Alternative Disclosure Methods For Accounting Changes And Analysts Earnings Forecasts

Main Article Content

Thomas A. Buchman
C. Patrick Fort

Keywords

generally accepted accounting principles, GAAP, changing accounting principles, earnings forecasts

Abstract

Generally accepted accounting principles (GAAP) require that firms changing accounting principles must report the change in one of three ways: the cumulative effect method, the retroactive restatement method, or a no-adjustment (prospective) method. The method a company should use is determined by the type of change being made. This raises the following question: can it be demonstrated that one of these methods is better, in some sense, than the other methods? A major problem in evaluating alternative methods of accounting of the same economic event and in deciding which one method should be adopted as GAAP is that it is impossible to objectively determine which of the alternatives is best. However, it is possible to rank alternatives on one dimension of interest-which method minimizes the income forecasts in years after the change. We obtained a sample of forms making accounting changes and formed three portfolios of firms based on the method they used to account for the change in accounting principle. We then compared financial analysts earnings forecast errors for the firms in the three portfolios. After controlling for relevant variables, we found that, in the year firms made accounting changes the firms making the changes requiring retroactive restatement had significantly larger forecast errors than the firms making changes requiring the other forms of disclosure, but in years subsequent to the year of change there were no significant differences in forecast errors. This leads us to the conclusion that, from an earnings forecast accuracy perspective, there is no advantage to calculating and presenting the cumulative effect of an accounting change or in preparing restated or pro-forma financial statements.

Downloads

Download data is not yet available.
Abstract 83 | PDF Downloads 133