Modelo de notação de risco para famílias portuguesas

Abstract

Credit risk and its measurement are the primary focus of a credit institution, especially since the start of the current financial crisis, which has brought credit constraints to the illiquidity of markets and a large number of defaults observed for firms and individuals. This failure arises from poor measurement of credit risk which translates in financing borrowers with high probabilities of default. This dissertation focuses on the measurement of the household probability of default. We estimate an econometric model to determine a credit rating, based on the data from the Household Finance and Consumption Survey held in 2010, in Portugal. It was possible not only to determine the household probability of default, but also to determine the exposure to default and the loss given default. The proposed model was based on a probabilistic regression including independent variables studied by other authors. To confirm the robustness of the model a thorough statistical analysis, and its was developed adjustment, accuracy and validation were tested. It was concluded that the independent variables included in the model, the number of children, credit declined and the deteriorating credit conditions contributed significantly and positively to the probability of default. On the other hand, homeownership contributes to the decrease of the household probability of default.

Date of Award	16 Mar 2015
Original language	Portuguese
Awarding Institution	Universidade Católica Portuguesa
Supervisor	Ricardo Cunha (Supervisor) & Paulo Alves (Co-Supervisor)