MEASURED · 28 SEPTEMBER 2026
The gold-silver ratio: 55 years of history, and when silver was the better buy
The gold-silver ratio is how many ounces of silver one ounce of gold buys. The standard advice is to buy silver when the ratio is above 80 and switch back to gold below 50. We tested that against every month since August 1971, when the US stopped exchanging dollars for gold, 661 months in all, and against every time the ratio has actually crossed 80.
The short version
Silver beat gold over 3 years, ratio 80+
90%
Against 44% for all months; random paths match it 1 time in 6
Times the ratio has crossed 80 since 1971
5
Each paid off in the end, after 3 to 10 years
Worst move against a silver buyer first
-28%
Feb 2018 to Apr 2020: the ratio went from 80 to 112
The 80/50 swap rule, from 1 oz of gold in 1971
1.99 oz
On 3 trades in 55 years; one of them did most of the work
What is the gold-silver ratio today?
On 29 September 2026 the London benchmark prices were $4,163 for gold and $60.87 for silver, a ratio of 68.4. The August 2026 average was 67.4, higher than in 58% of months since 1971. That is the middle of the range: by the classic rule, no signal either way.
It got there by one of the fastest swings on record. The ratio stood at 105.3 on 22 April 2025 and fell to 45.4 by 27 January 2026, as silver ran to its highest London benchmark price on record, $118.45 on 29 January 2026. Silver has since fallen back to about $61, and the ratio has risen with it.
What is silver’s all-time high?
Silver’s all-time high is $118.45 an ounce, the London benchmark price on 29 January 2026. Before that, the record was $49.45 from 18 January 1980, and it stood for 45.7 years until silver finally closed above it on 9 October 2025. Silver set a new record on 21 days in 2025 and 14 days in 2026. On 29 September 2026 it was $60.87, 48.6% below the record.
What is the historical gold-silver ratio?
From August 1971 to August 2026, the monthly ratio averaged 61.3, with a median of 63.3. Eight months in ten fell between 34 and 85. The lowest monthly average was 17.4 in January 1980, and the highest 111.5 in April 2020.
Single days went further. The lowest daily ratio since 1971 was 14.0 on 2 January 1980, two weeks before silver’s 1980 London benchmark high of $49.45 on 18 January 1980. Silver then collapsed as the Hunt brothers’ attempt to corner the market failed, ending on Silver Thursday, 27 March 1980. The highest was 123.5 on 17 March 2020, when silver dropped to $12.44 in the March 2020 market crash. The ratio has stood at 100 or more on 68 days, all in 1991, 2020 and 2025.
The ratio has drifted upward over the decades. It averaged 34 in the 1970s and 81 in the 2020s so far:
| DECADE | AVERAGE | LOWEST MONTH | HIGHEST MONTH |
|---|---|---|---|
| 1970s (Aug 1971 to Dec 1979) | 33.5 | 20.2 | 46.2 |
| 1980s | 52.7 | 17.4 | 75.7 |
| 1990s | 73.6 | 43.7 | 98.4 |
| 2000s | 61.5 | 47.4 | 79.2 |
| 2010s | 67.5 | 34.7 | 90.6 |
| 2020s (Jan 2020 to Aug 2026) | 81.1 | 51.6 | 111.5 |
Is there a “normal” gold-silver ratio?
Not one the market has kept to. Before 1971 the US set the ratio by law: the Coinage Act of 1792 made fifteen pounds of pure silver “of equal value in all payments, with one pound weight of pure gold”, and an 1834 act cut the gold in US coins, which moved the ratio to about 16 to 1. You will also see 8 to 1 quoted, from mining: the US Geological Survey estimates 2025 world mine output at 26,000 tons of silver and 3,300 tons of gold. Since 1971 the ratio has never been below 14, and it has averaged 61, about four times the old legal ratio. Neither number is a target the price returns to.
