The money supply and the gold price diverged significantly between 1980 and 2000. The US money supply M2 grew from an average of $1.54 trillion to $4.79 trillion, while the gold price fell from $608 to $279 per troy ounce. Anyone who deduces an automatically rising gold price from a long-term increase in the money supply needs to explain those twenty years.
Over the entire period from 1971 to 2025, the picture initially appears different. M2 and gold both rose sharply, and their nominal level series show a correlation of 0.93. However, two curves that trend upwards over decades can appear closely linked for this reason alone. When looking at the annual changes, almost nothing of this impression remains.
SHORT ANSWER
Does the price of gold rise with the money supply? Not automatically. The US M2 and gold rose sharply in the long term from 1971 to 2025. However, the correlation between their nominal year-on-year changes is only 0.09; after adjusting for inflation, it is -0.11. The money supply is part of the monetary environment of the gold market, but on its own, it provides neither a price formula nor a reliable annual signal.
Why the comparison initially seems plausible
M2 encompasses a broad range of readily available money. According to the Federal Reserve’s current definition, this includes cash and liquid deposits from M1, smaller time deposits, and holdings of private money market funds. This series is not synonymous with the Federal Reserve’s balance sheet total, nor does it simply represent the amount of newly printed banknotes. A significant portion consists of private sector bank balances.
The connection to gold is nonetheless obvious. When more liquid assets meet a slowly growing supply of the metal, a higher price seems like a logical consequence. In reality, the process is more complex. Money can remain in accounts, flow into consumption and investments, or increase the price of other assets. Crucially, it also depends on how frequently existing money is used for payments. This velocity of money is not constant.
Even the Federal Reserve treats money supply as just one indicator among many. In 2006, Ben Bernanke described how deregulation and financial innovations repeatedly destabilized the relationships between monetary aggregates, income, and inflation. Alan Greenspan had already pointed out in 1997 that M2 sent such unusual signals in the early 1990s that the central bank temporarily downplayed its significance.
For gold, there’s another layer to consider. The metal doesn’t pay ongoing interest. Its attractiveness therefore depends heavily on the real returns offered by safe investments and the level of investor confidence in the purchasing power of money. This relationship is explored in more detail in our articles on gold and inflation, as well as real interest rates and the gold price .
Which data are included in the evaluation?
The calculation combines three publicly available US dollar series. M2 is based on the Federal Reserve’s monthly, seasonally adjusted FRED series, M2SL . Gold prices are taken from the World Bank’s annual Pink Sheet . To adjust for inflation, the non-seasonally adjusted US Consumer Price Index (CPI-U) for all urban consumers is used.
Annual averages are calculated from the monthly values of the M2 and CPI-U; for gold, the annual value published by the World Bank is used. The analysis ends in 2025, the last fully completed calendar year. The significant price movements of 2026 are not considered because an annual average is not yet available for that year.
The starting point of 1971 was deliberately chosen. The Nixon shock of 1971 ended the dollar’s immediate convertibility to gold. At the same time, the long timeframe preserves the crucial phase following the gold peak of 1980 for the analysis.
The nominal series are set to 1971 = 100. For the real analysis, each index is divided by the cumulative change in the CPI-U since 1971. The levels and simple percentage changes compared to the respective previous year are then examined. The correlation values shown are Pearson correlations.
Real index = nominal index ÷ CPI index × 100. Thus, the values show the development of the money supply or the gold price after subtracting US consumer price inflation, each relative to 1971.
NOTE REGARDING THE DATA FOR 2025
The BLS file for 2025 contains eleven monthly values for the CPI-U; the October value is missing due to the interruption of US budget financing at that time. The annual average of 321.943 is therefore based on the available months. In June 2025, the World Bank changed its gold price definition from the average London afternoon fixing to the average of daily spot prices. It did not publish a parallel comparison series for this transition, so an isolated level effect from the file cannot be quantified. Neither of these peculiarities alters the long-term implications of the analysis.
Why the 2020 M1 change does not distort M2
Since May 2020, the Federal Reserve has classified savings deposits as “other liquid deposits” and thus as part of M1. According to the central bank, this caused M1 to jump by approximately $11.2 trillion, creating a clear break in the sequence. For M2, it was merely a shift between components, as savings deposits were already included there. The aggregate money supply of M2 remained unchanged by this reclassification.
