Book-tax conformity e a neutralidade fiscal das IFRS no contexto brasileiro

Abstract

In the process of convergence from Brazil to IFRS, which is still evolving, two regulatory environments were instituted from tax perspective: the RTT and Law n. 12,973/14. These two regimes preach different tax systems to accommodate the new accounting standard:while the first regime established the segregation of corporate accounting from that used for tax purposes and the tax neutrality as a generic principle, the second uses corporate accounting converged to IFRS as a starting point for determining taxable profit and specifically address the fiscal neutrality of this new accounting standard. Considering that these differences may theoretically have a strong impact on the book-tax conformity level (BTC), the present study aimed to verify whether the BTC’s level in Brazil was lower during the RTT period when compared to those observed in the pre IFRS and Law nº 12,973/2014 periods. In addition, the present study aimed to investigate whether there was a statistically significant difference between the BTC levels seen in the pre IFRS and post Law No. 12,973/2014. First, the model of Atwood, Drake e Myers (2010) was used to measure the BTC of each period. Then, the model of Tang (2014) was applied to assess the convergence of the results. Both models measure the level of BTC from the analysis of the root mean-squared error (RMSE) estimated in the regressions, preaching that a higher (lower) RMSE indicates lower (higher) BTC. Based on sample data from publicly-held companies listed on B3, obtained through the Economática® software, the RMSEs for pre-IFRS (2004-2007), RTT (2010-2013) and Law No. 12,973/2014 (2015-2018) periods were calculated and compared using the Means Comparison test (t de student test). From this test, it was found that there were no statistically significant differences between the levels of BTC in the three periods analyzed from the perspective of the two aforementioned methodologies. This contradicted the first hypothesis of this study - in the RTT the level of BTC was statistically lower than in the other two periods - and did not reject the second hypothesis - there were no statistical differences between the pre-IFRS and Law No. 12,973/2014 periods. Therefore, this study contributes to the national literature that researches the effects of IFRS adoption and has potentially relevant results for stakeholders as it elucidates the behavior of this important variable (BTC) with the implementation of international accounting standards.