What caused the inflation of 2021 to 2023, weighing the honest evidence

Here is a sentence that should embarrass every institution that said it: "inflation is transitory." It was said by officials, repeated by commentators and embedded in forecasts through most of 2021, and it was wrong in a way that can be dated, measured and explained. The consumer price index rose 9.1 percent over the twelve months ending June 2022, the highest reading in four decades.

I want to do something harder than gloat. I want to weigh the evidence honestly, because both sides of this argument have a tendency to pick a villain and stop. The supply-chain explanation has real force, and so does mine. Let me state the thesis, test the strongest opposing case, and see what is left.

The thesis, with terms defined

Inflation is a sustained rise in the general price level, not the rise of a single price. Demand is the total spending in the economy. Supply is the capacity to produce goods and services. A price increase in one sector can come from a supply problem. A rise in prices across the whole economy, in goods and services, rents and wages, for more than a year, requires that total spending grow faster than total supply. That is arithmetic.

My thesis has three premises.

One. In 2020 and 2021 the federal government and the Federal Reserve added an extraordinary amount of spending power to the economy.

Two. Supply disruptions were real, and they explain part of the timing and the composition of the price increases. But the economy-wide spread and the persistence of inflation require a demand explanation.

Three. The policymakers had the evidence and the warnings in time, and chose not to act.

Therefore the single largest cause of the 2021 to 2023 inflation was policy, specifically fiscal expansion and the Fed's near-zero rates and asset purchases, and the supply story is a supporting actor.

Premise one: the money and the spending

The fiscal side

Start with the laws. The CARES Act, signed on March 27, 2020, authorized roughly $2 trillion in relief. A further relief package of about $900 billion followed in December 2020. Then, on March 11, 2021, President Biden signed the American Rescue Plan Act, a $1.9 trillion package that included another round of direct payments of up to $1,400 per person, extended enhanced unemployment benefits, aid to states and localities, and other spending.

By that date the economy was already recovering. Vaccines were arriving, and the economy was reopening. A package of that size, delivered into an economy that was healing, was an injection of demand on top of a recovery. The point is not whether relief was warranted in the spring of 2020, which most conservatives accepted, but whether the third and largest dose was necessary in 2021. Lawrence Summers, who served as Treasury Secretary under President Clinton and as director of the National Economic Council under President Obama, warned in 2021 that the plan risked serious inflation. He was not a conservative. He was a Democratic economist doing arithmetic, and his warning is part of the record.

The monetary side

Now the Federal Reserve. In March 2020 it cut its policy rate to a range of zero to a quarter of a percent, and it began large purchases of Treasury securities and mortgage-backed securities, which expanded its balance sheet enormously. It kept the rate in that range until March 2022, when it made its first increase, to a range of a quarter to a half percent. In other words, the Fed held rates at zero for more than a year after the economy had begun to recover, and well into the period when prices were rising at their fastest pace in decades.

The broad money supply, M2, grew at the fastest rate in decades during the pandemic, and then, in 2023, it fell, a rare event in the postwar data. The money-based view of inflation, from Milton Friedman, holds that sustained inflation follows sustained excess growth in money, with a lag. The sequence here fits: a surge in money and spending in 2020 and 2021, inflation in 2021 and 2022, and a fall in money growth followed by a decline in inflation in 2023. The lag is not exact, and the theory is not a perfect clock, but the pattern is one that a monetarist would have forecast, and some did.

The best case for the supply-side explanation

I promised a steelman, and here it is, in words its defenders would accept.

First, the pandemic did disrupt supply. Factories closed, ports were clogged, shipping containers were in the wrong places, and semiconductors, which are in everything from cars to washing machines, were in short supply. The price of used cars, famously, rose sharply because new-car production was constrained. These are not fictions.

Second, Russia invaded Ukraine on February 24, 2022. Energy and food prices, particularly for wheat and for natural gas in Europe, rose sharply. A global energy price shock is real and cannot be blamed on any American policy.

Third, inflation was global. Many countries, with very different fiscal and monetary policies, experienced high inflation at the same time. If the cause was the American Rescue Plan, why did Europe have inflation too?

Fourth, the researchers who tried to quantify the shares, including economists in the Federal Reserve System, generally found that both supply and demand factors contributed. The question was how much of each, and estimates varied.

That is a strong case, and anyone who tells you that supply played no part is not weighing the evidence. Let me answer it.

The rebuttal, point by point

Point one: supply shocks change relative prices, not the whole price level

A shortage of semiconductors raises the price of cars relative to other things. It does not, by itself, raise the price of haircuts, restaurant meals, insurance and rent. When the price level as a whole keeps rising, the extra money has to be going somewhere. A one-time supply shock produces a one-time jump in the affected prices, and it fades when the bottleneck clears. What happened in 2021 and 2022 was different: goods prices rose, then services prices rose, then wages rose, and the pace stayed high long after the bottlenecks eased.

Point two: services and wages

This is the strongest answer. By 2022 and 2023 the persistent part of inflation was in services, including housing costs and a wide range of labor-intensive services, and in wages. Services do not depend on shipping containers. A restaurant meal is cooked and served in the same building. The rise in services prices and in wage growth reflects a labor market and a demand environment in which spending was strong, and employers competed for workers with money they had. A supply-chain story cannot explain a tight labor market with job openings far outnumbering the unemployed, which is a demand phenomenon.

