Stressed Value-at-Risk: Assessing extended Basel II regulation

This thesis investigates recently proposed enhancements to the Basel II market risk framework. The Basel Committee on Banking Supervision introduced a stressed Value-at-Risk, calculated from one year long period of financial stress, to be added to the current VaR as a reaction to the shortage in capital reserves of banks and thus their inability to cover extensive losses observed during the recent crisis. We present an empirical evidence that such an extension of the regulatory capital is not optimal. Firstly, supplementing an unconditional methods of VaR estimation, i.e. normal parametric VaR and historical simulation, by SVaR only lead to unnecessarily high capital requirements even in a low volatile periods whilst the same amount of capital during the crisis could be achieved using either the conditional GARCH VaR with student's-t innovations or the volatility weighted historical simulation. Moreover, we showed that all unconditional methods fail to capture volatility clusters such as the 2008 crisis