考虑环境影响的三阶段组合效率测度模型的改进及在研发效率测度中的应用

在比较现有考虑环境因素或/与统计噪音的三阶段组合效率测度模型的基础上，通过引入满足平移不变性的幅度调整测度（RAM）模型与Tobit回归相组合，构建了一个灵活的、无需强制性正向调整环境因素差异性影响的三阶段组合效率测度改进模型--RAM-Tobit-RAM.相对现有模型，改进的组合模型充分利用了RAM模型非径向与非定向测度特点，同时从投入与产出两个方向全面考虑径向与非径向松弛；更为重要的是，充分借助RAM模型平移不变性的优点克服了现有组合模型需借助最值强制性正向转化调整后的投入或/和产出以处理非正值的不足，降低效率估计的偏误.实证研究部分借助这一组合分析程序比较分析了中国省域研发效率.

An enhanced hybrid three-stage model to account for environment effects in measuring efficiency with an application to R&D efficiency

A flexible hybrid three-stage model for efficiency measures, RAM-Tobit-RAM, without arbitrary positive adjustment of the difference of environment effects is proposed by combining rang adjusted measure (RAM) model satisfying translation invariance with Tobit regression, which is based on the discussions and reviews of extant hybrid three-stage models for removing the differential effects in environment factors and/or of statistical noise. In contrast to the existing models, our proposing hybrid model takes advantage of RAM model with no-radical and no-oriented measures which can simultaneously presents radical and non-radical slacks on both inputs and outputs. More important, our model takes full advantage of the translation invariance of RAM model to deal with the non-positive values, which only can work for existing models by arbitrary positive transformation of adjusted inputs or/and outputs in virtue of maximum or minimum but with biased efficiency estimation. In the empirical study, this article applies this new model to measure and compare the research and development efficiency of China's provinces.