The 2020 Fed Framework changes around full employment were appropriate and an important part of achieving the full employment recovery seen after the pandemic. Blaming it for historically unique demand- and supply-side inflationary shocks misses the mark.
Executive Summary: In 2020, the Fed adopted important changes to its Monetary Policy Framework intended to prevent the labor market from falling short of full employment again. Last year, Romer and Romer (2024) blamed these changes for the Fed’s slow response to inflation in 2021 and 2022, and proposed rolling back these changes. In this piece, we argue that the 2020 Framework changes around full employment were appropriate and an important part of achieving the full employment recovery seen after the pandemic. While there is room for the Fed to improve their forward guidance communications and forward-looking approach to Fed policy, the counterfactual difference of hiking 2-3 meetings earlier to inflation outcomes was likely low, and offset by the accelerated speed of the Fed’s tightening in 2022 and 2023. Blaming the Fed’s commitment to maximum employment for historically unique demand- and supply-side inflationary shocks, some of which the Fed simply could not foresee, misses the mark.
At the end of this Summer, the Federal Reserve will complete its periodic framework review. The framework sets the strategies, tools and communications the Fed will use to conduct monetary policy, and is pivotal for the path of monetary policy over the next few years.
The current framework, which was implemented in 2020, contained two notable changes to the language around full employment. The first was to describe the maximum level of employment as a “broad-based and inclusive goal.” The second was to direct the Committee to “mitigate shortfalls of employment from the Committee’s assessment of its maximum level” rather than deviations from full employment. This meant that the Fed did not see low unemployment as a risk in and of itself, and would only respond to a tight labor market if it believed it posed an inflationary risk.
As the Fed revises its Framework this Summer, these changes to the full employment aspect of its framework are sure to be scrutinized. Some blame the “broad-based and inclusive” language and the emphasis on shortfalls over deviations for delaying the Fed’s response to inflation in 2021. The Fed opted to keep the federal funds rates at the zero lower bound until March 2022, by which time the unemployment rate had fallen below 4% and 12-month core PCE inflation broke through 5%.
One paper taking this view is Romer and Romer (2024), presented at the Brookings Papers on Economic Activity conference last September. They examine speeches by Fed officials and FOMC publications to attempt to explain how the 2020 Framework affected the Fed’s decision to keep rates so low for so long.
Three specific critiques they have of the 2020 framework—and my reading of the evidence—are as follows:
Romer and Romer (2024) conclude that the FOMC, guided by the new framework, “aimed for a robust (or even hot) labor market that would increase job opportunities for historically disadvantaged groups.” The result, they claim, is a Fed that adopted an “overly optimistic interpretation of maximum employment.”
It’s true that Fed officials championed the widespread benefits of a strong labor market, especially for those in marginalized communities. The new framework clearly emphasized and elevated the importance of the full employment component of the dual mandate. But did it really lead to a much more optimistic interpretation of what constituted maximum employment?
It’s hard to see that excessive optimism in the long-run projections of the FOMC. In the September and December 2019 Summaries of Economic Projections (SEPs), the FOMC’s median long-run unemployment rate projections were 4.2% and 4.1%, respectively; in the 2021 SEPs, this fell by just 0.1pp to 4.0%. The FOMC’s assessments of the destination of maximum employment, as measured by the unemployment rate, were similar before and after the 2020 Framework.
Still, as Romer and Romer (2024) observe, FOMC members did see a rapidly expanding and (by some measures) very tight labor market in 2021. They interpret the Fed’s decision to hold rates near zero during a “hot labor market” as evidence that the Fed was “interpreting the maximum employment goal in a more aggressive way in the past.” For example, they cite a Sept. 27th, 2021 speech by John Williams, in which he cites a high level of job openings and hires:
“Clearly, demand for workers is very high—we see this in an elevated number of job postings and hires. At the same time, people are leaving their jobs in large numbers, either to look for new work or exit the labor force altogether. These conditions reflect the extraordinary nature of the pandemic, and also illustrate that we still have a long way to go until we achieve the Federal Reserve’s maximum employment goal.”
