From the perspective of supervisions, this paper presents a practical path for and analyzes the necessity and distinctiveness of credit risk early warnings. We develop a forward-looking indicator system including regional economy indicators, industry operation indicators and indicators of enterprises' financial conditions, credit behaviors and connected guarantee by taking individual banks as the object of this study which is the basic unit taking credit risks, using the data of selected banks that the regulatory authority have monitored for a long time, as well as the off-site regulatory data and related economic data. We then measure credit risks of banks using a logistic model. . The empirical results show that this early warning model works. For the top ten risky customers of banks, the warning model capturing rate is 65%. To improve this early warning model, we narrow down its applicable zones to the top 200 enterprises in Shandong province, then apply this model to the data of enterprises of Shandong Province before June 30th 2014, and calculate the possibility that the normal class enterprises degrade in the next twelve months to win some time for credit risk preventions. Finally, this paper gives some suggestions on preventing credit risk as well as improving credit risk prevention mechanism for banks.%本文基于监管视角确定了信用风险预警的实现路径，辨析了基于监管视角进行信用风险预警的必要性及其特殊性；以信用风险发生的基本单元即银行客户为研究介质，以监管部门长期监测的银行客户数据、非现场监管数据和经济数据为基础，建立了涵盖客户财务指标、信贷行为、关联担保、区域经济、行业运行的前瞻性指标体系；利用Logistic模型对客户的信用风险进行度量。实证检验结果显示，模型预警效果良好、风险得分前十名的客户，预警抓获率高达65%。为进一步提高预警效率，使之能更好地用于实践，本文通过综合权衡模型的观察面和覆盖面，将风险得分排名前200名的客户确定为预警适宜区，再将2014年6月批次的山东省银行业客户数据代入预警模型，测算当前正常类客户在今后十二个月内变为不良客户的可能性，为信用风险防控提供抓手并赢得宝贵时间。此外，本文还根据预警分析提出监管部门进行信用风险防范、化解的意见建议，以推动银行业机构完善信用风险防范的长效机制。
展开▼
