The Post-Pandemic Inflation, 2021–2023

The 2021–2023 inflation episode is the most-actively-contested recent event in macroeconomic theory. US headline CPI inflation rose from approximately 1.5 percent in early 2021 to a peak of 9.1 percent in June 2022 — the highest reading in forty years — and fell to 3 percent by mid-2023, all without the recession that the Volcker disinflation framework would have predicted as the cost of comparable disinflation. The episode produced significant forecast failures for the dominant New Keynesian framework, partial vindication for supply-shock and Post-Keynesian conflict-inflation accounts, and ongoing debate about whether the working framework of mainstream macroeconomics needs revision or merely recalibration.

What happened

US CPI inflation, which had averaged below 2 percent through the post-2008 decade, began rising sharply in early 2021. The 12-month rate moved from 1.4 percent in January 2021 to 5.4 percent by June 2021 and continued upward to a peak of 9.1 percent in June 2022. Core inflation (excluding food and energy) followed with a lag, peaking at 6.6 percent in September 2022. By mid-2023 headline inflation had fallen to approximately 3 percent and core inflation to 4 percent; by the end of 2024, both were within striking distance of the Federal Reserve’s 2 percent target. The Federal Reserve raised the policy rate from 0–0.25 percent in March 2022 to 5.25–5.5 percent by July 2023, with no recession resulting.

Eurozone inflation followed a similar pattern with peaks in late 2022, driven additionally by the energy-price consequences of Russia’s February 2022 invasion of Ukraine. UK inflation peaked higher (above 11 percent) and persisted longer.

What had been predicted

In early 2021, before the inflation rose, mainstream forecasts (including the Federal Reserve’s) projected that the post-pandemic stimulus and reopening would produce moderate inflation that would prove transitory. As inflation rose through 2021 and into 2022, the official Fed framing remained “transitory” until November 2021, when Powell explicitly retired the term. By mid-2022, with inflation peaking, mainstream forecasts substantially underestimated the persistence of the inflation; by 2023, with inflation falling, they substantially underestimated the speed of the disinflation.

Larry Summers and a small group of mostly-Old-Keynesian-leaning economists publicly forecast in early 2021 that the size of the fiscal stimulus would produce sustained inflation requiring a significant Fed response. This forecast was directionally correct; Summers’s specific claim that disinflation would require Volcker-style unemployment was not borne out, but his overall warning against the transitory framing was vindicated.

The Post-Keynesian supply-shock and conflict-inflation framings produced one of the more empirically successful pre-2022 readings: pent-up demand combined with constrained supply chains and energy-price spikes would produce a sustained but bounded inflation that would resolve as supply normalised, without requiring a major recession. This reading, articulated by Isabel Schnabel, Olivier Blanchard (in his post-IMF capacity), and others, broadly tracked the realised path.

The supply-shock decomposition

Decomposing the 2021–2023 inflation between demand and supply factors has become its own substantial research literature. Standard decompositions (Bernanke-Blanchard 2023; Shapiro 2022) find that supply factors — pandemic-era supply-chain disruptions, the Russia-Ukraine energy shock, sectoral reallocation — accounted for the larger share of headline inflation through 2022, with demand factors becoming more salient in core inflation through 2022–2023. The relative weights are contested; the broad finding that supply played a larger role than the prior decade’s NK framework would have specified is widely accepted.

The conflict-inflation reading

The Post-Keynesian conflict-inflation framework reads the episode as a wage-price catch-up cycle following the supply-shock-driven price rises of 2021–2022. Workers, facing real-wage losses from the initial price spike, sought nominal-wage increases to recover their previous real-wage levels; firms, facing input-cost pressures, passed costs through. The cycle was bounded — the catch-up dynamic resolved once price levels stabilised at higher levels, without becoming a self-reinforcing spiral — partly because labour bargaining power has remained substantially below 1970s levels, partly because inflation expectations remained reasonably anchored.