John Williams September 27th, 2021
Why did the FOMC tolerate such a hot labor market for so long, and even projected that the unemployment rate would fall below its estimate of the natural rate for some time? The FOMC understood that the unemployment rate alone did not constitute a full picture of the labor market in 2021. Although the unemployment rate was falling rapidly throughout 2021, FOMC members still had good reason to believe that the labor market had not reached maximum employment. In 2021, labor force participation and employment rates remained depressed relative to their pre-pandemic levels. For example, at the September 2021 FOMC meeting, the prime-age employment rate was a full two percentage points below its 2019 average.
In the very same speech as above, just after the section quoted by Romer and Romer (2024), Williams said [emphasis added]:
“First, even if job postings are at a record high, job postings are not jobs. These vacancies won't be filled instantly—it takes time for employers to find the right workers. And second, a full recovery means a recovery in employment, not just lower unemployment. Employment dynamics are driven by both the unemployment cycle and the labor force participation cycle. Because employed workers are more likely to remain attached to the labor force, these cycles are interrelated, as lower unemployment raises participation by reducing labor force exits. This relationship also means that the participation cycle lags behind the unemployment cycle, which is an important feature to keep in mind in assessing the state of the labor market.”
John Williams September 27th, 2021
The Fed saw a tight labor market, but they believed that labor supply would recover after COVID subsided, people were more comfortable returning to work and leaving childcare activities, and extended unemployment benefits expired. Reading through the transcripts of the FOMC press conferences and the minutes of the FOMC meetings, this view of the labor market—not a desire to target the unemployment rate of any particular demographic group—was behind the desire to run a “hot” labor market.
STEVE MATTHEWS: Thank you. Chair Powell, I wanted to ask about the inclusive monetary policy framework…. I’m wondering if a 6.1 percent unemployment rate for African Americans is consistent with full employment or whether it would need to be lower as part of your inclusive growth strategy.
CHAIR POWELL: Right, so the point of the broad and inclusive goal was not to target a particular unemployment rate for any particular group.
...
CHAIR POWELL: We have a—really, I’ll call it a unique situation where, by many measures, the labor market is tight. And 11 million job openings, very widespread reports from employers all over the economy saying it’s quite difficult to hire people, wages moving up, and that kind—so, quite a tight labor market. So our view, I think—widespread view a few months ago was that several things were coming together in the fall, including kids back to school, which would lighten caregiving duties, including the expiration of extra unemployment benefits, and other things would come together to provide an increase in labor supply… And so what happened was, Delta happened. And you had this very sharp spike in Delta cases. And I think, so that affects—for example, when schools are open 60 percent of the time or when they’re always at a threat of being closed because of Delta, the Delta variant, you know, you might want to wait. Rather than going ahead and taking a job and starting work only to have to quit it three weeks later, you’re going to wait until you’re confident of that. So some of that may not have happened.
September 21st, 2021 FOMC Press Conference. Edited for clarity.
The topic of whether or not labor force participation would rebound was clearly on the minds of the rest of the Committee members as well, as can be seen in the September 2021 FOMC meeting minutes:
Some participants noted that the increase in labor force participation that they had expected had not yet materialized in the wake of the reopening of schools and the expiration of the extended unemployment benefits, and that this likely reflected in part concerns about the resurgence of the virus, childcare challenges, and the uncertainties generated by ongoing disruptions to in-person schooling. Participants expected the labor market to continue to improve in coming months. Several participants indicated that a rise in the labor force participation rate might lag the improvements in other indicators such as the unemployment rate—a pattern consistent with past business cycle recoveries.
