Advertiser disclosure
Gold $4,363.10 +27.2% on the year Silver $64.50 +66.2% on the year

11 Sept 2026 live spot

Market data · Real interest rates since 1962

Real interest rates and gold

The real 10-year interest rate is 1.45% as of the July 2026 close, which is the 4.75% Treasury yield minus the 3.30% rise in consumer prices over the twelve months before it. Real interest rates have been negative in 125 of the 774 months since January 1962, and gold has gained 107.3% since this one bottomed at minus 6.25% in March 2022. A scheduled job rebuilds both series each month from FRED.

Yield and inflation from the same month end, 2026-07-31 774 monthly observations since January 1962

Real 10-year interest rate

1.45%

July 2026 close. Treasury 4.75%, consumer prices 3.30% over the prior 12 months

Where that sits on 64 years

1.45%
-6.25%March 2022 low 1.99%64-year average 9.58%May 1984 high

Read this before the chart

This is the realised real rate, built from inflation that has already happened, rather than the expected real rate a bond is actually priced on. The two pull apart hardest when inflation surprises, which is the period a gold reader cares about most, and the size of that gap is measured in full further down.

The numbers

The real interest rate and gold, side by side

Every figure here comes from the July 2026 close. That is the last month the Treasury yield, the consumer price index and the gold fix can all be paired, because CPI publishes later than the other two, so the real rate always stops a month or so short of where the Treasury series already reaches.

Real 10-year interest rate

1.45%

1.63% in Jul 2025

The 10-year Treasury yield minus the change in consumer prices over the twelve months before it. A change in this number is measured in percentage points rather than percent, so the card prints the Jul 2025 level beside it instead of a return.

10-year Treasury yield

4.75%

+0.4% points since Jul 2025

The constant maturity yield at the July 2026 close, from FRED series DGS10. This is the half of the real rate that moves on Federal Reserve policy, Treasury supply and what the bond market expects from both.

CPI-U over the prior 12 months

3.30%

+0.6% points since Jul 2025

The other half, and the half that has already happened. Subtracting inflation the reader has lived through is what makes this a realised real rate rather than the expected one a bond is priced on.

Gold, July 2026 close

$4,026.60/oz

+22.1% since Jul 2025

The LBMA fix at the July 2026 close, which is the month every paired figure on this page uses. Gold has one further close, $4,562.75 in August 2026, and it carries no real rate yet because CPI for that month is unpublished.

Average real rate since January 1962

1.99%

The median across 774 months is 1.94%

Today's reading sits at the 39% percentile of that record, so it is below the long-run average without being anywhere near the ends of the range.

Months below zero

125 of 774

11 separate episodes

Negative real rates are rarer than the coverage suggests, at 16.1% of the record, and they arrive in long runs rather than scattered months. The longest ran 36 months from June 2020.

Sources: Board of Governors of the Federal Reserve System, market yield on 10-year Treasury constant maturity (H.15). U.S. Bureau of Labor Statistics, CPI-U (All Urban Consumers, SA). LBMA Gold Price (PM London auction, USD/oz). Retrieved 2026-09-09

The long record

Real interest rates chart since 1962

The chart plots the real rate itself rather than its two inputs, on a linear axis with a rule drawn at zero. It changes sign 22 times in 65 years, and an axis that cannot show the sign would hide the only feature of this series anyone argues about. It closed below zero in 125 months from 11 separate runs, and the 8 runs of 3 months or longer are shaded.

Treasury 10-year yield minus trailing 12-month CPI-U, monthly

The real interest rate, 1962 to today

July 2026

1.45%

Yield 4.75%, inflation 3.30%

The dashed rule is the 1.99% average across all 774 months and the solid rule is zero. Turn on the shading to see the 8 stretches of 3 months or longer spent below it. They cluster in the 1970s and again after 2020, with almost nothing in between.

May 1984 high9.58March 2022 low-6.2564-year average 1.99
197019801990200020102020
The 10-year Treasury constant maturity yield minus the change in consumer prices over the prior twelve months, at each month end from January 1962 to July 2026. It peaked at 9.58% in May 1984 and bottomed at -6.25% in March 2022, and it closed below zero in 125 of the 774 months. The shaded bands are the 8 stretches of 3 months or longer spent below zero. Linear axis with a rule at zero, because the series changes sign. Sources: FRED series DGS10 and CPIAUCSL.

