Quantificação de risco operacional

• Ano de defesa: 2012
• Autor(a) principal: Aaltonen, Alex
• Orientador(a): Sicsú, Abraham Laredo
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Palavras-chave em Português: Risco operacional; Abordagem de distribuição de perdas; Medida de Kullback-Leibler; Distribuições empíricas; Distribuição do capital
• Link de acesso: https://hdl.handle.net/10438/10246

Measuring operational risk is necessary as it affects the value and survival of companies. A central focus of researchers, professionals in the financial sector, regulators, and bank supervisors involves controlling this risk. For this study, we explored four applications of the loss distribution approach for quantifying operational risk. A set of operating losses spanning two years at a major Brazilian bank was used for the purpose of applying and testing this approach based on the four methods. The empirical distribution method was found to be the most appropriate for measuring operational risk and calculating economic capital from the available data. The operational risk quantification method based on fitting theoretical distibutions to losses revealed that the Johnson curves are particularly flexible and readily implemented. Further, the Johnson curves were fitted to the distribution of operational losses and to the empirical distribution of the economic capital amounts. Knowing the capital distribution provides us with a notion of the economic capital calculation accuracy and prepares the way for future theoretical studies on operational VaR. Rather than calculating a single capital amount, we determined the distribution of economic capital amounts. We compared two methods, used to establish capital amount distributions for the bank. Our study demonstrated the possibility of justifying verification points in internal audit procedures on the basis of operational risk data, modeling, and management. Based on these findings, we concluded by setting out recommendations for bank supervision and regulation.