2026-07-14
GLD vs SPY: did gold defend during stock market drawdowns?
By FinScope Research · Data as of 2026-07-16
The short answer: often, but not without pain
This comparison uses FinScope USD data from 2004-11-18, the first date available for GLD, through 2026-07-16. Across the five deepest SPY drawdown episodes since GLD launched, GLD had the better end-to-end return every time. It finished positive during the global financial crisis, the fourth-quarter 2018 selloff, and the 2025 correction.
Gold did not rise in every stock selloff. GLD returned -3.6% from SPY’s pre-COVID peak to its March 2020 trough, and -7.3% during the 2022 SPY bear market. During the financial-crisis window, GLD gained 23.9% from start to finish but still suffered a 29.4% drawdown inside that same period. Gold often cushioned the loss. It was not a painless hedge.
Comparison method
Both datasets end on 2026-07-16. The common comparison window begins on 2004-11-18 because SPY has a longer history than GLD. All prices and returns are in USD, on dividend-adjusted daily closes.
Each SPY episode runs from a prior peak to the subsequent trough. GLD’s return uses those same start and end dates. The GLD max drawdown inside the window column measures GLD’s fall from its own interim high during that period. That extra figure matters because start-to-finish returns can hide a rough path in between.
Current snapshot and observed maximum drawdowns
| Metric | GLD | SPY |
|---|---|---|
| Data date | 2026-07-16 | 2026-07-16 |
| First data date | 2004-11-18 | 1993-01-29 |
| Name | SPDR Gold Shares | State Street SPDR S&P 500 ETF Trust |
| Current price (USD) | $364.96 | $750.72 |
| Current decline from peak | -26.4% | -0.9% |
| Observed max drawdown (USD) | -45.6% | -55.2% |
| Peak before worst drawdown | 2011-08-22 | 2007-10-09 |
| Trough date | 2015-12-17 | 2009-03-09 |
| Peak-to-trough span | 1,578 days | 517 days |
| Trough-to-recovery time | 1,686 days | 1,256 days |
| Recovery date | 2020-07-29 | 2012-08-16 |
GLD’s observed maximum drawdown was -45.6%, about 9.6 percentage points shallower than SPY’s -55.2%. Its worst cycle lasted longer, though. GLD took 1,578 days to fall from its 2011-08-22 peak to the 2015-12-17 trough, followed by another 1,686 days to recover that old high.
A shallower loss did not mean a faster recovery. Gold can have a long bear cycle that does not line up with the stock market’s worst period.
What GLD did during SPY’s major selloffs
The table covers the five deepest SPY drawdown episodes observed since GLD began trading. GLD’s return uses the same peak-to-trough dates, while the last column captures GLD’s worst drawdown inside each window.
| SPY peak→trough | SPY drawdown | GLD return over same dates | GLD max drawdown inside window |
|---|---|---|---|
| 2007-10-09 → 2009-03-09 | -55.2% | +23.9% | -29.4% |
| 2020-02-19 → 2020-03-23 | -33.7% | -3.6% | -12.5% |
| 2022-01-03 → 2022-10-12 | -24.5% | -7.3% | -21.0% |
| 2018-09-20 → 2018-12-24 | -19.3% | +5.0% | -2.7% |
| 2025-02-19 → 2025-04-08 | -18.8% | +1.6% | -5.0% |
GLD beat SPY on an end-to-end basis in all five episodes. The difference was especially large during the financial crisis, when SPY lost 55.2% and GLD gained 23.9%. In the fourth quarter of 2018, GLD gained 5.0% while SPY fell 19.3%.
The COVID crash and 2022 were less tidy. GLD also fell, by 3.6% and 7.3%, respectively. The losses were smaller than SPY’s, but gold did not move in the opposite direction. In this sample, GLD behaved less like an asset that must rise when stocks fall and more like one that can move by a different magnitude.
Why the path matters
The 23.9% financial-crisis return can make GLD look smoother than it was. Inside that window, GLD had a 29.4% drawdown from its own interim high. During the 2022 SPY bear market, GLD’s internal drawdown reached 21.0% even though its start-to-finish loss was only 7.3%.
Investors experience the path, not just two endpoints. Drawdown depth and time to recovery belong next to total return. See What is maximum drawdown? and What is recovery time? for the definitions.
SPY is currently much closer to its peak
As of 2026-07-16, SPY was 0.9% below its peak, while GLD was 26.4% below its own. GLD often reduced losses during past stock selloffs, but its current cycle is different from SPY’s.
Gold should not be treated as a mechanical inverse of stocks. The evidence here is narrower: during the five largest SPY drawdown episodes since GLD launched, GLD performed better each time, but it still posted losses and sizable interim drawdowns in some of them.
How to view it in FinScope
Open the GLD page and SPY page to compare current decline from peak, observed maximum drawdown, trough dates, and recovery status. For the concepts, read What is maximum drawdown?, How to read underwater periods, and What is recovery time?. Korean investors may also want How FX changes drawdowns.
Methodology: drawdowns are peak-to-trough declines on adjusted daily closes; windowed figures measure GLD from its own interim high inside each SPY episode. Recovery counts calendar days from trough to the first reclaim of the prior high. See the methodology page.