Source: Board of Governors of the Federal Reserve System, market yield on 10-year Treasury constant maturity (H.15) and U.S. Bureau of Labor Statistics, CPI-U (All Urban Consumers, SA). Retrieved 2026-09-09

Download the full dataset as CSV

Average real 10-year rate in each half-decade of the record

The real interest rate by half-decade

Splitting the record into five-year blocks shows how slowly this number turns. It held above 3.5% for the twenty years from 1980 to 1999, without a single negative month after 1985. It then spent the twenty years after 2005 grinding back down toward zero. The block covering 2020 onward is the only one in 65 years whose average sits below it.

2.80%%1.95%%0.70%%0.14%%4.91%%5.17%%3.62%%3.67%%2.21%%1.51%%0.52%%0.69%%-0.84%%
1962-641965-691970-741975-791980-841985-891990-941995-992000-042005-092010-142015-192020-26
Mean realised real 10-year interest rate in each five-year block from January 1962 to July 2026, computed as the Treasury yield minus the trailing twelve-month change in CPI-U at every month end. Red bars are blocks whose average sat below zero. Hover or tab a bar for its month count and gold's annualised return across the same window.

Method, and the limit it carries

How the real interest rate on this page is calculated

Every real rate here is the 10-year Treasury constant maturity yield minus the change in CPI-U over the twelve months that already finished. That is the realised real rate, and it is not the number a bond is priced on, which is a limitation worth understanding before any of the figures below get used for anything.

A buyer in March 2022 was not accepting minus 6.25% on purpose. The 8.57% inflation inside that figure had already happened, and nobody was contracting to receive it for another ten years. What sets a bond price is the inflation the market expects over the life of the bond. The expected measure and the realised one pull apart hardest exactly when inflation surprises, which is the stretch a gold reader cares about most.

The Treasury publishes the expected version directly. A 10-year TIPS yield is a real yield by construction, it needs no inflation assumption, and it is what a bond desk means by the real rate. It also begins in 2003, which is 23 years of record against the 65 years available here, and it misses both of the inflations that made gold's reputation. We took the longer series and the weaker measure, and the cost of that choice is in the next paragraph rather than in a hedge.

Recomputing each month's real rate with the inflation of the following twelve months instead of the previous twelve gives the real return the same yield went on to deliver. Across 762 months the two versions differ by a median of 1.07 points, by more than 3.79 points in the widest tenth of them, and by 7.46 points in July 2008. They track each other at a correlation of 0.74, which makes the trailing measure a serviceable proxy and a poor substitute.

The months where the trailing measure and the earned return disagreed most
Month This page reports The yield actually earned Difference
July 2008 -1.51% 5.95% −7.46 pts
August 2008 -1.48% 5.31% −6.79 pts
September 2008 -1.10% 5.23% −6.33 pts
February 2021 -0.23% -6.50% +6.27 pts
January 2021 -0.26% -6.45% +6.18 pts

The middle column subtracts the prior twelve months of CPI-U and the third subtracts the following twelve, using the same Treasury yield in both

The inflation adjustment applied to gold itself is a separate operation and it lives on our inflation adjusted gold price tracker, which restates every monthly close in constant dollars rather than deflating a bond yield. Nothing on this page adjusts the gold column for inflation, so every gold figure below is a nominal price and a nominal return.

Drivers

What moves the real interest rate

Two numbers make this one, and they move for entirely different reasons and on entirely different clocks. The yield reprices every trading day on what the bond market expects. The inflation leg is a backward-looking statistic that arrives once a month with a two-week lag, so the real rate can jump on a CPI print that told nobody anything new about the future.

