Adequação das técnicas de validação dos modelos de probabilidade de default em carteiras simuladas

• Ano de defesa: 2013
• Autor(a) principal: Tsukahara, Fábio Yasuhiro
• Orientador(a): Kimura, Herbert
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Presbiteriana Mackenzie
• Palavras-chave em Português: validação de modelos; probabilidade de default; modelos de crédito
• Palavras-chave em Inglês: models validation; credit models
• Área do conhecimento CNPq: CNPQ::CIENCIAS SOCIAIS APLICADAS::ADMINISTRACAO::ADMINISTRACAO DE EMPRESAS
• Link de acesso: http://dspace.mackenzie.br/handle/10899/23422

The Basel II accord allows financial institutions to use internal models for measuring capital requirements. However, the use of internal models requires, among other approvals, a validation by an independent area. Thus, the development of techniques for validating risk models has gained much importance in the academic and in the financial markets. This study aims to evaluate some of the main techniques used for validate probability of default models, through the application of these techniques in PD models developed from simulated portfolios. Were evaluated traditional techniques such as KS, AR and area under ROC curve and newer techniques such as CIER, Information Value and measure M. The advantage of using simulated portfolios is that they allow the study of a large number of different situations, which would be impossible by using real portfolios. This study will provide an understanding of the limitations present in validation methodologies of PD models and an assessment of which techniques are more sensitive for each case analyzed.