The Federal Reserve operates under a dual mandate that requires it to target both price stability and maximum employment. But balancing the two objectives is hardly easy. The Fed has only a few policy tools, most importantly, changes in short-term interest rates and adjustments to the quantity and terms of the securities it owns. By contrast, inflation has infinitely many causes — not just monetary and financial crises, but also wars and pandemics, hurricanes and climate change, and more.

Across decades, the Fed has changed its inflation strategy several times. Here I offer a very brief review of 60 years of its successes and failures. And then I zero in on how the neglected problem of price “stickiness” derailed the most recent variant of the Fed’s strategy.

Sixty Years of Inflation in One Graph

Start with some historical context. We need to go back more than half a century to find a pattern of inflation close to that of 2022. From the late 1960s through the early 1980s, the U.S. economy was buffeted by three closely spaced waves of inflation, with price increases peaking at an annualized rate of 14.6 percent in April 1980. Those waves stand out clearly in the figure below, which charts the annual rates of change of the most familiar yardstick of inflation, the Consumer Price Index.

Before the 1960s, many economists were convinced there was a simple tradeoff, known as the Phillips Curve, between the rates of inflation and unemployment — the more of one, the less of the other. But that view was confounded by the experience of the next two decades. Far from falling as inflation rose, the jobless rate increased at successive inflation peaks. Something was off.

It was not long before two economists identified what was left out in the Phillips Curve model. Working separately in 1967 and 1968, Milton Friedman and Edmund Phelps pointed to the role of expectations. They hypothesized that each year of higher inflation caused people to expect even higher inflation in the next — a phenomenon that came to be known as “adaptive expectations.” The expectations became self-fulfilling, so that prices kept rising, at least temporarily, even after the Fed tightened policy and the unemployment rate edged up.

The implications for monetary policy seemed clear enough. To keep inflation under control, the Fed would need to keep inflation expectations firmly anchored. Attacking inflation sporadically and then letting it rise to a new peak was clearly counterproductive. But that thinking did not take hold right away. It would be more than a decade until Paul Volcker, appointed as chair of the Fed in 1979, would decisively bring inflation (and inflationary expectations) to heel, at the cost of nasty back-to-back recessions.

Even after the Fed achieved a degree of control over inflation, it did not immediately attempt to anchor inflation with an explicit inflation target. A history of inflation targeting published by the Federal Reserve Bank of San Francisco searched past minutes of the Fed’s policy-setting Federal Open Market Committee for mentions of inflation targeting. Among the earliest was a statement made in 1994 by Thomas Melzer, then head of the St. Louis Fed, who specifically identified greater policy credibility as a benefit of inflation targeting.

By the early 2000s, it was widely accepted that the Fed was following an implicit inflation target. However, it was not until January 2012 that it announced an official policy of targeting 2 percent inflation. The rate was calibrated to the pace of one of the widely used inflation measures, the Personal Consumption Expenditures index. (For reasons only nerds will want to read about, that measure of inflation runs about half a percentage point below the more familiar CPI.) At first, the Fed consistently undershot its target. For 2012 through 2020, PCE inflation averaged just 1.7 percent per year.

It seems only common sense to assume that if inflation is bad, then ultra-low inflation, or even deflation, would be welcome. However, economists do not agree. One reason is that when inflation is extremely low, the short-term interest rates that are the Fed’s main policy tool fall to — or close to — zero. Since it is all but impossible to push interest rates below zero, it then becomes difficult to apply monetary stimulus when needed.

In August 2020, in order to guard against the risk of undershooting the inflation target down the road, the Fed amended its 2012 version. The change made it clear that 2 percent was not a ceiling, but an average. According to the revised policy, “following periods when inflation has been running persistently below 2 percent, appropriate monetary policy will likely aim to achieve inflation moderately above 2 percent for some time.” The amended policy came to be known as Flexible Average Inflation Targeting (FAIT). But it was not long before FAIT was severely challenged. Not just challenged, but thrown off the rails.

Fait Not Accomplished

In the spring of 2021, the U.S. economy began to emerge from the Covid-19 pandemic. The recovery was turbocharged by massive government spending for pandemic relief and by pent-up spending by consumers who began to resume their normal habits after months of staying home. After little movement during the pandemic, prices began to rise. The figure on page 50 zooms in on the details, this time showing both 12- month inflation rates and one-month increases, expressed as annual rates.

