Macro Roundup

“…This analysis focuses on the evolution of household balance sheets over the course of the COVID-19 pandemic. Balance sheets have been buffeted by many factors, including weakness in labor-market income, strength in federal benefits, increases in asset prices, increases in household debt, and pandemic-constrained consumer spending. In aggregate, the result has been that households’ financial positions in 2022 are significantly improved relative to 2019. Savings held in deposits and other financial accounts have increased, real estate and stock market wealth has increased, and borrower distress has decreased. That improvement in financial strength will be a source of support to the aggregate economic recovery in coming quarters. The overall increase in after-tax and transfer income and the decrease in consumer spending have resulted in an extraordinary run-up in aggregate household savings. Although labor-market income was quite weak early in the pandemic, federal benefits more than compensated overall. In part as a result of relatively strong overall income growth,we estimate that households accumulated $2.5 trillion in excess savings (inflation-adjusted to 2020 dollars) between March 2020 and January 2022, much of which appears to have been deposited in checking and savings accounts.The difference is notable relative to the Great Recession, when income remained below its pre-recession trend for several years..”

Mitchell Barnes, Wendy Edelberg, Sara Estep, and Moriah Macklin, "Bolstered Balance Sheets: Assessing Household Finances since 2019," March 2022, The Hamilton Project, https://www.hamiltonproject.org/assets/files/ES_20220322_householdfinances_finalreport.pdf

“…Figure 11 shows various estimates of the distribution of excess savings that incorporate recent data on spending, income, and wealth. These estimates differ primarily by whether these models assume that the distribution of excess savings matches the pre-pandemic distribution and composition of savings, as the traditional Distributional Financial Accounts (DFA), released by the Federal Reserve, does. The traditional DFA based on historical patterns would suggest the share of excess savings held by households in the lowest two income quintiles to be just 4 percent. Estimates from Moody’s and Morgan Stanley, which more closely follow historical savings distributions, produce results that are similar to the traditional DFA, where roughly two-thirds of excess savings are held by the top income quintile….”

“…The JPMorgan Chase Institute documents flows into and out of 7.5 million active JPMorgan Chase checking accounts, aggregating the accounts across four income quartiles grouped by 2019 income. Figure 10 shows the medianchecking account balances for those quartiles grouped by 2019 income. The distribution of those changes is informative, with the caveat that the families holding those accounts are not representative of the U.S. population; those holding bank accounts and, moreover, those holding bank accounts at JPMorgan Chase over-represent financially secure households relative to the U.S. population as a whole. While outflows from accounts initially fell in early 2020 for each income quartile, that decline was largest and most persistent for those with higher incomes. As a result, checking account balances grew for higher-income account holders even though they experienced temporary declines in inflows. In contrast, spending rebounded for lower-income families and then rose above pre-pandemic levels. Those increases in outflows were more than offset by a surge in DPIs relative to 2019 (Greig, Deadman, and Noel 2021…”

“…Significant changes in labor-market income, government transfers, and the ability to consume goods and services have resulted in remarkably strong savings relative to prepandemic trends. Excess savings during the COVID period is defined as the difference between two measures. The first measure is the difference between actual disposable personal income (DPI) and the level of DPI if it had increased in 2020 and 2021 at the same average pace as over the 2018-19 period (“disposable income contribution” on figure 1b). The second measure is the difference between actual personal outlays and the level of outlays if they had continued at their pre-pandemic pace (“personal outlay contribution” on figure 1b). Because DPI was higher than trend in 2020 and 2021 and outlays were lower than trend (figure 1a), the result is excess savings out of income. From March 2020 through January 2022, elevated incomes contributed roughly $1.3 trillion while spending shortfalls contributed roughly $1.2 trillion to an estimated $2.5 trillion in excess savings held by households in inflation-adjusted 2020 dollars…”

