MEASURED · 29 SEPTEMBER 2026

Gold vs the S&P 500: 55 years of returns, with dividends counted

Which has been the better investment, gold or US stocks? The answer turns on two things most comparisons leave out: whether the S&P 500’s dividends are counted, and which year you start from. We compared them every month from August 1971, when the US stopped exchanging dollars for gold, to August 2026: 661 months, with dividends reinvested and inflation measured.

The short version

Has gold outperformed the S&P 500?

Since August 1971, no, once dividends are counted. $10,000 in gold grew to $1.03 million, an average of 8.8% a year. The S&P 500 with its dividends reinvested grew to $3.48 million, 11.2% a year. Without the dividends the S&P 500 returned 8.3% a year and ended at $793,020, less than gold. That is why price charts of gold against the S&P 500 often show gold ahead: they leave out about 3 percentage points a year that shareholders actually received.

The start date matters as much. From gold’s January 1980 peak, gold returned 4.1% a year and stocks 12.2%. From January 2000, near the top of the dot-com boom, gold won: 10.9% a year against 8.5%. From gold’s 2011 high (the September average), stocks returned 15.5% a year and gold 6.3%.

Average yearly return to August 2026, and what $10,000 became. Monthly average prices; dividends reinvested each month; no fees or taxes.

Starting in 1971 flatters gold in one way. Its official price had been fixed at $35 an ounce since 1934, and its re-pricing in the 1970s, 32.7% a year to December 1979, is the kind of move that happens once. Starting in 1980, at the top of that move, is just as unfair the other way. The fairest reading is the range: stocks won from most starting points, by a lot, and gold won from a few, also by a lot.

Gold vs the S&P 500 over the last 5, 10, 20 and 30 years

Gold won the last five years, 19.8% a year against 13.2%, including a rise of 63% from December 2024 to December 2025 (monthly averages). Over 10, 20 and 30 years the S&P 500 with dividends won. Over 20 years the margin was smallest, 11.4% a year against 10.2%, and without its dividends the index would have lost to gold.

Average yearly return to August 2026, with what $10,000 became underneath.

How often has gold beaten stocks?

A single start date can prove almost anything, so we used all of them. Starting in every month since August 1971, gold beat the S&P 500 with dividends over the next year 44% of the time, over five years 38%, over ten years 27% and over twenty years 19%. The longer the hold, the less often gold came out ahead. Gold pays no dividend, so it has to rise about 3 points a year faster than the index’s price just to draw level.

“Typical gap” is the median difference in average yearly return. Periods overlap, so read each share as the odds for someone who invested in a random month.

In calendar years the count is closer: gold came out ahead in 24 of the 54 years from 1972 to 2025. By decade, gold won three: the 1970s, the 2000s and the 2020s so far. Stocks won the 1980s, the 1990s and the 2010s, by 11 to 22 points a year.

Does gold go up when stocks fall?

In long bear markets, mostly yes. The S&P 500 fell 15% or more on monthly averages seven times since 1971. Gold rose during five of them, and by a lot in the slow ones: +183% in 1973-74, +31% in 2000-03 and +22% in 2007-09. It fell in two: -46% in 1980-82, a decline that began ten months after gold’s own January 1980 peak, and -7% in 2022.

Declines are measured on the S&P 500’s monthly average price, so they are shallower than the lows in daily closes: the 2020 crash, about a third in daily closes, shows here as -19%. “Back at its peak” is the month the index’s price regained its earlier high.

Month to month, gold was no hedge at all. The correlation between gold’s monthly return and the S&P 500’s since 1971 is 0.00: none. In the 66 worst months for stocks, one month in ten, gold rose in 38 (58%), against 50% of all months, a gap small enough to be chance. It gained 1.2% on average in those months, not far above its typical month of 0.7%. In the ten worst months it rose in five and fell in five:

Change in the monthly average from the month before.

The S&P 500 to gold ratio

Divide the S&P 500 by the price of an ounce of gold and you get how many ounces the index would buy. Since 1971 it has averaged 1.61 (median 1.37), from 0.16 in January 1980 to 5.42 in August 2000. In August 2026 it was 1.75, higher than in 61% of months.

On the record, high readings were followed by gold beating stocks over the next ten years, and low readings by the reverse. Splitting the months into fifths by the ratio:

Ten-year results with dividends, for months from August 1971 to August 2016. Band edges are fifths of all months since 1971.

