Modeling Total Accruals in an International Environment: The Impact of Alternative Measures of PPE

The role of accounting regulation hinges upon whether, how, and why managers manage earnings in the absence of regulatory constraints. Determining whether managers engage in earnings management to transfer wealth away from its most economic or socially effective/efficient distribution requires the researcher to first model predicted earnings, or accruals, so that unexpected variations are appropriately classed as earnings management. In this study, we examine whether refining the property, plant and equipment (PPE) variable in the Jones (1991) model of total accruals - by excluding assets not subject to depreciation or amortization, and including intangibles that are subject to amortization - improves the model's explanatory power. Results indicate that progressively refining the PPE variable improves the explanatory power of the total accruals model. The improvement in explanatory power is modest, however, and statistically significant only when the refinement is to use gross, rather than net, measures, and only when extreme observations remain in the sample. Therefore, the practical significance of this result for researchers is that written-down asset measures can be substituted for the more difficult-to-collect gross measures, particularly when extreme observations are removed from the sample. Adjustments for the effects of upward asset revaluations do not improve the model's explanatory power. © 1999 Cognitive Science Society, Inc.