By The Method Of Stochastic Frontier Analysis Starting Fund Performance Evaluation And Its Influencing Factors

[Objective](1) Using stochastic frontier analysis estimates the technical efficiency of the2007Beijing Medeicine Development Foundation, analyze the influencing factors of the performance.(2) To discuss the suitability of using stochastic frontier analysis to evaluate the scientific research foundation’s performance.[Methods](1) Using Indicators Equivalent Value method to transform the degree and the professional title of participants into comprehensive evaluation scores; Using Indicators Equivalent Value method to give different scores to different level, and then calculate the equivalent value of the7outputs score.(2) Using Factor Analysis method to integrate the average degree, the average professional title and the average age indicators; Using Factor Analysis method to integrate7outputs indicators.(3) Using stochastic frontier analysis to construct production function and stochastic frontier model, and to estimates the technical efficiency of the2007Beijing Medeicine Development Foundation, analyze the influencing factors of the performance.(4) Statistical analysis using the internationally accepted SAS (9.13) statistical software; Using the most commonly used stochastic frontier analysis software Frontier4.1to do the Stochastic Frontier Analysis.[Results](1) Sience the selection of indicators and measurement is an important prerequisite for performance evaluation, our study based on the principles of the scientific and practical to filter indicators. We selected the appropriation of funds, the number of participants, average degree of participants, the average professional title of participants and the average age indicators of participants as the input indicators and used Factor Analysis method to integrate the average degree, the average professional title and the average age indicators; We selected the numbers of thesis published, the numbers of books published, the numbers of patents, the numbers of awards, the numbers of talents trained, the numbers of follow-up projects that group members in charge of and other outputs as the output indicators; We selected the institution level of the project leader, project leader’s degree, leader’s professional title, leader’s age, participants’ average degree, participants’average professional title, participants’average age as the influencing factors.(2) Our stochastic frontier based on basic Cobb-Douglas production function and the Battese&Coall (1995) model. The Stochastic frontier model get though the hypothesis testing, Log-likelihood function value is252.18, Likelihood statistic value is47.84. p<0.01; y=0.7344, t=9.88, P<0.01, there is technical inefficiency exist in the model. It is to say that there is73.44%reason why the real output deviation from the frontier output could explaine by technical inefficiency. The coefficient of elasticity of funds P1=0.0096, t=7.63,P<0.01. The coefficient of elasticity of human ability β32=0.0046. t=2.28, P=0.0113. The quality of participants was not significant. t=0.51, P=0.3050. The elasticity of funds and human ability is less then1, the input-output is diminishing scale.(3) The average technical efficiency of the2007Beijing Medeicine Development Foundation is0.77, the median technical efficiency is0.84. The maximum of technical efficiency is0.97and the minimum of technical efficiency is0.21, there is a big gap between each project teams. The average output score of2007is0.18, the Frontier output could get0.23, there is0.05score of potential to enhance.(4) The performance of Independent innovation projecsts were the best, its average technical efficiency is0.85, and followed by the key supported projects with its technical efficiency0.76, Joint research projects were worst, with its technical efficiency0.44. The average output score of2007Joint research projects is0.25, the Frontier output could get0.57, there is0.32score of potential to enhance. The average output score of2007key supported projects is0.23, the Frontier output could get0.30, there is0.08score of potential to enhance. The average output score of2007Independent innovation projects is0.14, the Frontier output could get0.16, there is0.02score of potential to enhance.(5) Project performance is divided into four modes, there are189(65.85%) projects in Low-input high-performance,62(21.6%) projects in High-input low-performance,23(8.01%) projects in High-input high-performance,13(4.53%) projects in Low-input low-performance.(6) In the8influencing factors we concidered in our study, two factors reached the level of significance, they are project category and the education background of program leader. Parameter estimates of project category factor are as follow,51=-0.5272, t=-3.48, P<0.01. Parameter estimates of the education background factor of program leader are as follow.δ3=-0.0719, t=-1.95, P=0.0256. These two factors can explain the actual output deviates from the frontier output preliminary. After further analysis, we found that the institution level of the project leader, project leader’s degree, leader’s professional title, leader’s age have influence on technical efficiency.[Conclusion](1) The performance of2007Beijing Medeicine Development Foundation was good in general. However, the gap between the actual output from the frontier output is still exist. The main reason for this gap is the technical inefficiency, Including management control, staffing levels and the level of effort etc..(2) The scale of2007Beijing Medeicine Development Foundation input is ample. If wants to increase outputs, we can’t expect the output increase.(3) The main influencing factor is project category. Independent innovation projecsts were the best, followed by the key supported projects, Joint research projects were worse.(4) The institution level of the protect leader and the level of project leader are the influencing factors.(5) In joint research projects and the key support projects, we can find that the influence of the institution level of the project leader, project manager’s degree. professional title, age to technical efficiency is very obvious. This indicates that the higher inputs and most difficult the projects were, the requirement of quality and level were higher. On the other hand, the influence of the institution level of the project leader, project manager’s degree, professional title, age to technical efficiency is not much obvious in Independent innovation projecsts.(6) By using stochastic frontier analysis estimates the technical efficiency of the2007Beijing Medeicine Development Foundation, we confirmed that the stochastic frontier method used in scientific research foundation performance evaluation is scientific and feasible. This method can accurately describe the input-output level and identify factors that affect performance effectively, it has great significance to inprove the management level and the the level of performance of the future scientific research foundation. We get several enlightenments as follow.(1) We should track and investigat projects in a longer period of time.(2) We should improve the performance of projects, especifically for those high input projects.(3) The Independent innovation projecsts and the key supported projects should be taken by those who are possessed with a wealth of experience.