Housing deserves a sentence of its own. Rents are measured with a lag in the consumer price index, which is why the shelter component kept rising well after market rents began to cool. But the underlying rise in housing costs followed a period of very low mortgage rates and strong demand, both of which are connected to monetary policy.

Point three: the global argument cuts both ways

Other countries did have inflation. But other countries also spent heavily on pandemic relief, and their central banks also held rates low and bought bonds. The fact that many countries made a similar policy mistake does not show that the policy was innocent. It shows that the mistake was widespread, and that the same shocks, a pandemic and a war, hit everyone. The honest summary is that supply conditions explain why inflation was global and demand conditions explain why it was persistent and broad. I would add that the United States' fiscal response was large by international standards, and that the timing of its inflation, which rose early in 2021, followed its spending.

Point four: the Fed's own reasoning

The Fed, in its public statements through most of 2021, described the rise in prices as transitory, driven by supply bottlenecks and base effects that would fade. That was a judgment, and it was wrong. In 2020 the Fed had revised its strategy statement to say that, after periods when inflation had run persistently below 2 percent, policy would aim for inflation moderately above 2 percent for some time. That made the central bank slower to tighten by design. By the time it began raising rates in March 2022, inflation was already well above 7 percent. The Fed's mistake was not that supply shocks did not exist. It was that it treated an economy with surging demand as if its only problem were supply.

Premise three: they had the warnings

A policymaker who fails because of a surprise deserves sympathy. A policymaker who fails after the warnings does not. Summers was an early critic of the plan, and Olivier Blanchard also warned in 2021 that inflation could prove persistent. Other economists, left and right, made similar arguments. The administration and the Fed both considered and rejected them. That rejection is a policy choice, and it should be recorded as one.

That is the same pattern I described in The Great Inflation of the 1970s was a policy choice, not bad luck: a supply shock arrives, officials call it the cause, and the monetary and fiscal accommodation that turned a price shock into a price level shift goes unmentioned. The pattern repeated in the 2020s. And the claim that corporate greed was the real culprit, which I answer in greedflation is a story that skips the question of why greed took a break, fails by the same logic: greed did not change, but demand did.

What this means for a family

The real cost of inflation is not an abstraction. A price level that rises 9 percent in a year, with wages lagging, is a pay cut. The households that suffer most are those with fixed incomes, low savings or little ability to bargain for raises, and I count that arithmetic in inflation hurts the poor most, and the arithmetic of a fixed paycheck. A fiscal and monetary policy that treats inflation as a minor side effect of stimulus has, in effect, decided who pays for the stimulus: everyone who holds dollars. That is why inflation is sometimes called the cruelest tax, because it is levied without a vote.

It also bears on the central question of who should hold the monetary pen. The Fed's twin mandate, which I examine in The Fed's dual mandate and the 1977 law that gave it two masters, invites a tilt toward employment, and the 2021 episode is a case study in the risks of that tilt. And the theory that deficits do not matter as long as there is slack in the economy, Modern Monetary Theory, received an unwelcome test, which is the subject of Modern Monetary Theory says deficits do not matter, and 2022 answered it.

What I am not claiming

I am not claiming that the relief of spring 2020 was a mistake. The shutdowns were an emergency, and a bridge for households and small businesses was defensible, though its design had flaws. I am not claiming that supply shocks were imaginary. I am not claiming that I can assign exact percentages. Economists who have tried disagree, and an honest writer should say so.

I am claiming that the best explanation for why prices rose across the economy, for years, in services as well as goods, is that the government and the central bank poured spending power into an economy with a limited ability to produce, and that the supply explanation, taken alone, cannot account for wages and services. I think that conclusion survives the evidence. I think it also has a moral dimension: the people who set the policy had the authority to prevent it and a duty to listen to the warnings.

The Constitution assigns Congress the power "To coin Money, regulate the Value thereof," in Article I, Section 8, and the power of the purse in Section 9. Both of the levers that moved in 2020 and 2021 belong, in the Founders' design, to the people's representatives, and the repeated fiscal expansion and the delegated monetary policy were made by officials those representatives chose to empower. If we want accountability, it begins with voters knowing which lever is whose.

The case in brief

1. Inflation across the whole economy requires spending to outrun supply, and between March 2020 and the American Rescue Plan of March 11, 2021, the government and the Fed added enormous spending power.
2. The Fed held its rate near zero until March 2022, after prices were rising fast, and the consumer price index reached 9.1 percent year over year in June 2022.
3. Supply shocks, from clogged ports to the war in Ukraine, were real and explain the timing and composition of some price rises.
4. Supply cannot explain persistent inflation in services and wages, and the global nature of inflation shows a common policy error, not an innocent one.
5. The warnings, including Lawrence Summers's in 2021, were on the record, so the outcome was a policy choice.

What you can do: look up the monthly consumer price index tables at the Bureau of Labor Statistics, compare the goods and services lines for 2021 through 2023, and write your senators about which of the two levers they would restrain first.