Evaluation of the decision-making process for credit decisions at Preem AB

Abstract: The purpose of the following bachelor thesis report within mathematical statistics was to evaluate the decision making process at the credit department at Preem AB. The study used a logistic regression model to find a relationship between the probability of an application for credit being accepted and some quantitative and categorical factors about the applicant. These factors were both found in the applicant's financial statement and annual report as well as in data regarding risk level given to Preem AB by Upplysningscentralen. This data set was used to develop and train the logistic regression model with the aim of evaluating which factors have the biggest impact on the decisions being made after an application goes to trial at the credit department. The model was evaluated and perfected using different methods for variable selection and model evaluation. The study found that no statistically significant model could be created, and came to the conclusion there must exist further factors not covered by this study that affects a decision, or the decisions are taken randomly. Further research can therefore study which factors, such as financial security offered and level of knowledge regarding industry and financial statements among the credit controllers, affect the outcome of the manual trial of a credit application.