| 摘要: |
| 摘要：采用Prophet模型与支持向量回归（SVR）模型，并结合气候物理风险情景进行深入分析。通过设定基准、轻度、中度和重度四种压力情景，评估未来四个季度内气候变化引发的物理风险，以及对商业银行对公业务信贷资产质量的潜在影响。研究结果表明，随着气候物理风险的加剧，对公客户违约概率显著上升，导致银行信用风险成本和资本要求显著提升，特别是在重度情景下，银行的资本充足率已低于监管要求。基于此，提出商业银行需强化气候风险管理措施，优化调整信贷结构，并制定重度压力情景的资本缓冲方案，应对气候风险带来的潜在挑战。为商业银行提供科学的气候风险管理策略，并为后续的气候风险研究开辟新的视角和方法论。 |
| 关键词: 关键词：Prophet模型；支持向量机回归模型；压力测试；气候物理风险 |
| DOI： |
| 分类号:F830.33 |
| 基金项目:基金项目：国家自然科学基金面上项目“双层规划问题的约束规范与光滑函数方法的研究”（12071342）；河北省优秀青年基金项目“两阶段随机锥约束优化问题的稳定性分析”（A2020202030）。 |
|
| Study on the Impact of Climate Physical Risk on Credit Risk Quality of Commercial Banks——Stress Testing Methodology Based on Prophet and SVR |
|
ZHANG Zi-qing1，DUAN Qing-song21,2
|
|
1.（1.College of Science,Hebei University of Technology;2.Bank of China Limited）
|
| Abstract: |
| Abstract：Prophet model with Support Vector Regression (SVR) model is used and combined with climate physical risk scenarios for in-depth analysis.By setting four stress scenarios,namely, baseline, mild,moderate and severe,this study aims to assess the potential impact of physical risks arising from climate change on the quality of credit assets of commercial banks’ public business in the next four quarters.The results of the study show that the probability of default of public customers increases significantly with the intensification of climate physical risks,leading to a significant increase in the cost of credit risk and capital requirements of banks,especially in the severe scenario,where banks’ capital adequacy ratios have fallen below the regulatory requirements.Based on this,this study proposes that commercial banks need to strengthen climate risk management measures,optimize and adjust their credit structure,and develop capital buffers for heavy stress scenarios in order to cope with the potential challenges posed by climate risk.This study provides scientific risk management strategies for commercial banks in the face of increasingly severe climate risk,and opens up new perspectives and methodologies for subsequent climate risk research. |
| Key words: Key words：Prophet Model；Support Vector Machine Regression Model；Stress Test；Climate Physical Risk |
