On the Development of an Efficient Deflator for the Estimation of Accounting-Based Valuation Models

Abstract

Previous research documents a number of potential scaling problems when estimating accounting based valuation models using cross-sectional data. The differences in size across observations cast doubts over the robustness of estimated coefficients and measures of fit. Several solutions have been proposed without reaching a consensus. We demonstrate analytically that, under certain conditions, the standard deviation of market capitalization changes is an efficient deflator for the cross-sectional estimation of Ohlson's (1995) unbiased accounting valuation model. Our empirical tests confirm this, as we show that the largest observations do not unduly affect the estimation results, the number of influential observations decreases notably and we obtain estimated coefficients not significantly different from the theoretical values derived from the Ohlson (1995) information dynamics.

JEL Classification: G12, M41, C50