We study determinants of internal control reporting decisions under Section 404 of the Sarbanes‐Oxley Act (SOX 404) using a sample of restating firms whose original misstatements are linked to underlying control weaknesses. We find that only a minority of these firms acknowledge their existing control weaknesses during their misstatement periods, and that this proportion has declined over time. Further, the probability of reporting existing weaknesses is negatively associated with external capital needs, firm size, non‐audit fees, and the presence of a large audit firm; it is positively associated with financial distress, auditor effort, previously reported control weaknesses and restatements, and recent auditor and management changes. These results provide evidence that detection and disclosure incentives play a role in whether existing material weaknesses are reported, which has implications for the effectiveness of SOX 404 in providing investors with advance warning of potential accounting problems.

How do auditors behave during periods of market euphoria? To address this question, we study auditor going-concern opinions around the time of the wave of stressed Internet companies filing to go public on Nasdaq, a period many characterize as the ‘dot com bubble’. We focus on the day the auditor signs the opinion that appears in a stressed, Internet registrants’ IPO filing and document a sharp increase in the number of opinions with dates between January 1999 and April 2000. Contemporaneous with this jump in transaction volume, and for the duration of these 16-months, Big 5 firms were less likely to render going-concern opinions to their stressed, Internet IPO registrant clients. Upon conducting tests for determinants that could lead auditors to shift their decision criteria during this euphoric audit market, we find the presence of a going-concern opinion varies with variables that proxy for client reasons (financial distress, company age, venture backing, IPO cash burn) and for less auditor independence/skepticism (recent fees for clients without venture backing and a rush-to-market for clients with venture backing) by the Big 5 firms. These findings suggest a mixed conclusion regarding the Big 5's behavior; as the presence of a going-concern opinion varies inversely with variables that proxy for both client viability and auditor self interest. As for consequences to investors, our analysis of two, three and four-year post-IPO stock delisting provides some evidence of a decrease in the predictive content (early-warning value) of Big 5 opinions signed during the Internet IPO bubble.