Improving Auditors＇ Going Concern Judgment by Applying Statistical Failure Models as an Analytical Procedure

The main goal of this research is to enhance the auditor＇s judgment ability in going concern opinion by applying bankruptcy prediction models as an analytical procedure. Data for this research have been collected through questionnaires. The statistical population consists of auditors who are members of Iranian Association of Certified Public Accountants （IACPA）. The research results reflect that： （1） Auditors do not use statistical techniques for assessing going concern as an analytical procedure; （2） Auditors do not use these techniques as a tool to decrease the bias of judgments in assessing the going concern assumption; （3） Auditors do not use statistical techniques to assess audit risk in the planning stage; （4） Auditors do not use statistical techniques to assess audit risk in the final stage. Furthermore this research shows that auditors believe that the ＂standard concerning usage of analytical procedures needs more clarification＂ and ＂statistical bankruptcy predication models can help auditors in the planning stage＂. The other goal of this research is to show different auditor＇s judgments in assessing the going concern opinion with and without applying the bankruptcy prediction models as an analytical procedure. The result shows that the judgment of auditors toward the going concern assumption has improved by using statistical bankruptcy predication models.

A Statistical Analysis of Reliability of Audit Opinions as Bankruptcy Predictors

This research measures the reliability of audit firms in predicting bankruptcy for United States （US） listed financial institutions. The object of analysis is the going concern opinion （GCO）, widely considered as a bankruptcy warning signal to stakeholders. The sample is composed of 42 US listed financial companies that filed for Chapter 11 between 1998 and 2011. To highlight the differences between bankrupting and healthy firms, a matching sample composed of 42 randomly picked healthy US listed financial companies is collected. We concentrate on financial institutions, whereas the existing literature pays considerably greater attention to the industrial sector. This research imbalance is remarkable and particularly unexpected in the wake of recent financial scandals. Literature points out two main approaches on bankruptcy prediction： （1） purely mathematical; and （2） approaches based on a combination of auditor knowledge, expertise, and experience. The use of data mining techniques allows us to benefit from the best features of both approaches. Statistical tools used in the analysis are： Logit regression, support vector machines （SVMs）, and an AdaBoost meta-algorithm. Findings show a quite low reliability of GCOs in predicting bankruptcy. It is likely that auditors consider further information in supporting their audit opinions, aside from financial-economic ratios. The scant predictive ability of auditors might be due to critical relationships with distressed clients, as suggested by recent literature.