Reconciling the Conflicting Results of Prior Research on the Relation between Aggressive Book and Tax Reporting

I reexamine the conflicting results in Frank, Lynch, and Rego (2009) and Lennox, Lisowsky, and Pittman (2013) . Frank et al. (2009) conclude that firms can manage book income upward and taxable income downward in the same period, implying a positive relation between aggressive book and tax reporting. Lennox et al. (2013) conclude the relation is negative and aggressive book reporting informs users that aggressive tax reporting is less likely. I identify four key differences in the research designs across the two studies, including measures of aggressive book reporting, measures of aggressive tax reporting, sample time periods, and empirical models. I systematically examine whether each of these differences is responsible for the conflicting results by altering the key difference while holding other factors as constant as possible. I find the relation between aggressive book and tax reporting is driven by the measure of aggressive book reporting, as the relation is positive for some subsets of firms and negative for others. Firms accused of financial statement fraud have a negative relation while nonfraud firms exhibit a positive relation. Using discretionary accruals, I also look for, but do not find a “pivot point” in the relation between aggressive book and tax reporting. I provide a better understanding of the relation between aggressive book and tax reporting by identifying research design choices that are responsible for prior results. I show that measures of both discretionary accruals and financial statement fraud are necessary to gain a more complete picture of the relation between aggressive book and tax reporting.