Scholarlink

<p>This paper examines how equity markets value bank earnings quality through loan loss provisioning. Using a sample of 42 Chinese listed banks from 2012 to 2024, we propose a three-dimensional valuation framework that separates reported ROE (cosmetic earnings), provision-adjusted ROE (earnings quantity), and the coverage ratio (risk quality). In horse race regressions, we observe an opposite-signs pattern: holding True ROE constant, higher Reported ROE is associated with lower valuation, while higher True ROE predicts higher valuation. This pattern suggests that markets discount earnings increases driven mechanically by lower provisioning while rewarding underlying earnings capacity. The signaling value is stronger for non-SOE banks and banks with higher financing needs. A machine learning validation using SHAP analysis confirms the robustness of these results in a flexible, nonparametric setting. Overall, our findings highlight the value relevance of provision-adjusted earnings measures in bank valuation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Earnings Quality and Bank Valuation in China: Evidence from Provision-Adjusted ROE

  • 期刊论文

  • 发表日期：

    2026年6月26日

查看全文
  • Zhixing Lin

摘要

This paper examines how equity markets value bank earnings quality through loan loss provisioning. Using a sample of 42 Chinese listed banks from 2012 to 2024, we propose a three-dimensional valuation framework that separates reported ROE (cosmetic earnings), provision-adjusted ROE (earnings quantity), and the coverage ratio (risk quality). In horse race regressions, we observe an opposite-signs pattern: holding True ROE constant, higher Reported ROE is associated with lower valuation, while higher True ROE predicts higher valuation. This pattern suggests that markets discount earnings increases driven mechanically by lower provisioning while rewarding underlying earnings capacity. The signaling value is stronger for non-SOE banks and banks with higher financing needs. A machine learning validation using SHAP analysis confirms the robustness of these results in a flexible, nonparametric setting. Overall, our findings highlight the value relevance of provision-adjusted earnings measures in bank valuation.