Quantity Theory of Money: Fisher, Friedman and Their Limits

Money supply and prices: the equation that anchored a century of monetary theory — and why its predictive power broke down precisely when central banks needed it most.

MV = PT looks like a tautology and behaves like a law — until it doesn’t. The disconnection between money growth and inflation between 2008 and 2021 forced central banks to rebuild their toolkit from scratch.

Eurozone M3 grew 12.3% year-on-year in February 2021 (ECB) while HICP that same month printed 0.9%. The collapse of the historical money-to-prices relationship since the 2010s is one of the most consequential empirical breaks of modern macroeconomics — and the reason most central banks no longer publish monetary aggregate targets.

The equation everyone learned, and what it actually means

The Quantity Theory of Money (QTM) traces back to sixteenth-century Spanish scholastics observing the price effects of New World silver, but its modern algebraic form belongs to Irving Fisher’s The Purchasing Power of Money (1911): MV = PT, where M is the money supply, V the velocity of circulation, P the price level, and T the volume of transactions. Read mechanically, it is an accounting identity — true by construction. Read theoretically, with V and T treated as approximately stable in the short run, it becomes a causal claim: a 10% increase in the money stock translates, eventually, into a 10% increase in the price level.

The theoretical step from identity to law rests on two empirical premises. First, that velocity (V) is a behavioural constant or at least a slowly varying function of structural factors — payment technology, financial deepening, transaction conventions. Second, that real output (T) is determined by supply-side factors independent of money. If both hold, money becomes the residual driver of nominal prices. Most of macroeconomic teaching from the 1950s through the 1990s rested on this scaffolding, even when the precise mechanism of transmission was contested.

The Friedman moment: from identity to causal claim

Milton Friedman’s restatement of the QTM in 1956 (in Studies in the Quantity Theory of Money) sharpened the framework. His joint work with Anna Schwartz, A Monetary History of the United States, 1867-1960 (1963), built the empirical case: documented that every major U.S. inflation episode through 1960 had been preceded by sustained acceleration of monetary aggregates, and that the Great Depression itself reflected, in their reading, a one-third contraction of M2 between 1929 and 1933 that the Fed failed to offset. Friedman’s 1968 American Economic Association presidential address (“The Role of Monetary Policy”) delivered the operational corollary: long-run inflation is “always and everywhere a monetary phenomenon,” and central banks ought to target a stable money growth rule rather than discretionary interventions.

The intellectual victory peaked in the 1970s. The Volcker Fed, while not a strict monetarist, used M1 growth targets between October 1979 and 1982 as the operational anchor for the disinflation campaign. The Bundesbank ran explicit money-supply targeting from 1974 to 1998, with M3 as its operational variable. As we discuss in our framework on the two foundational logics of cost-push and demand-pull inflation, the QTM provided the analytical frame for diagnosing demand-pull episodes.

Velocity: the variable that broke the model

The empirical case for the QTM rested on velocity stability. From the 1960s through the early 1980s, U.S. M2 velocity oscillated around 1.65-1.85 (FRED, M2V series), tight enough that money-growth targets could plausibly map onto inflation forecasts. Then velocity started moving. M2 velocity declined from 2.20 in 1997 to 1.41 in 2019, then collapsed to 1.10 by Q2 2020 — a 50% cumulative decline over two decades. Once V is no longer a constant, the deterministic link between M and P breaks: the same money growth rate can produce wildly different inflation outcomes depending on the velocity regime.

The structural drivers behind the velocity collapse are well-identified: financial deepening (more transactions per unit of “money” narrowly defined), the rise of money-market funds and shadow banking, the growth of bank reserves held passively at central banks after 2008, and the demand for money as a precautionary asset rather than a transactional one. Each shrinks the share of the money stock that actually circulates against goods and services. Our piece on the differences between M1, M2 and M3 monetary aggregates explores how these definitional choices interact with the velocity puzzle.

The post-2008 anomaly: why QE didn’t produce 1970s-style inflation

The most visible test of the QTM came after 2008. The Federal Reserve’s balance sheet expanded from $0.87 trillion in August 2008 to $4.5 trillion by 2014 (Federal Reserve H.4.1), then to $8.97 trillion by April 2022. M2 doubled roughly between 2008 and 2021. By the simple QTM reading, this would have produced a major inflation surge. It did not — at least not until 2021, and even then the timing and channels did not match the monetarist prediction.

