Comparing GAAP with NIPA earnings

Abstract

The U.S. Bureau of Economic Analysis produces a measure of aggregate corporate profits (national income and product accounts [NIPA] earnings), which is an integral component of the accounting for gross domestic product (GDP). Interesting features of NIPA earnings include consistent accounting rules over time and determination with little or no managerial discretion. Thus, NIPA earnings provide a useful benchmark for generally accepted accounting principles (GAAP) earnings, especially in parsing out the effects of real-economy versus the accounting in explaining the documented temporal increase in volatility and decline in persistence of GAAP earnings. We find that GAAP and NIPA earnings are closely related in the early years, with similar means and standard deviations, and with earnings changes correlating at.90 during 1950-1983. This close relation substantially deteriorates, however, during the second half of the sample period, 1984-2016. Although the behavior of NIPA earnings remains roughly the same, the volatility of GAAP earnings increases 10-fold, and the correlation between GAAP and NIPA earnings changes falls to.39. Additional tests reveal that the increase in the volatility of GAAP earnings is mostly due to rapid earnings reversals, especially the effect of large transient items during economic downturns. The frequency and severity of such downturns, however, are roughly the same across the two periods. In addition, there is little change in the properties of aggregate cash flow from operations and revenue over time. Overall, this evidence suggests that in addition to changes in the real economy, changing GAAP rules and their application are significant factors in the changing properties of GAAP earnings.