Three essays in financial accounting

Abstract

Chapter three investigates how politics influence judicial outcomes from the perspective of corporate litigations. First, I find that firms and individuals receive lower monetary and non-monetary penalties if the Judge is from the political party that they support. This relationship is stronger for individual respondents than firm respondents, suggesting that politics plays an even bigger role when an actual person is involved. Second, I find that firms that make political contributions enjoy significantly lower monetary penalties than firms that do not make political contributions. Conditional on non-zero political contributions, Democrat firms receive lower monetary penalties than Republican firms holding all else equal. Third, Republican Judges tend to sentence higher non-monetary penalties than Democrat Judges, but they do not differ significantly on monetary penalties. In addition, the overall political environment moderates our findings. If the president in office is Republican, the effect of political contributions on reducing monetary penalties is stronger, and the tendency that Republican Judges have towards higher non-monetary penalties is more evident.

This dissertation consists of three distinct but related essays examining financial accounting issues by emphasizing the impact of the “human” element.

Chapter one examines the impact that the social networks of top executives and directors have on firms’ earnings management. I find that well-connected (i.e. central) firms are more aggressive in managing earnings through both accruals and real activities. Firm pairs that are socially connected to each other have similar magnitudes of earnings management and similar changes in earnings management over time compared to firm pairs that do not share a connection. Furthermore, the magnitude of earnings management decreases after a socially connected executive or director dies. In addition, among different types of social networks, professional networks have the strongest impact on both discretionary accruals and real earnings management; non-professional networks from activities such as charity and clubs only affect real earnings management but not discretionary accruals; school networks do not have significant effects on either. Finally, CEOs’ network centrality is associated with both high accrual- and real-based earnings management; CFOs’ network centrality has a even larger positive effect on accruals than CEOs’ network centrality but no influence over real-based earnings management. Collectively my findings suggest that social network is an important determinant of earnings management.

Chapter two studies the influence of CFO social networks on firm tax avoidance. I construct the social networks of the CFOs of US companies based on their employment history, education, and non-professional activities. I find that firms with central (well-connected) CFOs have lower effective tax rates (ETRs) than firms with non-central CFOs. Moreover, the ETR of a firm decreases if the CFO centrality of the firm increases. Furthermore, firm pairs have similar ETRs if their CFOs are socially connected, suggesting tax related information exchange among CFOs through social networks. The past ETRs of central CFOs predict the ETRs of non-central CFOs that they are socially connected with, suggesting that non-central CFOs follow central CFOs for tax planning. Overall, my findings suggest that socially well-connected CFOs have more information and resources to adopt aggressive tax strategies than not-well-connected CFOs.