Why US inflation surged in 2021 and what the Fed should do to control it

Inflation surged to 6 percent over the 12 months through January 2022, far above the Federal Reserve's target of 2 percent.[1] The jump was caused by strong consumer demand and a number of supply disruptions. The Fed, which eased policy in the 2020 recession, is signaling a gradual tightening this year, and it stands ready to tighten more rapidly if needed. Over the next couple of years, inflation will move down, but probably not all the way to 2 percent.

A key question for 2023 is whether the Fed should try to weaken the economy to eliminate any lingering excess inflation. Abundant evidence over the past two decades shows that the 2 percent target keeps unemployment excessively high. The Fed should raise its target to 3 percent.

Pandemic inflation

Three forces pushed inflation to a 40-year high in 2021. First, a series of fiscal support packages enabled consumer spending to exceed its pre-pandemic trend.[2] Second, many workers were either afraid to return to work or unable to do so because school or work closures forced them to stay at home. Third, spending shifted from in-person services to goods, exceeding the production capacity of the goods sector.

With the fiscal deficit shrinking dramatically, spending should gradually cool. The end of the Omicron wave should make employees more willing and able to return to work, boosting the economy's production capacity. It should also spur a shift in demand away from goods back toward in-person services, which have unused capacity. A factor pushing in the other direction is continued strong demand for housing, which is likely to push up the housing component of inflation in the coming months. But the net effect of modest cooling in overall demand and rising supply will cause inflation to fall over the next two years.

In a PIIE Policy Brief, David Reifschneider and David Wilcox project that core personal consumption expenditures (PCE) inflation (which excludes the volatile food and energy categories) will be 3.1 to 3.4 percent over the four quarters of 2022 and 2.1 to 2.6 percent in 2023.[3] The February projections by IHS Markit, a leading private forecaster, are within these ranges, at 3.4 percent for 2022 and 2.1 percent for 2023. The Fed is likely to revise its projections to somewhere in these ranges at its March 16 meeting.

Although these projections of inflation are not unreasonable, the risks are almost entirely on the side of higher inflation. One reason to worry about a higher path of inflation is that inflation likely responds more to tight labor markets than is assumed in standard forecasting models, including the model of Reifschneider and Wilcox.

The Fed, Reifschneider and Wilcox, and most private forecasters model inflation using a linear Phillips curve. The Phillips curve—a relationship first noted by A.W. Phillips in 1958—shows that inflation in either wages or prices is negatively related to the level of unemployment. The original Phillips curve is highly nonlinear. In recent research, some colleagues and I find that Phillips curves in the United States and many other countries are highly nonlinear—like the original curve—whenever inflation is very low. The solid line in figure 1 displays the US Phillips curve I estimated with Christopher Collins for the core consumer price index (CPI) when inflation is low. When unemployment is less than the non-accelerating inflation rate of unemployment (NAIRU), the unemployment gap is negative and the curve is steep. When the unemployment gap is positive, the curve is flat.

The most likely cause of this nonlinearity is downward wage and price rigidity. When inflation is very low, some workers and firms are pressured to accept wage and price cuts, which they resist even at the expense of losing their jobs and sales. Extremely high levels of unemployment are required to push inflation below, or even close to, zero. Low inflation effectively bends the Phillips curve to make it flatter when unemployment is high.

Figure 1 also displays the Phillips curve estimated by Reifschneider and Wilcox for core PCE inflation.[4] The US inflation rate was very low and the unemployment gap was positive for most of the period over which their curve was estimated. Because these are the conditions under which unemployment has little impact on inflation, their linear Phillips curve is relatively flat. Using data that start in the 1950s and include periods of negative gaps as well as periods of higher inflation, Collins and I estimate a Phillips curve that is steep when the gap is negative or inflation is high.