September 2021 FOMC Meeting Minutes
A better interpretation of the FOMC’s views during this period is that they recognized that while the speed of the labor market recovery was historically rapid, the ultimate destination of full employment was ultimately similar to their assessments of full employment prior to the 2020 framework review.
If the FOMC were guided by the unemployment rate alone, they might not have aimed for (and achieved) a full recovery in employment rates. That appeal to a broader set of labor market indicators beyond just the unemployment rate is the more salient impact of the addition of “broad-based and inclusive” to the Fed’s framework. The impetus to look beyond the headline unemployment rate was motivated by much more than the desire to see a fall in racial unemployment gaps. Take, for example, Powell’s response during the November 2021 FOMC press conference asking him to define full employment:
So maximum employment is, it’s a… broad-based and inclusive goal that’s not directly measurable and changes over time due to various factors. You can’t specify a specific goal. So it’s taking into account quite a broad range of things, and, of course, employment—levels of employment—[and] participation are a part of that. But, in addition, there are other measures of what’s going on in the labor market, like wages is a key measure of how tight the labor market is—the level of quits, the amount of job openings, the flows in and out of various states. So we look at so many different things, and you make an overall judgment.
Jerome Powell, November 3rd, 2021 FOMC Press Conference
In short, the importance of “broad-based and inclusive” was not that it distracted the Fed from its core mission by encouraging it to focus on reducing inequality and the unemployment rate of particular demographic or income groups. Rather, the Fed understood that full employment is a moving target that is not simply defined by the U-3 unemployment rate meeting some estimate of the natural rate of unemployment. The framework changes encouraged the Fed to correctly look beyond the headline unemployment rate and embrace a broader set of measures when judging the state of the labor market, most importantly recognizing that labor force participation and employment rates would lag the recovery in the unemployment rate.
Ultimately, the FOMC was vindicated in its “optimistic” view of maximum employment, as measured by employment rates. The prime-age employment rate broke through its prepandemic peak, and continued to climb until mid-2023. It has since settled into the range of 80.4 to 80.9%, above its 2019 average of 80.0% even as the unemployment rate has risen from its historic lows back to around 4.2%. There is broad consensus among the Committee that the current constellation of the data describes a labor market that is at or near full employment, and no longer an inflationary impulse to the economy.
There is a debate to be had about how rapid the labor market recovery should have been, but this is a separate question from where the ultimate destination of full employment is. To draw a metaphor, it may be the case that the Fed broke the speed limit en route to full employment, but it would be a mistake to then conclude that the Fed overshot the destination.
In September 2020, the FOMC provided explicit forward guidance about the conditions under which they would raise rates:
“The Committee decided to keep the target range for the federal funds rate at 0 to 1/4 percent and expects it will be appropriate to maintain this target range until labor market conditions have reached levels consistent with the Committee's assessments of maximum employment and inflation has risen to 2 percent and is on track to moderately exceed 2 percent for some time.”
September 2020 FOMC Statement
Romer and Romer (2024) argue that this forward guidance effectively tied the Fed’s hands during 2021, committing the FOMC to wait for observed developments in inflation and the labor market rather than forecasts. This commitment, they argue, delayed the Fed’s response to rising inflation as they waited to meet the “optimistic” goal of maximum employment.
Did the 2020 framework prevent the Fed from responding to an inflationary forecast? It seems reasonable to say that the strong forward guidance around the conditions for raising rates constrained Fed policy in 2021, but Powell’s comments at FOMC press conferences show that he believed the framework allowed the Fed to raise rates before reaching maximum employment if they believed that inflation was getting out of control:
…we actually have a paragraph in our framework, and something like this has been there for a long time. It’s, I think it’s paragraph six. And you’re talking about a situation in which the two goals are not complementary; they’re somehow in tension. And if we judge that’s effect—that is the case, what it says is, we take into account the employment shortfalls and inflation deviations and the potentially different time horizons over which employment and inflation are projected to return to levels judged consistent with the mandate.
