What causes inflation?

Few questions in macroeconomics have produced more sustained and structured disagreement than this one. The traditions agree that sustained inflation is a phenomenon to be explained; they agree that it has happened in identifiable episodes — the German hyperinflation of 1923, the postwar US through the 1970s, Latin American chronic inflation, the post-pandemic 2021–2023 episode — and that not every episode admits the same explanation. They disagree, often sharply, on what fundamentally drives the phenomenon, on what central banks and fiscal authorities can do to control it, and on how much of any given episode can be attributed to monetary, demand, supply, distributional, or expectational factors.

This page collects each tradition’s position. As primary-source testimony is added, individual claims will be linked to the dated statements that support them.

The fundamental positions

Monetarist: inflation is always and everywhere a monetary phenomenon

Sustained inflation requires sustained monetary expansion. Cost-push and conflict-driven accounts identify proximate channels but cannot explain why the price level rises over time unless accommodated by monetary policy. The empirical record through the 1970s — sustained money growth tracking sustained inflation across countries and decades — is the central evidence. The position’s most-quoted articulation is Friedman’s 1968 AEA address, and the framework’s central credit comes from the natural-rate prediction that the postwar Phillips Curve would break under sustained accommodation.

The strong claim has been substantially complicated by the post-1980 breakdown of the standard money-aggregate relationship and the post-2008 experience of large monetary-base expansions producing little immediate inflation. Contemporary Market Monetarist refinements (Sumner et al.) preserve the monetary-causality intuition while shifting the operative variable from monetary aggregates to nominal-GDP futures.

New Keynesian: inflation is a Phillips-Curve phenomenon conditioned on expectations and credibility

Inflation in the contemporary mainstream framework is governed by a forward-looking Phillips Curve in which current inflation depends on expected future inflation and the current output gap, with the slope of the relationship determined by the frequency of price adjustment in the underlying microeconomic environment. Central banks influence inflation through the policy rate, which moves real interest rates (because prices are sticky), which affects aggregate demand and the output gap, which affects inflation through the Phillips Curve. The framework’s key empirical anchor is the credibility of the inflation target: well-anchored expectations make the Phillips Curve flatter (because firms expect transitory shocks to dissipate without permanent inflation consequences) and stabilisation easier.

The framework’s performance through the Great Moderation (1985–2007) was substantial; its performance through the post-2008 zero-lower-bound period and the 2021–2023 inflation episode has been mixed. The principal active research questions involve why the Phillips Curve appeared to flatten in the 2010s, why standard NK Phillips-Curve estimates substantially under-predicted 2021–2023 inflation, and what role supply-side factors should play in the framework.

Post-Keynesian: inflation is distributional conflict mediated by institutions

Inflation is fundamentally a contest between wage-earners and profit-receivers over the real income share, mediated by labour-market institutions, indexation arrangements, central-bank credibility, and political settlement. Sustained inflation requires a sustained mismatch between claimants’ real-income demands; the apparent monetary-causality of the Friedman-Schwartz tradition is, on this reading, descriptive of the channel through which the conflict is sometimes accommodated rather than fundamental. The 1970s inflation tracked oil prices and labour-management conflict more closely than it tracked any standard monetary aggregate; the 1980s and 1990s disinflation worked principally by breaking labour bargaining power through unemployment.

The 2021–2023 episode produced one of the more empirically successful Post-Keynesian readings: supply-shock origin, wage-price catch-up dynamics, resolution without major recession when supply normalised. The framework’s weakness is its predictive thinness: it identifies many proximate sources of inflation pressure but generates few specific quantitative predictions.

Austrian: inflation is monetary distortion of relative prices

Sustained inflation is the consequence of central-bank credit expansion. The mechanism is not principally about the price level — that is the symptom — but about the distortion of relative prices and the structure of capital that monetary expansion produces. Inflation as conventionally measured (a CPI level) underweights the more consequential damage: the misallocation of investment toward long-horizon, capital-intensive projects that consumer time-preferences do not warrant. The 1970s stagflation and the 2008 financial crisis are both, on the Austrian reading, episodes of malinvestment liquidation following extended credit booms.

