Scholarlink

<p>This study examines the impact of corporate artificial intelligence (AI) adoption on cash dividend payouts in China. Employing a text-based measure derived from the Latent Dirichlet Allocation (LDA) algorithm to analyze annual reports, we construct a firm-level indicator of AI application intensity. The results reveal a significant negative association between AI adoption and cash dividend distributions. Further mechanism analysis indicates that this effect operates primarily through two channels: an investment crowding-out effect, reflected in heightened innovation activities and improved stock liquidity, and a risk-driven effect, manifested through increased operational risk and elevated cash holdings. These findings remain robust across a series of stringent tests, including difference-in-differences (DID) estimations and instrumental variable approaches, with additional validation drawn from a natural experiment surrounding the 2014 industrial robot policy. Cross-sectional analyses further show that the dividend-reducing effect of AI is more pronounced among private firms, particularly those with limited internal cash reserves and lower sensitivity to executive performance-based compensation.</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

Artificial Intelligence Applications and Dividends: Evidence from China

  • 期刊论文

  • 发表日期：

    2026年2月26日

查看全文
  • Xiamin Fan,
  • Shiwen Chen,
  • Junhang Chen

摘要

This study examines the impact of corporate artificial intelligence (AI) adoption on cash dividend payouts in China. Employing a text-based measure derived from the Latent Dirichlet Allocation (LDA) algorithm to analyze annual reports, we construct a firm-level indicator of AI application intensity. The results reveal a significant negative association between AI adoption and cash dividend distributions. Further mechanism analysis indicates that this effect operates primarily through two channels: an investment crowding-out effect, reflected in heightened innovation activities and improved stock liquidity, and a risk-driven effect, manifested through increased operational risk and elevated cash holdings. These findings remain robust across a series of stringent tests, including difference-in-differences (DID) estimations and instrumental variable approaches, with additional validation drawn from a natural experiment surrounding the 2014 industrial robot policy. Cross-sectional analyses further show that the dividend-reducing effect of AI is more pronounced among private firms, particularly those with limited internal cash reserves and lower sensitivity to executive performance-based compensation.