应用随机前沿法分析评价首发基金绩效及其影响因素

发布时间：2018-01-07 14:28

本文关键词：应用随机前沿法分析评价首发基金绩效及其影响因素　出处：《北京中医药大学》2013年硕士论文　论文类型：学位论文

【摘要】：目的：(1)应用随机前沿法分析评价2007年首发基金的绩效,探讨绩效的影响因素。(2)探讨应用随机前沿法对科研基金绩效的分析与评价。方法：(1)采用指标当量值法,将各项目团队参加人员的学位、职称转化为综合分值；采用指标当量值法,将不同等级的产出赋予分值,计算7项产出的当量值得分。(2)应用因子分析对课题组人员平均学位、职称、年龄进行综合；应用因子分析对7项产出指标进行综合。(3)采用随机前沿分析,构造生产函数及随机前沿模型,测算2007年首发基金项目的技术效率,分析影响技术效率的因素。(4)统计分析使用国际通用的SAS(9.13)统计软件包；随机前沿分析采用目前最常用的随机前沿分析软件Frontier4.1。结果：(1)指标的选取及测量是绩效评价的重要前提,本研究基于科学性和可行性原则对指标进行筛选。确定的投入指标为拨款金额、课题参加人数、项目组平均职称、项目组平均学历、项目组平均年龄,其中项目组平均职称、项目组平均学历、项目组平均年龄使用因子分析综合为人力质量指标。确定的产出指标为发表论文、出版著作、专利、成果奖、人才、课题组成员后续承担课题及其他产出成果,使用因子分析将这七项指标综合为产出综合指标。确定的影响因素指标为项目类别、项目负责人所在的单位级别、项目负责人职称水平、项目负责人学位水平、项目负责人年龄、项目参加人的平均职称水平、项目参加人的平均学位水平、项目参加人的平均年龄。(2)本研究选择最基本最常用的柯布-道格拉斯生产函数,采用BatteseCoall(1995)模型构造随机前沿面。随机前沿模型通过假设检验,对数似然函数值为252.18,似然比检验统计量为47.84,P0.01,模型具有较复杂的结构,有必要使用随机前沿法。Y=0.7344,t=9.88,P0.01,模型中存在技术无效率,实际产出偏离前沿产出的因素中有73.44%是由技术无效率造成的。资金的弹性系数β1=0.0096,t=7.63,P0.01。人员数量的投入弹性β2=0.0046,t=2.28,P=0.0113。人力质量的投入弹性系数未达到显著水平,t=0.51,P=0.3050。拨款金额弹性与参加人员数量弹性之和为0.01421,说明首发基金2007年投入产出处于规模效益递减状态。(3)2007年首发基金项目平均技术效率为0.77,中位数为0.84。2007年项目最大技术效率0.97,最小技术效率0.21,差距很大,各课题组的绩效水平很不均衡。2007年平均实际产出得分为0.18分,前沿产出可达到0.23分,在投入不变的情况下,技术效率还有0.05分的上升潜力。(4)联合攻关项目平均技术效率为0.44,实际产出得分为0.25分,前沿产出为0.57分,即在投入不变的情况下,联合攻关类项目的技术效率还有0.32分的上升潜力;重点支持项目平均技术效率为0.76,实际产出得分为0.23分,前沿产出为0.30分,即在投入不变的情况下,重点支持类项目的技术效率还有0.08分的上升潜力；自主创新项目平均技术效率为0.85.实际产出得分为0.14分,前沿产出为0.16分,即在投八不变的情况下,自主论新类项目的技术效率还有0.02分的上升潜力。(5)将项目绩效分为4种模式,其中低投入高绩效的项目有189个(65.85%)：其次是高投入低绩效的项目62个(21.6%)；高投入高绩效模式的项目只有23个(8.01%)。低投入低绩效的项目有13个(4.53%)。(6)本研究考虑的8项影响因素中,项目类别、项目负责人的学位这2个因素达到了显著性水平。项目类别参数估计值δ1=-0.5272,t=-3.48,P0.01。项目负责人学位参数估计值63=-0.0719,t=-1.95,P=0.0256。说明这两个因素可以初步解释实际产出偏离前沿产出的原因。结论：(1)2007年基金绩效水平总体较好,但实际产出距离前沿产出间仍存在23%的差距。(2)2007年首发基金总体投入水平比较充分,想进一步增加产出,不能依赖增加投入。(3)项目类别是导致实际产出与前沿产出间存在差异的主要影响因素,项目类别越高的项目技术效率越低。(4)项目负责人水平及负责人单位级别均是导致实际产出与前沿产出间存在差异的影响因素。负责入水平及单位级别越高,技术效率越高。(5)投入高、难度大的项目对负责人所在单位的基础条件、科研环境及科研人员素质水平的要求较高；自主创新项目难度小,对承担单位科研环境及负责人水平要求不高。(6)通过本研究将随机前沿法应用于首发基金2007年的绩效评价实例,证实随机前沿法应用于科研基金的绩效评价是科学可行的。针对本研究的结果及讨论,得到以下几点启示：(1)应在结题后更长时间进一步对结题课题进行追踪调查评价,以反映更真实全面的绩效水平。(2)应注重科研效率的提高,尤其是针对投入较大的课题,其绩效水平的提高对于首发基金总体绩效水平的提高是十分有意义的。(3)对于联合攻关、重点支持类项目,应该更多安排高级别的单位、高水平的负责人承担。(4)自主创新项目的投入少、难度小,且对环境影响要求较低,这类项目适合在各级单位及各种水平的负责人中推广
[Abstract]:Objective: (1) using stochastic frontier analysis to evaluate the performance of the first fund in 2007 and explore the influencing factors of performance. (2) to explore the application of stochastic frontier approach to the analysis and evaluation of research fund performance.
Methods: (1) the index of equivalent value method, the project team personnel in the degree title into a comprehensive score index; using equivalent value method, different levels of output given scores, calculate 7 output equivalent value score. (2) the application of factor analysis to research group average degree, title. The age of synthesis; the application of factor analysis to synthesize 7 output indicators. (3) using stochastic frontier analysis to construct the production function and stochastic frontier model, the calculation of technical efficiency in 2007 starting fund projects, analysis the influence factors of technical efficiency. (4) statistical analysis using the international general SAS (9.13) software package; the stochastic frontier analysis using the most commonly used stochastic frontier analysis software Frontier4.1.
Results: (1) the selection of indicators and measurement is an important prerequisite for performance evaluation, this study is based on the scientific and feasible principles of index selection. The input index to determine the amount of funding, the number of participating the project, project group average project title, average degree, average age group, the average group project title project the average degree, the average age of the project group using factor analysis comprehensive human quality index. The output index determined to published papers, publications, patents, awards, personnel, members of the project group subsequent bear project and other outputs, using factor analysis the seven indicators as a comprehensive index to determine the impact of factors of output. The index for the project category, unit level project manager of the project responsible person in charge of the project title level, degree level, project leader in the average age, post project Said the project level, average degree level in the average age of the participants in the project. (2) the selection of the most fundamental and most commonly used Cobb Douglas production function, using BatteseCoall (1995) model of stochastic frontier. The stochastic frontier model through the hypothesis test, the log likelihood function value is 252.18, the likelihood ratio test the test statistic is 47.84, P0.01 model has a complex structure, it is necessary to use the stochastic frontier