“…the 10 fastest growing counties last year accounted for nearly 80 percent of the national total, a testament not so much to the rapid pace of change in these places, but to the lack of significant growth in the rest of the nation. The bureau had previously called 2021 the slowest population growth year on record, with the nation growing by just 0.1 percent.....New York, Los Angeles, Chicago and San Francisco lost a total of over 700,000 people from July 2020 to July 2021, according to the Census Bureau. Meanwhile, Phoenix, Houston, Dallas, Austin and Atlanta gained more than a total of 300,000 residents....Despite the latest figures, many major cities are optimistic about growth. New York City lost more than 300,000 residents through June 2021, the census shows, which city planning officials said was consistent with their own analysis. But, they said, the city’s sharp population decline through mid-2021 resulted largely from temporary patterns earlier in the pandemic, including an uptick in residents fleeing to the suburbs and exurbs, fewer immigrants, deaths from Covid and lower birthrates. They said these patterns had most likely lessened or reversed during the second half of 2021, which is not captured in the data. “The city is on its way back,” said Arun Peter Lobo, the city’s chief demographer. “The initial indicators, at least for us, are in the right direction.”

Robert Gebeloff, Dana Goldstein and Winnie Hu, "Cities Lost Population in 2021, Leading to the Slowest Year of Growth in U.S. History,"New York Times, March 24, 2022, https://www.nytimes.com/2022/03/24/us/census-2021-population-growth.html

Cities Lost Population in 2021, Leading to the Slowest Year of Growth in U.S. History

Substantial population loss in some of the nation’s largest and most vibrant cities was the primary reason 2021 was the slowest year of population growth in U.S. history, new Census data shows.

Although some of the fastest growing regions in the country continued to boom, the gains were nearly erased by stark losses last year in counties that encompass the New York, Los Angeles and San Francisco metropolitan areas.

The pandemic played a role, as the number of people dying rose substantially and many Americans left cities for smaller places. But experts say that skyrocketing housing costs were also to blame, and that some of the changes are a continuation of fundamental shifts in American demographics that began before the pandemic, such as the steadily falling birthrate and steep drop in immigration.

New York, Los Angeles, Chicago and San Francisco lost a total of over 700,000 people from July 2020 to July 2021, according to the Census Bureau. Meanwhile, Phoenix, Houston, Dallas, Austin and Atlanta gained more than a total of 300,000 residents. And there was also substantial growth in some rural areas and smaller cities like Boise, Idaho, and Myrtle Beach, S.C.

Population loss, particularly of working-age adults and their children, can separate extended families and lead to funding cuts and labor shortages in schools, health care facilities and other services that are essential to the residents who remain.

The pattern is a notable contrast from a decade ago, when large cities were growing, bolstered by a decades-long boom in immigration and the rising popularity of urban living. At that time, most of the counties losing population were rural or experiencing economic decline.

In the years immediately preceding the pandemic, those factors began to shift. Immigration slowed, urban housing costs rose, and suburban and exurban growth began picking up steam, trends that continued through the pandemic.

The virus wrought other changes. Because Covid-19 caused so many deaths, only 828 counties had more births than deaths in 2021, the figures show, down from more than 1,900 a decade ago.

And the rise of remote work made it less of a requirement for many workers to live in expensive cities to take advantage of high-paying jobs.

The decline in fertility started a decade ago during the Great Recession, and reflects the ways in which women and men of the Millennial generation are prioritizing education and work, delaying marriage and parenthood, and struggling to gain their economic footing as they deal with student debt, slow wage growth and steep housing costs.

Alison Grady and Ernest Brown, both 31, moved to Atlanta from Oakland, Calif., in March 2021 after nearly five years of living with roommates to save money on rent. Most recently, they had paid $1,500 a month for one room in a three-bedroom, 1,200-square-foot apartment that they shared with two friends.

The couple, who plan to marry and start a family, had wanted to buy a home in Oakland, but they found that houses in move-in condition started at $700,000, which was more than they could comfortably afford.

As they worked from home during the pandemic, and the virus and wildfires curtailed their social life, they spent more time in their cramped apartment and questioned whether it was worth it. They eventually decided to return to Atlanta, where they had met in 2014. They were able to buy a 1,000-square-foot, two-bedroom apartment for $315,000. And they no longer have roommates.

“The price of living in the Bay Area was so out of sync with the quality of life we were getting,” said Ms. Grady, a public health strategist.

Cities like New York and Los Angeles remain alluring, as evidenced by bidding wars for homes and big companies investing in office space. But many residents have been pushed away by a lack of affordable housing, as well as by pandemic-related changes to how they order their priorities.

Across the nation, counties with more modest housing costs gained in population, the data showed.

