Research · Dataset

Gold in today's money, 1960–2026

Every other page on this desk quotes gold in nominal dollars. This one deflates it by CPI and asks the question the nominal price hides: what is gold actually worth, and does it behave like an inflation hedge once you strip the inflation out?

Last updated: 2026-10 (data through 2026-10) · Gold data updated November 03, 2026

Source and method

Gold price
World Bank Commodity Price Data (The Pink Sheet), Monthly Prices sheet, column 70 (Gold, $/troy oz), monthly average
CPI
FRED series CPIAUCSL (US CPI-U, All Items, index 1982-1984=100, seasonally adjusted), monthly
Coverage
801 months, 1960-01 to 2026-10
Real index
1980-01 = 100 — real(t) = 100 × (gold(t) / CPI(t)) ÷ (gold(1980-01) / CPI(1980-01))
Download
gold-real-monthly.csv (month, gold, CPI, real index, nominal and real returns, CPI inflation YoY)
Sources
World Bank commodity markets · FRED CPIAUCSL
Reproduction
scripts/gold_real_price_analysis.py — recomputes every table below from the workbook and the FRED CSV. Independently re-derived from the published CSV by scripts/verify_gold_real.py.

Gold is a monthly AVERAGE price and CPI is a monthly index, so the real series is a monthly approximation, not a daily inflation-adjusted price. No storage, financing or roll cost is included. The CPI base is 1982-1984=100; the real index base here is 1980-01=100.

Where real gold is now

Real gold index 140.6 1980-01 = 100 · 2026-10
vs 1980 in real terms +40.6% nominal +508.1%
Real peak 177.1 2026-02 · now -20.7% below
Nominal price $4,105 CPI 337.5

In nominal dollars gold is +508.1% above its January 1980 price of $675/oz. After inflation, the same comparison reads +40.6%. Both numbers are true; they answer different questions, and the second one is the one that matters if you care about purchasing power.

The all-time real high for this series is 177.1, set in 2026-02 — the same month as the nominal high, because in the last stretch gold outran CPI. From that peak, real gold is currently -20.7% below where it was, after 0.7 years. Nominal gold at the 1980 peak was $5,020; the real series and the nominal series now agree on the peak month, but they tell very different stories about the 45.1 years in between.

That recovery time is not an isolated curiosity. It is the reason the popular claim that gold cannot lose money over a ten-year hold does not hold up: a buyer who caught the 1980 spike was underwater in real terms for roughly four decades. Our long-horizon holding-period study runs every ten- and twenty-year window from every month of the sample and finds 132 losing ten-year windows.

Over the trailing twelve months (2025-09 to 2026-09, 12 calendar months) nominal gold is +17.7% and CPI is +3.7% (trailing twelve-month inflation 3.7%), so real gold is +13.5%. The window ends one month before the last observation because a true twelve-month comparison needs a CPI print on both ends, and the October 2025 CPI print was never published.

Real drawdowns, not nominal ones

An episode starts at an all-time real high and ends when real gold makes a new one. Episodes shallower than 5% are excluded as noise. Seven qualify.

Real peak Peak (real) Peak ($/oz) Trough Trough (real) Depth Decline (mo) Recovery
1960-01 13.8 35 1967-12 11.9 -13.6% 95 1969-02 (9.1 y)
1969-02 13.9 43 1970-08 10.4 -25.3% 18 1972-05 (3.2 y)
1972-08 18.5 67 1972-11 17.2 -7.1% 3 1973-02 (0.5 y)
1973-06 31.4 120 1973-11 23.9 -23.8% 5 1974-01 (0.6 y)
1974-04 41.3 172 1976-08 22.2 -46.3% 28 1979-05 (5.1 y)
1980-01 100.0 675 2001-04 17.0 -83.0% 255 2025-02 (45.1 y)
2026-02 177.1 5,020 2026-10 140.6 -20.7% 8 not yet recovered

The 1980–2001 episode is the one that matters

Real gold peaked at 100.0 in 1980-01 and fell to 17.0 in 2001-04 — a -83.0% decline that took 255 months to bottom. It then took 541 months (45.1 years) to make a new real high, finally doing so in 2025-02.

In nominal dollars the same period looks much less severe, which is exactly why the nominal chart is the one people quote. This table is computed on the deflated series only. For the nominal drawdown history, see Gold price history.

