Metro Area Client Bases, Metro Area Auditor Bases, and Client Auditor Choice

Abstract

Prior research finds no consistent association between metropolitan audit market competition (henceforth “competition”) and either audit fees or measures of audit quality (“AQ”). We revisit the association of competition with fees and AQ using a novel proxy for competition: the Office to Client Ratio (OCR). OCR equals the number of PCAOB inspected audit offices in a metro area divided by the number of public clients purchasing audit services in the area. We conduct a “contest” to determine which variable, OCR or the prevailing proxy for competition, the Herfindahl index (HERF), performs best (i.e. is significantly associated with audit fees and measures of AQ as theory suggests). To investigate the association of competition with fees, we employ a unique natural experiment: the exogenous shock to audit costs and fees in 2004 that accompanied implementation of auditors’ SOX 404(b) reports for their largest (i.e. accelerated filer (AF) clients). We document that audit offices appear to have imposed substantial fee increases in 2004 on their smaller non-accelerated filer (NAF) clients, although those clients were not required to obtain SOX 404(b) reports. Evidence suggests that these fee increases served to subsidize extension of the new SOX 404(b) services to the larger AF clients. In this setting, unusually high levels of conflict over audit fees should have existed between audit offices and managers of their NAF clients. The ability of NAF clients to resist paying fee increases should depend powerfully on competition. We find that higher levels of OCR are associated with smaller audit fee levels paid by NAFs in the shock year of 2004, and smaller fee increases from 2003 to 2004. HERF does not perform as a (reverse) proxy for competition in either year. We then investigate the associations of OCR and HERF with fee levels in three more recent years, 2010-2012. We again find that OCR outperforms HERF. In additional analyses we investigate the associations of OCR and HERF with three measures of AQ in the 2010-2012 period: absolute discretionary accruals, misstatements of audited data, and auditors’ propensity to issue first-time going concern modified opinions to potentially distressed clients. We find some evidence that competition, measured as OCR but not as HERF, is associated with improved AQ (i.e. reduced likelihood of misstatements and higher likelihood of going concern opinions). In summary, OCR appears to be an effective proxy for competition in the audit fees context. It also is significantly associated with AQ metrics although to a lesser extent. We find no evidence that HERF serves as a (reverse) proxy for competition.