I examine whether a reversal of an abnormal cut in discretionary investments is associated with the degree to which the cut is reflective of real earnings management (REM) and whether and how it predicts future operating performance. I define a reversal as occurring when a firm cuts discretionary investments to a below-expected level in one period and reverts back to at least the expected level of investment during the next period. Unlike accrual earnings management, REM involves deliberately altering the operations of the firm to influence reported accounting numbers. To the extent that such interventions diverge from optimality, they can expose the firm to real economic costs. I find that a reversal of an abnormal cut in discretionary investments in the year after the cut has taken place is indicative of REM. I further find that, on average, reversing cuts are associated with lower future operating performance, but that such results vary significantly depending on the various incentives to engage in REM, as well as other factors that affect its associated costs and benefits. These findings are of interest to investors, regulators, and academics with respect to the identification and consequences of REM.

Mean reversion in profitability and growth is a well-documented phenomenon; however, comparatively less is known about the level of analysis that best captures mean reversion. In this study, we find that analyzing firms by life cycle stage improves the out-of-sample accuracy of profitability and growth forecasts over analyzing firms pooled across the economy and analyzing firms by industry. The improved accuracy is robust to both short-term and long-term forecasts of profitability and growth. We also find that the improvement in accuracy of forecasts from life cycle analyses is greatest for firms in the introduction and decline stages and for firms with greater uncertainty. Finally, we examine market participants' use of life cycle information in forming expectations. We find inefficient use of life cycle information in analyst forecasts, but not in management forecasts. We also find that life cycle forecasts are associated with year-ahead abnormal stock returns.

No abstract