Gold-silver ratio by year since 1972
| YEAR | GOLD | SILVER | RATIO | LOW–HIGH |
|---|---|---|---|---|
| 1972 | $58 | $1.68 | 34.5 | 31–39 |
| 1973 | $97 | $2.55 | 38.2 | 33–46 |
| 1974 | $159 | $4.69 | 34.2 | 28–42 |
| 1975 | $161 | $4.41 | 36.6 | 32–42 |
| 1976 | $125 | $4.38 | 28.6 | 25–32 |
| 1977 | $148 | $4.63 | 31.9 | 30–34 |
| 1978 | $193 | $5.42 | 35.6 | 34–39 |
| 1979 | $307 | $11.11 | 29.7 | 20–37 |
| 1980 | $608 | $20.80 | 32.0 | 17–41 |
| 1981 | $460 | $10.53 | 44.4 | 38–49 |
| 1982 | $376 | $7.94 | 47.9 | 42–56 |
| 1983 | $423 | $11.42 | 37.4 | 34–43 |
| 1984 | $360 | $8.15 | 44.5 | 41–49 |
| 1985 | $318 | $6.13 | 51.9 | 49–55 |
| 1986 | $368 | $5.44 | 67.7 | 58–76 |
| 1987 | $446 | $7.01 | 64.5 | 54–74 |
| 1988 | $437 | $6.52 | 67.1 | 63–71 |
| 1989 | $381 | $5.52 | 69.3 | 65–73 |
| 1990 | $384 | $4.84 | 79.6 | 72–92 |
| 1991 | $362 | $4.04 | 89.8 | 83–98 |
| 1992 | $344 | $3.96 | 86.9 | 82–91 |
| 1993 | $360 | $4.31 | 83.9 | 77–92 |
| 1994 | $384 | $5.28 | 72.8 | 70–79 |
| 1995 | $384 | $5.21 | 74.0 | 69–83 |
| 1996 | $388 | $5.20 | 74.7 | 71–79 |
| 1997 | $331 | $4.91 | 67.9 | 50–74 |
| 1998 | $294 | $5.56 | 53.4 | 44–60 |
| 1999 | $279 | $5.22 | 53.4 | 49–58 |
| 2000 | $279 | $4.96 | 56.3 | 55–59 |
| 2001 | $271 | $4.39 | 61.8 | 56–67 |
| 2002 | $310 | $4.59 | 67.6 | 63–72 |
| 2003 | $364 | $4.88 | 74.6 | 72–79 |
| 2004 | $409 | $6.66 | 61.7 | 57–66 |
| 2005 | $445 | $7.30 | 61.0 | 59–64 |
| 2006 | $604 | $11.56 | 52.7 | 47–60 |
| 2007 | $697 | $13.38 | 52.1 | 48–56 |
| 2008 | $872 | $15.00 | 60.4 | 50–79 |
| 2009 | $973 | $14.65 | 67.0 | 61–76 |
| 2010 | $1,225 | $20.15 | 62.0 | 48–69 |
| 2011 | $1,569 | $35.23 | 45.0 | 35–54 |
| 2012 | $1,670 | $31.14 | 53.8 | 51–58 |
| 2013 | $1,412 | $23.86 | 59.7 | 54–65 |
| 2014 | $1,266 | $19.07 | 66.7 | 63–74 |
| 2015 | $1,161 | $15.72 | 73.9 | 71–76 |
| 2016 | $1,249 | $17.16 | 73.2 | 67–80 |
| 2017 | $1,258 | $17.06 | 73.8 | 69–78 |
| 2018 | $1,269 | $15.72 | 80.9 | 78–85 |
| 2019 | $1,393 | $16.22 | 85.9 | 83–91 |
| 2020 | $1,770 | $20.54 | 88.8 | 73–112 |
| 2021 | $1,800 | $25.17 | 71.8 | 66–80 |
| 2022 | $1,801 | $21.78 | 83.1 | 77–91 |
| 2023 | $1,943 | $23.41 | 83.1 | 80–87 |
| 2024 | $2,388 | $28.27 | 84.7 | 79–89 |
| 2025 | $3,442 | $39.80 | 88.1 | 69–101 |
| 2026 (8 months) | $4,581 | $74.59 | 62.0 | 52–69 |
Average prices in US dollars per troy ounce. The ratio is the average of the year’s monthly ratios; low and high are the lowest and highest monthly readings.
Is a high gold-silver ratio a signal to buy silver?