In February 2021, the monthly inflation rate reached 4.6 percent. At first, Fed Chair Jerome Powell greeted the news calmly. In a March 17 news conference, he explained, “We could also see upward pressure on prices if spending rebounds quickly as the economy continues to reopen, particularly if supply bottlenecks limit how quickly production can respond in the near term.” But his bottom line was reassuring: “These one-time increases in prices are likely to have only transient effects on inflation.”

The decision to forebear tightening monetary policy was consistent with Flexible Average Inflation Targeting as long as inflation seemed transitory. And a dip in monthly rates in the late summer of 2021 did seem to justify Powell’s outlook. But when the annualized inflation rate jumped to 11 percent in October and rose sharply again in the spring, the FAIT rationalization for not tightening looked increasingly untenable. The term “moderately above 2 percent” had never been specifically defined, but clearly, the rates did not fit the definition.

Accordingly, the Fed shifted course. In March 2022, it began the most aggressive tightening since the Volcker era. The target for the federal funds rate — the rate at which banks borrow from each other to meet dayto- day regulatory requirements — which serves as a key indicator of policy tightness — had been slashed to a range of zero to 0.25 percent early in the pandemic. And it had remained near the bottom of that range.

By the end of the year, the top of the fed funds target range had been raised to 4.5 percent. Monthly inflation briefly dipped below zero in December, but then started to rise again. Yet although the Fed had tightened policy by another full percentage point by May 2023, inflation remained well above the Fed’s 2 percent target. What has made it so persistent?

Sticky Prices: The Missing Element

A closer look at changes in the price index during and after the pandemic reveals a missing element in the Fed’s analysis: too little attention to the problem of price “stickiness.” Two aspects of this stickiness are relevant. One is that prices in some sectors of the economy respond much more quickly than others to underlying changes in supply and demand. The other is that prices in almost all sectors of the economy are more flexible upward than downward.

The Federal Reserve Bank of Atlanta publishes a set of monthly indexes that focus on the speed of price adjustment in various sectors. Its index of flexible prices is dominated by goods like oil, wheat and cattle that trade on commodity exchanges, where prices can change from minute to minute. Prices of clothing, jewelry and cars also change, on average, more often than once a month. The prices of a few services, such as hotel room rates, also fall in the flexible group.

Most prices for services, however, are relatively sticky. Prices of medical care, personal care and car insurance change, on average, less often than once a month. Government services like sewer fees and transit fares, and private services like apartment rents and college tuition, may change only once a year or even less often. Certain goods that are closely linked to services, such as medicines and personal care products, enter into the index of sticky prices.

There are several reasons why some prices are stickier than others. Price changes create direct costs for some businesses — for example, printing new menus for a restaurant or adjusting coin mechanisms for a laundromat — and thus are to be avoided. In other cases, firms fear that the first seller to raise prices might lose market share to those who hold back. Still others fear alienating loyal customers. And, of course, prices locked in by longterm contracts often can change only when the contracts expire.

As noted earlier, the second important aspect of stickiness is that in all sectors, prices are stickier going down than up. Firms are often quick to raise prices when rising costs squeeze profit margins, but slower to cut them when costs fall. Same deal with labor. Workers readily accept increases in pay but resist cuts.

The figure at bottom left illustrates both kinds of stickiness. This chart, unlike the first two, shows price levels, not rates of inflation. Rates of inflation can be inferred from the slopes of the lines.

In early 2020, as the first wave of Covid-19 cases were recorded, the U.S. economy dropped into a short but steep recession. Flexible prices dipped sharply, then began to increase again by May. The Russian invasion of Ukraine in February 2022 led to one last burst of flexible-price inflation. After reaching a peak in June, flexible prices fell a bit. Over the past six months, however, they have flattened out and remain nearly 30 percent higher than before the pandemic, as we would expect of prices that are less flexible on the way down than up.

Meanwhile, service-dominated sticky prices behaved very differently. During the brief 2020 recession, the sticky-price index leveled off but did not fall. Later in the year, the sticky-price index began to rise steadily, but more slowly than flexible prices. Even after flexible prices peaked in mid-2022, stickyprice inflation has continued, although the rate slowed slightly in March.

The ongoing increase of sticky prices in the service sector is gradually bringing them back toward a normal relationship with goods prices. A combination of supply chain disruptions and the war in Ukraine created an extraordinarily wide gap between the two price indexes. Indeed, from the pandemic recession in early 2020 to the inflation peak in mid-2022, the increase in the flexible index was 24 percent greater than that of the sticky index. The last time anything remotely like that happened was during the Arab oil embargo of 1973-74. Even then, the difference between the change in the sticky and flexible indexes was only half as large.

