基于Logit模型的中国系统性金融风险预警实证分析

发布时间：2018-06-05 02:25

本文选题：系统性金融风险 + 风险预警系统　；参考：《东北财经大学》2016年硕士论文

【摘要】：自上个世纪以来,全球发生了多次金融危机,对各国经济造成了难以挽回的损失。随着全球化的推进,金融危机的破坏力、传染性显著增强,任何国家都无法再别国的金融风险中独善其身,因此如何有效的监控国内外金融风险并准确的发出预警成为各国学者研究的热点。在2001年加入WTO后,中国在世界经济中扮演了更为重要的角色,对外部经济波动的敏感性也逐渐增加。同时中国的债务、房地产等领域均存在着较多不稳定因素,随着经济增长的减速,这些问题可能进一步爆发出来,给我国金融体系带来风险。因此,建立一个符合我国国情的金融风险预警系统具有十分重要、深远的现实意义。首先,本文基于国内外金融压力指数相关文献,综合考虑后续构建模型的需要,选取产出缺口作为金融风险的代理变量以确定金融风险的爆发时间点与影响长度,并以相关标志性金融事件作为金融风险时间点确定的辅助证据。其次,本文基于已有文献分析不同预警模型的优缺点与适应性,结合中国国情与数据特征选择Logit模型作为预警模型。同时,本文依据国内外预警模型构建经验,选择信贷膨胀缺口、房价指数等作为解释变量。在确定模型与指标后,以2000年到2015年的季度数据构建二元Logit金融风险预警模型。同时,为了研究政策制定与预警模型之间的关系,本文基于前文预警模型研究了不同政策偏好下预警模型的表现。再次,本文进一步探索了多元Logit模型在金融风险预警中多作用,以单因素、多因素两种形式构建了不同的多元Logit模型。同时还比较了二元与多元Logit预警模型各自的优势与特点。最后本文总结了研究结论,有以下四点:使用季度数据的多因素Logit预警模型适合作为中国金融风险的预警模型,该模型联动引入多个层次的指标,能适应中国不同时期不同金融风险的异质性对风险作出更好的预判。实证的结果显示,多因素Logit预警模型的预警正确率以及效率基本满足了目前国外早期预警模型的正确率与效率经验要求。本文根据理论研究选取的信贷、银行脆弱性、宏观背景的相关指标作为单独按预警指标均有效,但单独作为预警指标正确率与效率较差。房价作为预警指标本身几乎没有任何预警能力,但与信贷膨胀缺口指标联合使用能显著提高模型预警能力,显示信贷膨胀与房地产泡沫间的互相促进作用会带来较大的金融风险。实证显示多元Logit预警模型能够有效识别风险后的动荡时期,避免与一般时期混淆,能够很好的运用于中国的金融风险预警之中。但目前的统计数据无法满足该模型对数据量的要求,只能建立简单,不太稳定的预警模型。不同政策倾向的分情景讨论显示,任何模型的效用在风险权重比例与正常时期占全时期的比例相等时最高;同时TPR越高,模型在高风险权重中的表现越好,FPR越低,模型在低风险权重中的表现越好。
[Abstract]:Since the last century, the global financial crisis has caused irreparable losses to the economies of various countries. With the advancement of globalization and the destructive power of the financial crisis, the contagion has increased significantly. No country can be left alone in the financial risks of other countries. Therefore, how to effectively monitor financial risks at home and abroad and accurately issue early warning has become a hot spot of scholars all over the world. After China's entry into WTO in 2001, China has played a more important role in the world economy, and its sensitivity to external economic fluctuations has gradually increased. At the same time, there are many unstable factors in China's debt, real estate and other fields. With the deceleration of economic growth, these problems may erupt further and bring risks to our financial system. Therefore, it is of great importance and profound significance to establish a financial risk early warning system in accordance with the national conditions of our country. First of all, based on the domestic and foreign financial pressure index related literature, considering the need of building the model, the output gap is selected as the proxy variable of financial risk to determine the time point and impact length of financial risk. And related landmark financial events as the financial risk in time to determine the supporting evidence. Secondly, this paper analyzes the advantages and disadvantages and adaptability of different early warning models based on the existing literature, and selects Logit model as early warning model in combination with China's national conditions and data characteristics. At the same time, based on the experience of early warning model at home and abroad, this paper chooses credit inflation gap and house price index as explanatory variables. After determining the model and indicators, a binary Logit financial risk early warning model is constructed based on the quarterly data from 2000 to 2015. At the same time, in order to study the relationship between policy making and early warning model, this paper studies the performance of early warning model under different policy preferences based on the previous early warning model. Thirdly, this paper further explores the multi-role of multivariate Logit model in financial risk early warning, and constructs different multivariate Logit model in the form of single factor and multi-factor. At the same time, the advantages and characteristics of binary and multivariate Logit early warning models are compared. Finally, this paper summarizes the research conclusions, there are four points: the multi-factor Logit early-warning model based on quarterly data is suitable for China's financial risk early warning model, and the linkage model introduces multiple levels of indicators. It can adapt to the heterogeneity of different financial risks in different periods in China. The empirical results show that the accuracy and efficiency of the multi-factor Logit early warning model can basically meet the requirements of the foreign early warning model's accuracy and efficiency. According to the theoretical study of credit, banking vulnerability, macro background of the relevant indicators as a separate warning indicators are effective, but as early warning indicators of accuracy and efficiency is poor. As an early warning indicator, house prices have almost no early warning capability in themselves, but the combination of the credit inflation gap index can significantly improve the early warning ability of the model. It shows that the mutual promotion between credit inflation and real estate bubble will bring greater financial risk. The empirical results show that the multivariate Logit early warning model can effectively identify the turbulent period after the risk, avoid confusion with the general period, and can be used in China's financial risk early warning. However, the current statistical data can not meet the requirements of the model for the amount of data, can only build a simple, unstable early warning model. According to the situation discussion of different policy tendencies, the utility of any model is the highest when the proportion of risk weight is equal to the proportion of the whole period in normal period, and the higher the TPR, the better the performance of the model in the high risk weight. The model performs better in low risk weight.
【学位授予单位】：东北财经大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：F832.1;F224

【参考文献】