SD-07.6 — Scenario Analysis & Stress Testing

Business Domain: BD-07 Investment Risk (Middle office) · Applies: BOTH

Purpose

Runs hypothetical and historical scenarios across the total portfolio to estimate what it would lose under conditions the standard risk measures do not capture. Value-at-risk describes the loss in a normal market; scenario analysis and stress testing describe the loss in an abnormal one — a repeat of 2008, a rate spike, a correlation breakdown. SD-07.6 designs the scenarios, runs them, and works the relationship backwards in reverse stress testing. It carries both measurement paradigms: market-shock scenarios for liquid holdings, and private-markets and cash-flow scenarios for illiquid ones.

Service Operations

Design and run historical scenarios — replay defined past episodes (the 2008 crisis, the 2020 shock, the 2022 rate move) across the portfolio.
Design and run hypothetical scenarios — forward-looking shocks: rate spikes, spread widening, equity drawdowns, FX shocks, correlation breakdowns.
Run climate scenarios — run the climate transition and physical pathways (the NGFS scenario set) across the portfolio; climate is a scenario family in the governed library, not a separate capability — SD-07.8 Climate Risk Analytics consumes the scenario-conditioned outputs to compute its climate-risk measures.
Run sensitivity and shock stress tests — mechanically perturb single risk factors, by full revaluation or by sensitivity approximation.
Run private-markets scenarios — capital-call acceleration, distribution delay, denominator-effect and valuation-lag scenarios that the continuously-priced model does not see.
Run reverse stress testing — start from a defined adverse outcome and identify the scenarios that would produce it.
Govern the scenario library — maintain the scenario set — historical, hypothetical, private-markets and climate — calibrate severity, and review plausibility.

Inputs and outputs

Inputs: position holdings, the factor-shock and scenario definitions (Scenario, E-17), historical episode data, the pricing and revaluation models.
Outputs: the scenario profit-and-loss, the worst-case loss estimates and the reverse-stress scenario set — recorded as Risk Measurement (E-19), consumed by SD-07.7, SD-07.8 Climate Risk Analytics (the climate-scenario outputs) and the governance bodies.

Entities

Consumes: Holding / Position (E-04, book = ibor for intraday measures, book = abor for period-end), Price & Market Data (E-08), Portfolio / Mandate (E-03).
Owns: Scenario (E-17) — the defined stress and hypothetical scenarios — and E-19 Risk Measurement on the risk_type = scenario and risk_type = stress partitions — the scenario profit-and-loss, worst-case loss estimates and reverse-stress results. SD-07.6 is the sole authoritative source for these two partitions; SD-07.7 is the consumer for consolidated reporting and limits-breach detection.

Standards

The scenario-analysis and stress-testing methodology of the CFA Institute and GARP curricula; reverse stress testing as a named technique. No single external standard governs scenario design.

Open extensions

The third level decomposes the historical, hypothetical, sensitivity, private-markets and reverse-stress operations.
The interaction with SD-01 stress on the strategic allocation, and with the insurer’s regulatory stress testing.