Collar Backtest: 5 Years of Protection vs Buy-and-Hold
A protective collar — long stock, long put floor, short call cap — run on SPY from 2020 to 2024. Annualized return, the drawdown gap that defines the strategy, and the year the insurance earned its keep.
Simulated data for display. Illustrative narrative — not a verified live backtest. Build real backtests on the strategy builder.
All strategy backtests →The Test Setup
Underlying: 100 shares of SPY held throughout, with a rolling collar overlay.
Structure: long stock + long ~5% OTM protective put (the floor) + short ~5% OTM covered call (the cap, which finances the put). The baseline run targets a near-zero-cost collar.
Roll cadence: quarterly, with a 21-DTE roll rule; the put is rolled up in strong rallies to lock in gains, the call rolled out when tested.
Benchmark: the same 100 shares of SPY held unhedged (buy-and-hold).
Period: January 2020 – December 2024.
Annual Returns: Collar vs Buy-and-Hold
| Year | Regime | Collar | Buy-and-Hold | Edge |
|---|---|---|---|---|
| 2020 | Crash + recovery | +9% | +16% | -7 |
| 2021 | Melt-up | +12% | +27% | -15 |
| 2022 | Bear | ~0% | -18% | +18 |
| 2023 | Recovery | +10% | +24% | -14 |
| 2024 | Grind up | +8% | +13% | -5 |
The pattern is consistent: the collar trails in every up year because the short call caps the rally, then leaps ahead in the one bear year because the put floors the loss. Over the full 5 years it gives up roughly 3 points of annualized return for a dramatically smoother equity curve.
How Far OTM to Set the Strikes
| Put / Call Distance | Net Cost | Ann. Return | Max Drawdown |
|---|---|---|---|
| 2% put / 3% call (tight) | Credit | +4.2% | -5% |
| 5% put / 5% call ★ | ~Zero-cost | +6.1% | -8% |
| 8% put / 6% call (wide) | Small debit | +7.4% | -13% |
Tighter strikes give a near-flat ride at the cost of upside; wider strikes capture more of a rally but loosen the floor. The ~5%/5% zero-cost collar sat in the sweet spot — meaningful protection with the hedge fully financed by the call.
The Year the Protection Paid Off (2022)
2022 is the entire case for the collar. As SPY fell roughly -18% over the year, the protective put floored each leg down while the short calls expired worthless and were re-sold for fresh credit on every roll. The position finished close to flat — an 18-point edge over buy-and-hold in a single year.
There is a behavioral edge here too: a holder with a defined floor is far less likely to panic-sell at the lows than an unhedged holder watching a -22% drawdown. The collar's value is part math, part the discipline a known floor enforces.
Zero-Cost vs Net-Debit Collars
Zero-cost collar: the short call fully funds the put, so the hedge costs nothing in cash — but the cap sits closer to the money, surrendering more upside. Best when you mostly want protection and accept a tighter ceiling.
Net-debit collar: you pay a small premium so the call can be sold further out, raising the cap. Best when you still want meaningful participation in a rally and treat the small debit as the cost of a higher ceiling.
In the illustrative test the zero-cost collar produced the cleanest risk-adjusted result; the net-debit version captured more upside at the price of a deeper (but still floored) drawdown.
Five Takeaways
- A collar is insurance, not alpha. Expect to trail buy-and-hold on total return — the payoff is a far smaller drawdown.
- The edge is concentrated in bad years. The collar's entire outperformance came from the one bear year; in bull years it lags.
- ~5%/5% zero-cost was the sweet spot. Real protection with the put fully financed by the call.
- Roll the put up in rallies. Locking in a higher floor as the stock rises is what kept the drawdown shallow.
- Use it on stock you won't sell. Appreciated, high-conviction, or tax-locked positions are the ideal collar candidates.
Price your collar's floor and cap
Set the put and call strikes, see the net cost, protected floor, and capped upside before you place it.
Related Reading
Backtest narrative is illustrative — built from typical collar mechanics and historical regimes, not from live broker fills. Past performance, simulated or real, does not predict future results. See methodology.