Essays on Non-GAAP Reporting

Abstract

I investigate whether investor relations officers (IROs) serve as information intermediaries to help investors evaluate non-GAAP earnings. Using novel investor relations survey data for a large cross-section of US firms, I find that the value relevance of non-GAAP earnings is higher for firms with greater investor relations outreach, including having IROs who interact more frequently with capital market participants and have incentives tied to equity valuation. I also find that the increased value relevance of non-GAAP earnings for firms with active IROs coincides with a decrease in the value relevance of GAAP earnings. This is consistent with non-GAAP earnings being used to reduce reliance on certain GAAP income statement items. Finally, I provide evidence that firms with these IRO characteristics are generally not associated with more opportunistic non-GAAP reporting. Overall, the evidence suggests that the investor relations function plays a role in improving the value-relevance of non-GAAP earnings as a substitute for GAAP earnings. Five competing theories predict the relationship between takeover markets and non-GAAP reporting quality. My co-author and I test these theories using a novel source of variation in the strength of takeover markets for a subset of U.S. firms. Employing a difference-in-differences research design with U.S. and non-U.S. counterfactuals, we find that, when takeover markets weaken, firms more often 1) exclude recurring expenses, 2) exclude expenses incremental to those excluded by analysts, and 3) use non-GAAP earnings to meet or beat analysts’ forecasts. We further find accentuated effects when CEO pay is more tightly tied to non-GAAP performance. We conclude with evidence demonstrating that non-GAAP earnings persistence and value relevance decline when takeover markets weaken. Our evidence illustrates the disciplining role of the takeover market and supports the managerial entrenchment hypothesis. A wide variety of robustness tests corroborate this inference.