The measurement of quarterly earnings: Integral versus discrete method

We examine the measurement approach for quarterly earnings. U.S. GAAP utilizes an integral approach that smooths expected annual costs over the year with exceptions for certain transitory events. This hybrid approach has been in place for more than 40 years with little evaluation of its usefulness for decision‐making. We find the hybrid approach has less explanatory power for earnings four quarters ahead than earnings excluding discrete items, but similar explanatory power when predicting annual earnings. U.S. GAAP earnings also display lower explanatory power than IFRS earnings in predicting earnings for seasonal firms in peak quarters. Because IFRS earnings are determined using a fully discrete approach, we conclude the U.S. integral method’s quarterly expense estimates weaken its decision usefulness. We also consider a pro‐forma measure of fully integralized earnings that allocates discrete earnings amounts evenly over the year. We find fully integralized earnings are more useful than GAAP earnings in predicting both earnings four quarters ahead and annual earnings and are only marginally inferior to earnings excluding discrete items for our U.S. sample. Our evidence suggests current accounting for discrete items may create demand for pro‐forma earnings. We suggest that U.S. standard setters reevaluate the measurement of interim earnings.