IMPACTS OF THE IFRS 9 EXPECTED LOSSES MODEL ON BRAZILIAN BANKS

DID IT OCCUR AS FORESEEN?

Abstract

This study aimed to verify whether the estimates by Dantas, Micheletto, Cardoso and Freire (2017) that after the adoption of IFRS 9 the levels of credit losses of Brazilian banks in IFRS statements would be higher than those recognized under BRGAAP were confirmed. The empirical tests were based on the financial statements – BRGAAP and IFRS – from 2009 to 2020 of 41 Brazilian banks that disclose in the two accounting standards. Initially, the comparison between the BRGAAP and IFRS statements in the pre-IFRS 9 period revealed that the losses in BRGAAP were higher than those recorded in accordance with the incurred loss model of IAS 39. Then, the analysis of the IFRS statements in pre and post-IFRS 9 periods demonstrated that the expected loss model increased the levels of provisions compared to those recognized under IAS 39. Finally, it was found that despite the provisions being nominally higher in the expected loss model of IFRS 9 compared to the mixed model of the BRGAAP standard, this difference is not statistically relevant, not corroborating the estimates by Dantas et al. (2017). Empirical evidence suggests the presence of the anchoring effect of the loss model established by the regulator, pursuant to CMN Resolution nº. 2682/99, in the recognition of losses in IFRS 9. The findings allow for an understanding of how the expected loss model is being applied in Brazilians banks, in addition to serving as a basis for the adoption of an equivalent model in the BRGAAP standards from 2025, according to CMN Resolution nº. 4966/2021.