Quantity Theory of Money: What MV=PQ Actually Claims

The quantity theory of money says M × V = P × Q: money supply times velocity equals the price level times real output. That’s the equation of exchange. It’s true by definition, not by observation. It becomes a prediction about inflation only once you assume V and Q hold still — an assumption we can name but not verify. We hold M for 157 economies. We hold none of the other three terms, for any of them.

What does MV=PQ actually say?

M is the money supply. V is velocity of circulation — how many times a unit of money changes hands buying finished goods and services in a period. P is the price level. Q is real output. Multiply money by how fast it moves, and you get the same number as multiplying prices by how much got sold. Both sides count the same transactions two different ways. Irving Fisher formalized this in 1911, in The Purchasing Power of Money, as MV = PT — T for the volume of transactions. The modern version swaps T for Q, real output. Either way, the equation can’t be false. It restates “total spending equals total sales.” It doesn’t yet claim anything about cause and effect.

Is MV=PQ a prediction, or just true by definition?

By itself, an identity. Rearrange it and it says nominal spending equals nominal income — nothing more. The identity becomes the quantity theory, a testable claim, only when you add an assumption on top. That assumption: V and Q are stable, or at least predictable, in the short-to-medium run. Add it, and the identity turns into a mechanism. Raise M while V and Q hold still, and P has to rise to keep both sides equal. Drop the assumption, and the same equation tells you nothing about prices at all. M could rise while V falls by the same proportion, and P wouldn’t move. Both cases satisfy MV=PQ. Only one is what most people mean when they invoke “the quantity theory.”

Fisher vs Friedman: two different claims wearing one equation

Fisher’s version is mechanical. MV=PT holds at every instant, by construction. No behavioral claim attached, no forecast implied.

Milton Friedman’s 1956 essay, The Quantity Theory of Money: A Restatement, reframes the whole idea as a theory of money demand rather than a bookkeeping identity. It is the founding text of what became known as monetarism. Friedman’s claim: households and firms hold a stable, predictable amount of money relative to their income and a few other variables. If that demand function is stable, a sustained rise in the money supply beyond what it absorbs has to show up as higher spending. Once output can’t keep expanding to match, it shows up as higher prices instead.

That is testable and falsifiable in a way Fisher’s identity never was. It is also where Friedman’s most quoted line comes from — “inflation is always and everywhere a monetary phenomenon” (The Counter-Revolution in Monetary Theory, 1970). The distinction is the whole point of this section. Fisher’s MV=PQ can’t be wrong. Friedman’s claim about why it holds together can be, and whether it does is an empirical question rather than a matter of definition.

What does the theory assume about velocity — and does it hold?

The whole predictive version rests on V behaving. If velocity is genuinely stable, money growth maps onto nominal spending in a roughly fixed ratio — nominal GDP / M2, for the US case. The rest of the mechanism then follows. It isn’t stable. US M2 velocity ran near 2.2 in the late 1990s. It fell toward roughly 1.1 by 2020, about half its old level in two decades. It has since recovered to 1.412, the Q2 2026 reading (Federal Reserve Bank of St. Louis, FRED M2V, quarterly, released 2026-07-30). A term that swings by half is not the constant the load-bearing assumption needs. We trace why it fell twice — after 2008, and again after 2020 — at velocity of money.

The St. Louis Fed states the assumption and its consequence in one line: “The velocity of money is usually assumed to be constant; thus, any changes in the money supply imply a change in the price or quantity of goods” (St. Louis Fed, “Market Liquidity and Quantity Theory of Money,” 2022-08-29). Read that backwards and you have the whole conditional. The prediction holds if V holds. When V doesn’t, the identity stops saying anything about prices on its own. Money neutrality is built on the same floor. That proposition — a money-supply change moves only nominal variables like the price level in the long run, not real output — assumes the classical dichotomy between the real and nominal sides of the economy. It needs a stable V just as much.

What can our own data actually test — and what can’t it?

Only the M term. Across 157 economies with a full 10-year growth history, our own series shows the spread in money-supply growth alone. South Sudan runs 69.6%/yr (2026-06), the fastest of the 157 over a genuine ten-year window. Japan runs 2.61%/yr (2026-02). Dominica, the slowest of the 157, runs 1.07%/yr (2026-02). That is a 65-fold spread between the fastest and slowest full-decade rates.

Zimbabwe runs faster still, 338.2%/yr (2026-04) — but that one is annualised over the 47 months since a 2022-05 redenomination broke its series, so it belongs in a different column, not at the top of the same one. Rates measured over different windows are not comparable, and this page is about an identity that only works when every term is measured over the same window. We are not going to break that rule in our own table.

None of this tests MV=PQ by itself. Testing the identity against real behavior needs all four terms moving together, for the same country, over the same window. We don’t hold V, P or Q for any of the 185 economies we track — not Zimbabwe’s, not the United States’, not any other’s. A reader who wants the inflation side of any of these stories should go to the IMF’s International Financial Statistics database or a country’s own statistics office. We detail the full gap at money supply vs inflation. Every term used on this page — M, V, P, Q, CAGR, doubling time — is defined in full at our glossary, and how we source and compute the M side is documented at methodology.

Did the quantity theory predict the 2020s?

Partly. The one case where we can assemble something close to a full test is the United States. It only works because a separate, external CPI series lets us check our own M data against something.

One label first, because the numbers below are useless without it. Our US series is the IMF’s harmonised broad-money aggregate (MFS_MA/BM_MAI, quarterly), not the Federal Reserve’s headline M2. It is the wider of the two, by roughly $7-8T at current levels. When this page says the US money supply grew 23.5%, it means that series — not the ~$23T M2 figure a reader will find on FRED.

On that series, US money growth peaked at 23.5% year over year in Q2 2020, after the Fed cut rates to zero and began emergency bond purchases in March 2020. US CPI peaked at 9.1% exactly 24 months later, in June 2022 (BLS CPI-U, all items, 12-month change — not our data). Direction and rough timing line up with the transmission mechanism Friedman described: money growth, then a lag, then price growth.

Size doesn’t line up 1:1. Broad money rose 36.9% cumulatively from its pre-pandemic base of $19.88T (2019-12) to $27.23T by December 2021, then sat roughly flat for two years. CPI peaked at “only” 9.1% annually — nowhere close to that scale. The gap is exactly the two terms this dataset can’t supply. Real output kept growing through the recovery, absorbing part of the extra money: the Q term rising, not just P. Velocity collapsed and then partially rebuilt, absorbing more of it still: the V term falling, netting against M’s rise. One clean lag, on one country, with two of the four terms visibly doing real work in the background. That’s the honest read, not a law.

Is MMT debunked?

No, not on any evidence we hold — and “debunked” isn’t a term either side of that debate would accept as settled. Modern Monetary Theory rejects the load-bearing assumption above directly. It argues a currency-issuing government is constrained by real resources — labor, capacity, materials — not by the quantity of money it creates. New government spending funded by new reserves, on this view, need not raise prices while those resources sit idle. Quantity theorists argue the opposite: sustained money growth beyond what stable money demand absorbs eventually shows up as inflation, with a lag. Both camps point to the same 2020-2023 US episode as their evidence. Quantity theorists point to the 24-month M2-to-CPI lag above. MMT proponents point to inflation continuing to fall through 2023 even as M2 resumed growing, and to supply-side shocks doing real work in 2021-2022 that a pure money-supply story doesn’t explain on its own. We hold the M term for both arguments. We hold neither a CPI series nor a real-resources-slack measure to referee between them. That’s the honest limit, not a dodge.

FAQ