party (the Party hereafter) Committee, administrative departments, and teaching and scientific research units. In a system in which the university president takes responsibility under the leadership of the Party, the university-level Party Committee is the de facto leader of the administration of the university; that is, it leads and directs the operations of the university. In terms of organizational power, it decides the setup of internal organizations such as teaching, scientific research, and administrative units. The Party Committee also sets up branches in colleges or schools within the university. The administrative departments of the university have been categorized into three groups in terms of their corresponding duties. The first category of its administrative departments is responsible for teaching and learning affairs. It consists of departments such as the Dean’s Office, the Graduate School, the Enrollment Office, and the Examination Office. The second category of the administrative departments deals with the management of academic and scientific affairs, including offices such as the Department of Social Sciences and Humanities and the Department of Science and Technology. The third category of the university’s administrative departments is responsible for daily operations. These departments include the Human Resources Office, the Department of Finance, the Department of Foreign Affairs, and others. Teaching and research units are colleges, schools, independent divisions, or institutes. Some colleges or schools are composed of a number of departments. No 5

matter the Party Committee, the administrative departments, or the teaching and research units, they are all administrated in a hierarchically organized structure with corresponding administrative power at different levels (Figure 1).

Party Secretary President (President level) 1 (President level)

Party System Administrative and Teaching System

Vice-Party Secretary 2 2 Vice-President for 2 2 Vice-President for (vice-President level) Administrative Affairs Teaching Affairs

Department Director 3 3 College Party Secretary Department Director 3 3 College Dean (Division level) (Division level) (Division level) (Division level)

Vice Department Vice College Party Vice Department 4 4 4 4 Vice-College Dean Director Secretary Director (Vice-Division ) (Vice-Division level) (Vice Division level) (Vice-Division level)

Section Director Office Director Section Director 5 Department Chair (Section level) 5 5 5 (Section level) (Section level) (Section level)

Vice Section Director Vice Section Vice Section Director 6 Vice Department (Vice-Section level) 6 6 Director (Vice-Section level) 6 Chair (Vice-Section level) (Vice Section level)

Regular staff Regular staff Regular staff Regular professors 7 7 7 7

Figure 1 University Hierarchical and Administrative Structure University presidents and the Party secretary are government officials first; although the university and its staff have autonomy to conduct teaching and research activities, their autonomy is regulated and restricted by the control of the Party Committee, the administrative system, and the accountability mechanism. Academics are constituted, and constitute themselves, as enterprising subjects who endeavor to enhance their performance to secure their employment and promotion (Gao, 2012, p.253). This can be concluded by Yang el al.’s (2007) metaphor that even though their shackles have been removed, they are still “dancing in a cage”. Xiong (2009) has observed that for academic research and school teaching in universities, the administrative power has dominant control over resources, which has given rise to the omnipresent rent-seeking. There are more ethics and credibility

6 issues in grant applications, paper publications, and patent applications. Taking the Fifth National University Best Teachers Award as an example, 90 percent of the 100 award recipients were university presidents, university Party chiefs, school or college deans, department chairs, laboratory directors, or institute directors. More than 20 recipients had administrative posts at the university level. Roughly 10 recipients, accounting for 10 percent of the total, were frontline teachers without any administrative titles. 1The scarcity of frontline teachers receiving this award and the high positive correlation between administrative position and award-winning probability make one notice, once again, the value placed on education and the impact of the excessive administrative power in Chinese universities. Gu and Cao (2007) researched universities’ official-rank-oriented standards and bureaucracy in social sciences and they put forward the term “degree of officialdom”. They conducted three surveys and their statistics show that most chief experts of national research projects held very high administrative posts in their universities. They also found that for many research project awards, the more important the award, the higher the degree of officialdom. In the field of social sciences, education, law, and economics are the three disciplines that are impacted most by the official-rank-oriented standard. For NEDD, this impact has had an upward tendency over time. Other scholars (Yao and Zhou, 2007; Zhang et al., 2010) find that most NEDD winners are working with well-known professors in top-notch universities and the academic identities of advisors have diversified as they have more administrative and social services titles. The discussions and research findings summarized above, nevertheless, are based on news coverage or qualitative research; researchers conclude basically by qualitative description or just comparative analyses. There are few quantitative studies so far, so we still need quantitative research evidence. As a quantitative study, this paper measures how bureaucracy and administrative power impact the allocation of resources in universities by taking the following factors into account: First, the samples used are from the 2003 to 2010 NEDD award winners 7

(including nominees). 2The NEDD award started in 1999, and it has been awarded every year since then. As a very influential award, NEDD will provide winners and their advisors with large research grants and high scholarly distinction. Compared to other resources allocation mechanisms, NEDD relies upon strict and scientific selection procedures, and it is considered by many people to have higher academic standards and to be more impartial. In view of NEDD’s huge impact on the prestige of professors, students, and schools, almost everyone in academia thinks highly of its value. Second, in order to control the endogeneity of the administrative power’s impact on resources allocation, the academic ability of both the NEDD winners (including nominees) and their advisors, as well as the administrative power of the advisors, are all taken into account. Meanwhile, other variables, such as the disciplines and the ranking of schools, are also controlled. While our assumption is that although the administrative power of advisors should not have an impact on students’ academic performance, in reality, it may impact the ability of students to acquire resources. Third, we pair the NEDD winners and nominees. Given their features and the advisors’ features, we also take into account the research outputs of the winners and nominees after they receive the honor. This could make our observation more scientific.

