Outlier based early warning is especially useful when considering the many scenarios risk managers shock their positions with. A risk manager at a global bank explained that they run close to 200 daily scenarios against hedge fund counterparties alone, but that the amount of data was overwhelming and therefore largely ignored.
This insight gave rise to the StressGrades methodology to prioritize attention on escalating market based early warning signals. StressGrades are designed to complement the existing stress testing process by (a) drawing attention on escalating visible risk, as well as (b) highlighting abnormally low visible risk themes to search for hidden risk.
1. PStress = Market Implied Probability of a Stress Scenario, in bps per annum
2. DStress = Distance to stress scenario in standard deviations (e. g., a Z-score)
3. StressQ = Quantile (percentile) historical rank of stress scenario (e. g., StressQ = 0.82 implies stress levels have exceeded current levels 18 % of the time)