Jerome Powell, September 2021 FOMC Press Conference
In situations in which the pursuit of the maximum-employment goal and [the pursuit of] the price-stability goal are not complementary, we have to take account of the distance from the goal and the—and the speed at which we’re approaching it. And so that is, in effect, an “off-ramp” which could, in concept, be taken and it’s in our framework, it’s been in our framework a long time. I’ve talked about it on a number of occasions. It is a provision that would enable us to, in this case, because of high inflation, move before achieving maximum employment.
Jerome Powell, November 2021 FOMC Press Conference. Edited for clarity
Why did they not exercise this option until March 2022, despite high inflation readings? It’s important to remember that although the Fed saw high inflation readings in 2021, the inflation outlook during most of 2021 was relatively benign. The Fed’s projections for core PCE inflation in 2022 were just 2.1% and 2.3% in the June and September 2021 SEPs, respectively. That sentiment is echoed in their meeting minutes:
Looking ahead, participants generally expected inflation to ease as the effect of these transitory factors dissipated, but several participants remarked that they anticipated that supply chain limitations and input shortages would put upward pressure on prices into next year. Several participants noted that, during the early months of the reopening, uncertainty remained too high to accurately assess how long inflation pressures will be sustained. Some participants commented that recent readings of inflation measures that exclude volatile components, such as trimmed mean measures, had been relatively stable at or just below 2 percent.
June 2021 FOMC meeting minutes
It’s easy with hindsight to fault the Fed for its naïveté in believing that inflation would be “transitory,” but the Fed wasn’t alone in thinking that inflation would be quick to return to near-2%. As late as the 2021Q4 Survey of Professional Forecasters (released on November 15th, in between the November and December FOMC meetings), the median forecaster projected four-quarter core PCE growth in 2022Q4 at just 2.3%. It also seems unreasonable to expect the Fed to have forecasted some of the reasons behind inflation in 2022, such as the Russian invasion of Ukraine, which drove up prices in energy, food, and core services that rely on food and energy prices, like food services and airfares.
| Realized (Annualized) | Median SEP Projection | Median SPF Forecast | ||||||
|---|---|---|---|---|---|---|---|---|
| FOMC Meeting | 12-month | 3-month | 2021 | 2022 | 2023 | 2021 | 2022 | 2023 |
| Jun. 2021 | 3.1% | 5.0% | 2.3% | 2.0% | 2.1% | 2.3% | 2.0% | 2.1% |
| Sept. 2021 | 3.6% | 5.7% | 3.7% | 2.3% | 2.2% | 3.7% | 2.2% | 2.1% |
| Dec. 2021 | 4.1% | 4.0% | 4.4% | 2.7% | 2.3% | 4.1% | 2.3% | 2.1% |
| Mar. 2022 | 5.2% | 6.2% | N/A | 4.1% | 2.6% | N/A | 3.1% | 2.2% |
Arguably by the December 2021 FOMC meeting the inflation outlook had appreciably worsened (with the Committee projecting 2.7% core PCE growth in 2022, elevated but still far below the 5.2% that actually happened), but it’s also unlikely that raising rates in December 2021 instead of March 2022 would have done much to stem inflation. The difference of tightening just two meetings earlier on inflation is of dubious relevance; Reifschneider (2024) finds that an aggressive mid-2021 pilot would have only modestly reduced inflation at the cost of a significantly higher unemployment rate. A reasonable counterfactual must also account for (1) the global nature of the inflationary shock, (2) the more rapid pace of tightening the Fed might not have otherwise pursued if they hiked earlier, and (3) the scale of tightening they cumulatively engaged in as a result of hiking at a faster pace in 2022 and 2023.
The Framework did not tie the Fed’s hands in responding to an inflationary forecast. The problem was their forecast itself. The Fed, along with most forecasters and market-implied measures, was already offsides on the inflation risks brewing in late 2021. Faced with new shocks from the Russian invasion of Ukraine, they further underestimated the extent and duration of the inflationary impulses in the years that followed the pandemic.
