更有競爭力的學術遊戲？邁向頂尖大學計畫之政策效果分析

Abstract

1980起各國的科研政策深刻地影響了研究機構的發展，大量的國家科研經費也成為許多研究機構的主要研發經費來源。自2006年起，臺灣的教育部啟動了兩期長達十年的邁向頂尖大學計畫（俗稱五年五百億）。本研究從政府研究資訊系統的資料中，整理出各個大學及重要研究機構的公部門補助經費——著重在各機構在科技部研發經費的競爭成果，並且使用應用個體經濟學中的差異中差異估計計量方法，試圖估計出邁向頂尖大學計畫的政策效果。從回歸結果中，可以發現邁向頂尖大學計畫有以下五點政策特徵： 1. 頂尖大學比非頂尖大學每年可以多爭取到高達兩億的科技部學術補助經費； 2. 頂尖大學每多拿一億邁向頂尖大學計畫的經費，可以多獲得約莫三千萬的科技部經費； 3. 在醫學及工程領域上，邁向頂尖大學計畫才有顯著的政策效果； 4. 邁向頂尖大學計畫長達兩年的延遲效應及可能的後期累積效果； 5. 如果將頂尖大學分成三組，其中的頂尖組比其他兩組出現更為有效的政策效果。 從以上的計量結果中，本研究進一步推論出已經結束的邁向頂尖大學計畫可能有以下的政策意涵： 1. 差異中差異方法可以提供不錯的科研政策的政策效果評估； 2. 精準的計量政策評估應該被納入未來的政策規劃之中； 3. 未來對於年度評鑑及期中評鑑方法的改善應該要納入更長的政策延遲效果； 4. 學術社群的複雜性很難涵蓋在一個各領域、各大學標準一致的科研政策。 希望本研究以上的成果及結論，可以提供未來臺灣的科研政策及經費補助改善指引。

Since 1980s, research policies have dramatically influenced most of the research institutes, and abundant public research fundings has become the main financial support of many institutes, especially the university. In the 2006, Ministry of Education(hereafter MOE) in Taiwan had implemented the most important research policy, the project of World Class University(hereafter WCU), in the following decade. In this research, I sort out the annual funding, from Government Research Bulletin dataset, of each research institute and university, especially a public funding competed from Ministry of Science and Technology(hereafter MOST). As a result of Difference-in-Differences estimate, there are five distinguishing features: 1. The WCU project has a policy effect of two hundred millions in general 2. For every hundred million WCU funding input, it is expected to compete for about 30 million dollars in academic subsidy from MOST 3. Only two subfield receive a significant effect from the WCU project 4. There might be an accumulative effect on investment of public research funding 5. The policy effect of top group financially leading in a MOST ranking is a leap in magnitude comparing with other groups. Besides I conclude four policy implications from the previous results: 1. An intent policy assessment is possible under a careful econometric evaluation; 2. The accurate evaluation should be integrated into the planning of any new research policy; 3. The reform of both the annual and the midterm evaluation should take the delayed effect of the policy into consideration; 4. A homogeneous research policy can rarely sustain the complexity of our academic community. I wish that these results and implications can leave a more realistic assessment to the preceding funding project, and hopefully give some insight on the research policy to the following generations in Taiwan.