The sharp actual increase in M2 in 2020 cannot therefore be explained as a consequence of this statistical shift. Precisely for this reason, M2 is better suited for the long-term comparison chosen here than M1.
A long look tempts one to give a simple answer.
In 1971, the average annual value of M2 was approximately $674 billion, while the price of gold was $41 per troy ounce. By 2025, M2 had grown to $21.96 trillion, while gold reached an average annual price of $3,442. Using the common base year 1971 = 100, the M2 index ended at 3,257 and the gold index at 8,395.
| Row | 1971 | 2025 | Nominal Index 2025 | Real Index 2025 |
|---|---|---|---|---|
| US-M2 | $674 billion | $21.96 trillion | 3.257 | 410 |
| Gold | $41/oz | $3,442/oz | 8.395 | 1.056 |
Annual averages; indices 1971 = 100. Real indices after adjustment with the US CPI-U.
In nominal terms, gold has thus risen significantly more than M2 over these 54 years. Even after adjusting for consumer price inflation, a substantial increase remains: the real M2 index is projected to be around 410 in 2025, while the real gold index is at 1,056. The correlation between nominal levels is 0.93; for real levels, it is 0.72.
These figures describe the common long-term trend; they do not explain it. If two time series are driven upwards over decades by inflation, economic growth, and the expansion of the financial system, their correlation can be high, even though their short-term movements hardly match. In econometrics, this problem is known as spurious correlation of trending time series. Therefore, a level correlation alone should not be interpreted as a statement about cause, effect, or prediction.
Anyone who derives a theoretical gold price from the money supply and interprets the difference to the market price as evidence of manipulation is presupposing a stable and economically sound conversion relationship between M2 and gold. This is precisely what is not evident in the data. Our article ” Manipulation in the Gold and Silver Market ” contextualizes which interventions and trading practices are actually documented.
1980 to 2000: the necessary counter-examination
The average annual price of gold reached around $608 in 1980. The US M2 money supply stood at $1.54 trillion. Twenty years later, the money supply had grown to $4.79 trillion – an increase of 211 percent. By 2000, the average annual price of gold had fallen to just $279, a loss of 54 percent.
This alone could be dismissed as an unfavorable comparison of two endpoints, because 1980 was an exceptional year for gold. However, the intervening period confirms the discrepancy. M2 rose in every single year from 1981 to 2000. Gold declined in eleven of these twenty yearly comparisons and, according to published World Bank figures rounded to whole dollars, remained unchanged twice. The money supply expanded continuously without generating a corresponding upward trend in the gold price.
| Change 1980–2000 | Nominal | Adjusted for inflation |
|---|---|---|
| US-M2 | +211 % | +49 % |
| Gold price | -54% | -78% |
Comparison of the annual averages in 1980 and 2000. The CPI-U rose by approximately 109 percent during this period.
Adjusted for inflation, the contrast is even more pronounced. M2 gained around 49 percent in real terms, while gold lost about 78 percent of its 1980 purchasing power. The consumer price index more than doubled during these two decades.
The environment after 1980 was fundamentally different from that of the 1970s. The Federal Reserve under Paul Volcker combated high inflation with an exceptionally tight monetary policy; the federal funds rate temporarily reached 20 percent at the end of 1980. Inflation fell significantly in the following years, and the central bank regained credibility. For a non-interest-bearing metal, competition from high-yielding investments was considerable. Later, a long period of relatively stable consumer prices and a changed investment environment further contributed to this.
This means the decline in gold cannot be reduced to a single factor. Rather, the past two decades reveal what is missing from a purely M2-based narrative: real interest rates, inflation expectations, exchange rates, initial valuations, and investors’ willingness to hold gold. M2 can continue to grow while these forces work against the gold price.
The annual rates make the chart less clear.
For the period from 1972 to 2025, 54 annual changes are available. The correlation between nominal M2 growth and the change in the gold price is 0.09. A value close to zero means that there is practically no linear, simultaneous relationship discernible in these data. After adjusting for inflation, the correlation becomes slightly negative at -0.11.
| Comparison 1971–2025 | Pearson correlation |
|---|---|
| Nominal levels | 0,93 |
| Real levels | 0,72 |
| Nominal year-on-year changes | 0,09 |
| Real year-on-year changes | -0.11 |
Even the obvious objection of a delayed effect does not change the result. Comparing M2 growth with the gold price change of the following year, the nominal correlation is 0.03. With a two-year gap, it is 0.09. For the real changes, the corresponding values are 0.00 and 0.07.