What pushes each half of the real rate, and how it shows up in the series on this page
Driver Half it moves How it lands here
Federal Reserve policy expectations The yield The 10-year prices the average expected short rate over a decade, so it moves on the projections and the press conference rather than on the decision itself. It has run from 15.84% in September 1981 to 0.55% in July 2020, against a 5.81% average.
Treasury issuance and the term premium The yield More 10-year paper for sale lifts the yield without anyone changing their mind about policy, which raises the real rate here on supply alone.
Energy and food prices The inflation leg The most volatile part of CPI-U, and the reason a single month can swing this number by half a point without the yield moving at all.
Shelter, which is about a third of CPI-U The inflation leg Measured with a long lag against market rents, so the inflation leg keeps reflecting a housing market that has already turned.
Base effects from twelve months ago The inflation leg The denominator of every reading here is the CPI level one year back, so a large old month dropping out of the window moves the real rate with no new price data at all.
Annual seasonal-factor revisions The inflation leg The Bureau of Labor Statistics recalculates seasonal factors for the most recent five years each February, which can restate a reading on this page after it was published.

The yield is FRED series DGS10 at each month end and the inflation leg is FRED series CPIAUCSL, seasonally adjusted, on the 1982 to 1984 base

Why gold is supposed to care

Gold pays no coupon and no dividend, so the cost of holding it is whatever a safe asset would have paid instead, measured in purchasing power rather than in dollars. When a Treasury pays 9.58% above inflation, as it did in May 1984, that cost is enormous. When it pays minus 6.25%, holding metal costs nothing and the argument for owning some gets made for you.

Why the argument is weaker than it sounds

That reasoning describes a cost rather than a price, and the two sections below separate them. Gold also competes with equities, with the dollar and with whatever central banks are doing. Our gold against the S&P 500 tracker measures the same opportunity cost against the other asset most people actually hold.

Long history

Real interest rates and gold returns by decade

Gold's best two decades in this record are the two with the lowest average real rate, and its worst two are the two with the highest. That is where most of the relationship people describe actually comes from. It is a fact about how decades differ from each other rather than about what happens inside one, and the section after this takes the distinction apart.

The real rate and gold's annualised return, decade by decade
Decade Months Average real rate Range Months below zero Gold, a year
1960s 96 2.27% 0.7% to 3.4% 0 −6.1%
1970s 120 0.42% -4.7% to 3.5% 38 +31.1%
1980s 120 5.04% -4.2% to 9.6% 10 −4.9%
1990s 120 3.65% 1.8% to 5.3% 0 −3.5%
2000s 120 1.86% -1.5% to 5.5% 12 +14.5%
2010s 120 0.60% -1.9% to 2.2% 26 +3.5%
2020s 78 -0.84% -6.3% to 2.0% 39 +15.4%

The 1960s row starts in January 1962 and covers gold only from April 1968, and the current decade is incomplete. Past performance does not predict future results

Gold's annualised return in each decade of the record

Gold by decade, against the real rate of that decade

The gold bars are coloured by the decade's average real rate, with brass for the decades that averaged below 1% and grey for the rest. Every brass bar is positive and the three largest positive bars are all brass, which is the pattern the whole argument rests on and it is drawn from 7 observations rather than from 774.

-6.1%%31.1%%-4.9%%-3.5%%14.5%%3.5%%15.4%%
1960s1970s1980s1990s2000s2010s2020s
Gold's annualised return in each decade from April 1968, from LBMA monthly closes. Brass bars are decades whose average realised real 10-year rate sat below 1% and grey bars are the rest. The 1960s bar covers April 1968 to December 1969 and the 2020s bar runs to July 2026, so both are partial. Hover or tab a bar for that decade's average real rate and its count of negative months.

Where the 774 monthly readings landed

Distribution of the real interest rate

The busiest band sits between 1% and 2%, which holds 159 of the 774 months, and today's reading of 1.45% falls inside it. Readings below zero account for 16.1% of the record and readings above 5% for 8.9%, so both of the states people write about are genuinely unusual.

50751141591281146569
Below -2-2 to 00 to 11 to 22 to 33 to 44 to 5Above 5
Count of monthly readings of the realised real 10-year interest rate in each band, January 1962 to July 2026. The dark bar is the band the July 2026 reading of 1.45% falls into.

What we found in the data

How well real interest rates predict the gold price

Across 687 months the level of the real rate correlates with gold's previous twelve months at -0.40 and with gold's next twelve months at -0.11. Both figures come from the same months and the same two series, and the only thing separating them is which direction the return window points. The distance between them measures how much of this relationship is explanation and how much of it is forecast.

