基于多重信贷网络的银企间风险传染与控制研究

发布时间：2018-01-03 00:12

本文关键词：基于多重信贷网络的银企间风险传染与控制研究　出处：《东南大学》2016年博士论文　论文类型：学位论文

【摘要】：现代经济体系的一个重要特征是主体间具有较高的连接。为此,一个主体的违约可能会引发若干与其具有直接或间接关联的主体的违约,即所谓的多米诺骨牌效应。正如相关学者的研究,现代经济体系中的主体间的关联形成了一个网络,在这个网络格局中,一个主体的违约不仅受其自身财务状况的影响,同时亦受与其有关联(如借贷关系)的邻居主体的财务状况的影响,而其邻居主体的财务状况亦受其邻居主体的邻居主体的影响,依此类推。同样,一个主体违约,不仅会负向影响与其直接相连的邻居主体,同时亦会通过对其邻居主体的影响而将违约的负面效应传递给其邻居主体的邻居主体。企业和银行是现代经济体系的两大重要主体,二者间的关联也越来越紧密。企业主体的违约会产生银行坏账,当银行主体无法承受企业主体违约而产生的坏账时,银行主体会破产。银行作为极其重要的经济主体,其破产将会带来极大的负面效应。为此,对银企主体间的风险传染研究十分必要。基于此,本文在已有的研究基础上,构建了包含银行主体和企业主体的内生多重信贷网络模型,并基于该模型对银企主体间的风险传染与控制进行研究。首先,基于企业主体间的商业信贷连接、银行和企业主体间的银行信贷连接及银行主体间的拆借连接构建了包含银行主体和企业主体的多重信贷网络模型,且主体间信贷需求量及主体间信贷连接的建立均为内生。通过计算机仿真,所构建的模型能够刻画现实中主体及主体关联网络的一些典型特征：当企业规模大于一定阈值时候,企业规模服从幂律分布；银行规模可用具有幂律尾部的对数正态分布来拟合；银行-企业信贷网络和银行-银行信贷网络的银行入度服从双幂律分布。其次,基于多重信贷网络模型,从信贷网络结构和主体行为对银企主体间的风险传染进行研究。对于网络结构,分别研究了交易对手的随机选择数M、交易对手转移概率参数γ的变化对银企间风险传染的影响。研究表明M和λ对于主体的作用机理不同,随着二者的不断变动,相应的研究变量(银行主体、上游企业主体和下游企业主体累积破产数、社会产出、银行主体的信贷、坏账均值及标准差)表现出不同的变动路径,但较大的M、λ都将导致较大的研究变量,银企间风险传染加大,以经济的脆弱性换来社会产出的增大。另外,从企业主体和银行主体行为两方面,研究了主体行为对风险传染的影响。对于企业主体行为,研究表明随着企业主体产量参数φ与β取值的不断增大,相应的研究变量呈现不同的变化趋势。较大的φ值导致较大的研究变量,而较大的β值却带来相关研究变量的下降(银行主体的信贷均值除外)。对于银行主体行为,研究表明,随着银行主体风险厌恶系数Ψ的增大,银行主体累积破产数呈递减趋势,而上游企业主体累积破产数在高位波动后呈下降趋势,下游企业主体累积破产数则在低位波动后呈上升趋势。而随着银行主体分红比率上限d的不断增大,银行主体的累积破产数呈现明显递增趋势,上游企业主体和下游企业主体的累积破产数则呈现波动状态,无明显趋势。较大的存款比例下限υmin将带来低的银行主体和下游企业主体的累积破产数及较大的上游企业主体累积破产数,当达到一定阈值后,各部门主体累积破产数趋于稳定。而对于综合性参数α而言,随着α的不断增大,银行主体的累积破产数呈现递增趋势,上游企业主体的累积破产数则呈现先增后减的趋势,而下游企业主体的累积破产数则呈现先明显递增而后缓慢递增的趋势。以上参数的变动对其它研究变量的影响亦不同。最后,基于多重信贷网络模型,对银企主体间的风险控制进行研究。研究表明存款准备金率的提升并不是控制银企间风险的较好手段。存款准备金率的提升,虽然抑制了银行主体的风险投资比率,进而减少了银行主体的破产风险,但存款准备金率的提升亦减少了银行系统对经济主体的信贷供给,亦会导致经济主体因流动性不足而破产,并且宏观经济产出也未必一定增加。而资产最大银行主体与入度最大银行主体等系统重要节点主体的违约给银企系统中的相关部门主体带来较大的负向影响。对于银行主体和下游企业主体而言,资产最大银行主体违约所引发的累积破产数演化路径明显高于由度最大银行主体违约所引发的累积破产数演化路径,而由度最大银行主体违约所引发的上游企业主体每期累积破产数要高于由资产最大银行主体违约所引发的累积破产数。这说明,仅预防那些“最大而不能倒闭”银行主体“不倒”并不足够,亦应关注那些具有较大入度连接的银行主体。而对于那些具有较大入度连接的资产较大的银行主体,则更应重点监控。研究亦表明,债务透明度规则的实施可在社会产出变动不是很大的情况下减缓银企间的风险传染；中央银行的流动性供给利于银企系统稳定性的提高,并可促进社会产出的提升。另外,对于不同的市场参与度,中央银行的流动性供给效果相差不大。总之,本文基于银企主体间的信贷连接,构建了内生的银企多重信贷网络模型,基于该网络模型从信贷网络结构和主体行为两个视角对银企间风险传染进行了研究,并对网络视角下的银企间风险控制进行了研究,从理论上对银企主体间的风险传染与控制进行了深入系统的研究,为监管者提供了可供参考的模型和方法,具有较高的理论价值和现实意义。
[Abstract]:An important feature of modern economic system is the connection between the main body. Therefore, a subject of breach of contract may cause some subjects with directly or indirectly associated with the so-called Domino effect. As research scholars, related subjects in the modern economic system between the formation of a in the network, the network structure, the influence of default of a subject not only by its own financial situation, but also by its associated effects (such as lending) neighbors of the main financial status, influence, and its neighbor the main financial situation is subject to its neighbor neighbor subject and so on. The same subject a breach of contract, subject, not only would negatively influence directly connected neighbors subject, also through influence on its neighbor subject and transfer the negative effect of default to its neighbors and neighbors Subject. Enterprises and banks are two major subjects of modern economic system, relationship between the two is getting more and more close. The main body of the enterprise default will produce bad debts, when banks cannot afford enterprises subject and breach of bad debts, banks will go bankrupt. The main bank as a very important economic subject, its bankruptcy it will bring great negative effect. Therefore, it is necessary to study the risk contagion between the main bank. Based on this, this article on the basis of the study, constructed the bank subject and enterprise main endogenous multiple credit network model, based on the model of risk transmission and control between the main bank first. The connection between the main business enterprises, based on credit, banks and enterprises between bank credit and bank lending between the main connection connected by the multi subject and enterprise entity contains bank Credit network model, and the main credit demand and establish the connection between the subject of credit are endogenous. Through computer simulation, the model can describe some typical features of the subject and the reality subject Association Network: when the enterprise size is greater than a certain threshold value, the scale of enterprises follows the power-law distribution; logarithm of bank size available has a power-law tail of normal distribution to fit; Bank - enterprise credit network and bank - bank credit bank network penetration obeys the power-law distribution. Secondly, multiple credit network model based on the research of bank risk contagion between subjects from the credit network structure and behavior. For the random network structure. Choice of counterparties number M, counterparty transfer change probability parameter influence on the risk contagion between banks and enterprises. The research shows that the M and the main body of the different mechanism, With the