A carregar...
Publicação
Backtesting of a credit scoring system under the current regulatory framework
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 989.77 KB | Adobe PDF |
Orientador(es)
Resumo(s)
Desde a implementação do atual acordo de supervisão financeira internacional, os bancos podem usar as suas estimativas internas de avaliação de risco de crédito como base para o cálculo dos ponderadores de risco e requisitos de capital. Consequentemente, com vista a assegurar a estabilidade e solvabilidade das instituições de crédito, torna-se crescente a necessidade de um sistema de validação robusto, para garantir a consistência e precisão dos sistemas
de notação interna.
Existem vários estudos sobre o processo de validação de estimativas internas. No entanto, aprofundamento e acordo nesta matéria são ainda insuficientes, nomeadamente no que diz respeito à avaliação da precisão das estimativas internas para os parâmetros de risco de crédito, com o objectivo de atingir a estabilidade dos requisitos de capital. A calibração das probabilidades de incumprimento representa um dos procedimentos de validação quantitativa inerentes ao exercício
de backtesting.
Neste trabalho, será explorado o processo de calibração das probabilidades de incumprimento recorrendo a um modelo de scoring para exemplificar como é feita a avaliação da capacidade preditiva destas estimativas internas numa carteira de Crédito à Habitação. Para superar o desafio de desenvolver um sistema de validação adequado, o presente projeto tem em consideração o atual e amplo quadro regulatório proveniente do Comité de Basileia para a Supervisão Bancária (BCBS) e da Autoridade Bancária Europeia (EBA), alguns artigos relevantes nesta matéria e aquelas que são consideradas as melhores práticas de gestão do risco de crédito.
Since the implementation of the current regulatory framework within the global financial system, banks are allowed to rely in a system using their own estimates for credit risk parameters as inputs for the calculation of risk weights and capital requirements. Consequently, in order to assure the stability and soundness of credit institutions, the need for a robust validation system to ensure accuracy and consistency of internal rating systems is greater than ever before. Although several studies on validation processes already exist, a deeper understanding and agreement on this subject is required, namely in what concerns the accuracy assessment of internal estimates for credit risk parameters, in order to achieve capital requirements stability. Calibration of default probabilities represents one of the quantitative validatio procedures underlying the exercise of backtesting that must be performed on a regular basis. The present text discusses the probability of default (PD) calibration process using a scoring model to illustrate the assessment of the predictive power of these internal estimates in a residential mortgage portfolio. To overcome the challenge of developing an adequate validation scheme in compliance with the current regulatory framework, this project project keeps in mind the legislation from Basel Committee on Banking Supervision (BCBS) and European Banking Authority (EBA), some relevant studies developed on this subject and those that are consider to be the best practices of credit risk management.
Since the implementation of the current regulatory framework within the global financial system, banks are allowed to rely in a system using their own estimates for credit risk parameters as inputs for the calculation of risk weights and capital requirements. Consequently, in order to assure the stability and soundness of credit institutions, the need for a robust validation system to ensure accuracy and consistency of internal rating systems is greater than ever before. Although several studies on validation processes already exist, a deeper understanding and agreement on this subject is required, namely in what concerns the accuracy assessment of internal estimates for credit risk parameters, in order to achieve capital requirements stability. Calibration of default probabilities represents one of the quantitative validatio procedures underlying the exercise of backtesting that must be performed on a regular basis. The present text discusses the probability of default (PD) calibration process using a scoring model to illustrate the assessment of the predictive power of these internal estimates in a residential mortgage portfolio. To overcome the challenge of developing an adequate validation scheme in compliance with the current regulatory framework, this project project keeps in mind the legislation from Basel Committee on Banking Supervision (BCBS) and European Banking Authority (EBA), some relevant studies developed on this subject and those that are consider to be the best practices of credit risk management.
Descrição
Mestrado em Ciências Actuariais
Palavras-chave
Risco de Crédito Requisitos de Capital Sistemas de Notação Interna Backtesting Calibração Taxa Média de Incumprimento de Longo-Prazo Credit Risk Capital Requirements IRB Approach Calibration Long-Run Default Rate
Contexto Educativo
Citação
Santos, Maria Ana e Castello-Branco dos (2017). "Backtesting of a credit scoring system under the current regulatory framework". Dissertação de Mestrado, Universidade de Lisboa. Instituto Superior de Economia e Gestão.
Editora
Instituto Superior de Economia e Gestão