method.Y=0.7344, t=9.88, P0.01, the existence of technical inefficiency factors, actual output deviates from the frontier output in 73.44% is composed of technical inefficiency caused by the beta 1=0.0096, the coefficient of elasticity of funds t=7.63, P0.01. the number of inputs elastic beta 2=0.0046, t=2.28, the input elasticity coefficient P=0.0113. human quality did not reach significant level, t=0.51, and P=0.3050. in the amount of funding the number of personnel and the elastic elastic was 0.01421, indicating the first In 2007 the fund input and output in economies of scale diminishing. (3) the average technical efficiency of fund project starting in 2007 0.77, median 0.84.2007 project technical efficiency is 0.97, the minimum technical efficiency of 0.21, the gap is large, the research group's performance level is not balanced.2007 average annual real output score of 0.18 points, the frontier output can be achieved 0.23, investment in technical efficiency unchanged, there are 0.05 points up potential. (4) the average technical efficiency of joint research project 0.44, the actual output score of 0.25 points, the frontier output is 0.57 points, namely in the input unchanged, technical efficiency of joint research projects and 0.32 point rise potential; key support for the project the average technical efficiency is 0.76, the actual output score of 0.23 points, the frontier output is 0.30 points, namely in the input unchanged, key technical efficiency support project and 0.08 The potential to rise; the average technical efficiency of independent innovation project for the 0.85. actual output score of 0.14 points, the frontier output is 0.16 points, namely in the eight investment under the same technical efficiency on the new independent projects and 0.02 points up potential. (5) project performance is divided into 4 types, including low input and high performance project 189 (65.85%): the second is high input and low performance project 62 (21.6%); high investment and high performance project is only 23 (8.01%). There are 13 low input low performance project (4.53%). (6) 8 copies of the project category considering the influencing factors, and these 2 factors for the degree reached a significant level. The value of delta 1=-0.5272, estimates t=-3.48 project category parameters, P0.01. project leader in parameter estimation value of 63=-0.0719, t=-1.95, P=0.0256. shows that the two factors can explain the actual initial output deviation from the frontier output The reason.
Conclusion: (1) the 2007 fund performance is generally good, but the actual output from the frontier output still exist among the 23% gap. (2) starting in 2007 the fund's overall investment level more fully, to further increase the output, cannot rely on to increase investment. (3) the project categories are main factors which lead to the difference between the actual output and frontier output between the project, project technical efficiency of higher categories is lower. (4) the person in charge of the project level and unit level are responsible for the factors leading to the difference of effect between the actual output and the frontier output. In a single level and is responsible for the higher level of technical efficiency is higher. (5) input is high, the basic conditions the person in charge of the unit of the difficulty of the project, the higher requirements of the scientific research environment and scientific research personnel quality level; independent innovation project of small difficulty is not high on the undertaking unit of scientific research environment and the person in charge of the level of requirements (6) through the research. The stochastic frontier method is applied to the performance evaluation of the first fund in 2007, and it is proved that it is scientific and feasible to apply the stochastic frontier method to the performance evaluation of the scientific research fund.
According to the result of research and discussion, the following implications: (1) after the knot should be longer for further follow-up survey and evaluation of examination subject, to reflect more realistic and comprehensive performance level. (2) should pay attention to improve the efficiency of scientific research, especially for the large investment project, the level of performance the first to improve the overall level of performance of the fund is very meaningful. (3) for joint research, focusing on projects, should arrange more higher level units, the person in charge of high level of commitment. (4) independent innovation project investment is little, little difficulty, and low impact on the environment this kind of project, suitable for the promotion of responsible units at all levels and various levels of

【学位授予单位】：北京中医药大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F830.91;F224

【参考文献】