Sorted into bands the same split appears again. Gold's median return over the twelve months before a negative reading was +24.4% against −0.8% for the months above 2%, a spread of 25.2 points. Over the twelve months after those same readings the spread collapses to 2.0 points.

Band spread, looking back 12 months

25.2 pts

Band spread, looking forward 12 months

2.0 pts

Gold's median 12-month return before and after each band of the real rate

The 12 months beforeThe 12 months after
24.4%%3.1%%10.3%%14.8%%-0.8%%1.0%%-2.6%%-1.1%%
Below 0%0% to 2%Above 2%Above 4%
Median LBMA gold return over the twelve months before and the twelve months after every month in each band of the realised real 10-year interest rate, from April 1968 to July 2026. The bands separate by 25.2 percentage points looking backwards and by 2.0 looking forwards. The above 4% band is a subset of the above 2% band rather than a fourth exclusive group. Hover or tab a bar for its sample size and episode count.
Every band, with the months, the episodes those months came from, and gold either side
Real rate Months Episodes Gold, prior 12m Gold, next 12m Next 12m positive
Below 0% 125 11 +24.4% +3.1% 58%
0% to 2% 260 31 +10.3% +14.8% 76%
Above 2% 314 18 −0.8% +1.0% 53%
Above 4%subset 134 15 −2.6% −1.1% 43%

Medians on LBMA monthly closes. The above 4% row is a subset of the above 2% row, so the first three rows sum to the record and the fourth does not add to it

The episode column is the one that decides how much weight any of this carries. Those 125 months below zero came from 11 contiguous runs, so anyone reading the sample is reading 11 events rather than 125 independent draws. Two of those runs are the 1970s inflation that every gold advertisement is built from. The months above 2% are only slightly better served, at 314 months from 18 episodes.

The forward column also does something the standard story does not predict, because the best twelve months for gold did not follow the negative readings at all. They followed the middle band, where the real rate sat between 0% and 2%. That band returned a median of +14.8% against +3.1% after a negative reading and −1.1% after a reading above 4%. It is also the best-sampled of the four, at 260 months from 31 separate episodes.

One more thing worth knowing before quoting the headline correlation

The pooled correlation and the decade correlations disagree

Run the forward correlation inside each decade separately and every one of the 7 comes back positive, from 0.02 to 0.64, while the pooled figure across all 687 months is -0.11. Both are correct arithmetic on the same data, and the reason they disagree is that the negative pooled number is measuring the difference between decades rather than anything happening inside one. Decades with low real rates were decades gold did well in, and within any single decade a lower reading was not followed by a better year.

Seven decades is seven observations, which is a small enough number that the between-decade pattern is worth describing and not worth trading. It is also the reason every table on this page publishes an episode count beside the month count.

Where the relationship breaks

Why gold rose while real interest rates rose, 2022 to 2026

The real rate bottomed at minus 6.25% in March 2022, the lowest reading in 65 years, with gold at $1,942.15. It has risen +7.7 points since, and every version of the argument above says gold should have been taken apart. Gold gained 107.3% instead, or 18.3% a year, and it is the single largest thing this page has to answer for.

Real rate, March 2022

-6.25%

The lowest of the 774 months in the record

Real rate, July 2026

1.45%

A rise of +7.7 points across 4 years

Gold over the same months

+107.3%

$1,942.15 to $4,026.60, or 18.3% a year

Correlation between the 12-month change in the real rate and gold's return over the same months

The relationship decade by decade

Measured as a change rather than a level, the relationship is real. It comes back negative in 6 of the 7 decades, running as strong as -0.66 in the 1970s and -0.66 in the 2010s. The current decade is the only positive one in the record, at 0.09 across 66 overlapping windows.

-0.38-0.62-0.47-0.11-0.30-0.660.09
1960s1970s1980s1990s2000s2010s2020s
Pearson correlation between the change in the realised real 10-year interest rate over twelve months and gold's return across the same twelve months, computed separately for the windows starting in each decade. Blue bars are decades where the two moved in opposite directions, which is what the standard argument predicts, and the red bar is the decade where they did not. Hover or tab a bar for its window count.