continuous change of the two corresponding research variables (the banks of the main upstream enterprises and downstream enterprises, the main body of the cumulative number of bankruptcy, the main social output, bank credit, bad debts mean and standard deviation) showed changes in different ways, but the larger M, lambda will lead to larger research variables, bank risk infectious increase, the vulnerability of the economy to increase for social output. In addition, from the two aspects of enterprises and bank behavior, studied the effect of the main behavior of the risk of infection. For enterprise behavior, research shows that with the increasing of enterprises production parameter and beta value, study the corresponding variable trend different. Larger values of the study variables lead to larger, and the larger the beta value declines bring related research variables (except the main bank credit mean). For the main body of bank behavior, study shows, With the increase of the risk aversion coefficient is the main bank, the banks of the main cumulative number of bankruptcy decreased, while the upstream enterprises cumulative bankruptcy had decreased at a high level, the number of main downstream enterprises accumulated bankruptcy in the low volatility increased. With the increase of bank main payout ratio limit D, cumulative bankruptcy the number of the main banks showed increasing trend, the cumulative number of bankrupt enterprises and downstream enterprises upstream is fluctuated, no obvious trend. The proportion of deposits lower min will bring large main banks and downstream enterprises the main low cumulative number of bankruptcy and large enterprises upstream cumulative number of bankruptcy, when reaching a certain threshold after all, the main departments cumulative number of bankruptcy tends to be stable. For the comprehensive parameter, with the increasing of alpha, cumulative number of bankrupt banks subject increasing The cumulative number of upstream enterprise bankruptcy trend, subject showed a trend of first increase and then decrease while the cumulative number of downstream enterprises bankruptcy showed significantly increasing light and then slowly increasing trend. The above parameters change on other research variables are also different. Finally, the multiple credit network model based on bank risk subject the better way to control research. Research shows that the deposit reserve rate increase is not control between banks and risk. The deposit reserve rate increase, while suppressing the bank's risk investment ratio, thereby reducing the risk of bankruptcy of banks main body, but also improve the deposit reserve rate cut the supply of credit to the economy the main body of the bank the system, will also lead to the economic subject of bankruptcy due to lack of liquidity, and macro economic output may not necessarily increase. While the largest bank assets subject and the maximum principal bank penetration The main body of an important node system default to the relevant departments in the main bank system bring great negative effect. For the main banks and downstream enterprises subject, cumulative number of bankruptcy subject assets bank default caused by the maximum evolution path was significantly higher than that of cumulative number of bankruptcy by the largest bank degree caused by the breach of the main evolution path, and by most banks default caused by the main upstream enterprises of each subject should be higher than the cumulative number of cumulative bankruptcy bankruptcy by the largest bank entity assets caused by default. This shows that only prevent those "maximum fail" bank subject "down" is not enough, should also pay attention to those with larger banks. While the main connection degree for those with a greater penetration of connect assets larger banks subject, should be more focus on monitoring. The study also showed that the implementation of transparency rules can be in debt Social change is not great output under the condition of slow risk contagion between banks and enterprises; the supply of liquidity by the central bank to improve the stability of bank system, and promote the society to improve the output. In addition, for different market participation, the supply of liquidity effect of the central bank had little difference. In short, the bank credit subject based on the construction of the bank and credit connection, multi network model is endogenous, two aspects of the network model from the credit network structure and behavior are studied based on the risk contagion between banks and enterprises, and from the perspective of network between banks and risk control are studied, conducts a systematic research on the risk of infection and control of the main bank between the theoretical model and the method, provides the reference for regulators, has high theoretical value and practical significance.

【学位授予单位】：东南大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：F832.4

【相似文献】