The empirical break has several explanations. First, the new money sat largely on bank balance sheets as reserves, never reaching the broader economy as transactional money. Second, velocity collapsed in parallel with monetary expansion, mechanically offsetting the impact on PT. Third, deflationary structural forces — globalisation, ageing demographics, technology-driven price compression — kept the inflation trajectory subdued regardless of monetary aggregates. Our analysis of whether money creation always causes inflation develops this point in detail. The Fed effectively redefined “money supply” empirically: what mattered was not the monetary base but the credit supply transmitted to the real economy via the banking system.

The 2020-2024 partial vindication

The post-pandemic episode partially rehabilitated the QTM. U.S. M2 grew 26.9% in 2020 (FRED M2SL) — the fastest annual growth in the postwar series. CPI inflation accelerated from 1.4% in January 2021 to 9.1% in June 2022 (BLS). The temporal lag (12-18 months between M2 acceleration and CPI peak) is broadly consistent with the Friedman framework. Eurozone M3 expanded 12.3% YoY in February 2021 (ECB), and HICP peaked at 10.6% in October 2022 (Eurostat).

But the vindication is partial. The fiscal-stimulus channel (direct transfers to households, not bank reserves) explains the velocity rebound that allowed money to bite. Pure QTM reasoning would have flagged the inflation risk earlier than 2021; the fact that mainstream forecasters missed it — and that the Fed itself characterised the surge as “transitory” through 2021 — illustrates that the framework’s predictive power depends critically on how money is injected and what happens to velocity. The wave-like pattern of inflation through history reflects exactly these regime-dependent transmission channels.

What the QTM still tells us today

Even after the post-2008 anomaly, the QTM retains analytical value as a long-run anchor. Friedman’s claim that sustained inflation requires sustained money growth holds in episodes lasting more than 2-3 years. The Argentine, Turkish, Venezuelan and Zimbabwean episodes of recent decades all featured M2 or equivalent broad-money growth above 30% per annum, and inflation rates that closely tracked the monetary trajectory. The QTM survives at the extremes; it loses traction in moderate-inflation regimes where structural and behavioural factors dominate. This dynamic is documented in the structural drivers of inflation regimes.

The operational implication for central banks has been a shift away from money targeting toward direct inflation targeting (anchored on expected inflation rather than on money aggregates). The ECB dropped its M3 reference value in 2003. The Fed never formally targeted money supply post-1982. Yet the underlying intuition — that monetary policy is the long-run determinant of nominal magnitudes — survives in a more nuanced form. The data series on U.S. M2 money supply since 1959 remains a key macro indicator, even though its predictive power for short-horizon inflation has weakened.

The persistent intuition: why the QTM refuses to die

Despite the empirical setbacks, the QTM remains the implicit framework most market participants use to think about inflation. Every time central-bank balance sheets expand, the question “will this be inflationary?” is essentially a QTM question. The answer requires layering several other diagnostics: the share of new money flowing into the real economy versus financial assets (the Cantillon effect that determines who receives money first), the state of velocity, the credit transmission to firms and households, and the structural inflation regime. The pillar piece on the complete framework for inflation mechanisms, measurement, history and effects integrates these layers.

The QTM also reasserts its relevance during deglobalisation episodes. The structural shift documented in our work on how the inflation regime is changing under deglobalisation pressures suggests that the deflationary forces that absorbed monetary expansion from 1990-2020 are weakening. If true, the link between M and P could re-tighten over the next decade, restoring some of the QTM’s lost predictive power. The link between money supply and stock market returns illustrates that the financial-asset channel of money creation remains active even when the goods-price channel weakens.

Beyond the equation: a regime-dependent framework

The honest analytical posture is to treat the QTM as a piece of a broader puzzle, not as a self-sufficient model. The questions that determine whether monetary expansion translates into inflation today are: where does the new money go (banks reserves, asset prices, household consumption), how does velocity respond, what is the state of credit transmission, and what structural forces are pushing prices up or down independently. The 2025-2026 trajectory of major central banks — gradual balance-sheet runoff combined with stable inflation around 2-3% — illustrates that monetary aggregates and inflation can move independently for extended periods.

The reader who internalises this regime-dependence avoids both the monetarist trap (every QE round will cause hyperinflation) and the post-2008 complacency trap (money supply no longer matters). Both errors share the same root: treating a 100-year-old equation as if it generated context-free predictions.

Last updated — 7 May 2026