Skip to content

Stress Testing

Difficulty advanced

Overview

Stress testing evaluates portfolio performance under extreme but plausible market scenarios. Unlike VaR (which measures normal conditions), stress tests capture tail events.

Historical Scenarios

Major Historical Events

Event Date S&P 500 10Y Treasury VIX
Black Monday Oct 1987 -22.6% - -
Dot-com Burst 2000-2002 -49.1% - -
Financial Crisis 2008-2009 -56.8% +15% 80
Flash Crash May 2010 -9.1% (intraday) - -
COVID Crash Feb-Mar 2020 -33.9% - 82.7
SVB Crisis Mar 2023 -5% + -

Typing Shocks: Infrastructure vs Regulatory

Not all crisis events transmit through the same channel. In crypto (and increasingly in markets with concentrated infrastructure dependencies — DTCC, CHESS, single-cloud exchanges), the empirical evidence separates two regimes:

Shock type Examples Vol impact Return impact (CAR)
Infrastructure failure Exchange outage, custodian insolvency, oracle exploit, bridge hack ~5.7× larger vol response than regulatory −7.6%
Regulatory event Enforcement action, ban announcement, tax-rule change Baseline −11.1%

The headline finding from a 2026 event-study: vol impact differs by a factor of ~5.7×, but the return responses are statistically indistinguishable (p ≈ 0.81 for the difference). Both punch a roughly similar hole in price; one drives much more two-sided volatility on the way down and back.

Implication for hedge sizing

The two shock types call for different hedge instruments, not just different sizes:

Goal Infrastructure shock Regulatory shock
Protect against price drop Out-of-the-money puts work for both, similar notional
Protect against volatility itself Long-vol position (straddle, VIX-equivalent, long variance) Smaller vol hedge — directional dominates
Pre-position cost Higher (long vol bleeds theta) Lower
Trigger to add Concentration risk in infra · single-venue dependency Pending policy events, hostile jurisdiction exposure

Methodological caveat: the cited study is a 2026 event-study with a relatively small sample of categorized events and a 2018–2025 window. The qualitative ordering (infra > regulatory for vol impact) is robust across specifications; the precise multiplier (5.7×) is sensitive to the event window and the volatility estimator. Treat it as a direction, not a calibration target.

Practical Guidelines

  1. Multiple Scenarios — Test various types of stress
  2. Correlation Breakdown — Correlations increase in crises
  3. Liquidity Stress — Can you exit positions?
  4. Counterparty Risk — What if your broker fails?
  5. Leverage Amplifies — Small moves become large with leverage
  6. Regular Testing — Stress test quarterly or after major changes
  7. Action Plans — Have predefined responses for each scenario

Next Steps