节点文献
中国金融不稳定的计量研究
Financial Instability in China: an Econometric Study
【作者】 王妍;
【导师】 陈守东;
【作者基本信息】 吉林大学 , 数量经济学, 2014, 博士
【摘要】 国际金融危机的爆发引发了学术界和政府部门对金融监管政策的反思,过去以资本监管为核心的微观审慎监管政策,加之以控制通货膨胀为目标的货币政策,已经不足以维护金融体系的稳定。以防范系统性金融风险、维持金融稳定以及保证实体经济增长为目的的宏观审慎监管方式,成为后危机时代各国金融监管的主要发展趋势。加强宏观审慎监管,并将其与微观审慎监管进行有机结合,标志着金融监管理念和方式的重大调整。然而,目前关于金融不稳定、金融周期和系统性金融风险的概念尚未达成明确统一的认识,在刻画金融风险演化机制、量化金融不稳定以及建立金融-宏观关联模型等方面同样缺乏深入的研究,宏观审慎监管工具的开发和实施也因此受到了限制。虽然我国尚未出现内部风险引发的金融危机,但是近些年金融系统中的不稳定现象越发增多,金融风险呈现不断积聚的状态。因此,迫切需要在此背景下研究我国的金融不稳定,揭示风险在我国金融系统中的生成演化过程以及对实体经济的溢出影响,度量我国金融不稳定状态并识别金融周期特征,为设计有效的宏观审慎监管工具并实施金融监管政策提供依据。本文的主要研究内容和研究结论有:第一,金融不稳定的理论分析。第2章中通过梳理以往的研究,界定金融不稳定的内涵,阐述其产生原因、基本属性以及监管方式的历史演变。本章中对静态的金融不稳定和动态的金融不稳定给出明确的定义,并将金融不稳定的生成演变划分为金融脆弱性、金融压力和系统性金融风险三个主要阶段,介绍了各个阶段中金融风险的表现特征;给出金融不稳定产生原因的传统的传染理论和内生的金融不稳定理论,并主要从内生的金融不稳定理论的角度详细刻画了金融不稳定的产生过程;在金融不稳定基本属性的分析中,阐述了金融不稳定的基本特征、时间和横截面维度、监管政策以及直接和间接的经济成本;最后,以2008年始于美国的金融危机为分界点,介绍了金融不稳定的监管框架的历史演变。第二,金融不稳定的计量分析。本文的第3、4和5章,从金融不稳定的主要阶段,即金融脆弱状态、金融压力状态和系统性金融风险三个方面出发,分别采用计量模型度量金融不稳定并分析金融不稳定的动态变化特征。第3章的金融脆弱性计量分析中,对同步金融变量采用马尔科夫区制转移的状态空间模型,并通过传统的极大似然估计和贝叶斯吉布斯抽样估计,提取代表金融系统脆弱状态的共同因子,即金融脆弱指数;进一步分析我国金融系统脆弱状态的动态变化和区制特征;最后通过线性和非线性格兰杰检验,研究了金融不稳定与宏观经济之间的领先滞后关系。计量分析结果表明,我国金融系统具有“高金融脆弱“和“低金融脆弱”两个区制状态,并且具有非对称的区制持续特征,在“低金融脆弱”区制的持续概率要更大,持续时间要更长;金融脆弱性的周期变动幅度会受到各项金融政策以及财政政策和货币政策的影响,并且在异常的高金融脆弱状态后,将出现更加深度的低金融脆弱状态;我国金融不稳定会对宏观经济产生影响,并且两者之间存在复杂的非线性的影响和传导机制。第4章的金融压力计量分析中,以金融压力的表现特征为基础,选取银行部门、证券市场和外汇市场中的变量构建金融压力度量指标体系,并合成了代表金融系统压力状态的指数,即金融压力指数,以及银行压力指数、证券市场压力指数和外汇市场压力指数三个子指数,分析了我国金融压力的动态变化;进一步地,对金融压力指数建立持续期依赖的马尔科夫区制转移自回归模型,分析我国金融压力的区制特征和持续期依赖特征。计量分析结果表明,我国的金融系统具有“高金融压力”和“低金融压力”两种压力状态,并且更易从高金融压力区制转移到低金融压力区制;金融压力存在正的持续期依赖特征,即金融系统在某一金融压力区制持续时间越长,越容易转移到另一金融压力区制中。第5章的系统性金融风险计量分析中,使用基于极值理论的极端分位数回归方法度量了我国公开发行上市的33家金融机构的系统性金融风险贡献,并识别出我国的系统重要性金融机构;通过面板数据回归模型,分析了金融机构系统性金融风险贡献的主要影响因素。计量分析结果表明,单纯使用在险价值度量金融机构风险可能存在偏误,无法衡量金融机构与金融系统之间的关联影响;我国的系统重要性金融机构基本为银行类金融机构,并且股份制商业银行对金融系统的风险贡献相对更高;金融机构的规模越大,杠杆率越高以及股票的贝塔系数越高,其对系统性金融风险的贡献就越高,然而,具有高股票波动的金融机构不一定具有高的系统性金融风险贡献。第三,我国金融周期的计量分析。本文的第6章,采用滚动的H-P滤波方法,对金融系统高压力期出现前后的宏观经济和金融环境进行了分析,并通过对比金融脆弱期和金融压力期,给出金融周期的动态演变特征;对金融变量采用转折点分析方法,并通过联合分析金融变量的周期特征,划分出我国的金融周期;进一步地,通过与经济周期的对比分析,研究金融周期与经济周期的关联关系;最后,基于金融周期与经济周期之间的领先滞后关系,采用多元线性回归预测模型,评估了本文构建的金融不稳定综合测度指数对经济增长的预测能力。结果表明,高金融压力出现前的宏观经济处于高增长状态,金融机构各项贷款增加,同时房地产市场和股票市场也出现繁荣景象,但高金融压力出现后,这种繁荣的景象出现反转,经济增长快速回落,各项贷款减少并且资产价格迅速下降;从1999年1季度到2013年四季度,我国金融系统经历了三个主要的衰退期,分别为1999年2季度到2000年3季度,2006年1季度到2007年4季度以及2012年2季度到2013年2季度,并且我国的金融周期呈现衰退期较短而复苏期较长的非对称特征;金融周期一般要领先于经济周期,当金融系统出现衰退期后,宏观经济随后也会出现下滑,并且经济周期可能要经历不只一轮的下滑;金融脆弱指数和金融压力指数对于经济增长均具有良好的预测能力,尤其在中长期具有显著优于宏观经济变量的预测能力,同时,金融脆弱指数对于经济增长的预测能力在中期优于金融压力指数,而金融压力指数对于经济增长的预测能力在长期优于金融脆弱指数。
【Abstract】 The outbreak of the international financial crisis has triggered a reflection on the traditionalmicro-prudential supervision based on capital regulation, and the academia and governments havestarted to reconsider whether the monetary policy targeting controlling inflation is sufficient tomaintain the stability of the entire financial system. To prevent systemic financial risk andmaintain financial stability and economic growth, macro-prudential supervision is becoming themajor trends of financial regulation in the countries of the post-crisis era. Strengthening themacro-prudential supervision and cooperating with the micro-prudential supervision are majorrestructuring of the financial regulation. However, with regard to financial instability, financialcycle and systemic financial risk, there has not yet been a clear and consistent understanding. Theevolutionary mechanisms for the financial risk, the quantification of financial instability and thecreation of financial-economy model should be further studied, and more attention should be paidto the development and implementation of