Stressing Your Live Book
Running your current positions through a set of hypothetical shocks — a rate move, a vol spike, a credit widening — tells you what would happen before it actually happens, which is the whole point of doing it in advance.
Your risk report tells you today's P&L and today's exposures. It doesn't tell you what happens to your book if rates jump 50 basis points overnight, or if implied vol doubles, or if a sector you're long suddenly gaps down 10%. Stress testing your live book means deliberately running those hypothetical scenarios against your current positions — not history repeating, just "what if" — so you find out the answer at your desk on a quiet afternoon rather than during the event itself.
What a stress test actually does
A stress test takes your book exactly as it sits right now — every position, every hedge, every option with its current strike and expiry — and revalues it under a set of shocked market conditions. The shocks are usually a mix of things that have happened before (a repeat of a known crisis move) and things that haven't but plausibly could (a shock sized to some multiple of normal daily volatility). The output is a single number or small table: "if this scenario happens, the book loses roughly this much," which is a very different piece of information from your daily VaR, since VaR describes the typical range of outcomes while a stress test describes a specific, chosen bad day.
A trader running an options book that looked flat on delta and reasonably hedged on vega ran a scenario where the underlying gapped 8% overnight with implied vol jumping simultaneously. The stress run showed the book losing far more than the vega hedge alone would suggest, because the gap move pushed several short-dated options from out-of-the-money to near-the-money, where their sensitivity to further moves is highest — a nonlinearity that the book's static Greeks, measured at current prices, didn't capture. Rerunning the hedge to account for that gamma exposure before the event, rather than after, is the entire value of having run the stress scenario in the first place; the same information discovered live, during an actual 8% gap, would have arrived too late to act on calmly.
Stress tests are only as useful as the scenarios chosen — a shock list that only includes mild, familiar moves will always report a comfortable book, which is a false sense of safety rather than an actual one.
Stressing your live book means revaluing your current positions under hypothetical shocks — not "what has my book actually lost," but "what would it lose if this specific bad scenario happened" — so nonlinear risks like option gamma near strikes show up on a calm afternoon rather than during the real event. The exercise is only as good as the scenarios you choose to run.
Further reading
- Crouhy, Galai and Mark, The Essentials of Risk Management