The framework’s strongest claim is the empirical regularity of the wage-price-catch-up dynamic in 2022–2023; its weakest claim is the prediction of how the dynamic would resolve, which the framework specifies less precisely than retrospective accounts suggest.

The New Keynesian reading

The mainstream New Keynesian framework’s performance in this episode was substantially worse than its performance in the prior thirty years. Standard NK Phillips Curves estimated through the post-2008 period substantially underestimated the inflation that materialised in 2021–2022; the same models substantially underestimated the speed of disinflation in 2022–2023 (forecasting more persistent inflation than realised). The framework’s coefficient estimates have required substantial re-estimation; the underlying methodological apparatus has survived, but the model’s forecasting reliability has been visibly damaged.

The post-2024 NK research program is reconstructing around (i) explicit treatment of supply shocks in the Phillips Curve, (ii) more flexible treatment of inflation expectations than the rational-expectations-with-anchoring framework allows, and (iii) integration of sectoral price dynamics that aggregate frameworks had abstracted from. The successor program is recognisably continuous with NK rather than a break from it; whether the continuity will hold remains, in the mid-2020s, undecided.

The MMT reading

Modern Monetary Theory had long argued that inflation arises principally from real-resource constraints — labour, capacity, raw materials — rather than from “money supply” dynamics in the classical sense. The 2021–2023 episode is read by MMT proponents as broadly confirming this framing: the inflation was substantially driven by genuine supply constraints (energy, supply chains, labour-force participation), and the disinflation tracked the resolution of those constraints rather than the magnitude of the Fed’s rate hikes. The framework’s stronger claim — that fiscal stimulus did not contribute meaningfully to the inflation — is more contested; standard decompositions find a non-trivial demand contribution, particularly to core inflation through 2022.

The Monetarist reading

The contemporary monetarist and Market Monetarist reading (Sumner et al.) emphasises the unusually large M2 expansion of 2020–2021 (US M2 grew by approximately 25 percent in 2020 alone) and reads the subsequent inflation as the lagged consequence. The framework’s central evidence is the rough timing match between the M2 surge and the inflation that followed; its weakness is that the M2 expansion of 2008–2010 and the QE programs that followed produced no comparable inflation, suggesting the M-Y relationship is more conditional than the strong-form monetarist framework specifies.

What changed afterwards

The episode is too recent for its long-run consequences to be fully assessed. The principal preliminary lessons:

• The Phillips Curve depends substantially on the composition of shocks. The flat Phillips Curves estimated through the 2010s reflected an unusual decade in which supply factors were quiescent; the steeper curves apparent in 2021–2022 reflect a period in which supply factors were dominant. Neither curve is the structural relationship.

• Supply-side and sectoral factors matter for inflation more than the prior NK consensus had specified. The post-2024 research program is reconstructing around this finding.

• Disinflation can occur without recession under some conditions. Whether the conditions of 2022–2023 (anchored expectations, supply normalisation, fiscal contraction in the post-stimulus period) are reproducible or were unusually favourable is contested.

• Fiscal policy has macroeconomically-significant effects on inflation. This had been the operating assumption of MMT and a contested claim in the post-2008 New Keynesian framework; the 2021–2023 episode substantially confirmed it.

Unresolved questions

• How much of the inflation was demand-driven (fiscal stimulus, monetary accommodation) and how much supply-driven (pandemic, war, supply chains)? Standard decompositions vary; the question is methodological as much as empirical.
• How much did Fed rate hikes contribute to the disinflation, and how much did supply normalisation? The disinflation tracked supply normalisation more closely than the rate hikes’ transmission lag would predict; this is the principal embarrassment for the standard NK reading.
• Will inflation expectations remain anchored in the next inflation episode, or did 2021–2023 produce durable damage to the credibility of the 2 percent target? Active research question.
• Did the episode reveal a durable change in the Phillips-Curve relationship, or merely a regime-conditional shift the framework can absorb? The principal active debate in mid-2020s macroeconomics.