Five years ago, counties ranking above the 90th percentile for housing stress — a measure of housing costs as a share of income — accounted for a third of the nation’s population growth, suggesting that the high prices represented high demand.

But in 2021, those counties were net population losers, suggesting that costs have gotten out of hand.

In the Los Angeles area, for example, nearly half a million renters do not have access to an affordable unit, according to a 2021 report from the California Housing Partnership. For this group — including essential workers such as home health aides, janitors, child care professionals and medical assistants — homeownership can be all but impossible.

Regions where state and local governments do not make it easier to build affordable housing will face a troubled future, according to Dowell Myers, a demography and urban planning expert at the University of Southern California. There will be too few working-age adults to support a growing population of aging baby boomers. Jobs in nursing, utilities and other fields will go unfilled.

This demographic risk “is just deadly,” Professor Myers said. “We’re not building enough housing to keep our own kids.”

In Houston, like other growing Southern metro areas, the influx of relatively high earners has contributed to a real estate crunch that could erode the city’s reputation for affordability.

Kayli Thompson, 34, and her daughter, Analiese, 13, arrived in Houston last year from Ithaca, N.Y., where Ms. Thompson’s hours as a librarian were cut in half during the pandemic.

She was drawn to Houston by its stronger job market and warm climate. Still, relocating was a struggle. The rent money she set aside from savings, stimulus payments and a tax credit did not stretch as far as she had hoped. Her two-bedroom apartment costs $1,500 per month, nearly double her rent in Ithaca. She did not find a job right away, and received an eviction notice; in February, her church paid her rent.

Last month, she began working for a community newspaper, and she now feels the risk she took has paid off. She is delighted to be living downtown in a much more walkable neighborhood. When Ms. Thompson’s car broke down, it wasn’t as much the crisis it would have been in Ithaca.

“We’re happy,” Ms. Thompson said.

Gina Vargas, a 38-year-old consultant, joined the exodus from New York City in the spring of 2020. After 18 years in a city she thought she would never leave, she relocated to the Houston area, where she grew up, drawn mostly by the need to assist her aging parents.

The Houston she returned to was far different from the conservative place she remembered from her childhood, Ms. Vargas said. When her parents, an interracial couple, first moved decades ago to the working-class city of Pasadena, southeast of Houston, there had been a Ku Klux Klan meeting place near their home. Now, as a mixed-race Mexican American, she has felt welcomed into a cosmopolitan community, she said — though she noted there were more Trump flags.

And her New York salary has stretched a long way. Ms. Vargas was renting a one-bedroom apartment in Sunnyside, Queens, but she was able to purchase a three-bedroom house in a planned community in Cypress, 30 miles north of downtown Houston. She also bought a house for her parents, moving them to a more upscale neighborhood.

“The city is on its way back,” said Arun Peter Lobo, the city’s chief demographer. “The initial indicators, at least for us, are in the right direction.”The population estimates are the primary way the Census Bureau updates the population counts from the regular decennial census, which attempts to count every person living in every household and institutional setting. These estimates are largely based on administrative records, such as birth and death certificates and tax returns.

weeks of benefits claimed across all programs — traditional state-funded unemployment insurance (UI), the permanent law federal/state Extended Benefits (EB) program, and federally funded Pandemic Unemployment Assistance and other temporary federal benefit expansions — soared to a record 33 million in June 2020. That’s over two and a half times the pre-pandemic record of 12 million claims set in early 2010 in the wake of the Great Recession....That record level of claims significantly exceeded the number of unemployed individuals reported by the Department of Labor, which peaked at 23 million in April 2020 and by June 2020 had fallen to 17.6 million.... Overall, in the 18 months between when the pandemic struck in March 2020 and temporary federal unemployment programs expired on Labor Day 2021, a total of nearly 1.6 billion weeks of all unemployment benefit checks were paid out. That’s the equivalent of three full months of benefits for every one of the 121 million households in the US....State and federal unemployment benefits paid during the pandemic totaled around $900 billion, with over $700 billion in federal costs simply added to the deficit. Prior to the pandemic, the US in 2019 spent $27 billion on unemployment benefits, almost exclusively on state UI checks, which are supported by employer payroll taxes.

