Abstract
Abstract : Over the last decade, financial markets across the world have been devastated by operational risk-related incidents. These incidents were caused by a number of aspects, such as, inter alia, fraud, improper business practices, natural disasters, and technology failures. As new losses are incurred, they become part of each financial institution’s internal loss database. The inclusion of these losses has caused notable upward spikes in the operational risk Pillar I regulatory capital charge for financial institutions across the board. The inherent imperfections in people, processes, and systems–be it by intention or oversight–are exposures that cannot be entirely eliminated from bank operations. Thus, the South African Reserve Bank mandates South African financial institutions to reserve capital to cover their idiosyncratic operational risk exposures. Investors fund capital reserves that are held by financial institutions, and these stakeholders demand a viable return on their investment. Consequently, the risk exposure and capital held relationship should be fully understood, managed, and optimised. This thesis extends Sundmacher (2007)’s work through the use of one instance of the Standardised Measurement Approach data against that of the Advanced Measurement Approach, the Standardised Approach, and the Basic Indicator Approach to estimate the potential financial benefit that financial institutions in South Africa could attain or lose, should they move from a Basic Indicator Approach to a Standardised Approach, or from a Standardised Approach to an Advanced Measurement Approach, or from an Advanced Measurement Approach to a Standardised Measurement Approach. The Advanced Measurement Approach, a Loss Distribution Approach coupled with a Monte Carlo simulation was used. Parametric models were imposed to generate the annual loss distribution through the convolution of the annual loss severity and frequency distribution. To fit the internal loss data for each class, the mean annual number of losses was calculated and was assumed to follow a Poisson distribution. The Maximum Likelihood Estimator was used to fit four severity distributions: Lognormal;Weibull; Generalized Pareto; and Burr distributions. To determine the goodness of fit, the Kolmogorov-Smirnov Test at a 5% level of significance was used. To select the best fitting distribution, the Akaike Information Criterion was used. Robustness and stability tests where then performed, using bootstrapping and stress-testing respectively. Overall, we find that the Basel Committee on Banking Supervision’s primary consideration that postulates that there is value in a financial institution moving from the Basic Indicator Approach to the Standardised Approach, or from the Standardised Approach to the Advanced Measurement Approach is indeed valid, but fails in the movement from an Advanced Measurement Approach to a Standardised Measurement Approach. The best Pillar I Capital reprieve is offered by the Diversified Advanced Measurement Approach, whilst the second best is the Standardised Measurement Approach based on an average total loss threshold of €100k (0.87% higher than the Diversified Advanced Measurement Approach), closely followed by the default Standardised Measurement Approach based on average total loss threshold of €20k (5.63% higher than the Diversified Advanced Measurement Approach). To the best of our abilities, we could not find any work that is comprehensive enough to include all four available operational risk quantification approaches (Basic Indicator Approach, Standardised Approach, Advanced Measurement Approach, and Standardised Measurement Approach), for the South African market in particular. This work foresees South African financial institutions pushing back on the implementation of SMA, and potentially lobbying the regulator to remain in AMA – as the alternative might mean increased capital requirements leading to reduced Economic Value Added to shareholders (as more capital is required at the same level of profitability or business activity). The financial institutions are anticipated to sight advanced modelling techniques as helping management have a deeper understanding of their exposures – whilst the Scenario Analysis process allows them a method of identifying their key risks and quantifying them (adding to management’s tools set). However, if South African financial institutions want to compete at a global stage and wanted to be accepted among ‘internationally active’ institutions – their adoption of SMA may not be a choice but an obligation and an entry ticket to the game (global trade).
M.Com. (Financial Economics)