• accounting-based model
• Bankruptcy
• predictive ability

In recent years, due the economic and financial crisis, corporate financial distress has evolved dramatically. Bankruptcy negatively affects all firm’s stakeholders: management, creditors, investors, and regulators. Academic literature on accounting and finance has paid attention on bankruptcy prediction in order to find the variables that improve its predictive ability. Bankruptcy prediction models, also called credit-scoring models, allow assessing the future ability of firms to meet their obligations and to evaluate the firms’ credit risk. Prior researches find two approaches to predict the likelihood of firm’s default: accounting-based bankruptcy prediction models and market-based bankruptcy prediction models. These models use historical financial statements and market data to assess future firm’s performance. Since financial statements allow firms to share information about current and expected performance with stakeholders and market (Greco, 2010), the accounting-based bankruptcy models use information gathered from financial statements to predict the likelihood of firm’s default, by distinguishing between financially distressed firms and “healthy” firms. The latter approach is based on the assumption that the information provided by financial statements are reliable and enough to assess the financial health of firms. Nevertheless, there is no clear consensus in the literature on which variables are good bankruptcy predictors. In recent years, academic literature on accounting and finance highlights the limitation of financial ratios as predictors of bankruptcy (Beaver, 2005). Particularly, these studies provide evidence that other variables could affect the firm’s future performance and solvency. Given the limitation of the information gathered from financial statement (financial ratios) to predict default, this thesis attempts to investigate whether and how different financial and non-financial variables influence corporate financial distress and bankruptcy prediction. In this dissertation, I investigate the relevance of including “new” variables in traditional credit scoring models. I develop new models for public firms that predict financial distress and bankruptcy. Unlike previous studies, the models applied in this study use a combination of financial ratios, earnings management measures, intellectual capital performance proxies and external auditors attributes to analyze whether models containing these three kinds of variables are able to enhance the predictive power of bankruptcy prediction models. To the best of my knowledge, there are no studies, which augment the traditional accounting-based bankruptcy prediction models with the abovementioned variables. The purpose of present study is to fill this research gap providing models with higher predictive ability by reducing the misclassification between bankrupt firms and non-bankrupt firms. The dissertation provides evidence on the usefulness, in terms of predictive ability, of using different types of variables (e.g. accounting ratios, earnings management measures, intellectual capital performance indicators and external auditors’ characteristics) for bankruptcy prediction models. Firstly, in Chapter 1 I explore the historical evolution of the corporate financial distress concept analyzing the definitions provided by the Italian doctrine and the causes that generate distress situations. In Chapter 2, I examine the relationship between bankruptcy prediction and earnings management behaviors, taking into account both accrual earnings management measures and real earnings management indicators. Particularly, analyzing a sample of US public firms during the period 1998-2014, I investigate whether earnings management and real earnings management provide useful explanatory variables for bankruptcy prediction. I also study whether the earnings management measures significantly increases the predictive ability of accounting-based bankruptcy prediction models. Chapter 3 focuses on the role of intellectual capital to predict corporate bankruptcy. More specifically, examining a sample of US public companies during the period from 1985 to 2015, I test whether intellectual capital performance reduces the probability of bankruptcy. I use the VAIC as an aggregate measure of corporate intellectual capital performance. Chapter 4, provides a study on the relevance of external audit to predict bankruptcy. I suppose that bankruptcy prediction models can become more effective if complemented with audit data. Particularly, in this chapter I investigate the relationship between external auditor characteristics (e.g. auditors industry expertise, auditors size, auditors tenure, auditors fees) and the likelihood of bankruptcy. Using a sample of US public companies, I test whether the auditor’s attributes are associated with default. I also test whether the inclusion of such attributes in bankruptcy prediction models improves their predictive ability. In Chapter 5, I provide the conclusions and I review the contributions of my study. The key findings of my Ph.D thesis can be summarised as follows. The inclusion of earnings management measures, intellectual capital indicators and auditors features in an accounting-based model show that such variables include higher amount of relevant information, useful in predicting the likelihood of corporate financial distress that is not reported in the traditional financial ratios.
The dissertation makes several theoretical and practical contributions. The study can contribute to prior literature on accounting-based bankruptcy prediction models, by identifying new explanatory variables. The findings suggest that earnings management measures, intellectual capital performance indicators and external auditor characteristics can be successfully included in bankruptcy prediction models and can effectively complement traditional performance measures. I show that the inclusion of such variables improve the predictive ability of bankruptcy model, reducing the misclassification between bankrupt-firms and healthy firms. This study can have relevant practical implications for banks, investors, analysts and financial institutions in general which are interested in bankruptcy prediction. Hence, new explanatory variables related to managerial discretion, firm’s intellectual capital and characteristics of external auditors, improve the predictive power of credit scoring models. This implies a significant reduction of costs for financial institutions and creditors and improve the efficiency of investor’s decisions.