Sample Selection and Model Construction 1. Sample Selection Doctoral education is the highest level of education in China, which represents the national development direction of science and technology, and the allocation of doctoral education resources reflects the condition of universities’ allocation of resources. Among various ways of resources allocation, NEDD selection is considered one of the most efficient and fair because of its strict selection procedures. Wining NEDD can bring both doctoral students and advisors not only tremendous kudos but also a large number of research grants. 3More importantly, doctoral students will have “God-given” academic development opportunities from an

8 advanced starting point. Whilst students long for and fight tooth and nail to win NEDD, their advisors try to put their best foot forward to help students win NEDD, which is also consistent with their long-term interests. This paper takes NEDD selection results (including nominees) in humanities and social sciences from 2003 to 2010 as samples, 4 and focuses on how the administrative power of advisors impacts the selection results. The general sample covers 7 disciplines and their 20 sub-disciplines which totals 417 selection results for the consecutive 8 years, except for military science and management science (engineering management), in humanities and social sciences. The 7 disciplines are economics, philosophy, law, education, literature, history, and management. All data are from the official website of China Academic Degrees and Graduate Education Information at http://www.cdgdc.edu.cn.

2. Variable Construction and Definition To quantify key indexes, this paper gives objective numerical values for the academic ability of each awarded or nominated student, their advisors’ academic ability, the advisors’ in-school administrative power, and their out-of-school administrative power. The variables are measured as follows: 1). Dependent Variable: NEDD Award (Award) The Department of Degree Administration and Graduate Education of the Ministry of Education is responsible for NEDD selection and evaluation. The purpose of establishing NEDD was to encourage research originality and creativity and to promote the educational quality of graduate programs, especially the doctoral programs. NEDD is selected and evaluated every year, and each year there are no more than 100 dissertations that win the award. In the social sciences, the ratio of awarded and nominated dissertations is 1:3 each year, so this research set a dummy variable to describe a dissertation that is awarded or nominated. If it is awarded, it is coded as 1; otherwise, it is coded as 0. 2). Independent Variables

(1) Students’ Academic Ability（stuabil）

The international indexes to evaluate student’s academic ability contain 3 categories: the number of papers published in top journals, life-time citations, and status of being tenure-track professors after graduation. In China, the most authoritative method uses paper quality and academic contribution to judge a student’s ability. The paper quality dimension compares dissertations at vertical and horizontal levels, while the academic contribution analyzes literature statistics from SCI, EI, and other top international and domestic journals. This paper refers to international and domestic experiences, and sets evaluation criteria by taking the following factors into account. First, SCI and EI journals largely cover top journals in natural sciences and engineering sciences; since the Chinese evaluation criteria in humanities and social sciences have low integration with the international criteria, we rule out the SCI, EI and SSCI journal papers. Second, the criteria of life-time citations and status of being professors are used to evaluate a student’s ability after his or her graduation, so they are not precise in evaluating a new doctoral graduate. Third, it is technically difficult to compare international and domestic dissertations at both vertical and horizontal levels. If we use the Report of Doctoral Education Quality in China, data are only available after 2007. Taking all these factors into account, this paper uses paper publication conditions in domestic journals in social sciences to examine the 2003-2010 doctoral students’ ability. As for the statistical period of time, this study assumes that dissertation is the final output of a doctoral student’s work in his or her last three years of the whole program. Further, in order to ensure a reasonable period of time for statistics, this paper also takes other time factors into consideration. First, according to the NEDD Selection Rules, the candidate dissertation should be finished by a student who gained a doctoral degree one year prior to the selection year. 5 Second, given the fact the dissertation is usually the work finished in the last year of the program and it is published in the following year due to proof-reading and revision requirements, this paper takes the last 4 years (including current year) prior to the selection year as the statistical period of time to collect samples. 10

For the statistics of doctoral students’ papers published in domestic journals in the social sciences, the data of this paper are based on the indexes of core journals compiled by the China Academic Journal Network Publishing Database. 6 There is a huge variety of journals in the database, and the quality of the journals could be very different. So only checking the quantity of papers published in these journals may not be enough to show one’s academic ability. Thus, referring to the 2011 Index of Core Journals in Social Sciences of Sun Yat-sen University, 7this paper groups all journals into 3 categories: Grade 1, 2, and 3. For the papers published in the Grade 1 or Grade 2 journals, this paper gives them a weighted score of 3 or 1 respectively. For those papers published in the Grade 3 journals, because of their poor quality, the paper will not give them weighted scores. Therefore, a student’s academic ability can be simply described as within the last 4 years (including current year) prior to the selection year, the quantity of a student’s papers in Grade 1 journals multiplied by 3 plus the quantity of papers in Grade 2 journals multiplied by 1. The equation is: A student’s academic ability = quantity of papers in G1 journals×3 ＋ quantity of papers in G2 journals×1