One of the changes to the 2020 framework that has attracted much criticism is the move away from addressing “deviations” to addressing “shortfalls” of employment from its maximum level. After seeing the unemployment rate persistently fall below contemporary estimates of the natural rate of unemployment during the 2010s at the same time that inflation remained low, the Fed recognized that prematurely choking off a labor market expansion ran the risk of leaving labor market gains on the table.
There were limits to how permissive the Fed would be with the labor market. If the labor market threatened the inflation side of their mandate, they were prepared to respond. See, for example, Powell’s explanation of the change to “shortfalls” when he introduced the new Framework in 2020:
“The change to "shortfalls" clarifies that, going forward, employment can run at or above real-time estimates of its maximum level without causing concern, unless accompanied by signs of unwanted increases in inflation or the emergence of other risks that could impede the attainment of our goals.”
Jerome Powell, August 27th, 2020 speech
Romer and Romer (2024) do not blame the “shortfalls” addition to the 2020 Framework for the post-COVID inflationary episode specifically, given their view that the Fed’s estimate of maximum employment was overoptimistic to begin with. However, they and others propose reversing this change to the framework, arguing that surpassing maximum employment is likely to lead to poor macroeconomic outcomes.
For some cursory evidence, Romer and Romer (2024) compare Fed staff forecasts of the unemployment rate and estimates of the natural rate of unemployment from the Greenbook/Tealbook. They then find periods where the Fed staff’s forecasts of the unemployment rate fell below their estimates of the natural rate by more than 0.5pp, and note any adverse macroeconomic outcomes that followed.
As Romer and Romer (2024) say, this is a “small” amount of evidence. A full dissection of each episode is beyond the scope of both Romer and Romer (2024) and this response, but a list of poor macroeconomic outcomes that happened to follow periods where the Fed forecasted that unemployment rate would fall under its estimate of the natural rate is far from conclusive. Certainly there was more going on than the Fed’s support for low unemployment.
The connection between the labor market and inflation in some of these episodes is tenuous at best. Was the 1988 inflation the calamitous event that justified higher unemployment? What exactly is the labor market explanation of the dot-com bubble and the telecom investment boom-bust? Would it be better to have foregone the historic welfare and productivity gains of the 1996-2000 period? Would higher unemployment in 2017 and 2018 be an acceptable cost for insuring against an “undesirable macroeconomic outcome” that failed to materialize? There simply isn’t a consistent causal mechanism, just a retroactive assignment of these events to a strong labor market based on temporal proximity.
Turning to the current episode, it’s true that a period of high inflation did coincide and follow the rapid labor market recovery of 2021. And the labor market did play a role: cyclical segments of inflation were readily driven by the rapid scale of employment and wage gains tied to the reopening and recovery of the economy from the pandemic. But if a low unemployment rate itself was the primary cause, why then did disinflation begin in 2022 and 2023, even as the unemployment rate continued to fall below 4% and even remained below 3.6% until August 2023? If breaching some level of maximum employment was the primary reason for inflation, why did inflation begin falling even as prime-age employment rates continued to climb and reach record highs?
There was clearly more going on than a tight labor market, such as global supply chains disrupted by COVID closures, commodity price shocks, and the Russian invasion of Ukraine. As we have written elsewhere, the global dimension of the post-pandemic inflation points towards global supply pressures, pandemic-related demand and supply dynamics, and geopolitical shocks explaining the bulk of the excess inflation seen in the early 2020s. Bernanke and Blanchard (2025) only find a small role for the labor market in driving excess inflation in 2021 and 2022.
The presence of supply disruptions does not mean that the labor market is totally unconnected from inflation. As Skanda Amarnath argued in 2021, some of the inflation was due to the rapid addition of jobs (and therefore labor income), pushing up demand for supply-constrained goods and services such as housing.