Matt Weidinger, "Visualizing the 1.6 billion weeks of unemployment benefit checks paid during the pandemic,"American Enterprise Institute, March 23, 2022, https://www.aei.org/poverty-studies/visualizing-the-1-6-billion-weeks-of-unemployment-benefit-checks-paid-during-the-pandemic/

Visualizing the 1.6 billion weeks of unemployment benefit checks paid during the pandemic

As I reviewed in recent congressional testimony, the pandemic and government’s response to it unleashed unprecedented demand for unemployment benefits. Major federal benefit expansions included $600-per-week (and later $300-per-week) supplements to regular unemployment benefits and the provision of weekly checks to groups never before eligible for unemployment benefits. Those and other expansions contributed to record benefit payouts, as depicted in the chart below:

As the chart shows, weeks of benefits claimed across all programs — traditional state-funded unemployment insurance (UI), the permanent law federal/state Extended Benefits (EB) program, and federally funded Pandemic Unemployment Assistance and other temporary federal benefit expansions — soared to a record 33 million in June 2020. That’s over two and a half times the pre-pandemic record of 12 million claims set in early 2010 in the wake of the Great Recession.

Overall, in the 18 months between when the pandemic struck in March 2020 and temporary federal unemployment programs expired on Labor Day 2021, a total of nearly 1.6 billion weeks of all unemployment benefit checks were paid out. That’s the equivalent of three full months of benefits for every one of the 121 million households in the US.

State and federal unemployment benefits paid during the pandemic totaled around $900 billion, with over $700 billion in federal costs simply added to the deficit. Prior to the pandemic, the US in 2019 spent $27 billion on unemployment benefits, almost exclusively on state UI checks, which are supported by employer payroll taxes.

The Department of Labor Inspector General testified last week that improper payments of unemployment benefits during the pandemic totaled at least $163 billion. Other experts have pegged misspending at as high as $400 billion.

“…To what extent does an incremental increase in labour productivity growth translate into increased compensation growth?. We look at both the US and Canada, over the period 1961-2019 (updating the analysis for the US in Stansbury and Summers 2018). We regress labour productivity growth on real compensation growth - measured at the average, at the median, and for measures of the compensation of ‘typical’ workers in both the US and Canada - with the possible results falling on a spectrum from complete delinkage (a one percentage point higher productivity growth rate is not associated with any change in the compensation growth rate) to complete linkage (a one percentage point higher productivity growth rate is associated with a one percentage point higher compensation growth rate).To focus on the impact of productivity growth over horizons longer than the business cycle, we use three-year moving averages and control for unemployment….We find strong evidence of linkage in the US. In years with a one percentage point higher productivity growth rate, there tends to be close to one percentage point faster growth in compensation - whether for average compensation, median compensation, or for typical workers (which we proxy using average compensation for production/nonsupervisory workers)…For Canada, the evidence on the linkage between productivity and pay is more mixed.For average workers, we find moderate linkage - over 1961 to 2019, a one percentage point higher productivity growth rate is associated with about half a percentage point faster rate of compensation growth. Over 1961-2019 we also find evidence of very strong linkage for typical workers, whose compensation we proxy using a newly constructed measure of the compensation of hourly-paid workers in five industries.2 However, for more recent periods, estimates for both average and typical compensation are too noisy to rule out strong delinkage (no translation of productivity increases to pay growth). We find no evidence of linkage for median compensation, but caution that our measure of median hourly compensation likely incorporates significant noise, which may substantially attenuate results estimated from year-to-year changes. While the picture is clearer for the US than for Canada, these results overall suggest that in both countries, marginal increases in productivity growth raise compensation on average, including for typical workers (holding all else equal). This is despite the divergence between productivity growth and compensation growth in the US and Canada over the last five decades….”

Jacob Greenspon, Anna Stansbury, Lawrence Summers, "Productivity and pay: A comparison of the US and Canada,"Center For Economic and Policy Research, February 15, 2022, https://voxeu.org/article/productivity-and-pay-comparison-us-and-canada

Productivity and pay: A comparison of the US and Canada

Real pay of typical workers has grown much more slowly than productivity over recent decades in several developed economies. This growing gap has been well-documented in the US (Bivens and Mishel 2015, 2021), Canada (Williams 2021), the UK (Teichgräber and Van Reenen 2021), and elsewhere. This phenomenon - often referred to as a ‘divergence’ or ‘decoupling’ of the pay of typical workers from productivity growth - is all the more striking because it has not always been this way. Bivens and Mishel (2015), for example, document that in the US, labour productivity and the average pay of typical workers grew at the same rate from the late 1940s until the mid 1970s, when the decoupling began.