Rolling the same correlation through a moving ten-year window shows when the turn happened rather than which decade it landed in. That measure bottomed at -0.74 in March 1983 and held below zero for the whole of the 2010s. It reached -0.73 in March 2020, then crossed above zero in February 2024 for the first time since April 2008. It reads 0.07 at July 2026. Only 67 of the 568 windows in the record are positive at all, so the current reading is unusual rather than unprecedented.

Every stretch where the real rate rose 2 points or more over twelve months and gold gained 15% or more anyway
Episode Windows Widest window Real rate Gold
January 1971 1 January 1971 to January 1972 +2.02 pts +23.9%
November 1979 2 November 1979 to November 1980 +2.30 pts +49.1%
July 1982 1 July 1982 to July 1983 +2.28 pts +23.1%
March 2001 1 March 2001 to March 2002 +2.11 pts +17.0%
September 2005 2 October 2005 to October 2006 +2.98 pts +28.3%
October 2008 1 October 2008 to October 2009 +3.36 pts +42.3%
October 2022 2 October 2022 to October 2023 +5.29 pts +21.8%
April 2023 1 April 2023 to April 2024 +2.84 pts +16.4%

Overlapping monthly windows inside one run are grouped into a single episode and the widest of them is shown, so the count of events is 8 rather than the 11 windows they contain

There are 8 of these episodes in 65 years, which makes them uncommon rather than rare. The largest is November 1979 to November 1980, when the real rate rose +2.30 points and gold gained 49.1% across the same twelve months. The second is the year after October 2008, when a +3.36 point rise sat beside a 42.3% gain. In both of them something other than the opportunity-cost argument was setting the gold price, and 2 of them belong to the current break.

The relationship also goes missing without any drama at all. Through the 1990s the real rate never once closed below zero across 120 months and averaged 3.65%. The model says that should have been a decade of steady pressure on the gold price. Gold returned −3.5% a year across it, and the correlation inside the decade came in at -0.11. A variable that explains the 1970s and the 1980s and then explains nothing for ten years is a variable with a condition attached that nobody has written down.

What we do not know

We have no confident explanation for the current break

Central bank buying is the answer usually offered, and it is not one this page can test, because neither of the two series here measures official sector demand. The World Gold Council publishes those figures and they are worth reading, and the honest position is that we are pointing at a correlate rather than a cause. What the data supports is narrower: this relationship has worked in some decades, failed in others, and is currently in one of the failing ones.

The other thing worth holding on to is that 4 years is 1 observation. The record contains 8 episodes of this shape and it has always resumed afterwards, which is an argument for patience rather than for either conclusion anyone wants to draw from it.

Timeline

Real interest rates and gold, turn by turn

Every entry carries the Treasury yield, the inflation reading and the gold price of the month it describes, plus what all three did over the following year. The story and the numbers cannot drift apart that way. The last entry is rendered from the current data rather than written, which is why it carries no forward figures.

  1. August 1971

    Nixon closes the gold window

    The dollar stops converting into gold at $35 an ounce, and from that month gold has a market price that a real interest rate can push around. The 10-year Treasury paid 6.28% against consumer prices running 4.36%, so a bondholder was earning 1.92% in purchasing power while gold sat at $40.65. Over the following twelve months the real rate rose to 3.47% and gold gained 64.5% anyway, which is the first entry in a long list of months when this relationship did not hold.

    Real rate 1.92% Yield 6.28% Inflation 4.36% Gold $40.65 Rate 12 months on 3.47% Gold 12 months on +64.5%

  2. December 1974

    The real rate hits minus 4.70% and gold peaks with it

    Consumer prices had risen 12.10% over the previous twelve months while the 10-year paid 7.40%, which left a bondholder losing 4.70% of purchasing power a year and made the deepest reading of the 1970s. Gold closed the month at $186.50 after gaining 54.4% across the 25-month stretch of negative real rates that started in August 1973. The next twelve months took the real rate back above zero, and gold gave up 24.8%.

