engleski

Working together, ETL and VAR

The paper looks at the interaction between two measures of risk, the well established VaR and a “ coherent” risk measure ETL. The rivalry between these two measures has been going for a couple of years now, but in reality there is no reason why risk practitioners should have to choose between guiding their decisions based only on VaR or ETL forecasts. Knowing both figures leads to better understanding of risks a company is facing. Three most widely used VaR models are tested on NASDAQ index (Variance-covariance approach, Historical simulation and RiskMetrics) along with three ETL models, two of which use Extreme value theory and a bootstrapping approach using historical P&L distribution. The models are tested using four years of daily return data including the period of the latest US sub-prime mortgage crisis.

Expected tail loss, Value at Risk, Extreme value theory, NASDAQ

nije evidentirano