The fact that productivity has grown so much faster than the pay of typical workers raises the question: Does productivity growth actually benefit typical workers by raising their pay? Or do the fruits of any incremental increase in productivity growth flow only to those higher in the income distribution?

Understanding the strength of this productivity-pay relationship is of considerable policy significance. Given the recent productivity slowdown (Goldin et al. 2021), economic policymakers have focused on measures to accelerate productivity growth such as promoting investment, technological progress, or better functioning markets. If productivity-pay divergence has been accompanied by ‘delinkage’ - whereby year-on-year increases in the rate of productivity growth are not associated with pay growth for typical workers - then success or failure of these efforts to boost productivity will have little impact on middle-class wellbeing. In this case, only measures targeting the income distribution will be effective at raising middle-class incomes. In contrast, if despite divergence between productivity and middle-class pay there is still linkage between the two, then an incremental increase in productivity growth should act to increase the pay of typical workers - even if other variables (like the declining bargaining power of workers, or the increase in globalisation) are at the same time acting in the opposite direction.

In our recent paper (Greenspon et al. 2021), we ask the question: To what extent does an incremental increase in labour productivity growth translate into increased compensation growth?. We look at both the US and Canada, over the period 1961-2019 (updating the analysis for the US in Stansbury and Summers 2018). We regress labour productivity growth on real compensation growth - measured at the average, at the median, and for measures of the compensation of ‘typical’ workers in both the US and Canada - with the possible results falling on a spectrum from complete delinkage (a one percentage point higher productivity growth rate is not associated with any change in the compensation growth rate) to complete linkage (a one percentage point higher productivity growth rate is associated with a one percentage point higher compensation growth rate).1 To focus on the impact of productivity growth over horizons longer than the business cycle, we use three-year moving averages and control for unemployment.

We find strong evidence of linkage in the US. In years with a one percentage point higher productivity growth rate, there tends to be close to one percentage point faster growth in compensation - whether for average compensation, median compensation, or for typical workers (which we proxy using average compensation for production/nonsupervisory workers).

For Canada, the evidence on the linkage between productivity and pay is more mixed. For average workers, we find moderate linkage - over 1961 to 2019, a one percentage point higher productivity growth rate is associated with about half a percentage point faster rate of compensation growth. Over 1961-2019 we also find evidence of very strong linkage for typical workers, whose compensation we proxy using a newly constructed measure of the compensation of hourly-paid workers in five industries.2 However, for more recent periods, estimates for both average and typical compensation are too noisy to rule out strong delinkage (no translation of productivity increases to pay growth). We find no evidence of linkage for median compensation, but caution that our measure of median hourly compensation likely incorporates significant noise, which may substantially attenuate results estimated from year-to-year changes.

How can there be linkage between productivity and pay even as they diverge? If the productivity growth rate is consistently faster than the growth rate of median compensation - by a fixed amount, say a wedge of 1 percentage point - this leads to the gap between productivity and median pay expanding over time (i.e. divergence). But, conditional on this 1 percentage point wedge between the two growth rates existing, it may nonetheless be the case that in years when productivity growth was faster than average, median pay growth was faster than average too (and vice versa). This would tend to suggest that incremental increases in productivity growth rate feed through into higher pay growth for typical workers. Yet other, long-term factors have been pulling down pay even as productivity has been acting to raise it - so that on net, pay has grown much less than productivity.

This is illustrated by a simple metaphor. Consider a bucket containing some level of water (pay) being filled from a hose connected to a tap (productivity growth). The observed divergence between productivity and pay is analogous to the tap running, but the water level not increasing. This could have occurred because the hose does not actually connect the tap to the bucket, in which case increasing the water flow will not make any difference to the water level (delinkage - increasing the rate of productivity growth won’t affect pay). Or, the problem may not be with the hose but may be that there is a hole in the bucket. In this case, increased water flow from the tap cannot fix the hole in the bucket, but it nonetheless results in a higher water level in the bucket than would have occurred with a slower flow of water (linkage - faster productivity growth acts to increase pay relative to a world with slower productivity growth, even if other factors are acting to reduce pay at the same time).

