Modelling default risk: systematic literature review and model enhancements through event data, personal risk propensities and closeness to local communities

The main focus of this thesis is to provide a comprehensive assessment of default risk modelling: the relevant literature, the empirical implementations and the levers driving improvements in prediction’s accuracy. The research begins with an overview of the main features of such models and the analysis of the relevant theoretical framework through a systematic literature review. The first chapter’s contribution to the research field provides an updated comprehensive assessment of the existing literature and identifies best practices and choices to optimize model accuracy. The following chapter lays the ground for the empirical implementation of the models to validate the applicability of the results of the systematic literature review and select the base model with the highest accuracy. In the third chapter, different enhancements are tested on the base model to build a new integrated model. The enhancements leverage on complementary research fields (i.e. behavioral finance) and aim to feed the model additional layers of soft information, complementing the core evidence derived from accounting and financial data.