管理评论 ›› 2024, Vol. 36 ›› Issue (8): 28-38.

• 经济与金融管理 • 上一篇    

借贷便利工具、担保品渠道与小微企业贷款——基于双重机器学习的DID研究

欧阳志刚1,2, 李伟1   

  1. 1. 中南财经政法大学金融学院, 武汉 430073;
    2. 华东交通大学经济管理学院, 南昌 330013
  • 收稿日期:2022-10-17 发布日期:2024-09-03
  • 作者简介:欧阳志刚,中南财经政法大学金融学院、华东交通大学经济管理学院教授,博士生导师,博士;李伟,中南财经政法大学金融学院博士研究生。
  • 基金资助:
    国家自然科学基金面上项目(71973044);国家社会科学基金领军人才项目(22VRC016);国家高层次人才创新创业项目(20603);国家自然科学基金地区项目(72263008);全国统计学科重点项目(2022LZ09);中南财经政法大学基本科研业务经费。

Lending Facilities, Collateral Channels and Small and Micro Enterprises Lending—A Study of DID Based on Double Machine Learning

Ouyang Zhigang1,2, Li Wei1   

  1. 1. School of Finance, Zhongnan University of Economics and Law, Wuhan 430073;
    2. School of Economics and Management, East China Jiaotong University, Nanchang 330013
  • Received:2022-10-17 Published:2024-09-03

摘要: 为支持实体经济发展,缓解小微企业融资难现象,中国人民银行在2018年6月1日扩容了中期借贷便利工具的担保品范围,将优质小微企业贷款纳入中期借贷便利(MLF)担保品框架。本文利用这一准自然实验,将双重机器学习方法引入传统的双重差分模型(DID),研究借贷便利工具担保品扩容对小微企业贷款的影响。研究发现,中国人民银行将优质小微企业贷款纳入中期借贷便利(MLF)担保品显著增加了小微企业的贷款可获得性,且主要是缓解信息不对称、降低抵押担保要求实现的。异质性分析表明,东部发达地区的小微企业受到的促进作用大于中西部不发达地区的小微企业,专精特新类的小微企业受到的促进作用大于非专精特新类的小微企业。

关键词: 担保品扩容, 小微企业, 贷款可获得性, 双重机器学习, DID

Abstract: In order to support the development of real economy and alleviate the financing difficulties of small and micro enterprises (SMEs), on June 1, 2018, the People's Bank of China (POBC) expanded the collateral scope of the medium-term lending facility (MLF) to include high-quality SMEs' loans in the MLF collateral framework. Using this quasi-natural experiment, this paper introduces the Double Machine Learning method into the traditional DID model to study the impact of collateral expansion of MLF on SMEs' loans. The study finds that the inclusion of high-quality SMEs' loans as collateral for MLF by POBC significantly increases the availability of loans to SMEs, and it is mainly achieved by alleviating information asymmetry of SMEs and lowering mortgage guarantee requirements. Heterogeneity analysis shows that the promotion effect on SMEs in eastern regions is greater than that on SMEs in central and western regions, and the promotion effect on SMEs of specialized and special new type is greater than that on SMEs of non-specialized and special new type.

Key words: collateral expansion, SMEs, loan availability, double machine learning, DID