透過應收帳款與合約負債模型偵測營收管理行為： 以臺灣上市櫃公司為例

Abstract

本研究係針對企業透過收入認列進行盈餘管理之行為，建立可偵測異常裁量性收入之模型，以提升盈餘管理偵測之準確性與敏感度。現有文獻多聚焦於美國或其他已開發市場，針對臺灣企業進行營收管理偵測之研究相對稀少。而在IFRS 15在臺灣實施後，收入相關科目的裁量空間擴大，故本研究以台灣上市櫃公司為分析對象，具有補足文獻不足之意義。本文依循Giedt（2017）之模型設計，分別針對應收帳款、合約負債－流動與合約負債－非流動建立迴歸模型，以捕捉其正常變動，進而導出異常收入應計數作為營收管理之代理變數。研究結果顯示，相較於傳統應計模型，本文所建構之模型能更有效偵測企業透過應收帳款或合約負債進行之營收操縱行為，特別是在面對複合操縱策略下展現出較高之辨識能力。此外，於無操縱情境下，模型亦能維持較低之誤判機率，顯示其穩健性。整體而言，本研究提供一套具備實務應用潛力與研究貢獻之營收管理偵測工具。

This study investigates earnings management through revenue recognition and proposes a model to detect abnormal discretionary revenues, aiming to improve detection accuracy and sensitivity. While prior research mainly focuses on developed markets, studies on Taiwanese firms are limited—especially following IFRS 15, which expanded discretion over revenue accounts. Using data from listed companies in Taiwan, the study builds on Giedt (2017) by constructing separate regression models for accounts receivable and contract liabilities. The resulting abnormal revenue accruals serve as proxies for revenue manipulation. Empirical results show that the proposed model outperforms traditional accrual models, particularly under complex manipulation strategies, while maintaining low false positive rates in non-manipulated settings. The model thus offers a robust and practical tool for detecting revenue-based earnings management.