Choosing a stock hedge by volatility regime
Does picking the hedge by how expensive options are (put when cheap, put spread in between, zero-cost collar when rich) keep most of a put's protection for much less of its cost?
Result
Failed. The rule cut drawdowns by 24% but still cost 71% of what always buying a put costs.
- Median drawdown cut
- 24% (needed 25%)
- Yearly cost vs always-put
- 71% (needed ≤60%)
- Tail protection vs always-put
- 120%
- Robust pricing cells
- 0 / 9 (needed 7)
What it means
Protection was real: in the worst 10% of quarters the rule protected more than an always-on put, because crashes arrive when options are expensive and the collar gets the same floor for free. But hedging every quarter close to the money costs most of what it saves. No always-on hedge Signal ships from this; any refined rule is a new, separately locked Study.
Locked setup
- Universe
- 395 current S&P 500 members with full history, plus SPY as a control
- Window
- 2007-01-01 to 2026-09-30; rules locked 2026-10-22 before the run
- Method
- 3-month hedges rolled quarterly in three staggered tranches. Options priced with Black-Scholes from realized volatility plus a markup and put skew, with 5% friction per leg. Pass needed the rule to cut median max drawdown by 25%, cost at most 60% of always buying a 95% put, keep half its tail protection, and hold the first two in 7 of 9 pricing-assumption cells.
Breakdown
| strategy | median_cagr_pct | median_max_drawdown_pct | median_drawdown_cut_pct |
|---|---|---|---|
| Stock only | 15.1 | 41 | 0 |
| Put at 95% | 4.2 | 46.5 | -11.5 |
| Put at 90% | 5.5 | 48.7 | -15.6 |
| Put spread 95/80 | 10.2 | 35.1 | 15.2 |
| Zero-cost collar | 4.6 | 19.4 | 50.8 |
| Vol-regime rule | 7.5 | 31.3 | 23.6 |
Replay it
python scripts/study_hedges.pyOutputs: docs/studies/S-004_hedge_selection.md, data/research/hedge_study_summary.csv, data/research/hedge_study_by_symbol.csv, data/research/hedge_study_sensitivity.csv