Earnings Numbers Research Articles

Discussion of a Statistical Model of Earnings Estimation

I like idea behind Professor Greenball's paper-the idea that models of accounting measurement should be tested to see whether, in words of Beaver, Kennelly and Voss (1968), they capture relevant aspects of reality. My criticisms of paper lie in two areas. The first is theory behind Greenball's model of earnings estimation. The second is So-called validation tests. The stated purpose of Greenball's paper is to develop a method of earnings estimation which is independent of firm's accounting policy. Underlying Greenball's paper is assumption that such a method will yield numbers more useful to users than historical cost earnings numbers currently reported. However, Greenball does not tell us why he chose his particular method from total set of estimation methods which are independent of firm's accounting policy. Unless he has a theory which suggests that his method will be most valuable to users of earnings number, Greenball should compare all possible methods of estimation. The criticism is that Greenball does not present a theory which would suggest his specification of earnings. Greenball does tell us that the net realizable value or 'exit price' approach to valuation of tangible resources has substantial, though not unanimous, support in accounting theory and in economics. Such a statement, by itself, does not justify measurement of earnings in period t, x (t), as in Greenball's equation (2), i.e.:

Some Preliminary Findings on the Association between the Earnings of a Firm, Its Industry, and the Economy

The purpose of this paper is to investigate whether there is some significant -degree of association between the earnings of an individual finn, the earnings of other firms in its industry, and the earnings of all firms in the economy. A strong association has been found between the returns to a firm's stockholders, returns to stockholders of firms in the same industry, and returns to stockholders in toto. For example, Benjamin King's study of 63 firms over the period of June, 1927 to December, 1960 revealed . . that the typical stock has about half of its variance explained by an element of price change that affects the whole market, and that industry effects account for perhaps an additional ten per cent.' Since there is also both theoretical and empirical support for the contention that a concept of earnings is related to market valuation,2 earnings numbers might also reflect a roughly -similar degree of association.