隨著資本市場對資訊透明度要求提升，投資人對企業揭露品質的關注日益加 深。年報的可比性（Comparability）作為資訊品質的重要指標，能提升市場對企 業價值的解讀效率，進而影響資金配置決策。儘管過往已有文獻探討揭露品質與 市場反應之關係，但針對「年報文字可比性」與IPO市場行為的實證研究仍屬稀 少。 本研究以2005至2022年間台灣223家首次公開發行（IPO）公司為樣本， 擷取其IPO 前一年之年報，透過文字探勘技術與TF-IDF結合Cosine Similarity 計算年報可比性指標。進一步以多元線性回歸模型，探討可比性對四項IPO市場 反應指標之影響：承銷折價、蜜月期報酬、價格迴轉與股價波動性，並控制公司 財務特徵。 實證結果顯示，年報可比性越高，企業的IPO承銷折價、初期市場價格異常 報酬、股價波動性具顯著性且方向與預期一致，顯示市場較能精確評價資訊一致 性高的企業，進而降低資訊不對稱所致的價格錯估。雖價格迴轉未達顯著水準， 但方向與預期一致。綜合而言，本研究證實年報可比性可作為資訊品質的有效衡 量指標，對提升IPO市場定價效率與監理政策制定具有實質意義。;As capital markets place greater emphasis on transparency, annual report comparability has become a crucial indicator of disclosure quality and pricing efficiency. It reflects the textual similarity of financial disclosures across firms, allowing investors to more easily interpret firm value. However, empirical research on how textual comparability affects IPO market behavior remains limited. This study analyzes 223 Taiwanese IPO firms from 2005 to 2022, using their pre IPO annual reports. We apply text mining and TF-IDF with Cosine Similarity to calculate a comparability index across firms. We then evaluate its effect on four IPO outcomes: underpricing, honeymoon returns, price reversal, and stock price volatility, controlling firm-specific financial factors via multiple linear regression. Findings show that higher comparability is associated with significantly lower underpricing, reduced volatility, and smaller price ratios, suggesting that consistent disclosures help mitigate information asymmetry and pricing inefficiency. Though the impact on price reversal is statistically insignificant, the effect direction aligns with theory. These results support the use of textual comparability as a proxy for information quality and underscore its role in improving IPO valuation accuracy.