In what sense can we speak of a “normal” relationship between flexible and sticky prices? What prevents each of them finding its own level? It turns out that some powerful forces limit their ability to move independently.

One is the fact that producers of stickyprice services use many flexible-price goods as inputs. Food (flexible) is a major input for restaurants (sticky). Motor fuel (flexible) is an input for transportation services (sticky), and so on. When costs for flexible-price inputs rise, service providers quickly feel a squeeze.

The labor market also links the goods and service sectors. Any substantial divergence of wages between the sectors would cause workers to start switching jobs. Employers have to keep wages competitive in order to find staff. Normally, wages of nonsupervisory workers in the goods-producing sectors are a bit higher than those in the service sector, as was the case going into the pandemic. Since then, despite the gap that developed between output prices, no similar gap emerged for wages. In fact, from 2019 to early 2023, the wages of service workers actually rose slightly faster than those of workers in goods production.

As pressure on profit margins grew in 2021 and 2022, firms in the service sector struggled to cope. Some deferred maintenance on equipment. Some cut staffing at the cost of reduced service quality. Some borrowed against assets to meet payrolls. The stress is gradually being relieved as service-sector prices recover some of the ground they lost. Until that process is complete, though, it will continue to be difficult for firms in the lagging sector to stay solvent, provide quality service and attract investors.

Thinking About Monetary Policy

The story about relative prices told here has important implications for monetary policy — and for flexible average inflation targeting in particular. In a broad sense, recent events support rather than undermine the thinking behind FAIT: the Fed is right not to treat its inflation target as a ceiling. Its decision to balance periods of below-target inflation with periods of more rapidly rising prices, announced in 2020, was the right way to handle moderate disturbances.

However, when the economy is hit by oncein- a-generation shocks, the degree of flexibility itself needs to be flexible. In particular, policymakers need to take the extent of relative price movements into account in determining both the length of the period over which above-target inflation is accommodated and the price level at which the target inflation rate should ultimately be re-established.

As of the spring of 2023, three years of severe shocks and extreme relative price movements confront policymakers at the Fed with difficult choices. As Fed chair Powell noted in a March press conference, inflation in the goods sector is under reasonable control and housing inflation seems to be on track to stabilize soon. The question is what to do about ongoing inflation in what the Fed calls the non-housing service sector and others call the “supercore.”

A fully flexible approach would let current trends run their course. That would mean leaving policy on an even keel until sticky service prices have returned to something close to their normal alignment with goods prices. The gap between the two, which reached its widest in June 2022, had already closed by nearly half as of April 2023. It seems plausible that even with no further policy tightening, it would continue to close over the remainder of 2023, and that the rate of service-price inflation would slow as it did so.

However, taking that approach would mean allowing overall inflation to run well above the Fed’s 2 percent target for some time yet. There would be no hope of returning to the pre-pandemic trend for the price level. Instead, the 2 percent inflation target would have to be recalibrated to a permanently higher trend line.

Even though that is beginning to look inevitable, the Fed seems reluctant to say so out loud. Instead, Powell talks of bringing nonhousing service inflation under control by means of a “softening of demand and perhaps some softening in labor market conditions.” To ensure such an outcome, he indicates a willingness to raise rates further, if necessary.

The problem is that tightening policy enough to bring overall inflation back to 2 percent over the remainder of 2023 would require freezing the relative prices of services in place before they have closed the gap with goods prices. There is a significant risk that doing so would require an outright recession, perhaps a severe one, rather than mere “softening.” That, in turn, could undermine the long-term health of providers of both public and private services.

* * *

On the whole, it seems to me that the balance of arguments favors moderation. The Fed should announce an intention to return to a policy something like FAIT in the future. However, it should concede that this can be done only after the economy has more fully adjusted to the extraordinary shocks that have occurred during the pandemic and post-pandemic periods. It will not be possible simply to average away the inflation that has occurred. If FAIT is to be reestablished, it will have to be from a higher benchmark. If that is done in an explicit and transparent way, one that acknowledges the need for flexibility in the face of shocks of once-in-a-generation proportions, it should not disrupt the well-anchored long-run inflation expectations that the Fed is rightly proud of having established.