(2)Advisor Academic Ability （teaabil）

This paper also controls the advisors’ ability, and the scoring standards are similar to the standards set for students. The difference, however, is that an advisor’s academic ability should not be checked only at a given period of time, and all academic achievements should be taken into account. Thus, when calculating an advisor’s ability, this paper takes all years before the NEDD selection year (including the current year). An advisor’s academic ability can be simply described as within all years (including the current year) prior to the selection year, the quantity of a his or her papers in Grade 1 journals multiplied by 3 plus the quantity of papers in Grade 2 journals multiplied by 1. The equation is: An advisor’s academic ability = quantity of papers in G1 journals×3 ＋ quantity of papers in G2 journals×1

(3)Advisor In-school Administrative Power (schauth)

To measure an advisor’s administrative power, this paper grades the 11 administrative level and posts assumed by the advisor according to the rules defined by the National Public Servant Provisional Regulations. 8 For example, for the 32 universities at the vice-ministry level, 9their presidents and Party secretaries are at the vice-ministry level with a score of 7; the standing vice-presidents and standing vice-Party secretaries are at the department level with a score of 6, and so on. Other universities at the department level, their presidents and Party secretaries have scores of 6, and the standing vice-presidents and standing vice-Party secretaries are at the vice-department level with scores of 5, and so on. The detailed scorecard is listed as follows (Table 1):

Table 1 Scores of Administrative Power for Administrative Posts of Advisors

Score Administrative Universities at Vice-Ministry Universities at Department

Level Level Level

7 Vice-Ministry President, Party Secretary ——

6 Department Standing Vice-President and President, Party secretary

Standing Vice-Party Secretary

5 Vice-Department Vice-President, Vice -Party Standing Vice-President and

Secretary, Party Secretary of Standing Vice-Party Secretary

Discipline Inspection

Commission

5 Division Dean and College Party Vice-President, Vice-Party

Secretary Secretary, Party Secretary of

Discipline Inspection

Commission

4 Vice-Division Vice-Dean and College Party Dean and College Party

Secretary Secretary

3 Vice-Division Department Chair and Institute Vice-Dean and College Party

Director Secretary

2 Section Department Vice-Chair and Department

Institute Vice-Director Chair and Institute Director

1 Vice-Section —— Department Vice-chair

and Institute Vice-Director

0 Clerk Professor and Research Fellow Professor and Research Fellow

(4) Interaction Term for Academic Ability and Administrative Power（stuabil *schauth） To explore how a student’s academic ability contributes to win the award under the differentiated administrative power of advisors, this paper adds the interaction term for student’s academic ability and advisor’s administrative power in the model. If this variable’s coefficient is positive, that means the greater an advisor’s power, the more his or her student’s ability contributes to win the award. If the coefficient is negative, the conclusion is the opposite. 3). Control Variables: Schools in Project 98510 (sch985) and Disciplines In light of the fact that schools within Project 985 and those outside Project 985 have different abilities to gain education resources, which could give rise to a discrepancy in the educational quality of doctoral programs, this paper set a dummy variable to control for this factor. If a university is within Project 985, then it is coded as 1; otherwise, it is coded as 0. This paper selected seven disciplines: economics, philosophy, law, education, literature, history, and management. If a winner is from a certain discipline, then it is coded as 1; otherwise, it is coded as 0.

3 Model Construction

Given the research questions, this paper adopts the Logit model to evaluate how the administrative power of advisors and other explanatory variables impact on the probability of students to win NEDD. The model is constructed as equation (1) shows:

Logit() Awardit,1,2,3,, a stuabil it  schauth it  stuabil it * schauth it （1） 4,5teaabilit sch 985 6 Subject  7 year it ,

Here, the dependent variable Award represents whether the dissertation is awarded or not; if yes, the value is 1 and if not, the value is 0. Stuabil is the

student’s academic ability value. Schauth is the advisors’ in-school administrative power which is also the variable this paper focuses on. stuabil×schauth represents the interaction between the student’s academic ability and advisor’s administrative power. Teaabil is the value for the advisor’s academic ability. Other controlled variables include whether the student is from a “Project 985” university (shown by sch985), disciplines (subject), and the dummy variable for years (year).

Descriptive Statistics The features of the selected samples in this paper should be examined first, which are shown in Table 2.