Since individual gold years with very strong fluctuations can influence a Pearson correlation, the Spearman rank coefficient was also calculated. It is less sensitive to the magnitude of individual extreme values. With 0.06 for the nominal and -0.04 for the real annual changes, this test also remains close to zero.
In 35 of the 54 years, both series increased. At first glance, this also sounds like a connection. However, M2 increased in 53 of these 54 years. In 16 years, gold fell despite an increasing money supply; in 1995 and 2000, the World Bank’s annual value, rounded to the nearest dollar, remained unchanged. The only year with a falling M2 and an increasing gold price was 2023.
In 1976, for example, the money supply increased by 12.8 percent and gold lost 22.4 percent. In 2023, the opposite was true: M2 fell by 3.4 percent, while the price of gold rose by 7.9 percent. Such years are not mere statistical footnotes. They demonstrate that there are multiple pathways between the money supply and the gold market.
RESULT OF THE EVALUATION
The analysis identifies a strong common long-term trend, but no reliable correlation between annual changes. It proves neither that M2 is irrelevant to gold, nor that money supply growth drives the gold price. What is demonstrated is that M2 alone hardly describes the annual gold price development from 1971 to 2025.
How monetary impulses can still reach the gold market
Money supply growth can become relevant for gold if it alters expectations about inflation and currency stability. If, at the same time, the real interest rates on safe investments fall, the opportunity cost of owning gold decreases. In such an environment, additional money can meet a growing willingness to hold a portion of one’s assets in precious metals.
The same change in M2 may have no effect under other conditions. Rising deposits can reflect strong economic growth or higher demand for liquidity. If money circulates more slowly, inflation expectations remain stable, and interest-bearing investments are attractive, this does not automatically create buying pressure on gold.
The year 2020 demonstrates how multiple forces can operate simultaneously. M2 rose by 19.1 percent, gold by 27.2 percent. Pandemic-related uncertainty, extensive government transfers, a very loose monetary policy, and low real interest rates converged. By 2021, the correlation was weak: M2 grew by another 16.1 percent, but gold rose by only 1.7 percent on average for the year and lost real value after CPI-U adjustment.
Gold is also a globally traded asset. Demand from other currency areas, purchases by central banks, and political risks influence the dollar price independently of the US M2. Our articles ” Why Do Central Banks Buy Gold?” and “Is Gold a Safe Haven?” address two of these independent sources of demand.
Why the ECB M3 is not a second comparison series
For a German-speaking audience, a look at the European Central Bank’s M3 is a natural fit. However, M3 is defined more broadly than the US M2. In addition to the components of M2, it includes, among other things, repo transactions, money market fund units, and debt securities issued by monetary financial institutions with maturities of up to two years.
A direct comparison of Euro-M3 with the gold price in dollars would conflate currency and money supply effects. A proper euro-based analysis would require the gold price in euros, a price index for the eurozone, and a separate decision regarding the starting point: before or after the introduction of the common currency. This would constitute a second study with a different data basis. Therefore, here M3 remains an explanation of the different terms, not an additional series of calculations. The fact that even a change in the price currency alters the results is demonstrated by comparing the gold price in euros and US dollars .
What place does M2 deserve in gold analysis?
M2 describes the stock of liquid assets in the financial system. Its significance depends on why it grows, how the velocity of money develops, and what returns investors outside the gold market can achieve. As the sole metric, the aggregate remains too crude.
A chart from 2000 onwards appears convincing because gold and the money supply rose together during key periods. However, starting in 1971 changes the perspective: the twenty years after 1980, the gold weakness from 2011 to 2015, and the reverse trend in 2023 demonstrate the limitations of this simple formula.
For a reliable assessment, real interest rates, inflation expectations, and the dollar are just as important as investor and central bank demand. M2 complements this analysis with the stock of liquid assets in the US financial system.
Frequently asked questions about money supply and gold price
Does gold automatically rise in price when M2 increases?
No. From 1980 to 2000, US M2 grew by 211 percent nominally, while gold lost 54 percent. The correlation between the annual changes is also close to zero at 0.09.
Why is the long-term correlation still so high?
