【作者】 冯珂；

【导师】 王守清；

【作者基本信息】 清华大学 ， 管理科学与工程， 2018， 博士

【摘要】 股权结构优化是PPP项目融资和项目治理领域关注的核心问题之一。近年来,公共部门通过投入政府资金(股权投资和政府补贴)正日益紧密的参与到PPP项目融资中。在此背景下,结合政府资金对股权结构的影响,研究PPP股权结构在不确定条件、多目标约束和动态博弈等情形下的优化行为,既可拓宽对PPP股权结构优化理论的认识,也可为相关决策提供工具。论文首先通过文献调研和理论研究总结了PPP项目的主要资金来源及其特点,并分析了主要利益相关方在PPP项目股权结构优化中的诉求。随后,结合现金流量折现法识别了股权结构决策中的关键指标,并对政府补贴支付方式以及债权投资、社会资本股权投资、政府资金等不同资金的折现率进行了分析。在此基础上,本文提出了PPP项目选择最优股权结构的决策框架以及“三阶段”的决策流程,即“提出初步的股权配置方案—筛选可行的股权配置方案—综合评价方案及调整优化”,为后续优化模型的建立奠定了基础。其次,为考虑不确定性因素对PPP项目未来现金流的影响,论文采用蒙特卡洛模拟对影响PPP项目收益和成本的关键风险因素,包括建设成本、运营成本和客流量等,进行了模拟分析。将蒙特卡洛模拟嵌入到遗传算法适应度函数的计算模块可提高计算速度,保证算法收敛。通过模型在北京地铁四号线数值算例的应用表明:优化后的股权结构要优于项目实际融资安排,公共部门需适当增加财政资金投入以抵消不确定因素对项目收益的不利影响。再次,针对多目标约束的PPP项目股权结构优化,本论文分别以两个子目标,即社会资本股权投资净现值的最大化与公共部门投入的政府资金现值的最小化,代表社会资本和公共部门各自的利益诉求。采用带精英策略的非支配排序遗传算法对该问题进行求解,得出了一组满足双方利益诉求的帕累托解,从而提升了PPP项目管理的灵活性。随后,将模型在某轨道交通PPP项目中进行了应用,并采用逼近理想解法(TOPSIS)对结果进行了评价。最后,基于PPP项目的股权结构决策过程中公共部门与社会资本之间的相互博弈,构建了PPP项目股权结构的不完全信息动态博弈模型,求解了PPP项目的最优股权结构和最优特许价格,并以数值模拟方法分析了各关键定量因素变动对PPP项目最优股权结构的影响。更多还原

【Abstract】 Public-Private Partnership(PPP)mode has become a hot topic in the field of project finance and project governance.Recently,public sectors are increasingly involved in the financing of PPP projects through offering public equity investment or government subsidies.It is highly necessary to incorporate the influence of public funds and consider the optimization of PPP capital structure under circumstances such as uncertain conditions,multi-target constraints and dynamic gaming.Firstly,this paper reviews the main sources and characteristics of PPP financing through literature review and theoretical research.It then analyzes the main interest demands of key stakeholders in the financing of PPP projects.Combined with the method of discounted cash flow,this paper then calculates the key indicators in the optimization of capital structure and clarifies the way of offering government financial subsidy and determining discount rates for debt,private equity and public funds respectively.On this basis,it further puts forward a decision-making framework for PPP capital structure and defines a three-stage decision-making process,including proposing preliminary capital structure scheme,selecting feasible capital structure scheme,comprehensively evaluating and optimizing the capital structure.This framework would act as guidance for the establishment of optimization models in the following chapters.Secondly,to consider the impact of risky factors on the future cash flow of the projects,this paper uses Monte Carlo simulation to simulate the key risk factors that influence the future cash flow of PPP projects,including construction cost,operation cost and passenger flow.By embedding Monte Carlo simulation into the calculation of fitness module of genetic algorithm,the computation speed has been improved and the algorithm convergence is ensured.The application of the model in the numerical case of Beijing Metro line No.4 has proven the validity of the model.Results show that the optimized capital structure has better performance than the project’s original financial arrangement and public sector shall increase the investment of public funds in order to resist the adverse effects of uncertain risk factors.Thirdly,as for the optimization of PPP capital structure under multi target constraints,this paper uses two sub objectives,i.e.maximization of NPVe and minimization of present value of public funds,to represent the interests of private and public sector respectively.It then employs non-dominated sorting genetic algorithm-II(NSGA-II)to solve the problem.A series of non-dominated Pareto solutions is produced as a result for the decision makers to pick up,which largely increases the management flexibility of decision making on this issue.The model is then applied to a numerical case of an urban rail transit PPP project and the results are further selected through Technique for Order Preference by Similarity to an Ideal Solution(TOPSIS)method,which validates the applicability of the model.Finally,based on the mutual gaming process between public and private sectors,a dynamic game model with incomplete information is built.The optimal capital structure and optimal concession price are calculated.And the paper further analyzes the influences of key quantitative factors’ variations on the optimal capital structure.更多还原