Table 2 Distribution of Disciplines and Years of Sample Data

Discipline Year Total Proportion

2003 2004 2005 2006 2007 2008 2009 2010 （%）

Philosophy 5 6 2 4 4 3 7 4 35 8.4

Economics 6 10 10 2 4 6 12 11 61 14.6

Law 9 8 8 5 4 7 11 7 59 14.1

Education 4 4 4 4 8 5 10 11 50 12.0

Literature 16 9 14 10 10 11 19 18 107 25.7

History 5 5 7 6 12 10 9 10 64 15.3

Management* 5 2 3 3 7 6 8 7 41 9.8

Total Samples 50 44 48 34 49 48 76 68 417 100.0

Proportion（%） 12.0 10.6 11.5 8.2 11.8 11.5 18.2 16.3 100.0

As Table 2 shows, among all seven disciplines, literature takes the biggest proportion, accounting for 25.7% of the total samples. It is followed by history and economics, and they account for 15.3% and 14.6%. Law, management, and education account for 15.3%, 14.6%, and 12%, respectively. Philosophy has the smallest proportion, only 8.4%, in all samples. From the perspective of years, 2009 and 2010 account for 18.2% and 16.3% respectively, and they are ranked first and second.

Tables 3 and 4 show the quantity of the awarded and nominated dissertations by year and discipline. Table 3 Dissertations Awarded and Nominated in Each Year

Awarded or Nominated Year Total

2003 2004 2005 2006 2007 2008 2009 2010

Awarded 17 18 18 15 17 17 16 15 133

Nominated 33 26 30 19 32 31 60 53 284

Total Samples 50 44 48 34 49 48 76 68 417

Award Proportion（%） 34.0 40.9 37.5 44.1 34.7 35.4 21.1 22.1 31.9

As Table 3 shows, each year the absolute value of the awarded dissertation 11is around 17; there is only a marginal difference in terms of the absolute value from year to year. If the proportion is checked, 2006 and 2004 are the years with the biggest proportion of awarded dissertations, and they account for 44.1% and 40.9% of the total samples. 2009 and 2010 are the years with the smallest proportion of awarded dissertations, and they only account for 21.1% and 22.1%. The difference of the proportion of awarded dissertations from year to year is because there exists a significant difference in the number of the nominated dissertations in each year.

Table 4 Dissertations Awarded and Nominated by Disciplines

Awarded or Disciplines Total

Nominated Philosophy Economics Law Education Literature History Management*

Awarded 10 20 18 17 37 20 11 133

Nominated 25 41 40 34 70 44 30 284

Total Samples 35 61 58 51 107 64 41 417

Award Proportion（%） 28.6 32.8 31.0 33.3 34.6 31.3 26.8 31.9

If classified by disciplines, as Table 4 displays, the awarded dissertations’ proportion to the total dissertations in the same discipline remains highly constant, which is roughly 1/3 (except for management*) for most disciplines. It is smaller only for philosophy.

To better show the various features of NEDD winners, figures of important variables are displayed here. Figure 2 below compares students’ academic ability between the awarded students and the nominated students in different disciplines.

0 Philosophy Economics Law Education Literature History Management Awarded 6.2 12.35 4.77778 6.588 3.757 2.9 5.364 Nominated 3.68 4.585 3.05 4.824 1.671 2.136 3.767 Figure 2 Students’ Academic Ability It is self-evident that in all disciplines, both the average values and median values of the awarded students are greater than those of the nominated students, which to a certain extent indicates that the academic ability of winners is reliable. The statistical significance test for the difference of the average values of the students’ ability in the compared groups in every discipline reveals that there is a difference under the 1% statistical significance level in economics and literature. There is no difference in the remaining five disciplines. Figure 3 displays the advisor’s ability in the two compared groups.

4.5

3.5

2.5

1.5

0.5

0 Philosophy Economics Law Education Literature History Management Awarded 4 3.9 4.056 3.941 3.27 3.65 4.091 Nominated 3.52 3.073 3.875 3.676 2.714 3.068 3.633

Figure 3 Advisors’ Administrative Power

As shown in Figure 3, in all disciplines the average values of advisors’ administrative power for the group of the awarded students are greater than those for the nominated group. The statistical significance test for the difference of the average values of advisors’ power reveals that there is no difference in all disciplines. The quantities of the awarded and the nominated dissertations are compared; the proportion of awarded dissertations to the total under the different levels of advisors’ administrative power is also calculated. Figure 4 shows these data. The administrative levels at 4 and 5 (the vice-dean and dean in schools at the vice-ministry level) are significantly superior to other levels in terms of the quantities of awarded and nominated dissertations and the proportion of the nominated to the total. The sum of the nominated dissertations at these two levels accounts for 54% of the total nominated. One possible explanation is that the advisors as deans or vice-deans have reputations in academia for their scholarship, and they also have enough administrative power and social influence, so they have an advantage over other advisors. Therefore, the preliminary conclusion is that it is not solely the advisors’ administrative power that determines the award results but it is the combined effect of the students’ ability and advisors’ power that leads to the

17 result. Table 5 below shows the relationships between variables.

Table 5 Matrix of Important Variables

Award stuabil teaabil Schauth

Stuabil 0.243***

Teaabil 0.089 0.368***

Schauth 0.116** 0.069 0.162***

sch985 0.045 0.047 0.091 -0.009

Notes: * Significant at p .1 level ** Significant at p .05 level *** Significant at p .01 level Values in brackets are Z scores.

