Evaluation of market risk charge using internal model approach

Malindi, Mvula Michael
Journal Title
Journal ISSN
Volume Title
The revised valuation of capital requirement for market risk on non-securitizations in the trading book has become a critical field to model default events. The major reason for the process to take place is to reduce the variability in capital levels across banks and to address the flaws of the current measure. This paper provides an analysis of the impact of the new regulation framework on South Africa Equity Market across different sectors. The new measure is called the Expected Shortfall (CVar) which is the replacement of Value-at-Risk (Var) and Stressed Value-At-Risk (SVar) as suggested by Basel 2.5. The study is about the impact analysis of the new measure under Basel 4 against the current Basel 2.5 Var measure, the valuation of both metrics is done using Monte Carlo Simulation and Historical Simulation Approaches to evaluate which methodology is more efficient. The results are as follows. First, Value-at-Risk (Var) underestimate the potential large losses of equity benchmarks with fat-tailed properties and Expected Shortfall proves to be the robust metric to capture tail risks in the distribution. Second, observation from the findings is Expected Shortfall (CVar) can also disregard tail events, unless the metric is calibrated from the stressful period. Third, Monte Carlo Approach proves to be more efficient than Historical Simulation Approach. The methodology is also easy to calibrate and economical. Last, portfolio optimisation reduces the potential loss from both measures as a result reduces the market risk capital charge.
Thesis submitted in fulfilment of the requirements for the degree of Master of Management in Finance and Investment In the Faculty of Commerce, Law and Management at the University of the Witwatersrand, 2018
Malindi, Mvula Michael, (2018) Evaluation of market risk charge using internal model approach, University of the Witwatersrand, Johannesburg, https://hdl.handle.net/10539/28605