    Real rate -4.70% Yield 7.40% Inflation 12.10% Gold $186.50 Rate 12 months on 0.63% Gold 12 months on −24.8%

  3. June 1980

    Volcker turns the real rate positive and gold stops

    The real rate bottomed at minus 4.18% in June 1980 with the 10-year at 10.09% and inflation at 14.27%, five months after gold's January peak and with gold closing the month at $653.50. Paul Volcker then drove the federal funds rate to a 19.1% peak in June 1981, holding it in double digits for most of the following two years, the real 10-year rate climbed to 4.16% within a year, and gold fell 34.8% to $426. This is the single episode that every real-rate explanation of the gold price is built on, and it is the one where the two lines genuinely moved together.

    Real rate -4.18% Yield 10.09% Inflation 14.27% Gold $653.50 Rate 12 months on 4.16% Gold 12 months on −34.8%

  4. May 1984

    The real rate reaches 9.58%, the highest in 64 years

    Inflation had fallen back to 4.33% while the 10-year Treasury still paid 13.91%, so a bondholder locked in 9.58% above consumer prices, which remains the highest real yield anywhere in this record. Gold closed at $384.25 and lost another 18.3% over the following twelve months. Anyone holding metal was giving up close to a tenth of their money a year against a Treasury, and the gold price behaved the way that arithmetic says it should.

    Real rate 9.58% Yield 13.91% Inflation 4.33% Gold $384.25 Rate 12 months on 6.71% Gold 12 months on −18.3%

  5. July 2008

    The largest gap between the measured rate and the earned one

    The trailing measure read minus 1.51% in July 2008 because consumer prices had risen 5.50% over the previous year, but the 3.99% yield a buyer locked in that month went on to earn 5.95% in purchasing power, since prices then fell through the crisis. That 7.46-point separation is the widest in the record and it is the clearest single case for why this page names its method. Gold went from $918 to $939 across those twelve months while the number underneath it swung by more than 7 points.

    Real rate -1.51% Yield 3.99% Inflation 5.50% Gold $918.00 Rate 12 months on 5.48% Gold 12 months on +2.3%

  6. September 2011

    Gold's 2011 top arrives with the real rate at minus 1.89%

    The 10-year paid 1.92% against inflation of 3.81% and gold closed September at $1,620, a month after its own peak. The real rate stayed negative for another eight months and gold managed 9.6% over the following year before spending the rest of the decade going backwards. A negative real rate was in place for the top and for most of the decline that followed it, which is why the sign of this number is a poor entry signal on its own.

    Real rate -1.89% Yield 1.92% Inflation 3.81% Gold $1,620.00 Rate 12 months on -0.30% Gold 12 months on +9.6%

  7. March 2022

    The real rate hits minus 6.25%, the lowest reading on record

    Consumer prices rose 8.57% over the twelve months to March 2022 while the 10-year Treasury paid 2.32%, which left the deepest negative reading in 64 years. Gold closed the month at $1,942.15. Every version of the standard argument says this is the moment gold should run hardest, and instead gold managed 1.9% over the following twelve months while the real rate climbed almost 5 points back toward zero.

    Real rate -6.25% Yield 2.32% Inflation 8.57% Gold $1,942.15 Rate 12 months on -1.44% Gold 12 months on +1.9%

  8. May 2023

    The last negative month, and gold starts moving

    May 2023 closed the 36-month run of negative real rates that began in June 2020, which is the longest stretch anywhere in the record, with the 10-year at 3.64% against inflation of 4.13%. Gold sat at $1,964.40, roughly where it had been through the entire negative episode. From that month, with the real rate positive and rising, gold gained 19.5% over twelve months and kept going.

    Real rate -0.49% Yield 3.64% Inflation 4.13% Gold $1,964.40 Rate 12 months on 1.27% Gold 12 months on +19.5%

  9. July 2026

    Where the record currently ends

    The real 10-year rate reads 1.45% against a 4.75% Treasury yield and 3.30% inflation, which puts it at the 39% percentile of the record and inside the 1 to 2 band that holds 159 of the 774 months. Gold closed the month at $4,026.60. This entry is rendered from the data on every build rather than written, so it moves when the series does.

    Real rate 1.45% Yield 4.75% Inflation 3.30% Gold $4,026.60

Our view

Gold IRA Digest's perspective

Every entry carries the day it was written and the month it was reading, because an undated opinion sitting under a data table reads as current no matter how old it has become. Naming the basis month also stops a data refresh quietly restating an old view under its original date.