The variable Award is positively correlated to the students’ academic ability and advisors’ administrative power, which shows that these variables are important in selecting dissertations. The conclusion that the award results are determined by the combined effect of the students’ ability and advisors’ power is proven again. Students’ academic ability is closely related to their advisors’ academic ability, but it is not related to advisors’ administrative power. Advisors’ academic ability is positively correlated to their administrative power, and this demonstrates an old Chinese proverb that officialdom is the natural outlet for good scholars.

Empirical Data Testing and Analyses 1.Analyses of Factors Affecting Award Probability

First, the Logit model in Equation (1) is used to test the overall impact of the advisors’ administrative power and other variables on the award probability. As Table 6 shows, the coefficients of the students’ academic ability (stuabil) in the regression are all positive and they are statistically significant (.01 level). This means that, other variables being constant, the increase of the students’ academic

18 ability will increase their probability to win NEDD. It also proves that when the NEDD Selection Commission makes the award decisions, they consider seriously the students’ real ability, and the selection is with objective academic standards. The coefficients of the advisors’ administrative power (schauth) in the regression are all positive and they are statistically significant (.01 level). This means that, other variables being constant, the increase of the advisors’ administrative power will increase their students’ probability to win NEDD. Although there are no special terms and conditions for those advisors whose administrative levels are higher in the NEDD Selection Rules, in reality, if a candidate’s advisor happens to assume the administrative office at a high level in his or her school, then the advisor’s position will make the candidate better off in wining NEDD. This again proves that with the administrative power and social status, doctoral advisors can impact the allocation of academic resources to a great extent, which benefit them and their students. More importantly, the coefficient of the interaction between the students’ ability and advisors’ power（stuabil×schauth）in the regression is negative, and the interaction is significant at the .05 level. This means that when a student’s ability remains constant, as his or her advisor’s administrative power gets bigger, the contribution of student’s ability to win NEDD becomes less. So the influential factors are more administrative factors rather than academic factors. 12 The possible explanation is that if a professor takes an administrative duty at a high level, for example, the vice-ministry level, in a rigorous national administrative system, then his or her power of allocating resources, social reputation, and human connections will all be strengthened. In the process of allocating academic resources, they will become more able to bargain. Therefore, in contrast to improving their academic ability, it is easier for them to get resources by rent-seeking. Thus, for some students, if their advisors’ have great administrative power at a high administrative level, they are more willing to take good advantage of their advisors’ power and influence rather than their own ability to seek resources. Also, Table 6 shows that the coefficients of the variables for advisors’ academic 19 ability (teaabil) and if a school is in “Project 985” (sch985) are positive, but they are not statistically significant. Table 6 Advisors’ Administrative Power （teaauth）and Award Probability (1) (2) (3) (4) stuabil 0.116*** 0.108*** 0.110*** 0.234*** (3.82) (3.51) (3.58) (3.65) teaabil 0.006 0.004 0.005 (0.92) (0.62) (0.86) schauth 0.155*** 0.287*** (2.59) (3.41) -0.033** stuabil×schauth （-2.13） sch985 0.179 0.183 (0.69) (0.71) Year Yes Yes Yes Yes Constant -1.432*** -1.514*** -2.110*** -2.541*** (-3.20) (-3.26) (-4.14) (-4.91) N 417 417 417 417 Pseudo R2 0.074 0.077 0.089 0.098 Notes: * Significant at p .1 level ** Significant at p .05 level *** Significant at p .01 level Values in brackets are Z scores. 2. Robustness Tests Of the total samples, the nominated dissertations are 3 times that of the awarded dissertations. In different disciplines, or even in the same discipline, there could be a difference in the ability of nominated students. In order to control the influence of disciplines and students’ ability, this paper reassesses the regression model by pairing groups as the first robustness test. The pairing principle is to pair an awarded dissertation with a nominated one by a student with the best academic ability in the same year and from the same discipline. 13The purpose is to find a nominated student in the nomination group whose ability is the closest to the winner’s and to match them. Then, this paper re-runs the regression model with new samples (awarded samples and their matches) constituted by the awarded group and nominated group. The results are shown in Table 7. Table 7 Advisors’ Administrative Power and Award Probability (regression with the awarded samples and their matches only)