Why does there appear to be a weaker linkage between productivity and pay in Canada than in the US? Canada is a smaller, more internationally open economy than the US. One possible explanation is therefore that the smaller and more open the economy, the greater the degree to which productivity and real compensation may be determined abroad rather than domestically - and the less, therefore, one might expect researchers to be able to detect a process where domestic productivity increases translate into increases in real compensation. We present some evidence consistent with this hypothesis: there is weaker linkage between average compensation and productivity in regions of the US - which are similar in GDP and population to Canada, and could be considered small open economies - than for the US as a whole. The coefficients estimated for US regions are similar in magnitude to those estimated for Canada.

Another interesting point of comparison is that despite much slower labour productivity growth in Canada than in the US, the growth in real median compensation has been about the same in both countries since 1976 due to relatively larger increases in labour income inequality in the US.

While our focus has been on the relationship between productivity growth and wages, both are of course determined by deeper economic forces like the extent of capital accumulation and the nature of technical innovation. It would be valuable in future work to explore possibility that changes in productivity growth coming from different sources have different impacts on wage growth.

Overall, our findings support the hypothesis that productivity growth still matters for middle-class income growth. Measures to boost productivity growth are important for raising pay for the average and typical worker - and conversely, a continued productivity slowdown will reduce the likelihood of increased real compensation. This does not mean policymakers should ignore measures to improve redistribution, since productivity growth alone is not enough to substantially raise middle-class incomes, particularly in the face of other downward pressures on pay. However, it does mean that policy should not focus only on these issues and ignore productivity growth. Policymakers should utilise all measures in their toolkits - both those that boost productivity growth and promote inclusion - in order to have the greatest impact on middle-class living standards.

References

Bivens, J and L Mishel (2015), “Understanding the Historic Divergence between Productivity and a Typical Worker’s Pay: Why It Matters and Why It’s Real,” Economic Policy Institute

Bivens, J and L Mishel (2021), “The Productivity-Median Compensation Gap in the United States: The Contribution of Increased Wage Inequality and the Role of Policy Choices,” International Productivity Monitor 41.

Greenspon, J, A Stansbury and L H Summers (2021), “Productivity and Pay in the United States and Canada,” International Productivity Monitor 41.

Goldin, I, P Koutroumpis, F Lafond and J Winkler (2021), “Re-evaluating the sources of the recent productivity slowdown”, VoxEU.org, 31 May.

Stansbury, A and L Summers (2018), “On the link between US pay and productivity” VoxEU.org, 20 February.

Teichgräber A and J Van Reenen (2021), “Have productivity and pay decoupled in the UK?,” International Productivity Monitor 41.

Williams, D (2021), "Pay and Productivity in Canada: Growing Together, Only Slower than Ever," International Productivity Monitor 40.

Endnotes

1 Specifically, we regress the change in log of annual labor productivity (measured as net output per hour) on the change in log of hourly real labor compensation, controlling for the level and lagged level of the unemployment rate. In our baseline specifications, all variables are three-year moving averages, standard errors are Newey-West heteroskedasticity and autocorrelation consistent, and compensation is deflated by the PCE in the US and the HCE in Canada.

2 These are five industries that we believe are most likely to provide employment opportunities during times of growth and for which historical time series data is available from 1961 onward: manufacturing, mining, construction, laundries, and hotels and restaurants. These made up 34 per cent of total employment in 1961 and 24% by 2011.

“…Evidence from both recent worker surveys and historical data on quits shows that the “Great Resignation” is not as unusual as one might think.Waves of quits have been common during fast recoveries in the postwar period. In line with this historical evidence, the recent wave reflects the rapid rebound in labor demand for young and less-educated workers, largely driven by the retail, leisure and hospitality, and accommodation and food services sectors.... The relationship between high payroll growth and elevated quits rates is not only true across industries and occupations—it is also true over time. Figure 4 shows this, comparing the quits rate in the manufacturing sector with payroll job growth in the overall economy and in manufacturing. The figure shows that the quits waves in manufacturing in 1948, 1951, 1953, 1966, 1969, and 1973 are of the same order of magnitude as the current wave. All of these waves coincide with periods when payroll employment grew very fast, both in the manufacturing sector and the total nonfarm sector...Figure 2 shows that the increase in the quits rate is particularly pronounced for younger workers, including millennials (ages 25 to 34) in the current wave. That is reflective of a normal cyclical pattern in which the quits rate of younger workers responds more to business cycle conditions and drives a large part of the movements in the overall quits rate….”

