Summary Market risk governance is often described in terms of models — VaR, expected shortfall, sensitivities — but the decision-usefulness of those measures depends on the governance around them.
Limit architecture Limits should exist at the level at which trading and treasury decisions are made: desk, book and portfolio. Group-level limits alone do not constrain behaviour.
Escalation Limit breaches should be reported in real time to defined addressees with defined authority to act. Late reporting is a governance failure, not an operational quirk.
Model use vs model risk The models used for risk measurement are subject to independent validation. Their limitations — netting assumptions, correlation stability, tail behaviour — should be documented and reflected in overlays where appropriate.
Stress testing Stress scenarios extend the picture beyond the VaR horizon. They are useful when they are actionable: linked to specific hedging, position or capital decisions.
MI Board-level MI should focus on utilisation against appetite, concentration, sensitivities to defined shocks, and forward-looking indicators such as position build-up.
Limitations Governance cannot substitute for market liquidity. Where positions cannot be exited in stress, the limit framework must recognise it explicitly.
Related expertise See [Market, Liquidity & ALM](/expertise/market-liquidity) and [Banking Risk](/expertise/banking-risk).
Frequently asked questions
What should risk leaders know about limit architecture?
Limits should exist at the level at which trading and treasury decisions are made: desk, book and portfolio. Group-level limits alone do not constrain behaviour.
What should risk leaders know about escalation?
Limit breaches should be reported in real time to defined addressees with defined authority to act. Late reporting is a governance failure, not an operational quirk.
What should risk leaders know about model use vs model risk?
The models used for risk measurement are subject to independent validation. Their limitations — netting assumptions, correlation stability, tail behaviour — should be documented and reflected in overlays where appropriate.
What should risk leaders know about stress testing?
Stress scenarios extend the picture beyond the VaR horizon. They are useful when they are actionable: linked to specific hedging, position or capital decisions.
What should risk leaders know about mI?
Board-level MI should focus on utilisation against appetite, concentration, sensitivities to defined shocks, and forward-looking indicators such as position build-up.