基于DEA方法的科研机构方向规模收益研究

Returns to scale is an important issue in the analysis of organizational performance. It can help decision-makers decide if the size of the organization should be expanded or be reduced. Traditional definition of returns to scale in economics is based on the idea that changes of outputs are caused by those of all inputs in the same proportion (radial variation). But for non-traditional production process with multi-inputs and multi-outputs (such as research institutions), inputs are not always be expanded in the same proportion. In this case, how to describe returns to scale is an important issue. In response to this issue, this research firstly aims to define directional returns to scale and directional scale elasticity from the global and local perspective, and gives specific functional formulas to these two new concepts. Secondly, this research aims to use DEA to simulate the production function, introduce directional returns to scale and directional elasticity into DEA framework, and propose the estimation method for directional returns to scale of research institutions, including ① To extend the definition of returns to scale in the framework in the framework of DEA method, and shows how to estimate directional returns to scale and directional elasticity of research institutions using DEA models; ② To propose the method to detect the optimal input directions for research institutions; ③ To propose the qualitative methods for simple estimation of directional returns to scale for research institutions; ④ To investigate the relationship between the directional returns to scale and congestion.Also, this research will conduct empirical study about directional returns to scale of 16 basic research institutes in Chinese Academy of Sciences using the above methods from two dimensions to analyze their directional returns to scale and optimal input directions, which can provide solid base for decision makers to make developmental strategy in the future.

规模收益（Returns to Scale）是组织绩效分析所关心的一个重要问题，可以帮助决策者决策应该扩大还是减小组织规模。经济学中传统规模收益的定义是基于生产要素按相同比例变化而引起的产出的变化率（径向变化）。对于科研机构来说，往往并不是按相同比例扩张各类投入要素。针对这种现象，本项研究一是在帕累托偏好下，从全局和局部视角定义方向规模收益和方向规模弹性的概念，并给出具体函数表达；二是通过DEA方法模拟科研机构生产函数，并将方向规模收益的概念引入到DEA框架中，提出科研机构方向规模收益的分析方法，具体包括：①采用方向规模收益的思想，扩展DEA方法框架下规模收益的概念，并给出科研机构方向规模收益分析方法和方向规模弹性的定量表达；②研究提出科研机构最优投入方向的分析方法；③研究提出DEA方法框架下方向规模收益简便的定性判断方法；④研究提出方向规模收益与阻塞（Congestion）效应的关系。

规模收益（Returns to Scale）是组织绩效分析所关心的一个重要问题，可以帮助决策者决策应该扩大还是减小组织规模。经济学中传统规模收益的定义是基于生产要素按相同比例变化而引起的产出的变化率（径向变化）。对于科研机构来说，往往并不是按相同比例扩张各类投入要素。针对这种现象，本项研究一是在帕累托偏好下，从全局和局部视角定义方向规模收益和方向规模弹性的概念，并给出具体函数表达；二是通过DEA方法模拟科研机构生产函数，并将方向规模收益的概念引入到DEA框架中，提出科研机构方向规模收益的分析方法，具体包括：①采用方向规模收益的思想，扩展DEA方法框架下规模收益的概念，并给出科研机构方向规模收益分析方法和方向规模弹性的定量表达；②研究提出科研机构最优投入方向的分析方法；③研究提出方向规模收益与阻塞（Congestion）效应的关系。