That looks like a timing tool, and it is weaker than it looks. The top band’s 111 months fall almost entirely in one stretch, 1997 to 2007, and the bottom band’s 132 all fall between 1974 and 1988. So the table rests on about two episodes, not hundreds of tries. And a ratio compared with its own later change can look predictive inside one record even when it moves at random. To see how much of the result that explains, we built 2,000 imaginary histories from the real monthly moves of gold and the S&P 500, reshuffled in year-long blocks so that none of them has any pull toward a level, and ran the same test. The link between the ratio and the next ten years came out at least as strong as the real one in 12% of them, and a top fifth at least as one-sided in 18%, about one in 6. The record fits the idea, but it does not prove it.

Today’s 1.75 sits in the band where gold beat stocks over the next ten years 42% of the time: close to a coin toss, and the table is not proof either way.

Should you hold both? Stock and gold mixes since 1971

Holding some gold changed the ride more than the result. Rebalanced every January, 80% in the S&P 500 and 20% in gold returned 11.6% a year from August 1971, a little more than stocks alone (11.2%), and its worst fall after inflation was 39% instead of 52%. From January 1980 the same mix returned 11.1% against 12.2% for stocks alone: the smaller losses cost about 1.2 points a year.

What held from both start dates was the protection. The worst ten years for stocks alone left 54 cents of every dollar after inflation. With 30% in gold, the worst ten years left 92 cents.

S&P 500 with dividends. Worst falls and worst periods are from August 1971; each happened after 1980, so they are the same from either start. No fees, storage or tax.

The worst losses, and how long they lasted

After inflation, gold’s worst loss was 83%, from January 1980 to April 2001, and it did not get back to its 1980 level until February 2025: 45 years. The S&P 500 with dividends lost 52% after inflation from August 2000 to March 2009 and was back by May 2013, 13 years after the peak. Gold against inflation alone is covered in is gold a good hedge against inflation? Every fall of 20% or more in gold since 1971, and how long each took to recover, is in gold price history.

Gold vs the S&P 500, year by year

The S&P 500 with dividends fell in 12 of the 54 years from 1972 to 2025. Gold rose in 8 of those years, and both fell in 4.

December average to December average, so these differ from the calendar-year returns quoted from closing prices. The better of the two each year is brighter; losses are in red.

What this means for your money

• Compare total returns, not price charts. A price chart understates the S&P 500 by about 3 points a year since 1971, which over 55 years is the difference between $793,020 and $3.48 million.
• For long-term growth, stocks won most of the time. Gold beat them in 27% of ten-year periods and 19% of twenty-year periods.
• Gold earned its place in slow bear markets, not crashes. It rose in 5 of 7 big declines, but fell in the two worst months, and its monthly returns had no link to the stock market’s.
• A gold slice mainly buys a smaller worst case. 20% in gold cut the worst fall after inflation from 52% to 39%. Whether it also cost return depended on the start date.
• Check the tax before you choose. For US holders of gold coins or bars, a long-term gain counts as a collectible and is taxed at up to 28%, against 0%, 15% or 20% on most long-term gains from stock funds (IRS Topic 409).
• Trading gold is a different question. These are buy-and-hold results over decades. How gold moves day to day is in our gold trading statistics, and how it compares with silver in the gold-silver ratio, tested. What moves it month to month is measured in why is gold going up? How long a gold buyer has had to wait to break even, and whether buying after a big year helped, is in is gold a good investment?

How we measured it

Gold: the World Bank’s monthly commodity price data (the “Pink Sheet”), which averages the London benchmark price over each month. It is rounded to the nearest dollar; on unrounded LBMA prices the since-1971 result is 8.80% a year instead of 8.78%. S&P 500: Robert Shiller’s monthly S&P Composite data, where the price is the month’s average of daily closes. We reinvest one twelfth of the index’s annual dividend each month. His dividend figures run to June 2026; we carried that figure forward for July 2026 and August 2026. Inflation: the US Bureau of Labor Statistics consumer price index (CPI-U).

As a check, the S&P 500 averages matched our broker’s US500 prices within 0.15% on average over the 11 months from October 2025, and Shiller’s inflation figures match the BLS index. Everything uses monthly averages, which smooth out the daily extremes, and ignores fund fees, dealer premiums, storage and tax. Holding periods start in every month, so they overlap. The random-history test rebuilds gold, the S&P 500’s price and its total return together from their own monthly changes, resampled in 12-month blocks, and repeats every step of the ratio test 2,000 times. The method behind all our studies is published alongside them.

Sources: Robert Shiller, stock market data; World Bank commodity prices; LBMA precious metal prices; BLS consumer price index; IRS Topic 409.

Common questions