Risk measures [10]

Go to: Summary | Previous | Next   
Bullet points include: VaR is the (1-alpha) quantile of loss distribution. Often estimated using statistical methods based on historical returns/losses. Often, though, we will want VaRs for annual holding periods. Historical data often not available at this frequency, or if it is it is deemed out-of-date. Market-implied risk modelling, see e.g. Kemp (2005) and Kemp (2009). More elaborate (usually Monte Carlo) methods. Mimic evolution of value of individual assets (and liabilities) and then aggregate to get statistics relating to portfolio value including VaR / TVaR

Contents | Prev | Next | ERM Lecture Series

Desktop view | Switch to Mobile