Does the Money Supply Actually Explain Gold?

The aggregates run from most to least liquid:

• M1 — cash, coins, and checking/demand deposits. Money that's spendable immediately.
• M2 — M1 plus savings accounts, money market accounts, and small time deposits. "Near money" that converts to cash easily.
• M3 — M2 plus large time deposits, institutional money market funds, and repurchase agreements. The broadest measure. The Federal Reserve stopped publishing an official M3 in 2006; the ECB and Bank of Japan still do.

Here's where the three economies stand today, sized against GDP so the comparison means something:

GDP: US ~$32.4T, Eurozone ~€16.0T, Japan ~¥660T (2026 estimates). Japan's outsized ratio is the legacy of over a decade of BOJ balance-sheet expansion combined with structurally weak money velocity — a preview of the point below about aggressive money creation not mechanically producing broad money growth.

Over the last 25 years, growth in the three regions' broad aggregates looks like this, each indexed to 100 in 1999 so the different currencies are comparable:

US M2 grew fastest — roughly 5x — driven above all by the 2020–21 stimulus and, unusually, an outright contraction in 2022–23, the first sustained M2 decline since the 1930s. Japan's M2 grew slowest by far, despite the most aggressive QE and yield-curve control programme of any major central bank — the clearest sign that balance-sheet expansion doesn't mechanically translate into broad money growth when credit demand and velocity stay weak.

Laid against gold over the same period, the relationship is real but loose:

From 2011 to 2017, M2 kept climbing through three rounds of quantitative easing while gold fell roughly 35% from its peak — real yields and a strong dollar dominated instead. The most recent divergence runs the other way: since 2023 gold has gone parabolic while M2 growth has been comparatively modest, which looks less like a money-supply story than a central-bank buying and de-dollarization one — sovereign reserve managers (China, Poland, India, Turkey among the largest) have been the marginal buyer, not retail investors responding to inflation fears.

Twenty-five years is still a short window. Stretching to a century changes the picture again, because for the first 45 years of it gold's price was fixed by law, not set by markets: $20.67/oz under the classical gold standard, revalued to $35/oz by Roosevelt in 1934, held at $35 through Bretton Woods until Nixon closed the gold window in August 1971. Gold and the money supply were structurally decoupled by policy for nearly half the period, not by market forces.

To get real perspective we added two more lines: CPI, the inflation baseline every asset needs to beat to preserve purchasing power, and the S&P 500 on a total return basis — price appreciation plus dividends reinvested, which is the return an actual long-term holder would have earned (dividends have historically contributed roughly 40% of the index's long-run total return, so leaving them out understates equities substantially).

Good catch, and it's a fair question — the answer is just data availability: official BEA GDP and BLS CPI series both have clean, non-reconstructed values starting there, so that's the first year all five lines can be compared fairly.

Five things stand out once all five lines are on the same chart:

• Equities are the outlier, by a wide margin. The S&P 500 on a total-return basis is at roughly 2.24 million (indexed, 1926=100) — a 10.47%/yr compound rate for a century straight through the Depression, WWII, the 1970s, two lost decades, and multiple crashes. Nothing else in the chart is remotely close.
• Gold has clearly outperformed inflation — its index (~22,457) is roughly 12x CPI's (~1,893) over the full century. That's a genuine, durable real return, not just "keeping pace."
• But gold has just as clearly lagged M2 — its index sits at under half of M2's (~50,337). Both things are true at once: gold has been a solid long-run inflation hedge and a poor tracker of money-supply growth specifically.
• M2 and GDP have tracked each other far more closely than either has tracked gold — both grew on a similar order of magnitude across the century (50,337 vs. 35,587), which is the strongest evidence that money supply expands mainly because the economy and the credit system expand, not primarily through deliberate debasement.
• The 2000s→2010s stretch is a clean illustration of regime-dependence: even on a total-return basis the S&P 500 barely grew from 2000 to 2011 (265,370 → 287,600, the "lost decade" of the dot-com bust and GFC) while gold nearly sextupled over the same stretch.

If M2 tracks GDP so closely, the growth itself needs an explanation that isn't just "central banks printing money." Using the same century of US M2 data:

• 1926→2026 (full century): ~6.4%/yr compound growth
• 1926→1971 (gold standard era): ~6.3%/yr
• 1971→2026 (fiat era): ~6.5%/yr

Barely changed the long-run average pace, no — the striking part is how similar the growth rate is before and after 1971. What it changed was the variance: the 2020 stimulus, the 1970s inflation spiral. That undercuts the simplest "money printing accelerated once gold was abandoned" story.

The more complete explanation has several layers:

• Real economic growth needs it. A larger economy with more transactions needs more money to lubricate it — US nominal GDP has grown at roughly the same ~6%/yr CAGR over the same century.
• M2 is mostly bank credit, not printed cash. The overwhelming majority of M2 is deposits created when banks make loans, so M2 expands almost mechanically as the credit system grows.
• Financial deepening. Economies increasingly run on formal banking and credit over time, which pushes the M2/GDP ratio up on top of nominal growth alone.
• The 2% inflation target is, by construction, a standing instruction to expand the money supply roughly in line with nominal GDP — it isn't just tolerated, it's the policy.
• Crisis-driven step changes — WWII financing, 2008 QE, 2020 COVID stimulus — each produced a permanent upward jump the steady-state CAGR smooths over.

The last ten years give a cleaner, higher-resolution test of the GDP link specifically, since we have precise annual data rather than benchmark years a decade or more apart:

M2 and GDP track each other closely for most of the decade, both roughly doubling in index terms by 2026. The one clean break is 2020: M2 jumped to 145 while GDP barely moved, since COVID's output collapse was met with stimulus-driven money creation. M2 then stayed above GDP's line through 2021–22 before GDP caught back up as M2 actually contracted in 2023. By 2025–26 the two have converged again — M2 at 175 versus GDP at 173, almost back in lockstep. Gold, meanwhile, tracked neither for most of the period until it broke away entirely in 2024–26, nearly quadrupling while M2 and GDP each grew another 10–15% — the clearest recent evidence that the current gold rally is its own story, not a money-supply or nominal-growth story.

• Money supply growth is a slow-moving, structural signal for gold — useful over a decade, not over a trading year.
• Gold has been a genuine long-run inflation hedge (clearly ahead of CPI) but a poor tracker of M2 specifically (well behind it) — the two claims aren't in tension, they're about different baselines.
• The more consistent short/medium-term driver has been real interest rates: gold tends to rise as real yields fall, largely independent of the prevailing M2 growth rate.
• The current cycle's driver looks different again — sovereign central-bank buying and reserve diversification, a flow that doesn't show up in M2 at all.
• M2 itself grows mainly because the economy and the credit system grow, not primarily through deliberate debasement — the tightest evidence for that is how closely M2 and nominal GDP track each other, both over the last decade and over the full century.
• Equities, not gold, have been the dominant long-run wealth-compounding asset — the S&P 500 total return has outgrown gold, M2, and CPI alike, and by a wide margin.