· reading the July 2026 close · latest

The real 10-year rate is the variable most often named as the thing that sets the gold price, and across 687 monthly observations it correlates with gold's trailing twelve-month return at minus 0.40 and with gold's next twelve months at minus 0.11. Those two numbers describe different jobs, and this variable can only do the first one. We publish both because a page that shows the explanatory figure while calling it a forecast is running the same trick as a page that shows a price index and calls it a total return.

· reading the March 2026 close

Anyone who bought gold in March 2022 because the real rate had reached minus 6.25% got the right outcome for the wrong reason. The rate then rose 7.27 points over the next four years, which the standard argument says should have taken gold apart, and gold gained 137.3% instead. We have no confident explanation for that and we would rather say so than reach for central bank buying, which is the answer everyone offers and nobody has measured against this series.

· reading the July 2025 close

We chose 64 years of an approximate measure over 23 years of an exact one, and the cost of that choice is worth stating in a number rather than in a hedge. Comparing each month's trailing real rate against the real return the same yield went on to deliver, the two differ by a median of 1.07 points across the 762 months to July 2025, and the gap runs past 3.79 points in the widest tenth of them. The sample stops there because a month cannot be scored until the twelve after it have printed. The largest single separation is 7.46 points, in July 2008.

Written by the Gold IRA Digest Editorial Team and last reviewed 2026-09-09. Opinion, labelled as opinion, and never a recommendation to buy or sell gold or Treasuries.

What is scheduled

What to watch next

Windows are months rather than exact dates, because the later release dates are not published yet and inventing one would be the same mistake as inventing a verification date. The two marked as ours are dates we control.

  • Monthly, mid-month US CPI release The inflation half of this number directly, since the figure printed here is the change in CPI-U over the twelve months ending at each monthly close Bureau of Labor Statistics
  • Eight times a year FOMC rate decision The nominal half, though the 10-year yield reacts to the projections and the press conference more than to the decision itself Federal Reserve
  • Quarterly, from early February Treasury quarterly refunding announcement How much 10-year paper the Treasury intends to sell, which moves the yield through supply rather than through policy US Department of the Treasury
  • Sep 2026 Our monthly data refresh Ours Every figure on this page, including the 64-year average, the episode counts and the correlations Gold IRA Digest
  • Dec 2026 Our TIPS series addition Ours The method underneath this whole page. We intend to add the 10-year TIPS yield so the expected real rate sits beside the realised one from 2003 onward, which is the limitation set out in the method section Gold IRA Digest
  • Feb 2027 BLS annual CPI seasonal-factor revision The inflation leg for the most recent five years, which can restate a recent reading here after it has already been published Bureau of Labor Statistics
  • Feb 2027 World Gold Council full-year demand trends Central bank buying, which is the demand source most often offered as the reason gold stopped tracking real rates after 2022 World Gold Council

What this means if you are weighing a gold IRA

What real interest rates mean for a gold IRA

The opportunity-cost argument on this page is about how much gold to hold rather than about when to buy it, and the forward columns above are the reason. A real rate of 1.45% tells you what a Treasury is paying you to wait. The record says it tells you very little about what gold does over the next twelve months, so anyone timing a rollover on this number is using it for the one job it has never done.

A gold IRA also charges a setup fee, an annual custodian fee and a depository storage fee that a Treasury does not. On top of those it carries a dealer premium over spot at purchase, which runs between 5% and 12% on IRA-eligible coins. On a $50,000 purchase the difference between those two premiums is $3,500 before the account has done anything, and that is worth more than several years of the real rate this page tracks.

Which form of gold you hold changes the cost more than this chart does. A gold ETF and a mining stock both sit inside an ordinary brokerage IRA at brokerage cost. Bullion needs a self-directed IRA, a custodian and a depository, and the metal has to clear the 99.5% purity rule before a custodian will accept it. Everything on this page prices the metal at the spot price, so an ETF holder tracks that line minus an expense ratio and a mining stock does not track it at all.

