A carregar...
Publicação
Loan modifications and risk of default: a Markov chains approach
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 1.69 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
With the housing crisis, credit risk analysis has had an exponentially increasing importance, since it is
a key tool for banks’ credit risk management, as well as being of great relevance for rigorous regulation.
Credit scoring models that rely on logistic regression have been the most widely applied to evaluate
credit risk, more specifically to analyze the probability of default of a borrower when a credit contract
initiates. However, these methods have some limitations, such as the inability to model the entire
probabilistic structure of a process, namely, the life of a mortgage, since they essentially focus on
binary outcomes. Thus, there is a weakness regarding the analysis and characterization of the behavior
of borrowers over time and, consequently, a disregard of the multiple loan outcomes and the various
transitions a borrower may face. Therefore, it hampers the understanding of the recurrence of risk
events. A discrete-time Markov chain model is applied in order to overcome these limitations. Several
states and transitions are considered with the purpose of perceiving a borrower’s behavior and
estimating his default risk before and after some modifications are made, along with the determinants
of post-modification mortgage outcomes. Mortgages loans are considered in order to take a
reasonable timeline towards a proper assessment of different loan performances. In addition to
analyzing the impact of modifications, this work aims to identify and evaluate the main risk factors
among borrowers that justify transitions to default states and different loan outcomes.
Descrição
Dissertation presented as the partial requirement for obtaining a Master's degree in Statistics and Information Management, specialization in Risk Analysis and Management
Palavras-chave
Loan Modification Default Markov Chains Self-Organizing Maps Credit Risk