What Causes Inflation?

The consumer price index for all urban consumers (CPI-U) rose 9.1% in the 12 months to June 2022, the largest 12-month increase since November 1981. The Bureau of Labor Statistics (BLS) publishes it. The rate fell to 2.4% in January and February 2026, then rose to 4.2% in May 2026 as energy prices jumped. In August 2026, the latest month published, it was 3.4%. Core inflation, which leaves out food and energy, was 2.4%.

The Federal Reserve’s 2% inflation goal is measured with a different index, the PCE price index from the Bureau of Economic Analysis. The two indexes use different formulas: a Fisher formula for the PCE index and a modified Laspeyres formula for the CPI. They also weight items differently. And the PCE index counts spending made on households’ behalf as well as their own, including Medicare and Medicaid payments to health care providers.

In August 2026 both headline measures were up 3.4% on a year earlier. Core inflation was 3.0% on the PCE measure, against 2.4% on the CPI.

The rise began in goods. During the pandemic, spending moved from services to goods. By spring 2021, Americans’ spending on goods was about 10% above its pre-pandemic trend, and spending on services about 5% below it. So finds a 2025 Federal Reserve review of the episode by Ina Hajdini, Adam Shapiro, A. Lee Smith and Daniel Villar. Pandemic limits on supply still held, so inventories ran out, delivery times lengthened and goods prices rose.

Shortages hit inputs such as semiconductors, and markets for new and used vehicles were unusually distorted. Used car and truck prices rose 45.2% in the 12 months to June 2021 (BLS). The New York Fed’s Global Supply Chain Pressure Index, which combines supply chain indicators, peaked in December 2021, four standard deviations above its historical average.

Fiscal support added to the demand for goods. Comparing countries, Federal Reserve economists found that more generous fiscal support went with stronger spending on goods but not with more production (de Soyres, Santacreu and Young, 2022).

Energy prices gave a second push after Russian forces invaded Ukraine on 24 February 2022. Brent crude oil averaged $86.51 a barrel in January 2022 and $122.71 in June, according to the U.S. Energy Information Administration (EIA). U.S. regular gasoline averaged $5.006 a gallon in the week of 13 June 2022, its peak. The World Bank’s European natural gas price, from a Dutch trading hub, averaged $40.34 per million British thermal units in 2022, against $16.12 in 2021.

Within the consumer price index (CPI), the energy index rose 41.6% in the 12 months to June 2022. On BLS’s weights, that rise accounts for about 3 of the 9.1 percentage points. By December 2022 the energy index was 19% below its June level. Its rise over 12 months had slowed to 7.3%. Energy costs also pass through to other prices over time.

In 2026 another energy shock hit. Military action in the Middle East on 28 February stopped most shipping through the Strait of Hormuz (EIA). Brent crude averaged $117.29 a barrel in April, against $70.89 in February (EIA). The CPI’s energy index was 23.5% higher in May 2026 than a year earlier, and all-items inflation reached 4.2%. Core inflation stayed between 2.4% and 2.9% from January to August 2026.

Congress passed three large relief packages between March 2020 and March 2021. In a 2023 paper on the pandemic-era inflation, Ben Bernanke and Olivier Blanchard give their headline costs as follows.

A 2025 Federal Reserve study of the international experience, by Anna Lipińska, Enrique Martínez García and Felipe Schwartzman, draws on International Monetary Fund estimates. On those estimates, it puts U.S. support from January 2020 to September 2021 at about 25% of gross domestic product. That was “substantially larger than the fiscal support provided abroad”.

Estimates of what relief spending added to inflation differ about tenfold. The San Francisco Fed (Jordà, Liu, Nechio and Rivera-Reyes, 2022) put U.S. fiscal support’s effect at about 3 percentage points of core CPI inflation by the end of 2021. It added that this was “estimated with considerable uncertainty because the available sample is too short”.

A Federal Reserve Board note (de Soyres, Santacreu and Young, 2022) looked at the rise in U.S. inflation above its 2015–2019 average by the end of 2021. It linked domestic stimulus with about 2.5 points of that rise, and 1.8 points in the euro area. It called its results illustrative, “highlighting perhaps the higher end of potential price pressures”.

