Volume 3 — Valuation and Risk Metrics © Copyright 2002, CCRO. All rights reserved. 31 Results from sensitivity analysis, scenario analysis, and stress testing should be used to make informed risk-taking decisions during extreme events. The results can be used in a proactive sense to help measure risk and manage a portfolio. The results can be used for making the following business decisions: • Prepare for liquidity and funding issues that occur in stressful market environments by increasing credit, counterparty lines, limits, and funding sources and managing liability structure (See the Credit Risk Management White Paper.) • Unwind a position, mark, or business • Increase hedging or purchase insurance protection so the risk can be minimized. In general, companies evaluate the following: 1. Historical events that could recur today. Examples include the significant price spikes for Cinergy in June 1998 and Henry Hub Natural Gas in December 2000. Both price movements were unexpected by the market and achieved unprecedented levels. 2. Scenario analysis based on historical events updated to current conditions 3. Ad hoc institution-specific scenarios – scenarios based on position concentration, vulnerabilities, and worst-case loss events specific to the company, including potential causes of the events 4. Extreme standard deviation scenarios – extreme moves (5–10 standard deviations) and the conditions that can cause them to occur 5. Extreme moves in correlation – correlations to 1 or –1 6. Market curve moves a. Shifts in prices of commodities, currencies, interest rates, and volatilities b. Twists in the term structure (i.e., change in the shape of the curve) c. Basis changes (change in time or location spreads) d. Liquidity – widening in bids and offers and lengthening of time to hedge or liquidate 7. Contagion (“perfect storm”) – all prices in each market move in the wrong direction 8. Modeling assumptions that are not moved daily that can be stress-tested a. Choice of pricing model b. Volatilities c. Correlations d. Other variables that may produce unreasonable results when they reach extreme values