Answers are grounded in this paper and cite their sources.
Demonstrates that linguistic features in comment letters predict goodwill impairment risk in firms, suggesting early financial risk signals.
This study develops an interpretable machine learning framework to predict goodwill impairment using linguistic features of mergers and acquisitions comment letters. We extract thematic content via Bidirectional Encoder Representations from Transformers-based synonym self-generation with term frequency–inverse document frequency and capture readability, length and response time through text mining. Tree-based models with SHapley Additive exPlanations analysis reveal that valuation-focused, less readable letters signal higher impairment risk, whereas forward-looking, detailed letters with clear deadlines indicate lower risk. Predictive power strengthens in weakly governed firms. This study demonstrates that unstructured regulatory text serves as an early warning signal for financial risk and advances explainable natural language processing applications in accounting research.
No takes yet. Share an insight, caveat, or question.
Tang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: