Determinants for disclosure of the key audit matters in non-financial companies of B3

Abstract

This research analyzes the determinants of the quantity of key audit matters (KAM) reported in audit reports from 2016 to 2019 of the non-financial companies listed on B3. The longitudinal regression model for panel data - Poisson’s Log-Linear was applied to an annual sample of at least 272 companies, resulting in 1,198 reports and 3,295 KAM (dependent variable), considering as explanatory variables the aspects of the auditor, the auditee, and corporate governance mechanisms. The findings demonstrate that being audited by the big four is negatively associated with the quantity of KAM. This may be due to the big four or not reporting the same mean number of KAMs per company. Auditors of larger companies, less profitable and with greater operational risk, disclose more KAM, probably aiming to ensure their independence. The audit committees were not statistically significant, not impacting the quantity of KAM reported. The results contribute to the literature on factors that affect the disclosure of KAM in a non-European context (Bepari et al., 2022) with four years of data, and to companies, auditors, and regulatory and supervisory bodies, through Poisson’s multilevel regression, which is an appropriate approach for structured counting data in clusters of relatively similar observations (panels). This advances the studies by Ferreira and Morais (2020) and Guedes et al. (2021), who used ordinary least squares (OLS), and Colares et al. (2020), who used the Chi-square test in one- or two-year analyses of characteristics that tend to establish the quantity of KAM in Brazil.