A New Improvement to the Ohlson (1995) Model: Empirical Evidence from Japan

Abstract

This paper investigates the validity of the Ohlson (1995) information dynamics (Linear Information Model: LIM) and attempts to improve LIM. The difficulty concerning the empirical tests on LIM lies in identifying v, which denotes information other than abnormal earnings. Recent papers, such as Myers (1999), Hands and Landsman (1999), and Barth et al. (1999), all try to specify v by using various accounting information. Instead of tackling this difficult task, I circumvent it by dealing with serial correlation of the error terms that will be caused by omitting a necessary variable v from the regression equation. The results indicate it leads to the improvement on LIM.

Key Words: Ohlson (1995) model; Information dynamics; Other information vt; Serial correlation; Durbin's alternative test; GLS

JEL Classification: M41, G12