The irrelevant responses and stock liquidity: dual validation evidence from the Chinese stock market using machine learning and large language models
Purpose As China’s capital market advances toward higher-quality development, the quality of corporate information disclosure and the transparency of managerial communication have become central concerns for investors and regulators. This study examines the relationship between managerial irrelevant responses during site visits and stock liquidity, and further investigates the underlying mechanisms through which such responses affect market liquidity. Design/methodology/approach Drawing on Q&A records from site visits to A-share listed companies in China, this study constructs a quantitative measure of managerial irrelevant responses using machine learning techniques. In addition, online scoring based on a large language model (GPT-4) is introduced as a complementary indicator. Using regression analyses, we estimate the impact of irrelevant responses on stock liquidity and conduct mechanism and subgroup analyses to ensure robustness. Findings We find that managerial irrelevant responses are significantly negatively associated with stock liquidity. Mechanism analyses suggest that increased investor attention to financial issues reflects heightened uncertainty about firms’ future prospects, thereby exacerbating illiquidity. Furthermore, evasive or ambiguous managerial communication reduces analyst coverage, deteriorates the external information environment and further suppresses stock liquidity. Subgroup analyses reveal that this negative effect is more pronounced among firms without a price limit mechanism, firms with lower institutional ownership and firms receiving limited media attention. Originality/value This study contributes to the literature at the intersection of information disclosure and market microstructure by introducing a novel, text-based measure of managerial communication quality. The findings provide evidence on the real market consequences of evasive disclosure behavior and offer important policy implications for improving regulatory oversight and managerial communication practices in capital markets.
Authors
- Lingyu Mo (ORCID: https://orcid.org/0000-0002-8483-8315)
- Peng Zhou (ORCID: https://orcid.org/0000-0003-3802-8618)
- Changchun Tan (ORCID: https://orcid.org/0000-0001-8064-6917)
- Long Chen (ORCID: https://orcid.org/0009-0005-1383-9184)
Institutions
- Hefei University of Technology (CN)
- Chinese Academy of Meteorological Sciences (CN)
Publication Details
- Journal
- Journal of Accounting Literature
- Published
- 2026-11-13
- DOI
- https://doi.org/10.1108/jal-01-2026-0005
- Primary Topic
- Auditing, Earnings Management, Governance
- Type
- article
- Field-Weighted Citation Impact
- 0.00