The burst of US inflation in 2021 provides further evidence for a steep segment of the Phillips curve. The interaction of strong demand with weak supply is precisely the inflation mechanism the Phillips curve is designed to explain, provided that one can estimate the effect of the supply shocks on the NAIRU. Clearly, the COVID-19-induced labor supply shock increases the NAIRU. In addition, the shift in demand toward goods can be interpreted as an increase in the NAIRU.[5]

The Gagnon and Collins model is able to explain the increase in core CPI inflation in 2021 if the NAIRU is assumed to have risen about 6 percentage points, from around 4 percent to 10 percent in 2020 and early 2021, before gradually declining.[6] Figure 2 displays the model's projections of inflation starting in 2021Q1. Starting in 2022Q2, it includes two alternative projections, based on whether the NAIRU returns to its previous level of 4 percent or lasting scars on the labor market keep the NAIRU at 5 percent. (Unemployment is set equal to the February 2022 Consensus Forecasts projection.)

The Gagnon and Collins projection of inflation with the NAIRU returning to 4 percent is near the lower end of the Reifschneider and Wilcox projections if one adjusts for the tendency of CPI inflation to be a bit higher than PCE inflation.[7] However, if the NAIRU settles at 5 percent, inflation ends up above the top of their range for 2023. Measures of labor market slack such as job openings per unemployed worker and the quit rate point to a NAIRU well above the current unemployment rate of 4 percent, suggesting that it may remain above 4 percent indefinitely.

Even if the NAIRU does return to 4 percent, the Gagnon and Collins projection may be too optimistic, because it does not allow for any persistent adjustment in long-run expected inflation. Reifschneider and Wilcox show that measures of long-run expectations have been and remain remarkably stable. Yet, as Olivier Blanchard discusses in a PIIE blog, the inflation surge of 2021 is unlike anything seen in the past 40 years. This episode may prove salient for firms, workers, and households in a way that harkens back to the 1970s. Some upward adjustment of these expectations along the lines of the endogenous expectations simulations of Reifschneider and Wilcox seems likely. Such an adjustment would tend to shift the projection of 2023 inflation upward by about half a percentage point.[8]

Monetary policy

In response to the 2020 pandemic-induced recession, the Fed quickly adopted an ultra-loose policy stance, dropping short-term interest rates to zero and buying bonds to pull down long-term rates. In retrospect, it should have begun returning to a more neutral policy stance after the passage of the American Rescue Plan, in March 2021.[9] But neither the Fed nor most private forecasters began to predict an overheating economy until much later in the year.[10]

As the magnitude and persistence of the rise in inflation became apparent in late 2021, the Fed appropriately signaled a tightening of policy, beginning with a tapering of its bond purchases. These purchases will have ended by March 16, when the Fed is expected to announce the first increase in its short-term policy rates. The Fed should project a steady rise in policy rates to a neutral level, just over 2 percent, by January 2023. It should also announce that it will soon start to allow some of the bonds it purchased to run off its balance sheet as they mature. These runoffs should increase over the summer and reach a peak of $100 billion per month by the fall. That would be twice as fast as the reduction in bond holdings after the Fed's last bout of bond buying, and it would hasten a return to neutral conditions for longer-term interest rates.

Getting monetary policy to a neutral stance is not likely by itself to cause a recession. Indeed, with unemployment probably below the NAIRU and a large overhang of household savings from past fiscal packages, the Fed may need to raise rates a bit above neutral. Should the conflict in Ukraine lead to an upward surge in global energy prices, the Fed will face an unwelcome situation, with slower growth and inflation remaining elevated longer than otherwise. But the correct path of Fed policy is not much affected by energy shocks, as higher inflation that calls for higher interest rates is offset by lower growth that calls for lower interest rates.

Over the coming months, the Fed will have plenty of opportunities to adjust its policy path as new inflation and jobs data are posted. As always, it will have to be nimble, ready to stop tightening if the economy weakens significantly or to tighten further if inflation does not drop sharply.

If PCE inflation settles in much above 2 percent by the end of 2023, the big question will be whether the Fed needs to slow the economy further, risking a recession, to get all the way back to 2 percent.[11] As some economists have argued, the evidence of the past 25 years shows that an inflation rate of 2 percent is too low for many reasons, all of which lead to higher unemployment rates than necessary. It would be a mistake to cause or even risk a recession to get inflation down to a level that is too low.

Ben Bernanke argues that raising the target would damage a central bank's credibility, especially if it occurs when the central bank would have to tighten policy to achieve the original target. Choosing a suboptimal policy because it seems too hard to explain the benefits of the optimal policy is a cop-out. Central banks that are seen as always striving to deliver the best possible economic outcomes will have the greatest credibility in the long run. Moreover, no target or policy framework should be viewed as fixed forever. Rather, central banks should conduct regular reviews of their policy targets and frameworks in the light of developments in the economy and the understanding of how the economy operates. It is unfortunate that the Fed's first such review, in 2019–20, left a change in the inflation target off the table. The Fed should take this opportunity to raise its inflation target to 3 percent.

Notes

1. The Fed's preferred measure is the rate of change of the price index for personal consumption expenditures (PCE). Another widely reported measure is the rate of change of the consumer price index (CPI), which was 7.5 percent over the same period.

2. Nominal consumer expenditures in November 2021 were nearly 5 percent higher than a logarithmic trend estimated over the period from January 2010 through December 2019. Data are from the Bureau of Economic Analysis via https://fred.stlouisfed.org.

3. The Russian invasion of Ukraine has led to significantly higher energy prices. They will put upward pressure on headline inflation but will have only a small effect on core inflation, especially if global growth slows.

4. Phillips curves estimated on PCE inflation are typically flatter than those estimated on CPI inflation, for reasons that are not fully understood but probably reflect different weights on the component categories. The figure displays the short-run effect of unemployment on inflation. The long-run effect is roughly twice as high in each model, reflecting lag coefficients on inflation that sum to roughly 0.5.

5. Consider separate Phillips curves for goods and services. Starting from gaps of zero, a shift in demand toward goods and away from services puts the goods sector on the steep part of the curve, with a large increase in prices, and the services sector on the flat part of the curve, with only a small decline in prices. The net effect is an increase in average inflation with no change in average unemployment. Unemployment must increase to prevent higher inflation, which implies an increase in the NAIRU.

6. A similar exercise with the Reifschneider-Wilcox model would require an unrealistic increase in the NAIRU of 40 percentage points or more. The inability of the Reifschneider-Wilcox model to explain the recent surge of inflation is one reason Jason Furman worries about the reliability of that model's forecast. The performance of the Gagnon-Collins model provides some reassurance on that concern.

7. IHS Markit (February 2022) projects long-run core CPI inflation 0.3 percentage point higher than core PCE inflation.

8. Blanchard and Furman also raise the specter of a wage-price spiral. Reifschneider and Wilcox provide several reasons that such a spiral is unlikely, at least beyond that implied by an endogenous drift upward in long-run inflation expectations. Even during past periods of high US inflation, Phillips curve models similar to those discussed here worked well without any modifications for wage-price spirals.

9. The new operating principles the Fed adopted in 2020 in response to the problems raised by the zero lower bound on interest rates precluded a tightening in early 2021, given the small degree to which the Fed projected inflation to increase.

10. I and a few others warned of rising inflation in February and April. We based our warnings on expected strong demand from the fiscal measures. No one predicted that supply restrictions and their inflationary impact would be nearly as severe as they were.

11. Research shows that the neutral rate of interest fell notably over the past two decades. This decline is widely attributed to demographic and technological trends that are highly persistent. Central banks were slow to update their estimates of the neutral rate, leading to policy stances that were tighter than desired and that put downward pressure on inflation. If this trend continues, gradual downward pressure on inflation may reemerge even without intentional policy action.