Charta is on Google Play: practise on the go. Get the app
Charta

Describe where stress testing sits alongside other risk measures, especially when it is done enterprise-wide.

FRM® Part I · Valuation and Risk Models · Stress Testing · FRM1.VR.8.b

The idea

Stress testing sits beside VaR, expected shortfall and economic capital in the risk framework. Those measures describe losses with a probability attached under broadly normal conditions, while stress tests describe losses under named severe conditions. Done enterprise-wide, one coherent set of scenarios is applied across all business lines, legal entities and risk types (market, credit, liquidity, operational), and the results are aggregated. This matters because stress losses interact. Trading losses, credit deterioration and funding strain tend to arrive together, and diversification that works in normal times often disappears. Adding up separate silo stress tests built on different assumptions misses these interactions and usually understates the total. Enterprise-wide results are then compared with risk appetite, capital and liquidity buffers, and reported to senior management alongside VaR and expected shortfall.

A practice question

From the question bank, with the reason each wrong option is wrong.

Which description best fits enterprise-wide stress testing?

  1. Separate scenarios chosen by each trading desk or business unit, with the resulting stress losses simply added together to give a firm-wide figure.Why not: Desk-level, independent scenarios are silo stress tests; they miss the consistency and interactions that enterprise-wide testing captures.
  2. A consistent set of scenarios applied across all business lines, risk types and legal entities, with the results aggregated. - correct
  3. A stress test of market risk in the trading book only, run at the highest level of consolidation.Why not: Covers a single risk type; enterprise-wide testing spans market, credit, liquidity and other risks.
  4. A VaR calculation for the whole firm at a confidence level higher than the one used for desks.Why not: This is a firm-wide VaR, a statistical measure, not a scenario-based stress test.

Enterprise-wide stress testing applies a coherent scenario to the whole institution at once, so interactions between businesses and risk types are reflected in the aggregate result.

A harder one

A bank adds up stress losses calculated separately by its market, credit and liquidity teams, each using its own scenario assumptions, and reports the total as its firm-wide stress loss. What is the main weakness of this approach?

  1. Adding losses always overstates risk because diversification between risk types lowers losses in stress.Why not: Diversification typically weakens in stress, so the sum is not reliably an overstatement.
  2. The sum is valid only if each team uses a 99% confidence level for its loss figure.Why not: Scenario losses have no confidence level; consistency of assumptions is the issue.
  3. Firm-wide stress losses should include the trading book only, so the credit and liquidity figures ought to be excluded from the firm-wide total.Why not: Enterprise-wide stress testing is meant to span all material risks, not just the trading book.
  4. The scenarios are inconsistent and interactions between risks are not captured, so the sum can understate the combined loss. - correct

An enterprise-wide test uses one coherent scenario so that, for example, the same market fall hits trading positions, loan quality and funding together. Separate scenarios miss these links.

The hardest tier

Separate stress tests give losses of USD 6 billion (market), USD 9 billion (credit) and USD 2 billion (operational), a total of USD 17 billion. An integrated enterprise-wide test of a single consistent scenario gives USD 23 billion. What most plausibly explains the higher figure?

  1. The integrated test uses a higher confidence level than the separate tests, which mechanically raises the loss estimate above the simple sum of the desks.Why not: Scenario tests do not use confidence levels; the difference comes from interactions and consistency.
  2. Interactions across risks, such as one spread widening hitting trading and loan books together while funding costs rise, amplify the combined loss. - correct
  3. The separate tests double count some losses, so the integrated figure is the corrected, larger total.Why not: Double counting would make the sum of silo results larger than the integrated figure, not smaller.
  4. The integrated test adds expected losses on performing loans, which the separate tests exclude by design.Why not: Expected losses are not the reason; the point is that risks interact in a single stress event.

6 + 9 + 2 = 17, so the extra USD 6 billion is not in any silo. It arises from linkages: one shock hits several books, funding strain follows losses and diversification falls away.

In the app: 6 practice questions on this outcome, spaced review that brings it back before you forget it.

Start free →

More in Stress Testing