Predicting insolvencies: the contribution of the external financial audit

• Published in: 2022 17th Iberian Conference on Information Systems and Technologies (CISTI) pp. 1 - 6
• Main Authors: Ilheu, Joao M., Laureano, Raul M. S., Santos, Marcia R. C.
• Format: Conference Proceeding
• Language: English
• Published: ITMA 22.06.2022
• Subjects: Biological system modeling; Companies; Data analysis; Decision Trees; Economics; External Financial Audit; Insolvency; Prediction; Prediction algorithms; Predictive models; Text mining
• ISSN: 2166-0727
• DOI: 10.23919/CISTI54924.2022.9820058

In Portugal, since the financial crisis, the number of SME´s that have gone into insolvency is quite high and really worrying due the impacts they cause on the Portuguese economy and society. Although several insolvencies predictive models have already been developed whose predictors are, essentially, financial information, this matter is still critical today, so it is extremely important to continue to investigate and create models with greater precision than the previous ones. In this way, and as the company´s financial statements do not always show their true economic and financial reality, this study assesses the impact of the characteristics of the audit firm and audit opinion content in predicting insolvency. To this end, more advanced data analysis techniques are used, namely text mining and decision trees with the CART algorithm to analyse the audit report between the years 2016 and 2020 of a sample of 2.040 companies, 1.020 non-insolvent and 1.020 insolvent. The results obtained indicate the existence of a relationship between the audit opinion content as well as the characteristics of the audit firm and the company's insolvencies, achieving an accuracy of 93%. Therefore, the main empirical contribution of this investigation is to bring the best knowledge about the unfeasible of companies through the external audit activity, using new techniques have never used in predictive models.