【作者】 顾婧；

【导师】 周宗放；

【作者基本信息】 电子科技大学 ， 管理科学与工程， 2009， 博士

【摘要】 风险投资业的发展和壮大离不开完善的评价体系和正确的决策方法。风险投资家合理、正确的决策在很大程度上决定了风险投资机构的整体收益。然而目前国内现有的风险投资项目的评估与决策在理论研究和应用中普遍存在事前评估决策简单化、运作评估决策形式化、退出评估决策空白化等缺陷。对风险投资家决策过程的研究无论是从理论还是从实践方面均处于经验管理阶段，缺乏一套比较完善和合理的体系和方法。风险投资决策研究的滞后已严重制约了风险投资的发展，风险投资实践的发展呼吁着风险投资理论的指导。有鉴于此，本文在已有文献研究的基础上，主要针对风险投资家的事前评估决策和运作评估决策进行研究。本文首先提出了风险企业的信用风险识别方法，并在已知风险企业信用风险结果的基础上，构建了风险投资项目的总风险评估模型，进一步研究了风险投资项目的双方投资水平博弈问题，最后考虑和构建了基于信号学习的风险投资项目中止模型以及在此研究框架下的中止时机决策模型。首先，本文借鉴了可变精度粗糙集(VPRs)模型具有噪声数据的强适应能力和强抗干扰能力的优点，提出了一类基于vPRS的风险企业信用风险识别方法，并用已上市的部分新兴技术企业对其进行实证检验，检验结果表明了该方法具有较好的识别能力。该方法可作为风险投资项目的事前信用风险识别方法，为风险投资家的初始决策提供决策基础。其次，本文在信用风险识别方法提出的基础上，建立了基于不确定语言多属性决策的风险投资项目评估方法。该方法在已知风险投资项目的信用风险评估结果基础上，以风险投资项目的总风险为目标，构造了基于不确定语言变量的多属性决策方法。该方法可以对众多风险投资项目进行风险排序，也可将备选项目与已投成功项目进行比较，形成风险投资项目事前决策方法。再次，本文应用Stackelberg博弈模型，对风险投资项目投资双方的投资水平进行了研究。通过研究发现：风险投资家的总投资水平由债权“激励因子”和其边际投资机会成本决定，而股权投资水平由其绝对风险厌恶因子、主观收益率的参数以及“激励因子”决定。尽管风险投资家较企业家风险厌恶，但是后者可以通过激励前者，使前者对于风险投资项目的前景较后者乐观。最后，本文针对现有中止决策方法忽视了风险投资项目发展过程中信息的影响的缺点，归纳分析了风险投资项目发展过程中的信息；并以风险投资家观察到的好(坏)信号作为二元学习信号，就好(坏)信号随项目发展存在的三种状态，从贝叶斯后验估计的角度提出了信号学习模型，进而提出了依据蕴含了风险投资家风险态度的“心理阀值”确定外生的中止决策模型。另外，本文在此研究框架下，以风险投资项目的价值变化为内生变量，进一步提出了风险投资项目的中止决策时机模型。该中止方法的提出，为风险投资家及时做出中止决策提供依据，且有效地控制风险投资机构的风险。由于风险投资公司的投资过程较为保密，相关数据的获取较难，为此，本文无法对文中提出的外生和内生的中止模型进行实证分析，仅以算例说明。并通过在第六章中对案例的定性分析，对现有中止决策方法和本文提出的中止决策方法进行了比较。本文提出的初始决策和中止决策方法，改进了我国风险投资决策评价手段和评价方法，完善了风险投资项目投资规范和方法体系，提高了风险投资决策的效率，从微观运行机制方面改善我国风险投资事业的运行。更多还原

【Abstract】 The development and growth of venture capital industry are depended on aperfect assessment system and correct decision-making method. Venture capitalist’sreasonable and correct decision-making influences the overall revenue of venturecapital firms in a large extent. However, there are many defects in the domesticinvestment projects’ assessment and decision-making both in the theoretical researchesand applications at present, such as the simple in the decision-making of priorassessment, the formalization in the decision-making of operation, the blank in thedecision-making of abandon assessment. The study on venture capitalists’decision-making process is still in the stage of experience management whether intheory or practice, and it lacks a relatively perfect system and a sound method. The lagof the research on the venture capital’s decision-making has seriously hampered thedevelopment of venture capital, and the development of venture capital practice needthe advice of venture capital theory.Considering this, this paper studies the decision-making of prior assessment andthe decision-making of operation process about venture capitalist on the basis ofprevious studies. It proposes a credit risk identification method of venture firm firstly,and then constructs an assessment method of total risk about venture capital projects,discusses the game problem of venture capital both sides’ investment amount further,considers and proposes the model of venture capital project’s terminationdecision-making based on the signal learning and termination time model on thisresearch structure finally.Firstly, this paper proposes an identification method based on VPRS which usesthe advantage of VPRS in dealing with noise data, and the empirical testifies themethod by the emerging technology listing firms, the result demonstrates the approachidentifying the credit risk well. This method provides a reference for decision-makingas the identification of credit risk.Secondly, it constructs an assessment method based on uncertain linguisticMADA on the base of the giving of credit risk identification method. It builds a risk assessment method based on the total risk by knowing the result of credit riskidentification. This approach could arrange the venture capital projects by risk, and italso could compare with the invested projects and the projects that to be chooseed toinvestment, it forms a pre-decision-making method of venture capital projects.Thirdly, it analyses the investment amount of venture capital project by applyingStackelberg game model. From the analyses, we conclude that the total investmentamount is determined by the debt incentive scheme and marginal cost of venturecapitalist; and the equity investment amount is determined by the absolute riskaversion factor of venture capitalist, the parameters of subjective income rate andincentive scheme. Furthermore, since the capitalist is assumed to be more risk aversethan the entrepreneur, he is made to be more optimistic than the entrepreneur.Lastly, aiming at the problem that the present termination decision methods ignorethe information’s effect, we induce and analyse the information during the ventureproject development process, and then choose the good (bad) signal observed byventure capitalist as the binary learning signal, considering three stations that thegood(bad) signal observed along with project development, from the view of Bayesianposterior estimate, we propose the signal learning model, then decide the exogenousoptimal stopping point according to venture capitalist’s boundary point psychologically.On the basis of it, we propose a termination time model further which choose the valueof venture capital project as endogenous variable. The construction of the approachprovides the reference for venture capitalist to make termination decision timely, and italso controls the risk of venture capital finns efficiently.In this paper, the initial and termination decision-making methods improve theevaluation tool and assessment method of domestic venture investmentdecision-making, and perfect the venture capital investment system, and improve theefficiency of the venture capital decision-making and the operation of domestic venturecapital industry in the micro-operation mechanism. In the 6th part, we also comparethe existing termination decision-making methods and the methods proposed in thispaper by the analysis of the case.更多还原