At the low end, Barnichon, Oliveira and Shapiro (2021) studied the American Rescue Plan alone, working through the labor market. They estimated it would raise core personal consumption expenditures (PCE) inflation by about 0.3 point a year through 2022. The Fed’s 2025 review finds “little apparent consensus” on what fiscal and monetary policy contributed.

Studies also disagree on how much of the rise came from supply and how much from demand. The Fed’s 2025 review finds “no consensus on the relative importance of each”. In the San Francisco Fed’s split of 12-month PCE inflation, the supply-driven part peaked at 3.4 percentage points in June 2022. The demand-driven part peaked at 2.9 points in October 2022.

Inflation rose across rich countries in the same period, which points to shocks that crossed borders. The Fed’s international study finds that the surge “was set in motion by two global events”, the pandemic and Russia’s invasion of Ukraine.

But the same study also finds that U.S. inflation rose earlier, “likely due to an earlier reopening and larger fiscal support to income”. Countries with bigger pandemic stimulus had bigger rises in core prices from 2019 to 2023, a correlation other studies say may not show the true effect.

Shelter made up 35.6% of the consumer price index (CPI) basket in December 2025, by BLS’s relative importance weights. Its rise over 12 months slowed from a peak of 8.2% in March 2023 to 3.0% in August 2026.

By contrast, in August 2026 energy was the fastest-rising group, up 16.3% on a year earlier, “due in large part to the index for gasoline rising 27.4 percent” (BLS).

Housing costs in the CPI move slowly. For homes their owners live in, the CPI measures owners’ equivalent rent, “the implicit rent that owner occupants would have to pay if they were renting their homes, without furnishings or utilities”. BLS prices it from a sample of rented homes and collects each home’s rent every six months, so the index follows the rents all tenants pay, not only new leases.

Rents for new tenants move first. Rent inflation for new tenants leads the official BLS rent inflation by four quarters. So finds a study of BLS rent records published by the Cleveland Fed and BLS (Adams, Loewenstein, Montag and Verbrugge, 2022). Market rent growth returned to pre-pandemic levels in 2022, but CPI rent inflation “remained stubbornly high for years”, as existing leases slowly caught up to market rents (Hajdini and colleagues).

There was little evidence of a wage-price spiral. In such a spiral, wages rise in response to prices and then push prices higher again. Pay fell behind prices: real average hourly earnings of private-sector workers fell from $11.40 in February 2021 to $10.91 in June 2022, in 1982–84 dollars (BLS). Bernanke and Blanchard find that tight labor markets were not the main driver of the 2021–2022 rise, though their effects on inflation last longer than those of price shocks.

Real earnings have stayed above their February 2020 level, $11.01, since March 2023, when inflation was falling. The Fed’s international study finds that the recovery in real wages, as nominal wages grew faster than prices, “likely contributed to the persistence of inflation”.

All nine economies in the table peaked between June 2022 and January 2023. The United States, Canada and New Zealand peaked first, in June 2022.

The European economies peaked later and higher. The Fed’s international study attributes this “mainly” to “the larger exposure of European economies to cuts in energy supply from Russia”. However, Germany’s national consumer price index peaked lower than its harmonised index, at 8.8% in October 2022, below the U.S. peak. Japan’s peak, 4.3% in January 2023, was the lowest of the nine.

The 2026 oil shock raised inflation abroad too. In the euro area, inflation rose from 1.9% in February 2026 to 3.2% in August. In August, energy prices there were 14.3% higher than a year earlier.

Inflation began rising in the spring of 2021, but the Fed held its policy rate, the federal funds rate, at 0–0.25% until March 2022. In April 2021 the committee that sets the rate put the rise down to “transitory factors”. It dropped the word “transitory” only in December 2021. A 2024 Federal Reserve note by Jane Ihrig and Chris Waller recounts both decisions. Bernanke and Blanchard write that policymakers in the United States and other advanced economies “were accordingly slow to react”.

