Is gold a good hedge against a stock market crash?

In the four big selloffs since GLD launched, gold fell less than stocks every time, rising in 2008 (+24%) and late 2018 (+5%) and slipping in 2020 (-4%) and 2022 (-7%). It has been a shock absorber, not a guaranteed winner, and four crashes is a small sample.

Is gold a good hedge against inflation?

Over very long periods gold has broadly kept pace with inflation, but not reliably over short ones. In 2022, with US inflation near a 40-year high, gold's price still fell about 7% over the months stocks fell, because interest rates rose fast at the same time. Its link to inflation runs over decades, not years.

How much of a portfolio should be in gold?

There is no rule. Many diversified portfolios hold none, some lazy portfolios hold 5% to 10%, and a few designed around crash protection hold 20% to 25%. The right amount depends on what gold is meant to do in your mix and how long you can tolerate it lagging. Backtesting your own mix with and without gold is the best way to see the trade-off.

Should I buy a gold ETF or physical gold?

A physically backed gold ETF holds bullion in a vault and trades like a stock, with an annual fee and no storage or insurance to arrange. Coins and bars avoid the fund fee but come with dealer markups, storage and insurance costs, and are slower to sell. Mining stocks are a different thing again: they are companies, and they have tended to swing much more than gold itself.

Why does gold often rise when stocks fall?

Gold's price is not tied to company profits, so a recession does not hit it directly. In a panic, investors look for assets outside the financial system, and central banks often cut interest rates, which makes a holding that pays nothing more attractive. When a selloff is driven by rising rates instead, as in 2022, that support weakens.

Should You Own Gold? What It Did in Every Crash Since 2008

7 min read · Updated 2026-10-23

Add us as a preferred source on Google

Gold has a reputation as the thing that holds up when everything else falls apart. It is a big part of why gold ETFs are among the largest funds in the world, why central banks keep buying it, and why so many investors ask whether they should own some.

The reputation is partly earned and partly myth. This guide looks at what gold actually did, by price, in each stock market crash since 2008, how it compared with long-term government bonds, the deep falls it has had on its own, and how investors who hold it decide how much. Educational, not a recommendation.

Stress-test a portfolio with gold →

What gold is, and what it is not

A share of a company earns profits, and a bond pays interest. Gold does neither. An ounce of gold today is still an ounce of gold in twenty years, and its price depends entirely on what the next buyer will pay. That makes it a different kind of holding: not a growth engine, but something whose price tends to move for different reasons than stocks.

Those reasons include fear, the level of interest rates after inflation (gold pays nothing, so it looks more attractive when safe savings pay little), the US dollar, and buying by central banks. Because none of them are company profits, gold often zigs when stocks zag. Often is not always, and the waits can be long: after gold peaked in 1980, it took about 28 years for its price to get back to that level.

What gold did in every crash since 2008

The gold ETF most people know, GLD, has traded since late 2004, which covers four big stock market selloffs. Over the months the S&P 500 fell from its peak to its bottom in each one, here is how gold's price moved.

In the 2008 financial crisis the S&P 500 fell about 56% while gold's price changed by about +24%. In the late-2018 selloff, stocks fell about 20% and gold moved +5%. In the 2020 COVID crash, gold slipped about 4% while stocks fell about 34%, and in the 2022 bear market it fell about 7% against about 25% for stocks.

So gold rose in two of the four crashes and dipped in the other two, but it fell less than stocks every time. That is the honest version of the reputation: a reliable shock absorber in this sample, not a reliable winner.

Price change during each stock market crash: S&P 500 vs gold (percent)

2008 financial crisis: S&P 500

−56%

2008 financial crisis: gold

24%

2018 fourth-quarter selloff: S&P 500

−20%

2018 fourth-quarter selloff: gold

2020 COVID crash: S&P 500

−34%

2020 COVID crash: gold

−4%

2022 bear market: S&P 500

−25%

2022 bear market: gold

−7%

Price change of SPY and GLD over each crash's S&P 500 peak-to-bottom window, from the crash study at /studies/etf-crashes (snapshot of 2026-09-16). Prices only. Four crashes is a small sample, and past crashes do not predict the next one.

Gold or long bonds as the cushion?

The other classic crash cushion is long-term government bonds, and the two have protected against different kinds of crash. In panics where interest rates fall, long bonds have done the heavy lifting; in an inflation-driven selloff where rates rise, they have been part of the problem.

In the 2020 crash, long-term Treasuries (TLT) moved +14% while gold slipped about 4%. In 2022 it was the other way round: gold fell about 7% while TLT fell about 30% over the same months, further than the stock market. That is the case for owning a little of each in some portfolios: they have tended to fail at different times.

Price change during four stock market crashes (approximate)
Crash	S&P 500 (SPY)	Gold (GLD)	Long Treasuries (TLT)	Silver (SLV)
2008 financial crisis	-56%	+24%	+18%	-5%
2018 fourth-quarter selloff	-20%	+5%	+4%	+3%
2020 COVID crash	-34%	-4%	+14%	-29%
2022 bear market	-25%	-7%	-30%	-17%

Price change over each crash's S&P 500 peak-to-bottom window, from the crash study at /studies/etf-crashes (snapshot of 2026-09-16). Interest and dividends excluded, so total returns for SPY and TLT were somewhat better.

