Zero-shock reference

Baseline means zero shock, not most likely outcome

This reference sets equities, government rates, credit spreads and supported foreign currencies to zero change. It contains no prediction that markets will remain calm. Unlike a macroeconomic baseline with projected activity and policy paths, this baseline is simply a static check on the calculation's treatment of unchanged inputs.

There is no elapsed investment period, so coupons, dividends, cash interest, fees and financing costs do not accrue. A zero result is therefore not an expected total return. It also says nothing about inflation-adjusted purchasing power, future issuer payments or a probability that the hypothetical unchanged-market state will occur.

Check each factor branch before checking the total

Fixed equity beta multiplied by zero gives no local change. Fixed duration and spread duration multiplied by their zero shifts also give no change, and the squared rate term makes convexity contribute zero. Each modeled bond and equity row should therefore preserve value independently, not merely cancel another row's error.

Foreign holdings also remain unchanged because their translation multiplier is one. Starting values are already USD equivalents, so the engine must not convert them again. A stable total alone would miss offsetting mistakes; inspect row-level local returns, FX changes, USD P/L and ending values before accepting the control.

Unmodeled value remains unknown under zero shocks

An unmodeled holding does not become covered because all supported shocks are zero. Its return, P/L and ending value remain unestimated. The engine excludes its unknown P/L from the modeled sum while retaining its starting value in the portfolio denominator; that convention must not be described as a zero return assumption.

Add a hypothetical $1 million unmodeled asset to the $10 million default portfolio and coverage becomes approximately 90.91%. Modeled P/L remains zero, but no full ending portfolio value is supplied. The covered ending value is still $10 million; the unsupported asset's future value is deliberately not filled in.

Know what a passing identity check cannot validate

An incorrect beta, a misplaced duration or a wrong FX sign can all remain invisible when multiplied by zero. Follow the identity check with small, separate factor shocks and matched calculations. Keep unsupported nonlinear instruments outside the simple mapping rather than treating a passing baseline as evidence that the model handles them.

The Federal Reserve's revised model-risk guidance distinguishes conceptual assessment, testing and outcomes analysis [1]. This zero-shock control supports implementation testing only; it does not establish empirical accuracy, scenario plausibility or fitness for every portfolio. Even perfect reproduction cannot show that fixed sensitivities remain appropriate during a large market move.

Retain a reproducible control record

The Basel Committee's stress-testing principles include governance, methodology and documentation among the framework's core elements [2]. For this reference, preserve the holdings date, currency conventions, engine version, unsupported positions and the exact zero inputs. Those details allow later differences to be traced to a changed input or implementation.

Compare this control with other scenarios using the same starting portfolio, and distinguish rounded displays from underlying calculations. Do not relabel a zero-shock pass as regulatory approval, independent strategy validation or proof of safety. Its narrow success condition is that covered values reconcile while uncertainty about uncovered positions remains visible.

Review checkpoints

• Require zero change for every covered row, not just for the aggregate.
• Preserve unmodeled status and withhold a full ending value when coverage is incomplete.
• A zero-shock identity check verifies implementation behavior, not economic validity or lack of risk.

Source notes

1. Federal Reserve: Supervisory Guidance on Model Risk Management, revised interagency guidance of April 17, 2026
2. Basel Committee on Banking Supervision: Stress testing principles (2018)

Continue the analysis