Investors' differential reaction to positive versus negative earnings surprises

成果类型:
Article
署名作者:
Pinello, Arianna Spina
署名单位:
University System of Georgia; Georgia State University
刊物名称:
CONTEMPORARY ACCOUNTING RESEARCH
ISSN/ISSBN:
0823-9150
DOI:
10.1506/car.25.3.9
发表日期:
2008
页码:
891-+
关键词:
forecast accuracy risk preferences analysts expectations INFORMATION DISCLOSURES ambiguity
摘要:
Archival studies document an asymmetrically strong market reaction to positive vis-a-vis negative earnings surprises. This finding appears inconsistent with the well-known effect of loss aversion and remains unexplained. I posit that this pattern can arise when investors' earnings expectations do not coincide with analyst forecasts. Numerous studies document optimistic biases in analyst forecasts, so if investors perceive such optimism, their earnings expectations will be lower than analyst forecasts. Because the contrast between actual and expected earnings determines the degree of perceived surprise, when the earnings surprise is positive (negative), an investor expectation that is below the analyst forecast results in a larger (smaller) perceived surprise than would be expected based solely on the analyst forecast. Investors' lower expectations relative to analyst forecasts therefore result in a stronger reaction to positive than to negative reported earnings surprises of equivalent magnitude. I experimentally replicate the asymmetrically strong reaction to positive reported earnings surprises documented in archival studies, and I trace this reaction pattern to investors' perceptions of those surprises. I further show that when earnings surprises are measured on the basis of investors' expectations as opposed to on the basis of analysts' forecasts of earnings, the differential reaction pattern reverses: investors' reaction is asymmetrically strong to negative vis-a-vis positive perceived earnings surprises, consistent with loss aversion. My findings highlight the importance of reference points in the firm valuation process and demonstrate that caution must be exercised when making inferences regarding market participants' earnings expectations based on analyst forecasts.