Credit Modelling 50 Years after Altman's Z-score

Luca Sorlini

Student thesis: Master thesis

Abstract

This dissertation investigates the relationships between probability of default and firm-level covariates. Two hypotheses are tested by comparing different specifications. I find abundant evidence suggesting that some univariate relationships between firm-level covariates and probability of default and some ceteris paribus relationship between firm-level covariates and probability of default are not linear. The empirical results presented in this paper suggest also that, when a large sample is used, additive models outperform standard linear models. Finally, I also provide an explanation for any non-linearity that is presented.

EducationsMSc in Accounting, Strategy and Control, (Graduate Programme) Final Thesis
LanguageEnglish
Publication date2020
Number of pages80
SupervisorsJens Dick-Nielsen