A Reexamination of Auditor versus Model Accuracy within the Context of the Going‐Concern Opinion Decision*

Abstract

Tbe Coben Commission and previous researcb bave suggested tbat auditors' opinions are inferior indicators of bankruptcy relative to tbe predictions of statistical models. Tbis researcb reexamines tbis question in Iigbt of two important considerations tbat make tbe comparison between audit opinions and model predictions considerably more reOective of tbe auditors' real-world dedsion environment. First, tbe sample is partitioned into stressed and nonstrnsed observations and tbe importance of doing so is demonstrated; second, tbe statistical models and tbe forecast errors are adjusted so tbat tbey refiect tbe proportion of bankrupt firms actually fiaced by auditors. Tbe empirical results provide convincing evidence suggesting tbat tbe notion establisbed in previous researcb tbat auditors' opinions are inferior to models in predicting bankruptcy is unfounded. It sbould be noted, bowever, tbat neitber tbe auditors' opinions nor tbe bankruptcy prediction model are very good predictors of bankruptcy wben population proportions, differences in misdasification costs, and finandal stress levels are considered. Rtsumi. Les travaux de recbercbe de la Commission Coben et d'autres travaux qui les ont pr6cedes semblent indiquer que les opinions des verificateurs sont des indicateuTS de faillite moins efficaces que les predictions des mod&les statistiques. Les auteurs se pencbent

Auditor economic incentives and going-concern opinions in a limited litigious Continental European business environment: empirical evidence from Belgium

“…However, they did find evidence that Big 6 auditors issue more non-clean audit reports when financial problems are less explicit and voluntary liquidation follows.^ Following Hopwood et al (1994), it was checked whether the results might be biased because of including both financially stressed and non-stressed firms. To this end, the same logistic regression analysis was run for the subsample of bankrupt companies and financially stressed non-bankrupt companies.…”

“…Subsequently, a second sample was created containing 392 financially stressed non-bankrupt large Belgian companies. Using common criteria in the literature (Mutchler, 1985;Hopwood et al, 1994), a company is considered to be financially stressed if it has either suffered an operational loss, a bottom line loss, or negative retained earnings in the current year or previous two years or has a negative working capital in the previous two years.…”

Uzņēmuma darbības turpināšanas audits un riska novērtējums

“…Lai novērtētu bankrota iespējamību, pētnieki izmantoja Zmijevska modeli (1984) [21], Hopvuda modeli (Hopwood model) [22] vai arī pašmāju modeļus. Piemēram, Foroghi D., Shahhahani A.M [18], pētot auditoru slēdzienu pirms bankrota, izmantoja 54 Teherānas biržā reģistrētus uzņēmumus un pēc nacionālā modeļa veica bankrota iespējamības aprēķinus.…”

Financial analysis considering distress prediction models of telecommunications companies listed in Athens Stock Exchange: Hellenic Telecommunications Organization, Forthnet, Hellas Online

“…In such difficult business conditions, it becomes necessary to assess the financial status of a firm and estimate distress probabilities. Hence, distress prediction models, also known as bankruptcy prediction models (the two expressions are used interchangeably in this study, since financial distress is considered directly related to the probability of bankruptcy (Hopwood, McKeown and Mutchler, 1994) (Reckers, 2003). Moreover, distress prediction is essential for auditors, too, when they try to conclude about the viability of a firm (Anandarajan, Anandarajan and Srinivasan, 2004).…”