Gold has its own crashes

A hedge that rises over a crash can still give you a rough ride inside it. During the 2008 window, gold fell about 29% from its spring high to its autumn low before ending the period higher. In 2022 its deepest fall was about 21%. Silver, often sold as "gold's cousin", fell about 57% at its worst in 2008, so it is not a substitute.

Over longer stretches, gold can go nowhere for years. Investors who bought at the 1980 peak waited almost three decades just to get their money back in price terms, with no dividends or interest along the way. A gold holding needs a reason to be in the portfolio that survives a long dry spell.

How much gold, if any

There is no correct amount. Well-known lazy portfolios span the range: many hold none, some include 5% to 10%, and a few built around crash protection go much further, such as Harry Browne's Permanent Portfolio at 25% gold. A few questions help decide where you sit:

•What is it for? A small slice can cushion stock crashes and inflation shocks, but it adds no income and can lag for years. If the goal is growth, gold competes with assets that compound.
•What do you already hold? If your bonds are mostly short-term, the cushion they give in a crash is small, and gold plays a different role than if you already hold long bonds.
•Costs: gold ETFs charge an annual fee, and similar funds differ. Compare a few before choosing; the compare page shows GLD against IAU side by side.
•Taxes: in the US, gains on physically backed gold ETFs are generally taxed at the collectibles rate (up to 28%) rather than the usual long-term rate. Rules differ by country and account, so check yours.
•Rebalancing: gold's habit of zigging when stocks zag only helps if you trim what rose and add to what fell from time to time.

Test it on your own portfolio

The useful question is not whether gold is good, but what a slice of it would have done to your mix. Backtest your portfolio with and without, say, 10% gold and look at the maximum drawdown and the recovery time, not just the final value. Then replay both versions through 2008, 2020 and 2022 in the stress test, since the crashes behaved so differently.

Educational, not advice: whether gold belongs in your portfolio depends on your goals, your horizon and everything else you own.

Try it yourself

Backtest PortfolioRun your mix with and without a gold slice and compare drawdowns.Open tool →Stress Test / ScenariosReplay a portfolio with gold through 2008, 2020 and 2022.Open tool →Compare GLD vs BNDGold against the total bond market, on real data.Open tool →Compare GLD vs IAUTwo gold ETFs side by side, including what each costs.Open tool →

FAQ

Is gold a good hedge against a stock market crash?: In the four big selloffs since GLD launched, gold fell less than stocks every time, rising in 2008 (+24%) and late 2018 (+5%) and slipping in 2020 (-4%) and 2022 (-7%). It has been a shock absorber, not a guaranteed winner, and four crashes is a small sample.
Is gold a good hedge against inflation?: Over very long periods gold has broadly kept pace with inflation, but not reliably over short ones. In 2022, with US inflation near a 40-year high, gold's price still fell about 7% over the months stocks fell, because interest rates rose fast at the same time. Its link to inflation runs over decades, not years.
How much of a portfolio should be in gold?: There is no rule. Many diversified portfolios hold none, some lazy portfolios hold 5% to 10%, and a few designed around crash protection hold 20% to 25%. The right amount depends on what gold is meant to do in your mix and how long you can tolerate it lagging. Backtesting your own mix with and without gold is the best way to see the trade-off.
Should I buy a gold ETF or physical gold?: A physically backed gold ETF holds bullion in a vault and trades like a stock, with an annual fee and no storage or insurance to arrange. Coins and bars avoid the fund fee but come with dealer markups, storage and insurance costs, and are slower to sell. Mining stocks are a different thing again: they are companies, and they have tended to swing much more than gold itself.
Why does gold often rise when stocks fall?: Gold's price is not tied to company profits, so a recession does not hit it directly. In a panic, investors look for assets outside the financial system, and central banks often cut interest rates, which makes a holding that pays nothing more attractive. When a selloff is driven by rising rates instead, as in 2022, that support weakens.

Key terms in this guide

Plain-English definitions in the Learning Hub.

Correlation →Maximum drawdown →Stress test →Diversification score →Volatility (annualized standard deviation) →

Already computed · real data

60/40 vs 60/30/10 with gold on real data

Side by side, same dates

The 60/40 mix (VTI 60%, BND 40%) grew from $10,000 to $41,796 between May 2007 and Oct 2026 (7.6% a year). Its worst fall was -31%, from Oct 2007 to Feb 2009, and it took 39 months (about 3 years) to get back to that high. The same with 10% gold (VTI 60%, BND 30%, GLD 10%) grew from $10,000 to $48,339 between May 2007 and Oct 2026 (8.4% a year). Its worst fall was -30%, from Oct 2007 to Feb 2009, and it took 36 months (about 3 years) to get back to that high.

Real 2007-05 to 2026-10 data, dividends reinvested, rebalanced yearly. Past performance doesn't predict future results.

Growth of $10,000, May 2007 to Oct 2026

The 60/40 mix (VTI 60%, BND 40%)The same with 10% gold (VTI 60%, BND 30%, GLD 10%)

Run 60/40 with your own numbers →Run 60/30/10 with gold with your own numbers →

Keep watching one of them, free: 60/40 or 60/30/10 with gold. A free account saves it as your plan and checks it for you every month.

For educational and informational purposes only, not financial, investment, tax, or legal advice. Backtests are hypothetical; past performance does not guarantee future results.