macro-prudential supervision tools.Although China has not yet experienced the financial crisis triggered by internal risk,instability in the financial system is increasing and the financial risk is accumulating, meaning thatit is urgent to study China’s financial instability and its spillover effects on the real economy. Themeasurement of financial instability and the characterization of the financial cycle could providesupport for the design of effective macro-prudential tools and the implementation of financialregulatory policies. The aspects and main findings of the studies in this paper are the following:First, the theoretical analysis of the financial instability is given. In Chapter2, after reviewingthe previous studies, the connotation, the causes, and the essential nature of the financialinstability are elaborated. From the static and the dynamic perspective, the clear connotation of thefinancial instability is given. The evolution of the financial instability is divided into three stages,i.e., financial fragility, financial stress and systemic financial risk. The characteristics of thefinancial risk in each stage are introduced. Both the traditional financial contagion theory and theendogenous financial instability theory are used to explain the generation of financial instability.The essential nature of the financial instability, including the characteristics, dimensions,regulations, and economy costs, are discussed. Finally, the pre-crisis and post-crisis policies forfinancial instability are introduced.Second, the quantitative analysis of the financial instability is given. In Chapter3,4and5,the three main stages of the financial instability, namely the financial fragility, the financial stress,and the systemic financial risk, are measured and analyzed using econometric models respectively.In the quantitative analysis of the financial fragility in Chapter3, state-space model with Markov regime switching is used to model the co-movement of financial variables, and thetraditional maximum likelihood approach and Bayesian approach are used to estimate the modelrespectively. The common factor of the state-space model which is called the financial fragilityindex is extracted as a measurement for the financial fragility in financial system. Further analysisof the financial fragility and its switching features in China’s financial system are given. Moreover,the leader-lag relationship between financial instability and macro-economy is studied using linearand non-linear Granger test. Analysis results show that: two regimes of the financial fragility,“high financial fragility” and “low financial fragility”, exist in China’s financial system, and thetwo regimes are asymmetric, with the persistence probability being bigger and the persistenceduration being longer in the “low financial fragility” regime; the magnitude of the financialfragility cycle is affected by financial policy, fiscal policy, and monetary policy, and unusuallyhigh financial fragility is followed by much lower financial fragility; financial instability inChina’s financial system will have an effect on the macro-economy, but the transmissionmechanism and channel are complex and nonlinear.In the quantitative analysis of the financial stress in Chapter4, based on the characteristics ofthe financial stress, financial stress indicators are selected from the banking sector, the stockmarket and foreign exchange market, and they are constructed into a financial stress index as wellas three sub-index, namely bank stress index, security stress index, and foreign exchange stressindex. Duration-dependent Markov regime switching autoregressive model is used to characterizethe duration-dependent