M.Com. (Financial Economics) / This study investigates the best measure of extreme losses in the South African equity derivatives market, and applies this to estimate the size of a default fund for Safcom, the central counterparty (CCP) for exchange-traded derivatives in South Africa. The predictive abilities of historic simulation Value at Risk (VaR), Conditional VaR (CVaR), Extreme VaR (EVaR) calculated using a Generalised Extreme Value (GEV) distribution and stress testing are compared during historic periods of stress in this market. The iterative cumulative sum of squares (ICSS) algorithm of Inclan and Tiao (1994) is applied to identify significant and large, positive shifts in the volatility of returns, thus indicating the start of a stress period. The FTSE/JSE Top 40 Index Future (known as the ALSI future) is used as a proxy for this market. Two key periods of stress are identified, namely the 1997 Asian crisis and the 2008 global financial crisis. The maximum daily losses in the ALSI during these stress periods were observed on 28 October 1997 and 6 October 2008. For the VaR-based loss estimates, 2500 trading days’ returns up to 28 October 1997 and 2750 trading days’ returns up to 6 October 2008 is used. The study finds that Extreme VaR predicts extreme losses during these two historic periods of stress the most accurately and is consequently applied to the quantification of a default fund for Safcom, using 2500 daily returns from 5 June 2003 to 31 May 2013. The EVaR-based estimation of a default fund shows that the current Safcom default fund is sufficient to provide for market losses equivalent to what was suffered during the 2008 global financial crisis, but not sufficient for the magnitude of losses suffered during the 1997 Asian crisis.
| Identifer | oai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:uj/uj:11517 |
| Date | 11 June 2014 |
| Creators | Lourens, Karina |
| Source Sets | South African National ETD Portal |
| Detected Language | English |
| Type | Thesis |
| Rights | University of Johannesburg |
Page generated in 0.0031 seconds