Auditor Fees and Audit Quality

Abstract

The paper aims to examine the relation between fees paid to auditors and audit quality during the period of 2000-2003. The paper constructs a measure of auditor profitability that is used as a proxy for auditor independence. The methodology is grounded in the notion that auditor independence is influenced by effort and risk-adjusted fees, rather than the level of fees received from clients. Since, risk and effort are unobservable, the paper uses proxies based on client size, complexity and risk to estimate abnormal fees. Abnormal fees are derived using a fee estimation model drawn from prior literature. The paper employs two metrics to assess audit quality - the standard deviation of residuals from regressions relating current accruals to cash flows and the absolute value of performance-adjusted discretionary accruals. The paper documents a statistically significant negative association between total fees and both audit quality proxies over all years. These findings are robust to a variety of additional tests and several alternative design specifications. The results (pre- and post-SOX) are consistent with economic bonding being a determinant of auditor behavior rather than auditor reputational concerns. The possibility that the empirical tests do not completely capture the impact of unobserved risk cannot be ruled out, though the paper attempts to do so by employing alternative specifications and sensitivity tests. Policy makers should note that current restrictions on the provision of non-audit services may not sufficiently resolve the issue of economic bonding and its impact on auditor independence. In contrast to previous studies whose results are ambiguous, the paper finds a statistically significant positive association between several measures of total fees (it uses size-adjusted and abnormal fees) and two metrics of accruals quality in all years (2000-2003), consistent with economic bonding being a determinant of auditor behavior rather then auditor reputation concerns.