On the record since 1971, high readings were followed by silver beating gold more often than not, but the record is too short to call that proof. For every month we measured how silver did against gold over the following one, three and five years. The higher the ratio, the more often silver went on to beat gold:
| RATIO THAT MONTH | MONTHS | SILVER BEAT GOLD, 1 YR | 3 YRS | 5 YRS | MEDIAN, 3 YRS |
|---|---|---|---|---|---|
| All months | 625 | 39% | 44% | 38% | -5% |
| Below 50 | 184 | 29% | 30% | 22% | -15% |
| 50 to 80 | 359 | 38% | 40% | 37% | -8% |
| 80 and above | 82 | 58% | 90% | 84% | +16% |
When the ratio was 80 or above, silver beat gold over the next three years in 90% of months, against 44% for all months, and by a median 16%. Over one year the edge was much smaller: 58% of months. Over five years it was 84%.
That is weaker evidence than it looks. Even a ratio that moves completely at random seems to “come back down” after its highest readings when you measure it inside one fixed record. To see how much of the result that explains, we built 1,000 imaginary histories from the real ratio’s own month-to-month moves, reshuffled in year-long blocks so that none of them has any pull back toward a level, and ran the same test. The typical random history showed silver beating gold after its top readings 73% of the time, and about one in 6 did at least as well as the real record. The record fits the rule, but it does not prove it.
The other half of the rule is weaker still. Below 50, silver trailed gold over three years 70% of the time, against 57% for all months.
The record is also short: 82 months are not 82 separate chances. High readings come in clusters, and since 1971 the ratio has crossed 80 only 5 times. That is where the rule has to be judged.
What happened each time the ratio crossed 80?
Every one of the 5 episodes paid off for silver in the end. None paid off quickly. If you bought silver the first month the ratio averaged 80 or more:
| FIRST MONTH AT 80+ | THEN ROSE TO | SILVER VS GOLD BY THEN | AFTER 3 YRS | BACK TO 63 | SILVER VS GOLD BY THEN |
|---|---|---|---|---|---|
| Sep 1990 (81) | 98 in Feb 1991 | -18% | -4% | Nov 1997, 7 yrs | +35% |
| Feb 1995 (80) | 83 in Mar 1995 | -3% | +84% | Nov 1997, 3 yrs | +34% |
| Mar 2016 (80) | No higher | – | -6% | Jan 2026, 10 yrs | +56% |
| Feb 2018 (80) | 112 in Apr 2020 | -28% | +21% | Jan 2026, 8 yrs | +55% |
| May 2022 (84) | 101 in May 2025 | -16% | -16% | Jan 2026, 4 yrs | +64% |
In 4 of the 5 episodes the ratio kept climbing after it first crossed 80, so silver lost further ground against gold before it turned: up to 28% between February 2018 and April 2020. Three years after the first signal, silver was ahead in only 2 of 5. The payoff came when the ratio finally fell back to its long-run middle of about 63, which took 3 to 10 years. The last three episodes all resolved in the same month, January 2026, during silver’s spike to its record.
So the rule has paid off every time so far, but slowly, and five episodes are too few to tell a real tendency from luck. At best it tells you silver is cheap against gold, not when that will change. Anyone acting on it needed to hold for years, through a further loss against gold.
Does the 80/50 swap rule work?
We started with one ounce of gold in August 1971 and followed the rule: swap all of it into silver when the ratio reaches 80, back into gold when it falls to 50. Each decision uses one month’s average price and trades at the next month’s, so the rule never uses a price it could not have known.
| STRATEGY | TRADES | GOLD OUNCES AT THE END | WORTH IN AUGUST 2026 |
|---|---|---|---|
| Hold gold | 0 | 1.00 | $4,411 |
| Hold silver | 0 | 0.40 | $1,756 |
| 80/50 rule, no costs | 3 | 1.99 | $8,756 |
| 80/50 rule, 2% cost per swap | 3 | 1.87 | $8,240 |
| 80/50 rule, 5% cost per swap | 3 | 1.70 | $7,508 |
From $43 in 1971, holding gold grew to $4,411 and the rule roughly doubled that. But look at the trades. It bought silver in October 1990 at a ratio of 87, switched back to gold in January 1998 at 49, and bought silver again in April 2016 at 76, which it still holds. The first round trip, held for over seven years, gained 77% in ounces; the open position is up 12%. Three trades in 55 years is an anecdote, not a track record.