Forward correlation with gold

-0.11

Real rate level against gold's next 12 months, 687 months

Negative-rate episodes

11 in 65 years

Covering 125 months, or 16.1% of the record

Typical dealer premium

5% to 12%

Charged over spot on IRA-eligible gold coins at purchase

Common questions

Real interest rate questions

What is the real interest rate right now?
1.45% at the July 2026 close, measured as the 4.75% 10-year Treasury yield minus the 3.30% rise in consumer prices over the twelve months before it. That sits at the 39% percentile of the 774 months since January 1962, below the 1.99% long-run average and well inside the -6.25% to 9.58% range the record has covered. This is the realised rate rather than the expected one, and the difference is set out in the method section above.
How do real interest rates affect the gold price?
Gold pays no income, so the case for holding it weakens as a Treasury starts paying a positive return above inflation. The record supports that as a description of what has already happened rather than as a forecast. Across 687 months the level of the real rate correlates with gold's previous twelve months at -0.40 and with gold's next twelve months at -0.11. Anyone using this variable to decide what gold does next is using a relationship roughly 3 times weaker than the one they read about.
Does gold always rise when real interest rates are negative?
No. The real rate has closed below zero in 125 months since January 1962, and gold rose over the following twelve months in 58% of them, with a median gain of +3.1%. Those 125 months came from only 11 separate episodes, so the honest sample is closer to 11 observations than to 125. Gold's 2011 top and most of the decline that followed it both happened with the real rate below zero.
Why did gold rise in 2022 when real interest rates rose?
Nobody has a settled answer, and we would rather say that than pick one. What the data shows is the size of it. The real rate rose +7.7 points between March 2022 and July 2026 while gold gained 107.3%, or 18.3% a year. The rolling ten-year correlation between the change in the real rate and gold's return turned positive in February 2024 for the first time since April 2008. Central bank buying is the explanation usually offered, and it is not one this page can test against these two series.
What is the difference between the real interest rate and the TIPS yield?
The TIPS yield is a real yield by construction, so it needs no inflation assumption and it is what a bond desk means by the real rate. The figure on this page subtracts inflation that has already happened, which is a different thing and a worse one. TIPS also begin in 2003, against 65 years of record here, and they miss both of the inflations that made gold's reputation. Comparing our measure against the real return the same yield went on to deliver, the two differ by a median of 1.07 points across 762 months.
How often have real interest rates been negative?
In 125 of the 774 months since January 1962, which is 16.1% of the record, spread across 11 separate episodes. The longest ran 36 months from June 2020 to May 2023 and contains the -6.25% low of March 2022. The 1990s produced not one negative month in 120, while the 2020s have produced 39 so far.
How often does this page update?
The Treasury, CPI and gold series are rebuilt together, and every figure here is computed at build time from those files rather than written by hand. The last month all three can be paired is July 2026, because CPI publishes later than the other two. There is no live number anywhere on this page, and nothing here is dated to a month the underlying data does not reach.

How this page is built

The real rate is the 10-year Treasury constant maturity yield at each month end minus the percentage change in CPI-U over the twelve months ending that month, across 774 observations from January 1962. Gold is the LBMA fix at the same month ends, joined from April 1968. Every figure here is computed at build time from committed data files, so the numbers are in the HTML before any script runs and every chart works with JavaScript turned off.

The three series are joined on a year-and-month key rather than on a position in an array. That matters more than it sounds: a monthly file can be missing a month, and an index-based twelve-month lookback then measures thirteen months for everything after the gap and prints it as a year. The test suite rebuilds the whole series a second way and compares the two, so a regression of that shape fails rather than ships.

Three limits are worth knowing before anyone reuses these figures. The rate here is realised rather than expected, and the method section sizes that gap at a median of 1.07 points. CPI-U publishes later than the Treasury series, so the real rate stops at July 2026 while the yield already reaches further, and no figure on this page mixes the two dates. Returns everywhere carry no fees, no dealer premium, no storage cost and no tax, so a real gold IRA outcome is worse than the gold column shown by the size of all four.

Sources

Full dataset available as CSV, with the nominal yield and the inflation leg in separate columns, so anyone who prefers a different inflation measure can rebuild the real column themselves. Corrections go to the editorial team and are logged publicly.

Gold IRA Digest is an independent publisher. This page is information, not investment advice, and nothing on it is a recommendation to buy or sell gold, Treasuries or any fund. Past performance does not predict future results. Some links on this site earn us a commission, which never changes what we publish about a company.