Bart Hobijn, "“Great Resignations” Are Common During Fast Recoveries,"Federal Reserve Bank of San Francisco, April 2022, https://www.frbsf.org/economic-research/publications/economic-letter/2022/april/great-resignations-are-common-during-fast-recoveries/

“Great Resignations” Are Common During Fast Recoveries

The record percentage of workers who are quitting their jobs, known as the “Great Resignation,” is not a shift in worker attitudes in the wake of the pandemic. Evidence on which workers are quitting suggests that it reflects the strong rebound of the demand for younger and less-educated workers. Historical data on quits in manufacturing suggest that the current wave is not unusual. Waves of job quits have occurred during all fast recoveries in the postwar period.

The labor market recovery since the depth of the COVID-19 pandemic in the spring of 2020 has been the fastest in postwar history. This historic rebound is evidenced by the steepest decline of the unemployment rate and the fastest growth in payroll employment on record (Hall and Kudlyak 2021).

At the same time, the share of workers quitting their jobs—either to take new jobs or to exit the labor force—has also hit its highest level since 2000, when the data began being collected. This recent spike in quits has been referred to as the “Great Resignation.” Some have interpreted it as a wave of resignations, driven by people reconsidering their career prospects and work-life balance, that is drastically altering the labor supply. In this Economic Letter, I provide two pieces of evidence that cast doubt on this narrative and, instead, suggest that the high rate of quits is simply a reflection of the rapid pace of overall labor market recovery.

First, I find that the increase in the quits rate is driven by young and less-educated workers in industries and occupations that were most adversely affected by the pandemic. This is also where payroll employment growth has been high recently, offsetting the job losses incurred in 2020.

Secondly, the Great Resignation is not so great after all. Historical data on the quits rate in manufacturing point to waves of quitting being common during fast recoveries when employment growth was high. Thus, it is not an anomaly, but instead fits the pattern of many past rapid recoveries.

The quits rate and the Great Resignation

In the past few months, a record share of workers in the nonfarm sector quit their jobs each month. A quit is an employee who leaves their job voluntarily, either to go to another job, search for another job while unemployed, or drop out of the labor market. This share, known as the quits rate, reached 3.0% in November and December 2021—its highest level since it was first reported in December 2000 as part of the Bureau of Labor Statistics’ (BLS) Job Openings and Labor Turnover Survey (JOLTS) release.

The record level of the “Total” quits rate in Figure 1 is often pointed to as signaling a Great Resignation. This supposedly reflects a wave of people reconsidering their career prospects, opportunities for hybrid and remote work, and work-life balance—and ultimately resigning in such numbers that it is drastically altering the labor supply in the wake of the pandemic.

The earlier manufacturing data reveal several waves of quits that rival the current wave. Though the 2.3% rate in December 2021 is a record-high since 2000, the manufacturing sector hit comparable highs between 1948 and 1973.

Quits and rebound in demand for young and less-educated workers

One limitation of the JOLTS data is that they do not reveal what type of workers are quitting or whether they quit for career-changing reasons, such as switching industries or occupations, versus exiting the labor force. However, one can supplement the JOLTS information with evidence from the Current Population Survey (CPS; see Flood et al. 2021).

It is not possible to construct the direct equivalent of the JOLTS quits rate in the CPS because of conceptual, sampling, and scope differences between the two. With that in mind, I construct a proxy by measuring a quit in the CPS as someone who is payroll employed one month and, the next month, either reports to have switched job-to-job within payroll employment, switched to another type of employment, become unemployed and reports to have quit, or dropped out of the labor force and reports to have done so because of unsatisfactory working conditions.

The CPS data reveal four important facts. First, the industries with the biggest increase in their quits rate during the pandemic also saw the fastest job growth in 2021. Among these sectors are retail trade; arts, entertainment, and recreation; and accommodation and food services. These are also the sectors hit hardest by the pandemic. More generally, the correlation across industries between the change in the quits rate between November 2019 and November 2021 and payroll employment growth in the 12 months preceding November 2021 is positive in both the CPS (0.46) and JOLTS (0.49) data.

