Auditing models for the detection of qualified audit opinions in the UK public services sector

Abstract

We develop classification models that could assist auditors during the examination of firms operating in the public services sector. The sample consists of 228 firms from the Health and Social work and the Education UK sectors. The models, both sector-specific and general ones, are developed with discriminant and logit analysis through a 10-fold cross-validation procedure. The models achieve a satisfactory accuracy in classifying financial statements as qualified or unqualified. The variables that are important across all models are the logarithm of total assets, the return on capital employed, and whether the auditor is one of the big four or not. Copyright © 2006 Inderscience Enterprises Ltd

Predicting the type of auditor opinion: Statistics, machine learning, or a combination of the two?

“…Ordinary least squares (OLS) regression ◆ Low profitability as measured by profit margin to total assets ◆ Low liquidity as measured by current assets divided by current liabilities Gaganis and Pasiouras (2006) 114 qualified and 114 unqualified audit opinions from 2003-2004 (UK and Irish companies) Discriminant analysis and logistic regression ◆ Large size (assets) ◆ Low profitability ◆ Low growth ◆ A non-Big N auditor Gaganis, Pasiouras, and Doumpos (2007) 264 qualified and 3,069 unqualified audit opinions from 1997-2004 (public companies listed on the London stock exchange) Probabilistic neural network (PNN), artificial neural network (ANN), and logistic regression ◆ Low profitability ◆ Poor credit rating ◆ Extreme values (both high and low) of days payable outstanding ◆ Small companies Gaganis, Pasiouras, Spathis, et al (2007) 980 qualified and 4,296 unqualified audit opinions from 1998-2003 (UK companies) K-nearest neighbors (k-NN), discriminant analysis, and logistic regression ◆ High credit risk ◆ High liquidity (current ratio) ◆ High leverage ◆ Low growth ◆ Low profitability Kirkos et al (2007) 225 qualified and 225 unqualified audit opinions from [1995][1996][1997][1998][1999][2000][2001][2002][2003][2004] Characteristics of clients, both financial and non-financial, are the most important group of predictors used in prior studies, which is expected, given that the main reasons for issuing a modified audit opinion are materially misstated financial reports and substantial uncertainties related to the client's operations. 3 Characteristics of audit firms are somewhat less frequently considered as predictors, with the auditor's membership of Big N being the most researched effect.…”

“…The potential reasons for issuing a modified audit opinion, in order of decreasing frequency, are: a) materially misstated financial reports (reports do not conform to the international standards of financial reporting) or inadequate disclosure; b) substantial uncertainties exist about the entity's ability to continue operations (going concern) or some other important aspect of client's operations; c) auditor's inability to obtain enough evidence to verify the information presented in financial reports (scope limitation); and d) auditor's lack of independence. 4 One study(Gaganis and Pasiouras, 2006) reports that the auditor's membership of Big N decreases the chances of qualified opinion.…”