The framework’s empirical credit comes from its directional alignment with major boom-bust episodes; its empirical thinness comes from the difficulty of operationalising claims about capital-structure distortion in observable data.

New Classical / Real Business Cycle: inflation reflects the prevailing monetary regime

Inflation is determined by agents’ rational expectations of future monetary policy, conditional on the policy rule the central bank operates under. Phillips-Curve trade-offs are equilibrium artefacts of the regime; regime change moves the curve. Anticipated monetary policy has limited real effects; the operative policy variable is regime credibility, not the current policy rate. The framework’s strong claim that credible disinflations should be costless was disconfirmed by the Volcker disinflation; the methodological apparatus was substantially absorbed into the New Keynesian successor program.

Modern Monetary Theory (MMT): inflation is a real-resource constraint, not a monetary one

For monetary sovereigns issuing their own non-convertible currencies, inflation arises when aggregate spending presses against real resource limits — labour, capacity, raw materials. “Money supply” is endogenously determined by spending and lending; the relevant constraint is real-resource availability, not a financing constraint. Inflation control is therefore principally a fiscal-policy problem (taxing back excess spending power) and a sectoral-policy problem (managing real bottlenecks), not a central-bank policy-rate problem. The 2021–2023 episode was read by MMT proponents as confirming the supply-constraint reading; mainstream observers read the same episode as evidence that fiscal stimulus matters for inflation more than the prior decade had suggested but in ways the MMT framework had not specifically predicted.

Where the traditions converge

• Sustained inflation is unsustainable without monetary accommodation. Even Post-Keynesian and Austrian frameworks accept that pure cost-push or pure structural inflation, in the absence of central-bank monetary expansion, would resolve into recession or stagflation rather than continuing as inflation. The traditions disagree on the fundamental role of monetary factors but converge that monetary policy can in principle stop sustained inflation at sufficient cost.

• Expectations matter. The natural-rate critique was substantively absorbed across mainstream traditions; well-anchored expectations make stabilisation easier and central-bank credibility is a real macroeconomic resource.

• Different episodes admit different decompositions. The German 1923 hyperinflation is closer to the strong monetarist reading; 1970s US stagflation was substantially monetary plus supply shocks; the 2021–2023 episode involved meaningful supply-side and conflict-distributional components. The traditions converge on a plurality of inflation mechanisms even where they disagree on which is fundamental.

Where the traditions diverge most sharply

• What is the fundamental cause? Money supply (Monetarist), expectations and policy credibility (New Classical / New Keynesian), distributional conflict (Post-Keynesian), credit-driven malinvestment (Austrian), real-resource constraints (MMT). Each tradition’s answer is, by design, the answer the others reject.

• What is the central bank’s role? Indispensable inflation controller (Monetarist, New Keynesian), institution whose policy rule determines the inflation regime (New Classical), one of several institutional levers in a distributional-conflict framework (Post-Keynesian), source of the cycle and therefore of inflation (Austrian), accommodator of fiscal policy (MMT).

• What is the cost of disinflation? Low if credible (strong New Classical), substantial but worth it (Monetarist, mainstream New Keynesian), high and concentrated on labour (Post-Keynesian), justified as the necessary liquidation of malinvestment (Austrian), avoidable through fiscal-and-sectoral management (MMT).

The empirical record across episodes

The 1923 German hyperinflation, the 1973–1982 stagflation, the 2021–2023 post-pandemic inflation, and Latin American chronic-inflation episodes each apportion explanatory weight differently across the frameworks. Future Event Dossiers — beginning with stagflation and extending to others — will document each tradition’s reading on each episode in their own words, as primary-source testimony lands.