Second, a similar pattern holds across occupations. For example, the occupation with the highest increase in its quits rate is food preparation and service, which saw rapid employment growth over 2021. The cross-occupation correlation in the CPS between the quits rate change over the pandemic and 2021 employment growth is 0.38. Workers in these most-affected occupations tend to have less education.

Third, there was not a particularly large increase in the share of job quitters that changed either industry of employment or occupation. Such an increase would be expected if the current wave involved a broad reconsideration of career choices. The absence of such an increase is consistent, however, with the observation by David (2021) that there has not been a large net reallocation of payroll jobs across industries after the onset of the COVID recession.

The fourth fact is that the increase in the quits rate is largely due to younger and less-educated workers, as seen in Figures 2 and 3.

Figure 2 shows that the increase in the quits rate is particularly pronounced for younger workers, including millennials (ages 25 to 34) in the current wave. That is reflective of a normal cyclical pattern in which the quits rate of younger workers responds more to business cycle conditions and drives a large part of the movements in the overall quits rate.

Quits waves common during fast recoveries

The relationship between high payroll growth and elevated quits rates is not only true across industries and occupations—it is also true over time. Figure 4 shows this, comparing the quits rate in the manufacturing sector with payroll job growth in the overall economy and in manufacturing.

Thus, when the data are extended back prior to the JOLTS time frame beginning in 2000, the Great Resignation does not appear unusual. The current level of the quits rate is at a record high since 2000 because the recoveries of the 2001 and 2008 recessions were very slow by historical standards and thus did not put the same upward pressure on the quits rate as during other recoveries.

Wages and the “Great Renegotiation”

The link between employment growth and the quits rate that I document across sectors and occupations and over time is consistent with the model commonly used to assess the U.S. labor market conditions for low-skilled workers (Burdett and Mortensen 1998). That model explains how the quits rate is related to the frequency of job offers from other employers.

After a temporary spike in layoffs, like the one during the pandemic, employers post many vacancies to restaff and satisfy the rebound in demand for their goods and services. This explains the highly elevated job openings rate that has coincided with the record quits rate.

These vacancies do not only attract applicants who are unemployed. They also draw workers from other employers. Workers who receive a better offer will either quit their current job for another higher-paying job or renegotiate their wage, benefits, and work arrangements with their current employer. So, the “Great Resignation” is actually better interpreted as a “Great Renegotiation.”

As Manning (2006) points out, this can be interpreted as a temporary reduction in the bargaining power of employers that puts upward pressures on wages. His interpretation is consistent with recent wage pressures being highest in the retail and the leisure and hospitality sectors, which have also seen the highest employment growth and increases in job openings and quits rates.

These upward pressures on wages wane as the fast employment recovery slows down, the job openings rate tapers off, and the fraction of workers being poached by competitors declines.

It might take until late 2022 before this happens, however. Respondents to the November 2021 Survey of Professional Forecasters (Federal Reserve Bank of Philadelphia 2021) generally expect nonfarm payroll growth to remain high in 2022. If these forecasts are correct, the job openings and quits rates are likely to remain elevated and wage growth is likely to remain strong for the rest of the year.

Conclusion

Evidence from both recent worker surveys and historical data on quits shows that the “Great Resignation” is not as unusual as one might think. Waves of quits have been common during fast recoveries in the postwar period. In line with this historical evidence, the recent wave reflects the rapid rebound in labor demand for young and less-educated workers, largely driven by the retail, leisure and hospitality, and accommodation and food services sectors.

This analysis covers data through February 1, 2022.

“….Were wages flat during the Obama administration? Did they rise appreciably faster in the first three years of Trump...the answer is “no”...Annualized growth rate of real (PCE deflated) wage was about 0.6% during Obama administration. It was about a 0.6 percentage point higher during the first three years of the Trump administration, but that difference was nowhere near statistically significant.…”

Menzie Chinn, "Average Hourly Earnings, 2000-2022M02,"Econbrowser, March 31, 2022, http://econbrowser.com/archives/2022/03/average-hourly-earnings-2000-2022m02

Average Hourly Earnings, 2000-2022M02

Were wages flat during the Obama administration? Did they rise appreciably faster in the first three years of Trump (as asserted by this commentator)? 30 seconds of data retrieval reveals the answer is “no”.

Here’s the picture: