The Potential for Gains in Predictive Ability Through Disaggregation: Segmented Annual Earnings
推导了将企业整体收益分解为分部门收益后，能够提高年度收益预测精度的条件，为投资者和监管者评估分部门披露的价值提供理论依据。
Both the Financial Accounting Standards Board and the Securities and Exchange Commission have issued guidelines for reporting annual earnings on a subentity basis (FASB [1976; 1977] and SEC [1977; 1978]). Presumably, these disclosures can be used by investors for various purposes, including the prediction of total-entity profits. Indeed, researchers have assessed this potential use by examining the relative predictive ability of consolidated (CN) versus segmented (SG) financial disclosures. The results of previous CN-SG comparisons are mixed. While there is evidence that CN forecasts are generally less accurate than SG forecasts (e.g., Kinney [1971] and Collins [1976]), there is also evidence that such differences may be situation-specific rather than general (e.g., Barefield and Comiskey [1975], Fried [1978], Ang [1979], and Barnea and Lakonishok [1980]). In these studies, however, there was no attempt to determine what general conditions are necessary for disaggregation to result in improved forecasts.' Here, we derive the conditions under which the disaggregation of entity earnings into subentity earnings can lead to gains in predicting annual earnings before extraordinary items and discontinued operations. Simu-