When the Cat's (Far) Away: The Effect of Control Centralization and Compensation Interdependence on Performance Misreporting

Abstract

This study investigates how control features influence employees' perceptions of controls and, subsequently, their reporting behaviors. I examine whether implementing a more centralized versus decentralized monitoring control influences employees' perceptions of whether the control will detect misreporting. In an experiment that holds the true detection rate constant, Ifind that participants perceive a relatively decentralized control as more likely to detect misreporting than a centralized control. Because decisions about monitoring controls and compensation systems are interrelated (O'Donnell 2000) and made contemporaneously (Brenner and Ambos 2013), I also examine whether compensation interdependence influences the effect that control centralization has on misreporting. Prior research shows that compensation interdependence leads to more aggressive reporting behavior (Sutter 2009). This aggressive reporting behavior is likely to reduce sensitivity to differences in subjective assessment of a control's detection likelihood resulting from control centralization. Consistent with this reasoning, I find that the difference in the extent of performance misreporting between a centralized and decentralized control is greater when compensation interdependence is low than when compensation interdependence is high. This study contributes to the accounting literature by defining control centralization and providing the first academic research on a consequence of control centralization. Moreover, this study provides theory and evidence of the joint influence of control centralization and compensation interdependence on misreporting, and it provides evidence that control design choices can influence employees' perceptions of control effectiveness.ii ACKNOWLEDGEMENTS