The most-posted macro chart in the world.
Here it is: the Fed’s balance sheet and the S&P 500, both soaring since the 2008 crisis. Add the money supply and it climbs too. On a log scale — the only honest way to compare percentage moves across years — the three lines look joined at the hip. It feels like proof that the “money printer” drives stocks.
It isn’t. This is the oldest trap in macro: almost anythingthat trends up for fifteen years — nominal output, house prices, the price of a sandwich — overlays the S&P this neatly. A chart of two things that both go up is not evidence that one drives the other.
The fix is simple, and it’s why the popular threads never do it: stop comparing levels and start comparing growth— how fast each thing is changing. Watch what happens next.
overview
In growth terms, the relationship dissolves.
The level chart works because both sides trend up together over decades. Strip that shared trend out — ask not “how big is the money supply” but “how fast is it growing” — and the perfect fit turns into a shapeless cloud.
Then we test the harder question the story really makes: does liquidity move before markets, the way a cause moves before its effect? If it did, the line-up curve below would bulge to the left. It doesn’t.
growth_scatter
leadlag_curve
The Fed reacts to markets — it doesn't lead them.
The lead-lag curves already hinted at it; here we test it head-on. The one liquidity gauge with a positive-looking tie to stocks is the Fed’s balance sheet. So we ask, both directions at once: do stock moves come before the Fed’s next balance-sheet change, or does the Fed’s change come before the stock moves?
The answer is lopsided, and it’s the heart of this report. By far the stronger pattern is that markets move first and the Fed responds. The Fed expands its balance sheet after a crash — so the famous “QE lifted stocks” overlay is the Fed reacting to a falling market, mistaken for the Fed driving the recovery.
causality_bars
Same returns whether the Fed printed or shrank.
The cleanest test the liquidity story has to pass is also the simplest. Since 2008 the Fed has spent years expanding its balance sheet (printing) and years shrinking it (quantitative tightening). If liquidity drove stocks, returns should be visibly better while the printer ran.
They weren’t. And when we run the only test that truly counts — train on the past, score on months never seen — every liquidity gauge lands at zero forecasting power. The impressive in-sample fits simply don’t survive contact with the future.
qe_vs_qt
walkforward_r2
Gold has a real link — small, and same-month.
A debunk that finds nothing anywhere would be suspicious. So here is the honest exception: of every pairing tested, one genuinely tilts upward — gold against global-liquidity flow. When liquidity surges, gold tends to rise that same month.
But notice what it is and isn’t. It’s modest, it’s coincident (not a forecast you can trade ahead of), and it leans on a short, author-defined liquidity index. A real nudge, in the moment, for gold — not the all-powerful tide the threads promise.
gold_coincident
Mid-expansion — a mild tailwind for gold, not stocks.
This is not advice to buy or sell anything on the state of liquidity. It is a map of what the money-tide story does and doesn’t survive, and where the gauges sit today. Observation, not advice.
liquidity_now
Proves nothing. Liquidity, output, and asset prices all trend up together for decades, so any level-on-level overlay looks like a perfect fit. Difference to growth and the fit evaporates — which is exactly why the popular threads only ever show levels.
No. Across the money supply, the Fed's balance sheet, and a global-liquidity index, growth has no reliable power to move before stock returns. The lead-lag curves hug zero; out of sample, forecasting power is at or below zero.
Markets move the Fed, not the reverse. Stock moves come before the Fed's next balance-sheet change far more strongly than the other way. The Fed prints after crashes — so 'QE lifted stocks' is the Fed reacting, mistaken for the Fed driving.
Identical. Post-2008 the S&P returned +0.99%/mo while the Fed printed and +1.00%/mo while it shrank its balance sheet. The market rose regardless of direction — the natural experiment kills the causal claim.
Gold, coincident, modest. Global-liquidity flow and gold move together about +0.26 in the same month — the only genuine positive link found. Real but small, not predictive, and resting on a short, author-defined index.
- —Three liquidity gauges (M2 money supply, the Fed's WALCL balance sheet, a global-liquidity index) plus the S&P 500 and gold. No credit, flow, or positioning data.
- —The two better-looking gauges are the shortest: the Fed's balance sheet has ~280 monthly points and the global index ~200, against ~800 for M2 — so the more interesting numbers are also the least statistically trustworthy.
- —The global-liquidity index is a black box: its constituents and weights are author-defined and undocumented, only its growth is meaningful, and its best result spans roughly one liquidity cycle.
- —'M2 contraction precedes weak returns' is a sample of one — the money supply has shrunk in only 15 months ever, all in 2022–2024. That scary conditional is a single bear market, not a repeatable law.
- —Post-2008 confounding: stocks rose ~13% a year through both QE and QT alongside zero rates, fiscal stimulus, and mega-cap earnings. Crediting liquidity specifically is unidentifiable from this data.
- —Descriptive, not predictive. These tests describe what happened; they don't tell you what's next, and the current readout is context, never a forecast.
Descriptive research, not financial advice. Stocks and gold can both fall hard and stay down for years regardless of what the money supply is doing — as both have more than once in this dataset.
Appendix — every chart in one place →