The Valuation Consequence of Accounting Changes: A Multi-Year Examination

Abstract

We examine the valuation impact of changes in the accounting procedures and estimates underlying reported financial data in the year of the change as well as in the postchange years. Since most accounting changes have undisclosed effect on financial variables in subsequent years in addition to the earnings impact disclosed in the year of the change, an accounting change might be motivated by its long-term valuation effect, even if investors are cognizant of the initial earnings impact and fully account for it in the year of the change. This conjecture is empirically examined for the first time in this study. Our tests are based on a cross-sectional examination of the valuation impact of the earnings effect of accounting changes in the year of the change and a longitudinal examination of the behavior of returns in the postchange years. We also provide descriptive evidence indicating that earnings management is a managerial motive for changing accounting techniques. Cross-sectionally, for income-increasing accounting changes, our results show that investors' valuations seem to reflect a concern for the reduced quality of earnings, as reflected by smaller earnings response coefficients and R2s. However, the decline is not attributable specifically to the earnings effect of the accounting change. Similarly, the earnings effect of income-decreasing changes does not have valuation impact in the year of the accounting change. Although investors appear to largely ignore the accounting changes in the year they are made, our longitudinal test does show that firms undertaking accounting changes experience different long-term returns relative to other firms in the postchange period. However, income-increasing accounting changes are associated with negative valuation changes in the postchange period, rather than the positive impact expected from the conjecture stated above. Over the five years following the year of the accounting change, abnormal returns of income-decreasing firms exceed those of income-increasing firms substantially, with the latter firms experiencing large negative returns over the period. We demonstrate a trading rule that, ex post, exploits the information contained in the accounting changes to yield large abnormal stock returns. The results suggest that income-increasing accounting changes are perhaps the first visible sign indicating other hidden, fundamental problems that get revealed in subsequent years.