The short answer
From 30 August 2000 to 29 October 2026, gold beat the S&P 500: COMEX gold futures rose 1,372% (10.8% a year), against 732% (8.4% a year) for the S&P 500 with dividends reinvested and 414% (6.5% a year) on price alone. Volatility was almost identical (about 17% a year for both). But the winner changed often: gold won 13 of 25 full calendar years, and in 2026 so far stocks are ahead.
The correlation is close to zero over the whole period (0.06 for weekly returns) but swings between −0.52 and +0.61 on a rolling one-year basis. Gold rose 24.5% through the 2008 crash, yet fell in the 2020 Covid crash (−2.5%) and the 2022 bear market (−6.8%). Gold is a diversifier, not a guaranteed hedge.
Returns, volatility and drawdowns
| Measure | Gold (COMEX futures) | S&P 500 total return | S&P 500 price only |
|---|---|---|---|
| Start value (30 Aug 2000) | $273.9 | 2,083 | 1,503 |
| End value (29 Oct 2026) | $4,031.4 | 17,325 | 7,720 |
| Total change | +1,372% | +732% | +414% |
| Annualised return | +10.8% | +8.4% | +6.5% |
| Annualised volatility (weekly) | 17.3% | 17.4% | 17.4% |
| Worst peak-to-trough fall (daily closes) | −44.4% | −55.3% | −56.8% |
| Peak → trough | 22 Aug 2011 → 17 Dec 2015 | 9 Oct 2007 → 9 Mar 2009 | 9 Oct 2007 → 9 Mar 2009 |
| Back at the old peak | 23 Jul 2020 | 2 Apr 2012 | 28 Mar 2013 |
Three points stand out:
- Dividends matter. On price alone the S&P 500 returned 6.5% a year; with dividends reinvested, 8.4%. Comparing gold with the price index alone, as many charts do, flatters gold by about two percentage points a year.
- The start date matters even more. August 2000 was near the top of the dot-com boom and gold, at about $274, was only a little above its August 1999 monthly average of $256.7, the lowest since 1979 in World Bank data. Starting in 2011, near gold’s then-record, would give a very different answer, which is exactly what the drawdown row shows.
- Similar volatility, different pain. Gold’s worst fall was smaller than the stock market’s, but it lasted much longer: gold needed until July 2020 to regain its August 2011 peak, while the S&P 500 total return index was back at its 2007 peak by April 2012.
Year by year
| Year | Gold | S&P 500 (total return) | Better |
|---|---|---|---|
| 2001 | +2.5% | −11.9% | Gold |
| 2002 | +24.7% | −22.1% | Gold |
| 2003 | +19.6% | +28.7% | S&P 500 |
| 2004 | +5.2% | +11.0% | S&P 500 |
| 2005 | +18.2% | +4.8% | Gold |
| 2006 | +22.8% | +15.8% | Gold |
| 2007 | +31.4% | +5.5% | Gold |
| 2008 | +5.8% | −37.0% | Gold |
| 2009 | +23.9% | +26.5% | S&P 500 |
| 2010 | +29.8% | +15.1% | Gold |
| 2011 | +10.2% | +2.1% | Gold |
| 2012 | +7.0% | +16.0% | S&P 500 |
| 2013 | −28.2% | +32.4% | S&P 500 |
| 2014 | −1.5% | +13.7% | S&P 500 |
| 2015 | −10.4% | +1.4% | S&P 500 |
| 2016 | +8.5% | +12.0% | S&P 500 |
| 2017 | +13.6% | +21.8% | S&P 500 |
| 2018 | −2.1% | −4.4% | Gold |
| 2019 | +18.9% | +31.5% | S&P 500 |
| 2020 | +24.7% | +18.4% | Gold |
| 2021 | −3.5% | +28.7% | S&P 500 |
| 2022 | −0.1% | −18.1% | Gold |
| 2023 | +13.4% | +26.3% | S&P 500 |
| 2024 | +27.5% | +25.0% | Gold |
| 2025 | +64.4% | +17.9% | Gold |
| 2026 (to 29 Oct) | −7.1% | +13.8% | S&P 500 |
Gold won 13 of the 25 full years from 2001 to 2025. It tended to win when stocks struggled (2001, 2002, 2008) and to lose in some strong equity bull years (2013, 2021). The years in which both rose strongly together, such as 2019, 2020, 2023, 2024 and 2025, are a reminder that gold and stocks do not have to move in opposite directions.
Correlation: usually low, sometimes not
| Period | Weeks | Correlation of weekly returns |
|---|---|---|
| Sep 2000 – Dec 2007 | 382 | −0.04 |
| Jan 2008 – Dec 2009 | 105 | −0.09 |
| Jan 2010 – Dec 2019 | 521 | 0.04 |
| Jan 2020 – Dec 2021 | 105 | 0.42 |
| Jan 2022 – Oct 2026 | 252 | 0.16 |
| Whole period | 1365 | 0.06 |
Correlation measures how often two weekly returns move together: +1 means always together, 0 no linear link, −1 always opposite. Over 1,365 weeks it was 0.06, effectively independent. But the rolling one-year correlation was positive 59% of the time, peaked at 0.61 in April 2020 when both markets were driven by the same liquidity and central-bank news, and stood at 0.45 at the end of October 2026. When correlation is high, gold diversifies a stock portfolio less than its long-run average suggests.
We used weekly rather than daily returns on purpose: gold futures settle at 1:30 pm New York and US stocks close at 4:00 pm, so daily returns would be measured over different windows and would understate the link.
