CEO overconfidence and subsequent firm performance an indirect effect via earnings manipulations

Empirical evidence on the indirect effect of chief executive officer (CEO) overconfidence on future firm performance are missing. Corroborating upper echelons theory, this study identifies this effect exploiting the earnings manipulation methods as a mediator. Tobin's Q is used as a proxy for firm performance and financial press-based measure is used to capture CEO overconfidence level through articles counts. Data from 4 European samples listed on Stoxx Europe 600 Index: AEM sample (2305 observations and 461 firms), CFO-REM sample (1770 observations and 354 firms), PR-REM sample (1350 observations and 270 firms) and RD-REM sample (1295 observations and 259 firms), for the period 2010 to 2020 are used to test the mediation model of Baron and Kenny (1986). Evidence reveals that AEM/REM practices partially mediate the relationship between CEO overconfidence and subsequent firm performance. Findings from this paper can be of interest for accounting regulators, since that CEO overconfidence level can impact future performance through AEM and REM. Increasing scrutiny over AEM does not eliminate other modalities. It changes the preference of an overconfident CEO for different strategies even if they are costly for investors. © 2023 Elsevier Inc.