Central supervision and earnings management: Quasi-experimental evidence from China

Abstract

Earnings management is costly to society because it decreases the informativeness of earnings and hence distorts capital market efficiency. Drawing upon a natural experiment generated by the staggered random on-site inspection programme initiated by China's central government between 2013 and 2017, this paper finds that highly intensive central supervision significantly decreases local firms' earnings management behaviours. Moreover, the effect of central supervision is found to be more pronounced in provinces with severe GDP exaggeration, provinces with local governors facing impending promotion, and firms controlled by the government. These findings suggest that on-site inspections by the central government may alleviate local officials' political incentives and ability to pressure local firms to engage in earnings management. However, the estimation results of timing tests indicate that this monitoring effect is short-lived, calling for a more comprehensive strategy to enhance the supervision of local officials and consequently improve the reliability of firms' financial reporting quality. These findings highlight the importance of addressing the agency problem between central and local governments in curbing firms' earnings manipulation to improve the capital market efficiency of economies characterized by strong government intervention.