基于改进迁徙率模型的金融工具预期信用损失估值研究——以B公司为例
Research on Expected Credit Loss Estimation of Financial Instruments Based on Improved Migration Rate Model:Taking Company B as an Example
耿界翔 1祝叶 1周圆兀 1赵玉玲 2刘彩云 2李莉 2陆文2
作者信息
- 1. 广西科技大学,广西柳州 545006
- 2. 北京中同华资产评估有限公司,北京 100073
- 折叠
摘要
新版CAS22中最大的变动是金融资产的减值计提调整,由"已发生损失"变为"预期信用损失",因此需要企业前瞻性地关注未来潜在风险来确保资产质量,避免减值计提不充分或者不及时.本文在此背景下,以非金融类环保行业B企业作为案例,结合该企业自身、相关行业和国家数据构建前瞻性评价指标体系,通过机器学习DT-LSTM模型对迁徙率模型进行改进,评估该企业金融工具预期信用损失,并与迁徙率模型结果对比,满足会计谨慎性原则,并给出建议.
Abstract
The biggest change in the new version of CAS22 is the adjustment of financial assets'impairment provision,which changes from"incurred loss"to"expected credit loss".Therefore,enterprises need to pay forward-looking attention to potential risks in the future to ensure asset quality and avoid insufficient or untimely impairment provision.In this context,this paper takes enterprise B in non-financial environmental protection industry as a case,combines the enterprise itself,relevant industry and national data to build a prospective evaluation index system,improves the migration rate model through machine learning DT-LSTM model,evaluates the expected credit loss of the enterprise's financial instruments,and compares the results with the migration rate model.Meet the accounting prudence principle and make recommendations.
关键词
迁徙率模型/机器学习/预期信用损失/案例研究Key words
Migration rate model/Machine learning/Expected credit losses/Case study引用本文复制引用
基金项目
研究生教育创新计划项目(YCSW2022444)
出版年
2024
NETL
NSTL
万方数据