资本流动风险主要预警模型述评

Review of Warning Models of Capital Flow Risks

通过对常用的资本流动风险评价模型,如KLR模型、FR概率模型、STV模型、神经网络模型、马尔科夫区制转换模型及其他模型进行分析发现,不同的预警模型均存在着预警效果的优势与不足。未来的预警模型构建需要注意以下几点:一是预警指标的选择范围要广,指标阈值的确定需根据不同国家的国情来设定;二是在考虑预警指标之间线性关系的同时,重视风险发生的非线性关系;三是关注危机爆发的连续性、传染性及时滞性。

By analyzing commonly used capital flow risk evaluation models such as KLR model, FR probability model, STV model, Neural Network model, Markov model, and other types of models, the paper discovered that there exists advantage and disadvantage in different types of waning models. The establishment of future of warning models should highlight the following three points. Firstly,the selection of warning indicators should cover a wide range and index threshold should be based on situation of different countries. Secondly, non-linear relation in the occurrence of risks should also be watched while considering the linear relations. Thirdly, the continuity, contagion and delay of the outbreak of crisis should be watched out.