Two further checks look reassuring at first. The exact thresholds hardly matter: we re-ran the rule for 57 pairs of levels, from 60/30 to 100/60, and 55 of 57 ended with more gold than holding gold did (median 2.34 oz). The best pair, 75/55, reached 5.9 oz, but picking it after the fact is exactly the hindsight to avoid. And a version with no fixed levels also did well. It swaps into silver when the ratio is above the top fifth of its own previous ten years, and back into gold when it drops into the bottom fifth. From August 1981 it made 8 swaps and ended with 2.12 oz of gold for every ounce held (holding silver: 0.68).
Then we ran the same rules on the 1,000 random histories. The 80/50 rule did at least as well as it really did in 12% of them; 55 or more of the 57 threshold pairs beat gold in 14%; and the ten-year version matched its real result in 22%. A rule that swaps at the edges of a range wins whenever the price happens to wander back across that range, and random paths do that often enough. The real record beats most random ones, but not by enough to rule out luck.
Both rules depend on the same thing: since 1986 the ratio has stayed between 35 and 112 on monthly averages, and kept coming back toward the middle. Nothing guarantees that it will keep doing so. The two rules also disagree right now. The 80/50 rule has held silver since April 2016; the ten-year version switched to gold in January 2026, when the ratio fell to 52.
Is silver or gold the better investment?
Over most of the last 55 years, gold. Silver has been the better buy mainly when bought at a high ratio.
| FROM | RATIO THEN | GOLD A YEAR | SILVER A YEAR | GOLD, AFTER INFLATION | SILVER, AFTER INFLATION |
|---|---|---|---|---|---|
| August 1971 | 27 | 8.8% | 7.0% | +4.7% | +3.0% |
| January 1980 | 17 | 4.1% | 1.1% | +0.9% | -2.0% |
| January 2000 | 55 | 10.9% | 10.0% | +8.1% | +7.2% |
| April 2011 | 35 | 7.4% | 2.8% | +4.6% | +0.2% |
Silver beat gold in 39% of one-year, 38% of five-year and 34% of ten-year holding periods since 1971. It was also the rougher ride. On monthly averages, silver’s price swung about 1.7 times as much as gold’s (28% a year against 17%). After inflation, it lost 95% of its value between January 1980 and November 2001. Gold’s worst fall was 83%.
Silver’s January 1980 average of $38.90 is worth $167 in August 2026 dollars. It passed that in dollars in April 2011, and its highest monthly average was $92.10 in January 2026. After inflation it has never come back: at $65.40 in August 2026, silver is still 61% below its 1980 level. Gold passed its 1980 level after inflation in 2025.
What does the ratio do in a recession?
It usually rises. In 6 of the 7 US recessions since 1973, the ratio was higher at the end than at the start, and silver did worse than gold in 6.
| RECESSION | GOLD | SILVER | RATIO |
|---|---|---|---|
| Nov 1973 – Mar 1975 | +87% | +48% | 33 → 41 |
| Jan 1980 – Jul 1980 | -5% | -58% | 17 → 39 |
| Jul 1981 – Nov 1982 | +2% | +14% | 48 → 42 |
| Jul 1990 – Mar 1991 | 0% | -20% | 74 → 93 |
| Mar 2001 – Nov 2001 | +5% | -7% | 60 → 67 |
| Dec 2007 – Jun 2009 | +18% | +3% | 56 → 64 |
| Feb 2020 – Apr 2020 | +5% | -16% | 89 → 112 |
Recession dates are the NBER’s business-cycle peak and trough months; prices are those months’ averages. Seven recessions are a small sample.
The usual explanation is that silver is also an industrial metal: industry used 657.4 million of the 1.13 billion ounces of silver demand in 2025, about 58%, according to the Silver Institute. We did not test that explanation.
What this means if you trade or buy silver
- Treat a high ratio as a reason to hold for years, not a trade. Three years after first crossing 80, silver was ahead in only 2 of 5 episodes. The payoff took 3 to 10 years.
- Size for the move against you. After the first signal, silver lost up to 28% more against gold before it turned. A position you cannot hold through that is too big.
- Expect silver to lag in a downturn. The ratio rose in 6 of 7 recessions, so a high reading can go higher when the economy weakens.
- Count the cost of each swap. At 5% a swap, the 80/50 rule’s 1.99 oz became 1.70. For US holders of coins or bars, each swap is also a sale: the IRS taxes gains on collectibles at up to 28%, and since 1 January 2018 collectibles no longer qualify for tax-free like-kind exchanges. On a broker, check the overnight financing on a position held for years.
