This study proposes a longitudinal text analysis framework combining Japanese-language NLP metric extraction with paired testing and shift function analysis to evaluate qualitative changes in corporate risk disclosures. Applied to Japan's 2019 disclosure reforms, the approach analyzes 19,770 firm-year observations over ten years to capture multidimensional dynamics often masked by single-indicator methods.

  • The framework incorporates a cross-section relevance indicator to measure topical alignment between risk disclosures and management strategies.
  • Analysis of FY2015-FY2024 data reveals that while disclosure volume increased substantially, it was accompanied by a decline in readability.
  • Overall information structure improved, but specific descriptive quality stagnated during the period.
  • The degree of adaptation to reforms varied across different market segments.

The joint analysis demonstrates that conventional single-indicator methods frequently obscure complex shifts in disclosure patterns, highlighting the necessity of multidimensional evaluation for accurate assessment.