Bank Market Risk — From VaR to Expected Shortfall
A tightly-bounded move to CVaR: estimators, backtesting and limit/capital impact, one clean handover.
A capable but stretched market-risk team needs to move its trading books from VaR to CVaR without opening the door to scope creep. We would scope a tightly-bounded project: Expected Shortfall estimators suited to the book, CVaR-appropriate backtesting, and impact analysis on limits and capital — documented for internal audit and regulatory review, with one clean handover.
The Challenge
The team knows what it needs — Expected Shortfall on the trading books — but is fully occupied running the existing framework. A migration like this can easily sprawl into a multi-quarter programme, and the team can't afford one.
The move also has to withstand scrutiny. Estimators, backtesting design, and the effect on limits and capital all need to be documented well enough to satisfy internal audit and regulators, not just to work in a notebook.
The Approach
We would define the boundary first, then deliver inside it — estimators, testing and impact — with the documentation that lets the work stand up to review.
1. Fit the estimators to the book
We would select and implement Expected Shortfall estimators appropriate to the specific book — its instruments, tails and data — rather than a generic textbook recipe.
2. Backtest for CVaR, not VaR
We would design backtesting suited to Expected Shortfall, whose properties differ from VaR, so the validation actually tests the measure being adopted.
3. Quantify the limit and capital impact
We would analyse how the switch changes limits and capital, giving the team and its stakeholders a clear view of the consequences before adoption.
Scope of the engagement
Expected Shortfall estimator selection and implementation, CVaR-appropriate backtesting, and impact analysis on limits and capital — all documented for internal audit and regulatory review, and delivered as a single, tightly-bounded project with one clean handover.
The Likely Outcome
The team would move its trading books from VaR to Expected Shortfall on a fixed scope and a clean handover, with estimators fitted to the book, validation designed for the measure, and a documented view of the limit and capital impact. Because everything is written up for audit and regulatory review, the migration would be defensible from the day it goes live.