The current episode (2026-02 → 2026-10, -20.7%) has not recovered. It is the shortest decline in the table and the only one still open.

Nominal versus real, decade by decade

Period Gold start Gold end Nominal Real Nominal CAGR Real CAGR CPI CAGR Real vol
1960-1969 35 35 +0.0% -22.1% +0.00% -2.49% +2.55% 5.4%
1970-1979 35 455 +1,200.0% +540.7% +29.52% +20.60% +7.40% 20.9%
1980-1989 675 409 -39.4% -62.6% -4.93% -9.44% +4.98% 24.0%
1990-1999 410 283 -31.0% -47.9% -3.67% -6.36% +2.87% 10.4%
2000-2009 284 1,135 +299.6% +211.3% +14.99% +12.13% +2.55% 13.6%
2010-2019 1,118 1,479 +32.3% +11.2% +2.86% +1.08% +1.76% 11.6%
2020-2029 (to 2026-10) 1,561 4,105 +163.0% +101.9% +15.61% +11.12% +4.04% 13.2%

The gap between the nominal and real columns is inflation, and in some decades it is the whole story. The 1970s delivered +1,200.0% nominal and +540.7% real. The 1980s and 1990s lost money in both terms — -62.6% and -47.9% in real terms against -39.4% and -31.0% nominal. The 2010s are the quiet case: +32.3% nominal became +11.2% real, because CPI ran +1.76% a year and took most of it.

The final row is a partial decade (81 months to 2026-10) and is labelled as such. Real volatility is the annualised standard deviation of monthly real returns.

Does gold actually hedge inflation?

The claim is easy to test and the answer is not the comfortable one. Correlating every monthly gold return against the trailing twelve-month CPI inflation rate over 798 months gives 0.0330 for real gold and 0.0801 for nominal gold. Both are indistinguishable from zero. Monthly gold returns carry essentially no information about the inflation rate of the same month.

Splitting the sample into months with inflation of 4% or more (255 months) and months below it (543 months) does not change the picture. Real gold averaged 0.348% a month in the high-inflation months (4.18% annualised) and 0.398% a month in the low-inflation months (4.77% annualised) — a difference of -0.050pp a month in the wrong direction for the hedge claim.

Test Result Reading
Naive difference in monthly real returns -0.050pp p = 0.906 — no difference
Non-overlapping subsample (every 12th month) -0.175pp p = 0.896 — still not significant
Moving-block bootstrap (12-month blocks) -1.001 to 1.042pp 95% interval crosses zero

The point estimate flips sign between the naive test and the non-overlapping subsample (-0.050pp versus -0.175pp), which is the signature of noise rather than a small effect. The bootstrap interval is the honest summary, and it crosses zero.

Looking forward twelve months instead of back one month gives the same message with different numbers: from high-inflation months, real gold returned 4.67% over the next year on average (nominal 11.52%); from low-inflation months, 6.40% real (nominal 9.28%). Real gold did better when inflation was low. Those forward windows overlap, so read them as description, not significance.

The honest summary: gold may or may not be a good long-run store of value, and this dataset supports the second claim. It does not support the specific claim that gold protects you in the months when inflation is running hot. If you are hedging inflation over a horizon of years, the deflated return column is the relevant one.

Calendar years

Year Gold Real gold CPI YoY Real index (end)
2024 +30.2% +26.9% +2.87% 96.3
2023 +6.7% +3.9% +3.32% 75.8
2022 -1.0% -6.4% +6.40% 69.5
2021 -4.1% -10.3% +7.17% 73.7
2020 +19.0% +17.7% +1.32% 81.9
2019 +14.5% +11.8% +2.32% 66.1
2018 -6.1% -7.5% +2.00% 57.1
2017 +6.0% +4.2% +2.13% 58.9
2016 +5.4% +3.2% +2.05% 55.1
2015 -14.0% -15.1% +0.64% 52.3
2014 -3.5% -3.9% +0.65% 58.7
2013 -26.9% -27.9% +1.51% 60.2
2012 +1.9% +0.4% +1.76% 84.2
2011 +20.6% +17.4% +3.06% 83.4
2010 +24.4% +22.7% +1.44% 72.9
2009 +32.1% +28.8% +2.81% 60.3

Last 16 calendar years shown; the full series is in the downloadable CSV. Year figures are January-to-December changes in monthly averages.

