While value-at-risk , calculated on a daily basis, supplies forecasts for potential large losses under normal market conditions, we also perform stress tests in which we value our trading portfolios under extreme market scenarios not covered by the confidence interval of our value-at-risk models.
The quantification of market risk under extreme stress scenarios forms the basis of our assessment of the economic capital that we estimate is needed to cover the market risk in all of our positions. Underlying risk factors (market parameters) applicable to the different products are stressed, meaning that we assume a sudden change, according to pre-defined scenarios. We take the worst predicted losses resulting from applying these scenarios to the various portfolios as the economic capital for those businesses. We derive the stress scenarios from historic worst case scenarios adjusted for structural changes in current markets.
For example, we calculate country-specific event risk scenarios for all Emerging Markets and assess these event risk results daily. A committee reviews the country risk ratings and scenario loss limits bi-weekly. In addition to the country-specific event risk scenarios for Emerging Markets, we also run regular (weekly) market stress scenarios on the positions of every major trading portfolio .
Our stress test scenarios include:- Price and volatility risks for interest rates, equity prices, foreign exchange and commodity prices for industrialized countries. This covers both trading and nontrading securities and investments, as well as trading book derivatives portfolios and includes many basis risks.
- Emerging Markets' risks, including equity price declines, strong interest rate movements and currency devaluations.
- Credit spread risks for bonds and traded loans of both industrialized and Emerging Markets countries.
- Underwriting risks in debt and equity capital markets .
We calculate economic capital by aggregating losses from those stress scenarios using correlations that reflect stressed market conditions (rather than the normal market correlations used in the value-at-risk model). These calculations are performed weekly.
In 2003 we implemented a refined aggregation process for trading market risk in order to better reflect correlations among market risk factors in stressed market conditions. Based on this refined aggregation, our economic capital usage for market risk arising from the trading units totaled € 1.0 billion on December 31, 2003 and € 0.8 billion on December 31, 2002, compared with € 0.9 billion for year-end 2002 based on our previous aggregation process.