(1) (2) (3) (4) (5)

stuabil 0.0318 0.0206 0.1581**

(1.18) （0.72）（2.07）

teaabil 0.0102* 0.0087 0.0112*

(1.79) （1.44）（1.83）

schauth 0.1538** 0.1498* 0.3371***

（1.99）（1.92）（2.74） stuabil×schauth -0.0367**

（-2.02）

sch985 0.0279 0.0079

（0.09）（0.03）

Constant -0.2986 -0.2800 -0.5510 -0.9931 -1.6109**

(-0.52) (-0.52) （-0.94）（-1.40）（-2.03）

Discipline Yes Yes Yes Yes Yes

Dummy Variable

Year Dummy Yes Yes Yes Yes Yes

Variable

Sample Size 246 246 246 246 246

R-squared 0.004 0.010 0.012 0.023 0.036

Notes: * Significant at p .1 level ** Significant at p .05 level *** Significant at p .01 level Values in brackets are Z scores. As Table 7 indicates, according to the one-to-one pairing test, the coefficient of the student’s ability is not significant any more. On the other hand, the coefficient of the advisors’ administrative power is still significant at the .05 level. This demonstrates that when the disciplines and the students’ ability are the same, the advisors’ administrative power does help their students to win NEDD. After researching the “degree of officialdom” of the award results in all social science disciplines, Gu Haibin et al. (2007) found that education, law, and economics

21 are the top 3 disciplines with the highest degree of officialdom. They are followed by management, philosophy, and history. The second robustness test in this paper is to classify all samples into two groups based on the degree of officialdom and re-run the regression model to see if there is a difference in features of the two sample groups. The test results are shown in Table 8. Table 8 Regression Results based on Degree of Officialdom Total Samples Paired Samples Degree of Officialdom High Low High Low （1）（2）（3）（4） stuabil 0.352*** 0.186** 0.296** 0.120 (3.12) (2.34) (2.14) (1.55) teaabil 0.005 0.006 0.007 0.005 (0.65) (0.61) (0.91) (0.45) schauth 0.396*** 0.241** 0.454** 0.261** (2.63) (2.32) (2.01) (1.97) stuabil×schauth -0.054** -0.023 -0.057* -0.032 (-2.21) (-1.00) (-1.90) (-1.46) sch985 0.655 -0.008 0.601 -0.117 (1.56) (-0.03) (1.26) (-0.28) Constant 0.159 0.084 -0.508 -0.639 (0.59) (0.61) (0.72) (0.73) N 170 247 109 147 Pseudo R2 0.038 0.040 0.047 0.050 Notes: * Significant at p .1 level ** Significant at p .05 level *** Significant at p .01 level Values in brackets are Z scores. As Table 8 shows, in the four tests of the two sample groups, the coefficients of students’ academic ability and advisors’ administrative power remain significant. The coefficients in the two groups with a high degree of officialdom are greater than those in the groups with a low degree of officialdom. This means that in the groups with a high degree of officialdom not only does the students’ ability contributes more to the award, but so does the administrative power, and the administrative power plays a more and more important role as the students’ ability improves. In the groups with a low degree of officialdom, the interaction is not significant, which indicates that in these groups students are not prone to rely on their advisors’ power to win the award as their advisors’ power increases. However, the advisors’ administrative 22 power undoubtedly still plays an important role in order for students to win. These phenomena are consistent with the main test results. Generally speaking, the higher the degree of officialdom, the greater the administrative power’s impact on winning the award. 3. Further Consideration on Endogeneity The above analyses have not considered one possibility: are the students with stronger academic ability inclined to choose to work with an advisor with stronger administrative power? If this is the case, it can affect our test results. Therefore, we tracked the academic outputs of the awarded and nominated students before and after winning the award to find whether there are differences in the academic ability and potential for students in these two groups. According to the data collected, our test results show that, with other controlled variables, the advisors’ administrative power is significantly correlated to the students’ ability and potential. The regression tests are shown in Table 9. Table 9 Are Good Students Inclined to Choose Powerful Advisors? Outputs Before Outputs After Difference with the Difference with the Award Award Best Best Nominee Before Award Nominee After Award (1) (2) (3) (4) Schauth -0.018 -0.014 -0.286 0.054 (0.14) (0.12) (0.46) (0.35) Teaabili 0.054** 0.039*** 0.024 0.065*** (0.02) (0.01) (0.04) (0.02) sch985 -0.981 -0.144 -0.671 -0.991 (0.79) (0.45) (1.85) (1.27) Teaaward -0.297 -0.116 (0.37) (0.26) Award 2.163 1.671* (1.44) (0.95) Discipline Dummy Yes Yes Yes Yes Year Dummy Yes Yes Yes Yes Constant 1.629** 1.361* 1.913 -0.511 (0.77) (0.79) (2.88) (2.03) N 262 262 99 99 R-squared 0.235 0.213 0.197 0.169 F 5.658 3.679 1.648 1.594 Notes: * Significant at p .1 level

** Significant at p .05 level *** Significant at p .01 level Values in brackets are Z scores. The results in Table 9 show that there is no evidence to prove that there is significant correlation between the students’ ability and the advisors’ power. It means that the endogeneity issue we were concerned about does not exist, so the conclusion that the advisor’s power is conducive to students’ winning NEDD is solid.