Both nominal series rise over long periods. Inflation, economic growth, and the expansion of the financial system create trends that favor a high level correlation. However, this does not necessarily imply a stable relationship between their annual movements.
Was there a series break in M2 in 2020?
The reallocation of savings deposits from May 2020 onwards created a break in the M1 money supply. Since these deposits already belonged to M2, the total M2 money supply remained unchanged according to the Federal Reserve.
What does inflation-adjusted gold price mean?
The nominal dollar price is adjusted for changes in the US Consumer Price Index. This allows us to see whether gold has actually gained purchasing power compared to 1971, and is not simply quoted in more, now less valuable, dollars.
Would the ECB-M3 lead to the same result?
This cannot be deduced from this analysis. M3 has a different composition and would need to be examined using gold in euros and a suitable eurozone price index. A comparison with the dollar gold price would be methodologically flawed.
Conclusion
From 1971 to 2025, US M2 and the price of gold rose far more sharply than the general price level. The shared long-term trend is unmistakable. Equally unmistakable is the long divergence after 1980: While M2 continued to grow, gold lost more than half its nominal value and around 78 percent of its real value over two decades.
The year-on-year changes therefore provide the more important correction to the impressive long-term chart. Their correlation remains close to zero even with a one- or two-year delay. Money supply growth can influence gold via inflation expectations, real interest rates, exchange rates, and confidence. M2 only becomes meaningful in conjunction with these variables.
About the author
Dirk Wasserthal writes for Wasserthal RareCoin.Store about precious metals, certified rare gold coins, and the history of their markets. His focus is on verifiable sources, a clear distinction between statistical correlation and causality, and the understandable explanation of complex market mechanisms.
Transparency note
Wasserthal RareCoin.Store deals in rare, certified gold coins and therefore has a financial interest in the gold and collector coin market. The calculation of this figure is based on publicly available data from the Federal Reserve, the World Bank, and the Bureau of Labor Statistics. It is for informational purposes only and does not constitute individual investment, legal, or tax advice.
Sources and data sources
- Federal Reserve Bank of St. Louis, FRED: M2 (M2SL) – monthly, seasonally adjusted US money supply in billions of dollars; source of the initial data is the Board of Governors of the Federal Reserve System.
- Board of Governors of the Federal Reserve System: H.6 Money Stock Measures – About – current components and historical availability of M1 and M2.
- Board of Governors of the Federal Reserve System: H.6 Technical Q&As – Reclassification of savings deposits from May 2020, series break in M1 and unchanged aggregate M2 money supply.
- Board of Governors of the Federal Reserve System: What is the money supply? Is it important? – Definition of US monetary aggregates and their role as part of a broader set of indicators.
- World Bank Prospects Group: Commodity Markets – Pink Sheet Data – annual gold prices in US dollars per troy ounce.
- World Bank: Commodity Price Data, Annual Prices – annual file used and description of the gold price series.
- US Bureau of Labor Statistics: Historical Consumer Price Index for All Urban Consumers (CPI-U), December 2025 – monthly and annual index values adjusted for inflation.
- US Bureau of Labor Statistics: CPI-U Historical Tables for US City Average – historical index and note on the missing October 2025 value.
- Clive WJ Granger and Paul Newbold: Spurious regressions in econometrics, Journal of Econometrics, 1974 – fundamental article on spurious relationships in trending time series.
- Ben S. Bernanke: Monetary Aggregates and Monetary Policy at the Federal Reserve – A Historical Perspective, November 10, 2006 – historical instability of the relationships between monetary aggregates and other nominal variables.
- Alan Greenspan: Rules vs. discretionary monetary policy, September 5, 1997 – changed velocity of money and limited explanatory power of M2 in the early 1990s.
- Federal Reserve History: Volcker’s Announcement of Anti-Inflation Measures – monetary tightening, interest rates and disinflation from 1979/1980.
- Federal Reserve Bank of St. Louis, FRED: Velocity of M2 Money Stock – Ratio of nominal gross domestic product to the average M2 money supply.
- European Central Bank: Monetary aggregates – definition and components of M1, M2 and M3 in the Euro area.
Calculation and last source verification: August 29, 2026. The correlation values were calculated from 55 annual averages for 1971 to 2025 and 54 annual changes for 1972 to 2025. The lag tests use 53 or 52 pairs of observations, depending on the interval. Data series may be revised retrospectively.
