Managing through a pandemic

: ex-ante earnings uncertainty and income-decreasing earnings manipulation in 2020

Abstract

The realization that economic conditions can affect the value-relevance of reported earnings has motivated a series of contextual studies in the field of earnings management. Less than one year after the first reported case of COVID-19 in the U.S., this thesis investigates how and why the magnitude and direction of income manipulation by North American firms changed in 2020. This study is split into a univariate analysis and a regression analysis part. In the former I run expectation models for normal accruals within each industry-quarter and proxy for accruals management through the regression residuals. The significance of the difference in the magnitude of abnormal / discretionary accruals between the comparison period (2018-2019) and the crisis period (2020) is estimated through a bootstrap procedure. In the latter I regress discretionary accruals against a proxy for high-order uncertainty (dispersion in analyst forecasts) in both the comparison period and the crisis period. First, I find that 2020 brought about a significant increase in the magnitude of income-decreasing manipulation (relative to the comparison period, firms underreported earnings by a greater percentage of assets in 2020). Second, I find that in 2020 higher ex-ante earnings uncertainty levels are correlated with higher absolute negative discretionary accruals (associated to income-decreasing manipulation). Third, I conclude that the underreporting trend in 2020 is at least partially explained by earnings shifting incentives since the effect of positive surprises on announcement returns is significantly mitigated by ex-ante earnings uncertainty.

Date of Award	28 Apr 2021
Original language	English
Awarding Institution	Universidade Católica Portuguesa
Supervisor	Fani Kalogirou (Supervisor)