What this dataset does and does not support

Supported: the real price of gold is +40.6% above its 1980 level, not +508.1%; the real drawdown from the 1980 peak took 541 months to recover; real and nominal returns diverge sharply in high-inflation decades; and monthly gold returns are essentially uncorrelated with the contemporaneous inflation rate.

Not supported: that gold is a reliable inflation hedge over a year or less. The regime test's interval crosses zero and the forward means point the other way. And nothing here says what gold does next.

If you size positions on this market, the deflated series is the one to keep next to your nominal chart — the drawdown-to-size converter and the position size calculator both take a drawdown figure, and the real drawdowns here are much deeper than the nominal ones.

Frequently asked questions

What does "inflation-adjusted gold" actually mean here?

Gold prices are deflated by the US CPI (FRED CPIAUCSL, index 1982-84=100) and rebased so that the real index is 100 in January 1980. The series covers 801 months from 1960-01 to 2026-10. As of 2026-10 the real index is 140.6, which is +40.6% against the 1980 base, while the nominal price is $4105 per ounce — +508.1% above the January 1980 nominal price of $675. Inflation took the larger share of that gap.

Was gold an inflation hedge?

Not on this evidence, and the test is unusually direct. The correlation between the monthly real gold return and trailing 12-month CPI inflation is 0.033 over 798 months — the nominal correlation, the one that makes gold look like a hedge, is only 0.0801. Splitting the months at 4% inflation, the high-inflation half (255 months) averaged 0.348% real return per month against 0.398% in the low-inflation half, a difference of -0.05 pp with p = 0.906. The honest check is the block bootstrap, and its interval runs from -1.001 to 1.042 percentage points, crossing zero. Read the interval, not the point estimate.

Why does the point estimate change sign between the two statistical tests?

Because they measure different things, and because the split variable comes from the data. The naive split gives -0.05 pp (high inflation slightly worse). The non-overlapping subsample of 67 months gives -0.175 pp (high inflation slightly better) with p = 0.896. A point estimate that flips sign between two honest tests is the signature of noise, not of a small effect. The forward 12-month view points the same way: from a high-inflation month the average real 12-month return was 4.7% against 6.4% from a low-inflation month — real gold did better in the calmer inflation regime.

How bad was the worst real drawdown in gold history?

Far worse than the nominal chart suggests. From the real peak of 100 in 1980-01 ($675 nominal) the real index fell to 17 by 2001-04 ($260 nominal), a decline of -83% that took 255 months. It then took 541 months — 45.1 years — to set a new real high, finally doing so in 2025-02. There are 7 real drawdown episodes in this sample. The one still open runs from the 2026-02 real peak of 177.1 to 2026-10 at 140.6, a decline of -20.7% over 8 months.

How much did inflation actually change each decade of returns?

A great deal, and in both directions. In the 1970-1979 decade gold rose +1200% nominally but +540.7% in real terms, with CPI compounding at 7.4% a year. In the 1980-1989 decade the nominal loss was -39.4% while the real loss was -62.6% — inflation deepened a bad decade rather than cushioning it. In the 2010-2019 decade gold rose +32.3% nominally but only +11.2% in real terms, with CPI running at 1.76% a year.

Where does gold stand relative to its real peak right now?

The real index peaked at 177.1 in 2026-02 and is -20.7% below that 8 months later. The nominal peak fell in the same month (2026-02) because over the last few years gold rose faster than CPI did. Over the trailing 12 months — the window 2025-09 to 2026-09, 12 calendar months — gold is up 17.7% nominally and 13.5% in real terms, with CPI at 3.7% year on year.

What are the limits of this series?

Gold here is a monthly average price and CPI is a monthly index, so the real series is a monthly approximation, not a daily inflation-adjusted price. No storage, financing or roll cost is included anywhere. The CPI base is 1982-84=100 while the real index base is 1980-01=100, which is a rebasing choice and not a claim. And the joint series ends 2026-10, one month before the gold price series ends, because CPI does not run as far forward. None of this changes the drawdown arithmetic, but it does mean the real numbers are approximations of a monthly average, not executable values.

Sources

Data accessed 2026-11-09. Prices are nominal US dollars with no carry, financing or storage cost. Nothing on this page is investment advice or a forecast.

Educational content only. Not investment advice. Trading forex and CFDs carries substantial risk of loss.