自結財務資訊申報與盈餘品質之關聯性

Abstract

為提升財務資訊之時效性並縮短資訊空窗期，臺灣主管機關自 2002 年起推動「自結損益資訊申報制度」，並自 2021 年起分階段實施「年度自結財務資訊申報制度」，要求特定條件下之上市櫃公司揭露完整自結財報，並於 2023 年起全面適用。本研究以 2009 年至 2023 年之臺灣上市櫃公司為研究對象，並以自結財務資訊制度為分析基礎，探討在雙重制度架構下，企業之資訊揭露行為與盈餘品質之關聯性，並進一步解析企業於資訊空窗期間之揭露策略，是否反映其會計資訊處理能力與盈餘管理傾向。 實證結果顯示，有申報自結損益資訊之公司，其盈餘品質顯著優於未申報者，特別是在盈餘高估之樣本中效果更為顯著。進一步分群分析亦發現，無申報自結損益資訊但須申報年度自結財務資訊之公司，其盈餘操縱幅度與總審計調整幅度顯著較高，顯示資訊揭露主動性不足或會計資訊處理能力較低之企業，盈餘品質相對較差。 本研究更進一步以 IFRS 穩定施行後之年度作為替代樣本期間進行測試，亦另行排除因公告財務預測而強制申報自結損益資訊之公司樣本，重新檢驗企業資訊揭露行為與盈餘品質之關聯性，結果皆與假說一致，進一步強化結果之穩健性。惟年度自結財務資訊申報制度之實施並未顯著改善資訊揭露與盈餘品質間之關係，顯示制度成效仍待後續觀察與落實。

In response to the growing demand for timely and transparent financial information, Taiwan’s financial supervisory authority introduced the “Self-assessed Earnings Disclosure System” in 2002, followed by the phased implementation of the “Annual Self-reported Financial Information Disclosure System” beginning in 2021, which became fully applicable to all listed firms by 2023. This study investigates the association between voluntary and mandatory financial disclosure behavior and earnings quality, using a sample of Taiwanese listed and OTC-listed companies from 2009 to 2023. Specifically, the analysis examines whether firms’ disclosure practices during the financial reporting gap reflect their accounting information processing capabilities and tendencies toward earnings management. The empirical findings demonstrate that firms voluntarily disclosing self-assessed earnings exhibit significantly higher earnings quality, as indicated by lower discretionary accruals, particularly in cases of income-increasing earnings management. Further, a disclosure-based classification reveals that firms which do not disclose self-assessed earnings but are required to submit annual self-reported financial statements tend to exhibit greater earnings manipulation and larger aggregate audit adjustments ratio, suggesting that lower disclosure initiative and weaker financial reporting capacity are associated with poorer earnings quality. Two sets of robustness tests were conducted—one using data from the post-IFRS stabilization period, and the other excluding firms required to disclose self-assessed earnings due to financial forecasts. Both results support the main hypotheses and reinforce the robustness of the findings. Nevertheless, the implementation of the annual self-reported financial information disclosure system has not yet led to a statistically significant enhancement in the relationship between disclosure behavior and earnings quality. This outcome implies that the policy’s effectiveness in enhancing information disclosure and earnings quality remains uncertain and may depend on further institutional development and practical enforcement.