【作者】 何祖玉；

【导师】 韩玉启；

【作者基本信息】 南京理工大学 ， 管理科学与工程， 2004， 博士

【摘要】 论文是国家社会科学基金项目《中小企业信用担保研究》(99CJY003)的后续研究。中小企业(Small and Medium-sized Enterprises,SMEs)在当今经济生活中作用日益重要，信用不足和融资困难是其发展过程中面临的主要问题。在此背景下，一种由政府主导建立、专门破解SMEs融资问题的第三方信用工具—中小企业信用担保(Small and Medium-sized Enterprise Credit Guarantee,SMECG)机制应运而生。SMECG在我国尚属新生事物，SMECG机构的功能发挥和自身的持续发展，涉及到运作体系结构、管理经验、价格决定、风险分担与防范等一系列问题，其中信用担保风险管理与控制是最为核心的问题。信用担保是国际上公认的高风险行业。SMECG机构是为中小企业提供信用担保的专门机构，且在一定程度上体现出国家政策的扶持性，其高风险特征尤其明显。如果缺乏有效的风险管理措施，SMECG机构就有可能成为信用风险的“集合体”，赔偿损失的“归集地”，导致信用担保机构陷入难以为继的困境。因此，建立专业化的信用风险管理能力，对于SMECG机构的健康持续发展和功能发挥至关重要。本文针对SMECG风险管理的特点，立足于解决重点和难点问题，主要研究了以下几方面内容： (1) 定性分析了中小企业信用担保风险管理的基本内容。通过“商业银行—SMECG机构—SMEs”与“商业银行—SMEs”交易成本的比较，分析了市场选择运用SMECG这一工具对于其风险管理的基本要求，及所包含的担保项目和机构总体(项目组合)两个层面的风险管理任务。根据SMECG的基本运作模式，分析了SMECG风险管理的目标与原则、组织形式、过程控制、制度设计等，从定性角度研究了SMECG风险管理的基本内容。 (2) 研究了SMECG项目风险评价指标体系及数据处理。基于SMECG机构对SMEs的扶持宗旨，论文在SMECG项目风险评价的指标选择上，注重SMEs的发展性、成长性和创新性，并充分考虑SMECG的操作风险。以SMEs的财务风险、道德风险、市场风险创新能力等10个指标和若干个子指标构建SMECG项目风险评价指标体系。对于指标体系中的不完备数据和主观数据，采用D-S证据融合方法和群决策技术，有效利用专家信息。基于指标体系的复杂性和指标之间的相关性，采用多元统计方法进行降维处理，避免了分析问题的虚假性和复杂性，可提供客观、科学和合理的决策信息。 (3) 建立了基于贝叶斯网络(Bayesian Networks,BN)的SMECG项目风险评价模型和动态贝叶斯网络(Dynamic Bayesian Networks,DBN)的在保项目风险预警模型。基于BN的SMECG项目风险评价模型有效克服了SMECG项目风险评价的小摘要博十论文子样、信息多源、缺损问题，实现了SMECG风险评价的定性与定量结合:根据SMECG项目风险评价结果，建立了担保项目的分类决策与控制标准。与其它风险评价的模型化方法的比较分析表明:基于BN的SMECG项目风险评估方法数据需求量小，数据精度要求低，评价结果更可靠。在运用BN进行SMECG项目风险评估的基础上，进一步引入DBN建立了SMECG在保项目的风险预带模型，实现对在保项目的风险控制。 (4)运用组合风险管理的思想，研究了组合风险的度量和优化技术。通过从机构总体层面度t和优化风险，有助于信用担保机构作好以下几个方面的风险管理:a.为机构的运行有效性(经济、财务评价)提供可量化分析的工具:b.为担保定价、担保方案设计提供客观依据，进而建立机构的可持续发展模式;c.增强担保评审的透明性，减少评审人员的道德风险。 (5)实证分析表明，本文提出的信用管理方法客观、系统、有效。对于SMECG机构充分发挥信用中介、信用扶持作用，有效管理控制担保风险，合理利用有限资源，推动SMECG机构的持续健康发展具有重要意义。更多还原

【Abstract】 The dissertation is the follow-up study of the Credit Guarantee for Small and Medium-sized Enterprise (SMECG) project supported by national social science fund. Small and Medium-sized Enterprises (SMEs) become more and more important in current society and economics, while lacking of credit and capital is main difficulty in the development of SMEs. Based on the background, an important credit instrument (SMECG), has been established by Chinese government for resolving the difficulty of SMEs’s financing. Institution for SMECG (SMECGI) is a new thing in China, its running and development of SMECGI involves system structure, management experience, policy decision, sharing and avoiding risk. Credit guarantee is high-risk industry, while SMECG is much higher risk because of the characteristics of national policy assistance for SMEs. Without effective risk management measure, SMECGI will become SMEs’s credit risk aggregated institute, and go into bankruptcy. So, credit guarantee risk management is the key to SMECGI. specializing capability is important for SMECGI effective and healthy running. In order to resolve the important and difficult problem, the primary work in the dissertation includes:(1) The qualitative analysis for fundamental content of the SMECGI risk management. SMECG mechanism involves three participators, which are commercial bank, SMECGinstitute and SMEs. Comparing with the cost of the tripartite-participators transaction and two-participators’s, we analysis the demand for credit guarantee management. At the same time, we get the conclusion that SMECG project risk management and project portfolio aggregated risk management are two main tasks for SMECGI. The target together with rule, form of organization, process control, rule and policy of the SMECGI risk management based on the operational mode is provided.(2) Index of the SMECG project risk evaluation was built.As SMECGI taking on the task of supporting SMEs, together with considering SMECGI operation risk, the index selection focuses on SMEs’s development, continuous increscent and innovation abability. We use ten indexes to evaluate SMECG project risk such as finance risk, moral risk, and market risk. Every index includes several branch indexes. For imperfect data and qualitative analysis, we adopt D-S Proof Fusion and Group Decision Making Technology to make use of expert knowledge effectively. According to the index system characteristics is complex and correlative, multi-statistical method wasused to reducing the complexity, difficulty and iteration.(3) Model of SMECG project risk evaluation was studied.To the characteristics of SMECG risk evaluation, such as little sample, information diversity and shortage, SMECG risk evaluation model based on Bayesian Networks(BN) was built, which can integrating qualitative analysis and quantitative analysis into one framework. Based on the results of SMECG project risk evaluation, SMECG risk decision criteria has been built. The SMECG project risk evaluation model based on BN has several advantages against other risk evaluation methods, such as less data demand, imperfect data and experts knowledge using, self-learning and more reliable. Based on the SMECG risk evaluation model, Dynamic Bayesian Networks (DBN) was used to build precautionary model for guarantying project, which can evaluate risk dynamically and reduce compensative loss.(4) Portfolio risk measurement and optimization of SMECGI have been studied.We look SMECGI as a whole to evaluate and optimize portfolio risk, which is helpful for the risk management of SMECGI, includes:a. Rate SMECG mechanism operation validity (Economy and Financial).b. Provide the method of valuation the price of the guarantee business, design the strategy. Then we can establish the continuous development pattern for SMECG institute.c. Supply SMECG institute for risk control objective, Criterion that can be measured. (5)From the example shows that the methods of the dissertation can realize theSMECG risk management. Which can advance healthy更多还原