- Today the ratio says nothing. At about 68 it sits in the middle of its range, and the two rules we tested disagree about which metal to hold.
The ratio is a long-horizon measure. How gold behaves day to day is covered in our gold trading statistics, and how it has held up against rising prices in is gold a good hedge against inflation? Gold against the stock market is in gold vs the S&P 500.
How we measured it
Prices: the World Bank’s monthly commodity price data (the “Pink Sheet”), which averages the London benchmark prices for gold and silver over each month. The ratio is the month’s average gold price divided by its average silver price. Inflation: the US Bureau of Labor Statistics consumer price index (CPI-U), used only for the after-inflation figures.
The World Bank publishes silver to the nearest 10 cents and gold to the nearest dollar, which matters when silver cost $1.50. So we re-ran the whole study on unrounded daily LBMA prices, the London benchmarks themselves. The monthly ratio moved by 0.52% on average, and the conclusions did not change. The 80/50 rule ended with 1.99 oz on one series and 1.99 on the other, and 55 against 53 threshold pairs beat gold. We also checked both metals against our own broker’s prices over 151 months from February 2014. Silver differed by 0.28% on average and gold by 0.12%.
“Silver beat gold” means silver’s price rose more, or fell less, than gold’s over the period, which is the same as the ratio falling. Every month from August 1971 is a starting point, so periods overlap: read each share as the odds for someone who acted in a random month. To judge luck, every result is compared with 1,000 imaginary ratios built from the real one’s own monthly changes, reshuffled in 12-month blocks: they move like the real ratio from month to month but have no pull toward any level. An earlier version of this page used a weaker test that made the above-80 result look much less likely to be luck; the figures here are the corrected ones. The results ignore dealer premiums, storage and tax unless a cost is stated, and monthly averages hide the daily extremes, which are quoted separately. The method behind all our studies is published alongside them.
Sources: World Bank commodity prices; LBMA precious metal prices; NBER business cycle dates; Coinage Act of 1792 (FRASER); USGS silver and gold summaries, 2026; Silver Institute, World Silver Survey 2026; IRS Topic 409 and like-kind exchanges.
Common questions
What is a good gold-to-silver ratio to buy silver?
80 or higher is the usual answer, and the record since 1971 fits it, though not strongly enough to rule out luck. In months when the ratio averaged 80 or more, silver beat gold over the next three years 90% of the time, against 44% for all months, but random price histories produce a record that good about one time in 6. It also worked over years, not weeks: after first crossing 80 the ratio kept rising in 4 of 5 episodes, by up to 28% against a silver buyer, and took 3 to 10 years to fall back to its long-run middle of about 63.
What is the gold-silver ratio today?
68.4 on 29 September 2026, from the London benchmark prices (gold $4,163, silver $60.87). The August 2026 monthly average was 67.4, higher than in 58% of months since 1971: the middle of its range.
What is the historical average gold-silver ratio?
From August 1971 to August 2026 the monthly ratio averaged 61.3 (median 63.3). Eight months in ten fell between 34 and 85. It was below 50 in 28% of months and 80 or above in 16%.
What was the highest gold-silver ratio?
Since 1971, 123.5 on 17 March 2020, when silver fell to $12.44 with gold at $1,536. The highest monthly average was 111.5 in April 2020. The lowest was 14.0 on 2 January 1980, during silver's 1980 spike.
Is there a normal gold-silver ratio?
Not one the market keeps to. US law set it at 15 to 1 in 1792 and about 16 to 1 from 1834, and mining produces about 8 ounces of silver per ounce of gold (USGS, 2025). Since 1971 the market ratio has averaged 61.3 and never fallen below 14.
Is silver a better investment than gold?
Over most periods since 1971, no. Gold returned 4.7% a year after inflation to August 2026, silver 3.0%, and silver beat gold in only 34% of ten-year periods. Silver is still 61% below its January 1980 average after inflation. It has been the better buy mainly when the ratio was high.
What happens to the gold-silver ratio in a recession?
It usually rises. In 6 of the 7 US recessions since 1973 the ratio was higher at the end than at the start, and silver did worse than gold in 6 of them. Its highest monthly average, 111.5, came in April 2020. Seven recessions are a small sample.