Conclusion and Discussion Chinese universities have been haunted by the problem of excessive administrative power and bureaucratization. Ma and Sun (2011) have analyzed the syndromes of the resource dependence caused by excessive administrative power. They hold that it is an education disease for a school with excessive administrative power, which includes the external administrative disease and the internal administrative disease. The former disease is caused by universities’ dependence on the government departments of education, thus making universities subordinate to government departments. The latter disease is caused by the academics’ resource dependence on the administrative staff, thus making sure that academic affairs are dominated by the administrative staff which causes them to lose their proper development direction. In its Outline of State Plans for Medium and Long-term Reform and Development of Education (2010-2020), 14 the Chinese national government put forward in 2010 that the de facto administrative levels and the administrative management model should be removed from universities and unnecessary administrative intervention should be decreased. This by nature indicates the future reform direction and the way to resolve the problem of the excessive administrative power in Chinese higher education. The major finding of this paper is that the administrative power of doctoral advisors holding an administrative post has important influence on allocating academic resources. Specifically, even though a doctoral student's academic ability to win the award matters, the advisor’s administrative power does play an important role in helping the student to get the award. The administrative power also impacts

24 how the students’ academic ability is weighed in the NEDD selection procedures. In other words, the greater an advisor’s administrative power, the lower the standard of evaluating his or her students’ academic ability to win the award. This conclusion is particularly true for those highly bureaucratized disciplines or those with a high degree of officialdom. This is to say, to some extent, the platform of resources allocation has been transformed into an arena of power competition. The administrative power’s impact on allocating academic resources has encroached upon the platform of choosing research projects which is considered fair and impartial by the public. Then it can be imagined that in other areas or on other platforms, the administrative power will play a more outstanding role in allocating academic resources. The allocation of academic resources dominated by the administrative power will seriously impede the sacred academic freedom, distort the pursuit of the truth by academic professionals, and also breed academic corruption. In fact, the root problem is not in Chinese universities per se, but in the relationship between universities and government. If the state remains a central player in allocating resources, appointing personnel, and conferring status on bases other than merit, what besets Chinese universities today will continue to do so tomorrow with or even without a formal administrative rank. The possible policy proposition is to separate the universities’ resources allocation system from the administrative system. Government only plays the role of sponsor; the role of umpire is played by the purely academic institutions or the third-party social agencies. Each level of administrative departments and scholars holding administrative posts do not participate in the selection procedures of research projects and the administration of research grants. The selection and approval of research projects are administrated by the committees made by professors with good academic reputation, but the number of those who hold administrative posts should be controlled at a certain ratio. The special professional agencies and enterprises authorized by government are responsible for supervising and administrating research grants to ensure money is used in a transparent and rational way. The administrative agencies should treat the 25 academic evaluation committees and social supervision agencies with enough respect. Thus, the separation of power among sponsors, professional approvers, and social supervisors, who do their respective jobs well, can ensure academic resources are used effectively. The academic achievements accomplished under this system are able to benefit the long-term national development and promote significant social progress. In one word, all these reforms imply a fundamentally new method of defining the relations between universities and the state. While for a school, schemes of studies may be laid down by governmental officials, scientific research cannot be regulated by decrees of the ruling powers, but can only thrive in full liberty. This was the view taken by the illustrious German educator Wilhelm von Humboldt in a memorial published in 1810: “On the whole the state should not look to them (universities) at all for anything that directly concerns its own interests, but should rather cherish a conviction that, in fulfilling their real destination, they will not only serve its own purpose, but serve them on an infinitely higher plane, commanding a much wider field of operation, and affording room to set in motion much more efficient springs and forces than are at the disposal of the state itself ”(Paulsen, 1908, p.186-187).

Notes

This research project is supported by the National Natural Science Foundation of China (No.70902024 and No. 71003108) and research fund from Sichuan University (skqy 201312).

1 More information at http://news.xinhuanet.com/edu/2009-09/11/content_12032603.htm

2 We selected our samples from the year 2003 because before 2003 only the NEDD winners list was available but the nominees list was undisclosed. Thus, we do not have a control group. The NEDD selection rules and regulations can be retrieved from the website of the China Academic Degree and Graduate Education Information at http://www.cdgdc.edu.cn/xwyyjsjyxx/zlpj/yblwpm/257695.shtml.

3 The winners of NEDD will be financially supported by the NEDD research grant for 5 years.

4 The NEDD selection started in 1999, and it is awarded every year. The awarded dissertations from 1999 to 2002 were publicized without the nominated ones being disclosed.

5 The Article 7 of NEDD Selection Rules defines that the candidate dissertation should be

completed by an author who gains the doctoral degree one year prior to the selection year.

6 The China Academic Journal Network Publishing Database is the largest Chinese academic journal database in the world which covers all disciplines including natural sciences, engineering, agriculture, philosophy, medical science, humanities, and social sciences. Over 96% of the core journals are collected in the database.

7 The 2011 Index of Core Journals in Social Sciences of Sun Yat-sen University is compiled by the Department of Social Sciences of Sun Yat-sen University. In China, universities have their criteria to judge the quality of journals; we adopt Sun Yat-sen University’s index because it is similar to others and universally accepted.