From March 2022 to July 2023 the Fed raised the rate by 5.25 percentage points, to 5.25–5.5%. Ihrig and Waller call it “much faster tightening of policy than had been seen in more than 30 years”. The average 30-year fixed mortgage rate rose from 3.22% in the first week of January 2022 to a peak of 7.79% in the week of 26 October 2023 (Freddie Mac).

Partly, though how much is hard to measure. The Fed’s 2025 review finds that “Monetary policy tightening and a fading of fiscal support reduced demand, while a healing of supply chains, strong productivity growth, and an increase in the labor force helped expand supply.” It adds that monetary policy helped keep longer-term inflation expectations “generally well anchored”. By June 2023 the consumer price index’s rise over 12 months was down to 3.0%.

One estimate of the Fed’s share comes from Rami Najjar and Adam Shapiro of the San Francisco Fed (2025). They find that without the rate rises of 2022 and 2023, core personal consumption expenditures inflation “would have been approximately 3 percentage points higher on average between 2022 and 2024”. They add that “Easing labor shortages, dissipating supply chain bottlenecks, and the fading effects of fiscal stimulus also likely played crucial roles.”

Before the $1.9 trillion American Rescue Plan was signed in March 2021, Olivier Blanchard warned that it could overheat the economy. Working from the Congressional Budget Office’s estimate of potential growth, he put the output gap at about $900 billion at the end of 2020. The output gap is the shortfall of output below what the economy could produce.

If each dollar of relief spending added a dollar of demand, the American Rescue Plan and the December 2020 package together would add $2.8 trillion. That is “nearly 3 times an overly generous estimate of the output gap”, Blanchard wrote.

Bernanke and Blanchard report that Lawrence Summers also argued in 2021 that inflation could rise much more than conventional calculations predicted.

U.S. inflation also rose before inflation abroad did, likely due in part to larger fiscal support, and the Fed held its rate near zero until March 2022. Ihrig and Waller consider guidance like the Fed’s of December 2012. Under it, they write, the conditions for a first rise “would have been met in the spring of 2021”, about a year earlier.

Against this, estimates of what relief spending added run from about 0.3 to 3 points. The two highest, from the San Francisco Fed and a Federal Reserve Board note, carry their authors’ warnings. The United Kingdom, Sweden and the euro area gave less fiscal support yet peaked higher, mainly because of Russian energy cuts, the Fed’s international study finds. And in 2026 U.S. inflation rose again after an energy shock, while core consumer price index inflation stayed below 3%.

All nine economies in the international comparison peaked within seven months of each other, after the pandemic and Russia’s invasion of Ukraine. The Fed’s international study uses a model that relates core inflation to expected inflation, spare capacity and supply factors. It suggests the 2021–22 surge was “mainly driven by ‘cost push’ factors, such as supply shortages and relative price shifts”.

Against this, the same study warns that if its supply measures were partly driven by demand, its estimate of demand’s part “should be viewed as a lower bound”. Bernanke and Blanchard conclude that excessive demand “set off the inflation” but did so mainly through price shocks. In their account, excessive demand raised commodity prices, which the war in Ukraine also pushed up, and the demand for goods whose supply could not keep up.

Supply and demand together. No single cause fits the record of 2021 and 2022. Inflation rose in the U.S. and abroad in the same period, after the same global shocks, and European economies with less fiscal support peaked higher. Yet U.S. inflation also rose earlier, likely due in part to larger fiscal support, and estimates of what relief spending added run from about 0.3 to 3 percentage points. The San Francisco Fed’s split puts the peaks of supply-driven and demand-driven inflation at 3.4 and 2.9 points.

Rate rises helped bring inflation down, together with healing supply chains, though how much each did is hard to measure. In 2026 an energy shock raised inflation again while core consumer price index inflation stayed below 3%.

Download the conclusion

The verdict above as one image, with the two numbers that frame it. They are the peaks of the supply-driven and demand-driven parts of U.S. personal consumption expenditures (PCE) inflation in the San Francisco Fed’s split. The supply-driven part peaked at 3.4 percentage points in June 2022, the demand-driven part at 2.9 in October 2022. The image is 1,200 by 630 pixels, the size social networks use for link previews, and anyone may download it and post it.

Download the card (PNG, 1,200 × 630)