Abstract
Data mining applications in accounting and finance areas are increasing; however, there are still not many classification or prediction studies in the auditing area. This article revisits the prediction of auditor changes upon the receipt of a qualified opinion using data mining approaches in U.S. companies. Overall, the prediction rates of several data mining approaches show reasonably well using financial and other data. The authors hope to see more applications of data mining tools in accounting or finance areas in the future. However, a qualified audit opinion does not add significant incremental information value in predicting auditor changes.
TopBackground And Prior Research
Financial distress variables in predicting bankruptcy are useful for auditors to predict going concern evaluation (Gepp, Linnenluecke, O’Neill, & Smith, 2018). A stream of research (Sundgren & Svanstrom, 2014; Fernández-Gámez et al., 2016) in financial distress predicts audit opinion decisions. Sundgren and Svanstrom (2014) find that the phases of auditor careers are related to audit reporting quality, proxied by the auditors’ propensity to issue a going-concern opinion prior to bankruptcy filings. Fernández-Gámez et al. (2016) utilized neural networks to predict a qualified audit opinion using corporate governance and financial ratios. Nevertheless, non-Big 4 firms tend to be reluctant to use big data techniques to determine going concern of a firm (Read & Yezegel, 2016).
Key Terms in this Chapter
Qualified Audit Opinion: A qualified opinion signals a lack of conformity with Generally Accepted Accounting Principles (GAAP) and casts doubt on the credibility of financial statements.
Support Vector Machine (SVM): SVM is primarily used in the application of classification problems, though it can also be used for regression analysis.
Data Mining: Classification or prediction is the most common data mining method, which uses a set of pre-tagged datasets to train the model for future prediction. Fraud detection, auditor changes, and risk analysis are classic examples of applications in this technology.
Auditor Changes: Firms may change auditors after receiving a qualified opinion in favor of their favorable audit reports called “opinion shopping.”
Logit Analysis: Logit analysis is used to model the probability of a certain group such as the auditor change group in statistics.
Decision Tree: The decision tree summarizes the patterns using the format of the binary tree and it is commonly used in operation management.