8 According to the National Civil Servants Provisional Regulations which was enacted in 1993, there are 15 levels of public servants. The correspondence between the position and level is: …… (3) ministerial chiefs, provincial chief: three to four; (4) and ministerial deputies, provincial deputy: four to five; (5) The Secretaries chief, departmental heads, inspector: five to seven; ) The Secretary-level deputy departmental level deputy assistant inspector: six to eight; (7) division level chiefs, county heads, researcher: seven to ten; division level deputy, county , assistant researcher: eight to eleven; (9) Section chief level, township level heads, and director of Clerks: nine to twelve; (10) Section deputy, township deputy, deputy director of Clerks: nine to thirteen; 11) Clerks: nine to fourteen; (12) Clerks: 10 to 15.

9 The 32 universities at the vice-ministry level include schools such as Tsing Hua University, Peking University, Renmin University of China, Beijing Normal University, Nankai University, Fudan University, Shanghai Jiaotong University, Xi’an Jiaotong University, Zhejiang University, Nanjing University, Xiamen University, Sichuan University, Wu Han University, Sun Yat-sen University, Lanzhou University, and others. They are all in Project 985.

10 Project 985 is a project that was first announced by Chinese President Jiang Zemin at the 100th anniversary of Peking University on May 4, 1998 to promote the development and reputation of the Chinese higher education system. The project involves both central and local governments allocating large amounts of funding to certain universities to build new research centers, improve facilities, hold international conferences, attract world-renowned professors and visiting scholars, and help Chinese teachers attend conferences abroad. Now there are 39 universities sponsored by this project. 11 Each year, the awarded dissertation is selected from all nominated ones. For clear illustration, we use the winner group (awarded) to refer to those nominated and awarded samples and use the nominee group (nominated) to refer to those dissertations nominated only.

12 If we only check the regression results, we can see that as the student’s academic ability gets better, the role of the advisor’s administrative power gets smaller. However, this paper holds that this causal relationship does not exist in reality or it barely exists. So the regression results show that as the advisor’s power gets stronger, the coefficient of the student’s ability gets smaller.

13 If in one year, in one discipline there are awarded dissertations without sufficient qualified nominated ones to match, we then rule these awarded samples out. We eventually match 123 awarded dissertations with 123 nominated ones, so there are totally 246 paired samples.

14 More information at http://www.moe.edu.cn/publicfiles/business/htmlfiles/moe/moe_177/201008/93785.html

References Cheng, K. M. (1998). Reforms in the administration and financing of higher education. In M. Agelasto, & B. Adamson (Eds.), Higher Education in Post-Mao China, 11-27. Hong Kong:

Hong Kong University Press.

Gao, J. (2012). Higher education policy in China 1992-2010: A genealogical analysis of the enterprise university. Ph.D. dissertation, Faculty of Education, Queensland University of Technology.

Gu, H. & Cao, F. (2007). Understanding the degree of officialdom in Social Sciences (jie du she hui ke xue zhong de guan wei du). Social Sciences Weekly. September, 20.

Ma, J. & Sun, K. (2011). Analysis of the syndromes of the resource dependence in universities (gao xiao xing zheng hua de zi yuan yi lai bing li fen xi). Journal of Beijing Normal University ( Social Science), Vol.3, 40-46.

Murphy, M. (2009).Bureaucracy and its limits: accountability and rationality in higher education. British Journal of Sociology of Education 30:6, 683-695.

Paulsen, F. (1908). German education: past and present (translated by T. Lorenz). New York: Charles Scribner’s Sons.

Samier, E. (2002). Weber on education and its administration: Prospects for leadership in a rationalized world. Educational Management and Administration 30, No. 1: 27-45.

Weber, M. (1958). Religious rejections of the world and their directions. In From Max Weber, ed. H. Gerth and C.W. Mills, 323-59. London: Kegan Paul.

Weber, M. (1974). Max Weber on universities: The power of the state and the dignity of the academic calling in Imperial Germany. Chicago: University of Chicago Press.

Weber, M. (2004). Science as a vocation. In Vocation lectures, ed. D. Owen and T. Strong, 32-94. Indianapolis, IN: Hackett Publishing.

Xiong, B. (2009). The Drawbacks of bureaucratization of universities (gao xiao xing zheng hua zhi bi duan) , Study Monthly, Vol.21, 28-29.

Yang, R., Vidovich, L., & Currie, T. (2007). “Dancing in a cage”:Changing autonomy in Chinese higher education. Higher Education Vol.54, No.4,575-592.

Yao, Y. &Zhou, X. (2007). The yearly output analysis of national excellent doctoral dissertation in education discipline (li nian jiao yu xue xue ke quan guo you xiu bo shi lun wen chan chu fen xi). Journal of Higher Education Vol. 11, 66-73.

Zhang, S., Peng, L., & Gu, J. (2010). The Analysis of advisor of the national excellent doctoral

dissertations(quan guo you xiu bo shi xue wei lun wen zhi dao jiao shi te zheng fen xi). Academic Degrees and Graduate Education, Vol.5, 1-5.