An Evaluation of the Ability of Logit-Based Financial Statement Analysis to Identify Market Mispricing
Posted: 23 Sep 1996
Date Written: Undated
Using logit-based financial statement analysis, Ou and Penman estimate an earnings predictor (Pr) that also predicts returns. We develop a forecasting model that purges measurement error (i.e., expected earnings) from the Ou and Penman proxy for the market's unexpected earnings. If market mispricing explains the returns to their trading strategy, then these returns should increase when we estimate Pr to predict our proxy for the market's unexpected earnings. Alternatively, if the market is efficient with respect to this information, then Pr can only predict the market's expected earnings and abnormal returns should decrease when the analysis is aimed at our unexpected earnings proxy which contains less expected earnings. None of our results support the mispricing explanation. The results support the alternative explanation that Pr predicts abnormal returns because it predicts expected earnings which are associated with expected returns omitted from the model used to estimate abnormal returns.
JEL Classification: G14, M41, G12
Suggested Citation: Suggested Citation