Why a scenario loss is bigger than a modelled one calculator
Why a scenario loss is bigger than a modelled one, worked through in your browser.
Interactive: why a scenario loss is bigger than a modelled oneMedium
Two exposures and a shock to each. A scenario says both happen, so the losses add. A distribution-based measure asks how often they happen together, so the correlation does the work — and at any correlation below one it returns a smaller number for the same two positions.
- Loss on A
- —
- Loss on B
- —
- Scenario loss, both at once
- —
- Aggregated at that correlation
- —
- What the correlation assumption is worth
- —
- The same at correlation one
- —
Linear exposures, one shock each, and the standard square-root aggregation — a demonstration of why the two numbers differ, not a risk model. The correlation is an assumption, it is estimated from calm periods, and it is the parameter least stable in the stress the exercise describes. Information and education only.
The idea behind it
This calculator belongs to Stress Testing, where the mechanism it computes is explained step by step, with the formula, what each input means and where the simple model stops being right. Everything is computed in your browser; nothing you type is sent anywhere.