【作者】 史云鹏；

【导师】 赵黎明；

【作者基本信息】 天津大学 ， 技术经济及管理， 2013， 博士

【摘要】 能源是人类生存与发展的物质基础，是经济发展的重要驱动因素，同时能源安全也是国家安全的一个重要方面。鉴于能源问题的复杂性，基于系统论的观点从投入产出角度对我国现阶段所面临的能源问题进行了概述，从能源消费总量及能源效率两方面对经济发展与能源消费之间的关系进行了研究。论文的主要成果及创新点如下所示：（1）全要素能源效率指标体系的构建。将环境污染及人力资本因素引入全要素能源效率框架，构建了全要素能源效率指标体系。相比较传统的能源效率指标体系，一方面将环境污染视作生产过程中的一种负产出，考察了能源利用过程中的生态影响；另一方面将人力资本因素作为一项投入要素显明的包含在全要素能源效率框架之中，体现了人力资本对于生产过程的重要性。基于所建立的投入产出指标体系，进一步考察了基于节能、增产、减排联合目标并且能源与其他要素之间不具有可替代性情况下的各地区能源效率水平。（2）能源效率测算模型的构建。在对经济发展优先、节能优先以及能源价格管制等情况进行分析的基础上，确定相应偏好锥的结构，并将其引入CCR模型，构建带有偏好锥的DEA模型，对各地区的全要素能源效率水平进行测算，将传统DEA模型的Pareto解扩展到了非支配解，体现了不同假设前提下，决策者对各投入产出要素的偏好。除此之外，为考察方法不同对测算结果的影响，在全要素能源效率及生产率变化率的测算中，同时使用了数据包络分析方法及随机前沿分析方法，并对两种方法的测算结果进行了比较。（3）能源效率影响因素分析模型的构建。构建了包含单要素能源效率及全要素能源效率两种指标在内的具有联立方程组形式的能源效率影响因素计量分析模型。在考虑能源效率与其影响因素之间的关系时，通过对以往文献研究成果的总结及对影响路径的讨论，将所有影响因素区分为直接因素和间接因素。同时把全要素能源效率与单要素能源效率两种形式的效率指标纳入模型，与直接影响因素一同构建联立方程组模型，并运用似不相关回归（SUR）及PCSE方法，讨论了各个因素与不同的能源效率指标之间的关系以及这种关系的差异。该计量模型具有两方面的优点：一是能够从模型设定的角度降低内生变量的影响；二是避免了以往文献中仅考虑影响因素与单一定义形式的能源效率指标之间关系的局限，有利于消除理论及实践两方面存在的差异。更多还原

【Abstract】 As material basis of human existence and development, energy is an importantdriving factor of economic development, and its security is also an aspect ofnational security. Due to the complexity of energy system, a survey on the energyissues which china faced is made in the view of input and output based on systemtheory, and a research is made on the relationship between the economic developmentand the energy consumption in aspects of both total energy consumption and energyefficiency. The main work and innovations of this dissertation include:(1) Construction of the total-factor energy efficiency input-out index system.Environmental pollution and human capital are introduced into the total-factor energyefficiency index system. Compared to the traditional total-factor energy efficiencyindex system, this dissertation has improvements in the following two aspects:considering environmental pollution as a negative output of the productive process tostudy the ecological effect of energy utilization and embodying the importance of thehuman capital. Then based on the index system, the performance of each DMU isstudied in the situation that the joint goal includes energy conservation, economicgrowth and emission reduction, and there is no alternative effect between energy andother input factors.(2) Construction of the energy efficiency evaluation models. The structure ofthe relative preference cones are defined on the basis of the discussions on thefollowing situations: economic growth is preferred or emission reduction is preferredand the energy price regulation exists. Then the preference cones are introduced intothe CCR model, so the DEA models with the preference cones are built up to measurethe efficiency of each DMU. Through this, the Pareto solutions are developed into thenon-dominated solutions and the different preference to the input and output factorsare considered. Besides, in order to study the affect of evaluation method on themeasurement results, both DEA methods and SFA methods are used to evaluate theenergy efficiency performance and the change rate of total-factor productivity, and thedifference between the results is examined by using the statistical test(3) Construction of the Influencing factors analysis model of the energyefficiency. The model is built up which includes both single-factor energy efficiencyand total-factor energy efficiency and has the form of simultaneous equations. When consider the relationship between the energy efficiency and its influencing factors, thelatters are divided into direct and indirect factors. Then the single-factor energyefficiency and total-factor energy efficiency are both introduced in to build up thesimultaneous equations model with the direct influencing factors, and the coefficientsare estimated both by SUR and PCSE method to study the relations between theinfluencing factors and the different energy efficiency indexes and the difference ofthe relations. This model can both consider the influence of the definition of energyefficiency on the relation between the variables and great possibly eliminate the affectof endogenous variable in model construction.更多还原