feature and regime switching feature of the financial stress in China’sfinancial system. Analysis results show that: China’s financial system has two regimes of thefinancial stress,“high financial stress” and “low financial stress”, and it is more likely to transferfrom the “high financial stress” regime to the “low financial stress” regime; positive durationdependence exists in both regimes, that is to say, the financial system is more likely to transfer tothe other regime as one regime lasts longer.In the quantitative analysis of the systemic financial risk in Chapter5, the extremal quantileregression method based on extreme value theory is used for the measurement of the systemicfinancial risk of33financial institutions and their contributions to the systemic financial risk.Through the panel data regression model, factors contributing to the systemic financial risk offinancial institutions are analyzed. Analysis results show that: the value-at-risk approach may bebiased in measuring the financial risk of the institutions, since it cannot measure the correlationbetween financial institutions and the financial system; systemically important financialinstitutions in China are mainly from the bank sector, and joint-stock commercial banks haverelatively higher systemic risk contribution; the larger the size of the financial institution, thehigher leverage and the higher beta of the stock, the higher its risk contribution to the financial system, however, financial institution with a higher stock volatility does not necessarily have ahigher systemic financial risk contribution.Third, the quantitative analysis of the financial cycle is given. In Chapter6, the scrolling H-Pfiltering method is used to analyze the macro-economy and financial environment before and afterthe period of high financial stress. By comparing the financial fragility and financial stress inChina’s financial system, the characteristics of the financial cycle are given. Through the turningpoint analysis of the financial variables and the studies of their cyclical features, China’s financialcycle is determined. Furthermore, by comparing with the business cycle, the relationship betweenthe financial cycle and the business cycle is analyzed. Finally, multivariate linear regression modelis used to assess the predictive power of the financial instability measures given in this paper forthe economic growth. The findings are: before the occurrence of the high financial stress, theeconomy is in a high-growth path, with the loans of financial institutions increasing, and the realestate market and stock market booming; however, when the high financial stress occurs,everything reversal, the economic growth declining to a lower level, the loans reducing and theasset prices declining rapidly; from the first quarter of1999to the fourth quarter of2013, China’sfinancial system has gone through three major recessions, i.e.,1999Q2to2000Q3,2006Q1to2007Q4and2012Q2to2013Q2, and China’s financial cycle shows an asymmetrical feature with ashorter duration of recession but a longer duration of recovery; the financial cycles lead thebusiness cycles generally, and after the recession of the financial cycle, there will always be one ormore downturns of the business cycle; the financial fragility index and the financial stress indexhave good predictive ability for the economy growth, especially in the medium-term andlong-term; the financial fragility index has better predictive power than the financial stress indexin the medium-term, while the financial stress is better in the long-term.
【Key words】 Financial Instability; Financial Fragility; Financial Stress; Systemic Financial Risk; FinancialCycle;