An Improved Methodology to Assess Value-relevance of Earnings and Book Values on Corporate Equity Securities

Abstract

This paper advances Minimum Sum (MS) nonlinear iterative regression to model price as a function of earnings and book values per share. The MS methodology improves upon OLS methodology in three ways. First, MS allows for nonlinear estimation of price to eps and bvps. Second, MS allows for modeling Minimum Absolute Percentage Error (MAPE) as the objective function instead of Minimum Sum of Squared Errors. Third, MAPE mutes the negative effects of outlying observations and non-normal data compared to OLS. Better price estimates are provided which will aid participants in the primary and secondary markets, or financial services in general.