基于SVAR模型的商业银行压力测试研究

Research on Stress Test of Commercial Banks Based on SVAR Model

宏观经济变量间往往存在多重共线性,运用SVAR模型可把变量转化为同期独立的结构冲击.选用房价,M2,CPI,PMI作为宏观风险因子,不良贷款率度量银行信用风险,建立SVAR(4)模型进行商业银行压力测试,把宏观风险因子转化为同期独立的结构冲击,测试时期为2016年第三季度,测试方法选用历史情景法,历史情景为使测试时期遭受样本期内最不利的结构冲击,结果显示五种结构冲击造成的不良贷款率分别为0.0076,0.0044,0.0058,0.0095,0.0048,结合当期脉冲响应函数,短期影响最大的是房价结构冲击和PMI结构冲击,两者占比53.25%.若5种最不利结构冲击同时发生,则不良贷款率为0.0320,高于自2008年第四季度以来的所有历史不良贷款率,这是十分严重的.长期中,由累积脉冲响应函数可知M2、CPI结构冲击为影响银行不良贷款率的关键因素,想抑制银行信用风险,央行需要实施积极的货币政策.

There are multiple linear among macroeconomic variables, and the use of SVAR model can transform the variables into the structural shocks that independent in current period. Selection of house prices, M2, CPI, PMI as macroeeonomie risk factors, default rate as the bank credit risk measurement, establishing SVAR（4） model for commercial bank stress tests. The macroeeonomic risk factors are transferred into structural shocks independently in current periods, the test period is the third quarter of 2016. The test method is historical scenario method. The historical scenarios for the test period suffered during the sample period the most unfavorable shock. The results showed that five default rates caused by the shocks of default rate are 0.0076,0.0044, 0.0058, 0.0095, 0.0048. Connected to impulse response function in the current periods, house price structural shocks and PMI structural shocks is the largest impact in the short term, both accounted for 53.25%. If the five most adverse structural shocks occur simultaneously, the default rate will go to 0.0320, higher than that rate since the fourth quarter of 2008, which is very serious. However, in the long term, according to accumulated impulse response function, M2 structural shocks and CPI structural shocks are the largest price structure impact. To control credit risk of commercial banks, central bank need to put into active monetary policy.