Firm and industry effects in accounting versus economic profit data

Abstract

This article presents estimates of firm and industry fixed effects on profit rates for large US corporations, using Economic Value Added (EVA), the popular measure of profits produced by Stern Stewart & Co., and simple (unadjusted) accounting measures as the dependent variables. We find that the improvement in explanatory power of the fixed-effect model is substantially greater when using EVA than has been documented with alternative measures.

Notes

¹What we call simple profit rate measure is net income plus interest over total assets, directly from the firm's reported income statement and balance sheet. EVA, as we describe below, is a much more complicated measure of profits.

²Kapler (2000) actually finds a higher R ² than we do, but the improvement in fit between the accounting and economic measures is larger in our study.

³Some differences between our data and Kapler's are that (1) firms in her sample were manufacturing firms, whereas our firms are large corporations; (2) she has 562 firms in her sample compared with our 331; and (3) we use more recent data and have 15 years of data compared with that of 6 years.

⁴Firms were grouped into industries based on their three-digit SIC code.

⁵To be precise, model (3) includes all industry dummies (but one). However, to avoid multicollinearity when estimating (3), we dropped the first dummy in cases where industries were represented by only one firm. For the industries represented by more than one firm, we also dropped one firm's dummy per industry (on the basis of whichever firm came first in alphabetical order). Model (1) then includes only those firm dummies that were included in (3).