You have seen the chart. Bitcoin price overlaid on M2 money supply, shifted forward a few weeks, and the two lines hug each other so tightly it looks like proof of something. Over certain windows the correlation is real and statistically significant. The problem is that a correlation with no mechanism behind it cannot warn you when it is about to break, and this one breaks more than people admit.
What M2 measures and what it misses
M2 is currency in circulation plus demand deposits, savings deposits, money market funds, and other near-money. It is the broadest measure of money most people bother with. When it goes up, there is more cash sloshing around to be allocated, and some of that lands in financial assets, crypto included.
What it leaves out matters just as much. M2 says nothing about credit creation by shadow banks, nothing about leverage in the system, and nothing about velocity, which is just how fast money changes hands. Two economies with identical M2 but different velocity have very different effective liquidity. If M2 rises while velocity falls by the same amount, effective liquidity is flat, and the M2-to-price relationship quietly stops holding.
That is basically the 2020 to 2023 story. M2 ballooned in 2020 and 2021 on fiscal stimulus, a lot of it flowed into assets, and the chart looked bulletproof. Then in 2022 and 2023 M2 growth slowed while velocity stayed depressed, and the link to crypto fell apart. Anyone reading the M2 line by itself missed that velocity was the thing actually driving the divergence.
The lag is not fixed
The correlation is usually tightest with a lag of about 10 to 14 weeks. That is roughly how long it takes new money to work through the economy and reach the speculative end of the pool. But the lag is not a constant. It moves with how fast monetary policy transmits, how healthy the banking system is, and how much risk people feel like taking.
When risk appetite is high, the lag shrinks, because investors dump fresh liquidity into risk assets almost immediately. When appetite is low, the lag stretches out or the relationship just snaps, because the new money gets parked in something safe instead of chasing crypto.
Global M2, and the dollar problem
For crypto, global M2, the sum across major economies, matters more than US M2 on its own. Crypto trades everywhere and capital from any jurisdiction can wander in. Chinese M2 expanding while US M2 sits flat still adds to global liquidity and can hold crypto up.
The catch is that converting global M2 into one currency, usually dollars, drags exchange rates into the picture. When the dollar strengthens, global M2 measured in dollars shrinks even if not a single central bank tightened anything. That is a purely mechanical effect, and it can smear the M2-crypto correlation enough to throw off a false signal.
How I would actually use it
Treat M2 as a background condition, not a stopwatch. When it is expanding, the macro backdrop is friendly to risk and you can lean into exposure. When it is contracting, the backdrop is hostile and you tighten up. That is the whole job it does well.
Do not try to time exact entries and exits off it. The lag wanders too much, and too many other things move crypto over a few days or weeks. M2 tells you whether the tide is coming in or going out. It will not tell you which wave to take.
The most useful cut of the data is rate of change, not the level. A deceleration, meaning M2 still growing but slower, often shows up 2 to 3 months ahead of asset weakness. An acceleration tends to lead strength by about the same stretch. Watching the second derivative, whether growth is speeding up or slowing down, gives you an earlier read than staring at the level. When I am building macro context into Blockcircle screens, that is the version I lean on, because a level chart tells you where you have been and the rate of change hints at where you are headed.