What happened in the big sell-offs
| Episode | Dates | S&P 500 total return | Gold |
|---|---|---|---|
| 2008 crisis | 9 Oct 2007 → 9 Mar 2009 | −55.3% | +24.5% |
| Lehman month | 12 Sep 2008 → 10 Oct 2008 | −28.0% | +12.5% |
| 2020 Covid crash | 19 Feb 2020 → 23 Mar 2020 | −33.8% | −2.5% |
| March 2020 dash for cash | 6 Mar 2020 → 19 Mar 2020 | −18.9% | −11.5% |
| 2022 bear market | 3 Jan 2022 → 12 Oct 2022 | −24.5% | −6.8% |
The pattern behind these numbers is explained in does gold go up when stocks crash?. In short: in the first, forced-selling phase of a crash (the March 2020 “dash for cash”, gold −11.5% in two weeks), investors sell what they can, including gold. Over the long 2007–2009 bear market as a whole, gold rallied. In 2022, a bear market that came with sharply rising real yields (the 10-year TIPS yield went from about −1.0% to +1.7%), gold fell too, but by much less than stocks. In the 99 weeks since 2000 when the S&P 500 fell more than 3%, gold rose in 54, with a median weekly change of +0.51%.
How we built this comparison
- Gold: COMEX gold futures, Yahoo Finance symbol GC=F, daily closes from 30 August 2000 (the first day in the series) to 29 October 2026.
- S&P 500: the S&P 500 Total Return index (^SP500TR), which reinvests dividends, and the price index (^GSPC), both from Yahoo Finance.
- Returns: close to close; annualised returns use calendar time. Volatility is the standard deviation of weekly returns × √52.
- Drawdowns: the largest fall from any previous closing high to a later low, using daily closes.
- Correlation: weekly (last close of each week) returns, whole period, by regime, and rolling over 52 weeks.
- Crisis windows: dates of the S&P 500’s peaks and troughs in each episode, applied to both series.
Limitations
- One window. Results are sensitive to the start and end dates; 2000 favoured gold.
- Futures, not physical gold. GC=F is a continuous front-month series: roll jumps add small noise, and a real futures holder would also earn interest on cash collateral (not included), while a physical holder would pay storage.
- No costs or taxes for either asset, and the S&P total return index assumes dividends are reinvested without tax.
- Correlation is linear and backward-looking. It can change quickly, as 2020 showed.
What traders can do with this
- Don’t assume a stock sell-off means a gold rally. Check why stocks are falling: forced selling and rising real yields have hit gold too. The gold correlations tool shows current correlations with stocks, the dollar and more.
- Watch real yields and the dollar. They explain more of gold’s moves than equities do; see gold vs real yields and the dollar and our gold vs bonds study.
- Size for gold’s own volatility. Gold and the S&P 500 have similar annual volatility, but gold’s daily ranges in 2026 have been unusually large; the position size calculator keeps risk per trade constant.
- Use regimes, not slogans. When the rolling correlation is high, gold and stocks can fall together.
For more comparisons, see gold vs real estate, gold vs bonds, gold vs Bitcoin and gold vs silver, all collected on our gold data studies hub.
Discuss the big picture with other traders
Inside the XAUUSD Global community, members discuss how stocks, yields and the dollar are feeding into gold each week, and share their own plans. No signals. Your first month is free, then $19.99 a month.
Sources
- COMEX gold futures (GC=F), S&P 500 Total Return index (^SP500TR) and S&P 500 index (^GSPC) daily data via Yahoo Finance, downloaded 30 October 2026; all returns, drawdowns and correlations are our own calculations
Educational content only, not financial advice. Past performance does not predict future returns. Trading gold with leverage carries a high risk of loss.
Frequently asked questions
Has gold beaten the S&P 500?
Over our window, yes. From 30 August 2000 to 29 October 2026, COMEX gold futures rose 1,372% (10.8% a year) while the S&P 500 returned 732% with dividends reinvested (8.4% a year) and 414% on price alone. The result depends heavily on the start date: 2000 was close to a stock-market peak and a gold low, and gold won only 13 of the 25 full calendar years.
Is gold more volatile than stocks?
About the same, in this period. Annualised volatility of weekly returns was 17.3% for gold and 17.4% for the S&P 500 total return index. Gold’s worst fall from a peak was smaller (−44.4% from August 2011 to December 2015) than the S&P 500’s (−55.3% with dividends, October 2007 to March 2009), but gold took almost nine years to recover, against about three for stocks.
Does gold go up when the S&P 500 falls?
Often, but not reliably. In the 99 weeks when the S&P 500 lost more than 3%, gold rose in 54 (55%), with a median gain of 0.51%. Gold rose 24.5% during the 2008 crash, but fell 2.5% in the 2020 Covid crash and 6.8% in the 2022 bear market.
What is the correlation between gold and the S&P 500?
Close to zero over the long run: 0.06 for weekly returns since 2000. It moves around a lot, from −0.52 to +0.61 on a rolling one-year basis. In 2020–2021 it was 0.42, and at the end of October 2026 the one-year figure was 0.45, unusually high by historical standards.
Is it fair to compare gold futures with the S&P 500?
With caveats. Gold pays no income, so we compare it with the S&P 500 both with dividends (the fair long-run comparison) and without. Futures prices include small roll effects and ignore the interest a futures holder earns on cash collateral; neither series includes costs or taxes.
Should I hold gold or stocks?
We cannot advise on your portfolio. The data show two assets with similar volatility, very different drivers and a correlation that is usually low, which is why many investors hold both. For traders, the more useful finding is that the link changes by regime, so a stock-market sell-off is not an automatic signal for gold.