Practically speaking, arguing that the Fed should not allow unemployment to fall below its estimates of the natural rate presumes some unwarranted confidence in the Fed’s estimate of the natural rate of unemployment. As former Chicago Fed President Charles Evans wrote in defense of the 2020 framework,
…focusing on the quadratic loss in unemployment around its sustainable rate invites too much arbitrariness in deciding at what unemployment rate inflation pressures begin. The uncertainty is high around when u - u* turns inflationary.
Charles Evans (2024)
It’s important to keep in mind that the Fed has a track record of overestimating the sustainable level of unemployment, especially in recent decades. It’s precisely because of that poor track record that the language around shortfalls was added to the 2020 framework.
Athe Fed’s estimates of the natural rate kept falling as well. Romer and Romer (2024) write that if the 2020 Framework’s changes to the employment goal were reversed, “it would be essential to continually and carefully reassess the evidence about maximum employment.” But in both the 1990s and 2010s, much of the reassessment around what was attainable under maximum employment came about precisely because the Fed allowed the labor market to surpass its own estimates of maximum employment.
“As the unemployment rate moved lower and inflation remained muted, estimates of u-star were revised down.”
Jerome Powell, August 27th, 2020 speech
If there is a key Fed failure to identify from the 1990s, it was not in allowing the unemployment to fall below its natural rate in 1996, but in the tightening actions of 1999 and 2000. Low unemployment and firm wage gains were still coinciding with stable prices during this period, and especially so amidst robust productivity growth. The Fed’s anxiety about low unemployment and high wage gains posing inflation risk is nevertheless evident in the Fed transcripts of this period. It’s why the Fed chose to raise the target for the Fed Funds Rate from 4.75% in June 1999 to 6.50% in May 2000, and kept interest rates at that elevated level until the economy already was headed for recession in January 2001.
The ensuing recession may still have materialized, but the Fed’s actions exacerbated the investment and stock market bust and limited the prospect for a more expansionary rebalancing of aggregate investment away from technology & telecommunications. The costs to the Fed’s worries about an inflationary labor market ultimately led to dynamics that were effectively irreversible. Even after nearly 500 basis points of interest rate cuts in 2002, and bringing their interest rate to a historically low 1% in 2003, the unemployment rate kept rising swiftly. The recovery from this downturn proved sluggish. For key labor market metrics like the prime-age employment rate, the heights achieved in the late 1990s were never re-established.
In response to the pandemic, the Fed aggressively pursued a return to full employment. The ensuing labor market recovery—the fastest in American history—is partly thanks to those actions. Unlike in previous recoveries, the Fed did not underestimate the extent to which the labor market could rebound. Their 2020 Framework played an important role in the Fed’s decision to pursue a full labor market recovery to its pre-pandemic peaks. In this regard, the Fed was ultimately proven right.
This doesn’t mean that the Fed was perfect throughout this period. They, like many others, were blindsided by global inflation shocks that persisted longer than they expected. This inflation was not primarily due to the labor market, but the labor market did play some role and disentangling the causes of inflation proved to be immensely challenging.
Some critics put the blame for those developments on the 2020 Framework’s heightened emphasis on maximum employment, and argue that those changes should be rolled back. This would do little to address the underlying issue: the Fed’s strategy remains deficient when dealing with sharp “one-time” adjustments that nevertheless occur over multiple quarters and years and involve both supply- and demand-side interactions.
Instead of rolling back progress on their full employment strategy, the Fed should modify their framework to complement their inflation measures with nominal consumer spending and labor income, which provide a better real-time measure of aggregate demand and strip out one-time and supply-side effects that distort the inflation data. These measures were signaling to the Fed in 2022 and 2023 that there was still space to tighten policy to slow the economy without breaking the labor market. If the Fed wants to achieve more soft landings, as they functionally did in 2024, they would be wise to formally incorporate these measures into their framework and communications strategy.
