硬信息与软信息:孰重孰轻?——非上市中小企业贷款违约风险实证研究  

Hard and Soft Information： Which One Is More Important？ An Empirical Study on Risk of Loans Default for Non-listed SMEs

作　　者：过新伟[1] 王曦[2] 

机构地区：[1]广发银行博士后科研工作站及中山大学岭南学院 [2]中山大学岭南学院

出　　处：《金融学季刊》2015年第2期61-92,共32页Quarterly Journal of Finance

基　　金：国家社会科学基金重大项目(15ZDA014);国家社会科学基金项目(11BJ022);广东省高等学校高层次人才项目(2012);广东省高层次人才(珠江学者)项目;教育部高校重大课题培育项目(2012);中国博士后科学基金面上项目(2013M542221)的资助

摘　　要：本文利用国内某商业银行的中小企业贷款违约数据，采用Logistic模型和随机效应Logistic模型对非上市中小企业贷款违约风险的影响因素进行实证分析。结果表明：非上市中小企业的硬信息指标特别是营运资金比率、负债比率对贷款违约有较大影响；软信息指标中的企业特征特别是企业在人行的信用记录、经营稳定性、水电税费变化趋势和银行信用对违约风险有显著影响；非上市中小企业企业主的个人特征，例如受教育程度、家庭年收入、持股比例和资产抵押率等对违约风险也具有较大影响。预测对比研究发现，软信息指标对贷款违约的预测能力高于硬信息指标，而随机效应Logistic模型的表现要强于Logistic模型；综合使用软信息和硬信息指标建立随机效应Logistic模型具有最佳的预测效果。研究结论对商业银行开展中小企业贷款业务和信贷风险评估具有参考价值。Based on a unique set of SMEs loans data from a domestic commercial bank, we employ the logistic model and random effects logistic model to analyze the determinant of loans de- fault risk for non-listed small and medium-sized enterprises （SMEs）. The results show that： for non-listed SMEs, the hard information indictors especially for working capital ratio and total liabil- ity/total asset, could influence loan default of SMEs significantly; the soft information indictors es- pecially for credit records filed in PBC, operating stability, expense variation tendency of utilities andtaxation, bank credit, could also affect default risk of SMEs greatly ; some personal characteris-ties related owners of SMEs, such as education, annual household incomes, share proportion, ratio of assets mortgage, also have significant influence on default risk of SMEs. Forecast comparisons show that ： the soft information indicators exhibit stronger prediction power than the hard information indictors and random effect logistic model outperforms random effects logistic model; and optimal forecast is obtained when both soft and hard indicators are incorporated in random effect logistic model. Our findings contribute to the credit business of commercial banks to SMEs in both risk measurement and business management.

关 键 词：中小企业 贷款违约 硬信息 软信息 随机效应Logistic 

分 类 号：G33[文化科学]