By Lambert Strether of Corrente.
My previous round-up on Covid and “the economy” focused mainly on macro economic effects like GDP, or the total cost of Long Covid (“17% of pre-COVID US GDP”). In this post, I want to dig deeper into Covid and the labor market. First, in what may look like a diversion, I’ll look briefly at the January jobs report from the Bureau of Labor Statistics (BLS). (It’s not really a diversion, because I learn by writing, so to master these sources I must write about them. And read your commentary, of course. But please, readers, don’t fix on the jobs report! Move on down to the Covid material.) Finally, since I’m dependent now on the horridly crapified Google, I invite readers with sources I’ve missed to add them in comments. The labor market is a big, big topic, and I’m sure I haven’t got my arms around it yet.[1] Again: “This will in no sense be an exhaustive or even an expert post, but I hope it will serve you to at least create a coherent narrative about where we are, and even, perhaps, what to expect.”
So first, I’ll look at that jobs report, then at two new large studies on Long Covid and labor market participation. Next, I will consider Long Covid as a mass disabling event. Finally, I’ll look at how changes in the labor market affect other aspects of the economy, like commercial real estate (CRE), and conclude.
About That Jobs Report
The BLS report on the “Employment Situation” (“jobs report”) is one of the most highly anticipated document drops in the world. From trading firm Pepperstone:
The Employment Situation report, to use its official name, is by far the most important US, and global, economic indicator released every month. While its impact has waned over time, there is no other indicator to which FX, equities, and bonds typically display such a significant and violent reaction.
And:
The headline measure of the labour market report is the Change in Nonfarm Employment, often shortened simply to NFP. This gauge measures the absolute month-on-month change in employment, thereby representing the number of jobs added to, or lost from, the economy over the past month.
Importantly, NFP excludes agricultural workers, employees of non-profit organisations, and serving military personnel, meaning that the data typically accounts for around 80% of the entire US workforce at a given time. To collate the data, the BLS use the aforementioned establishment survey, surveying around 130k businesses on a monthly basis.
Furthermore, the NFP metric is a seasonally adjusted data point.
Of this month’s jobs report, the Wall Street Journal wrote:
The labor market is in robust health—and that’s scrambling rate-cut expectations. This morning’s jobs report showed an acceleration in hiring, with employers adding a far larger-than-expected 353,000 jobs, a big upward revision to December’s payrolls and unemployment holding at an unexpectedly low 3.7%. That sent investors scurrying to adjust bets on when, and how fast, the Federal Reserve might cut rates. Bond yields soared.
This being an election year, the jobs report is also political. NBC News:
The White House leaned into the numbers, as President Joe Biden seeks re-election this year. Officials view the data as vindication of Biden’s policies….
Or as the Council of Economic Advisors wrote:
So, while we wouldn’t hang our hat on any one monthly number, this morning’s data clearly reveal the continued strength of the US job market, which is at the heart of both the current recovery and Bidenomics!
(About as politicized as you can get. Why not just have Karine Jean-Pierre write something up?) As we saw this morning, however, Brad DeLong, with Paul Krugman and others cited, hang their hats on the entire report[2]. But not everyone does. One sore spot is those pesky “adjustments” that Pepperstone wrote of. CNN, “What’s with all the revisions?”
Friday’s whopper of a jobs number was double expectations, and December’s data was one of several months to get heavily revised. Why are economists’ forecasts frequently so wrong?
It’s nearly impossible to say with certainty how much prices rose or how many people were hired at a given point in time across an entire country’s economy. Finding out how many new hires there were in a given month would involve asking every employer how many people were on their payrolls. That’s why the government and other economic data providers often rely on surveys[3] to make sophisticated estimates.
The BLS, Census Bureau and other government agencies that conduct surveys that inform economic reports do rigorous work to make the best possible estimates with the information they gather. And more often than not they do a tremendous job at it.
But surveys, by nature, are imperfect.
In the same way that election polls don’t always predict the candidate who ends up winning, surveys don’t capture the exact true picture.
Not only was mere data revised this month, but methodology and data structures were revised. From The Hill:
[The BLS] applied its estimation of new business start-ups compared to business failures, the so-called birth-death model that has long been targeted by critics as subject to manipulation and leaps of faith.
In addition, BLS updated the North American Industry Classification System, which shifted about one-tenth of workers into different industries, resulting in “major revisions” to sectors like retail trade and information and less important changes in industries including manufacturing and financial services.
[The BLS also] adjusted “the sample-based payroll jobs numbers based on a census of employment.” As they state in a footnote, the census adjustment resulted in a loss of 266,000 jobs from last year’s March report.
Finally, when you cross-check other BLS data with the jobs report data — not to mention the real world — discrepancies emerge. From alert reader Chris:
So am I wrong in thinking one way to puncture the claims that it’s “just vibes” is to begin with the labor participation rate, then show the current population of working age people, and the follow those with the numbers of people with second jobs? Because that data shows there are roughly 209.3 million working age people in the US, but only 130.8 million are employed or actively seeking work, and about 10.5 million of those have two or more jobs. Which means the tiny variations in the unemployment rate we’re seeing are dwarfed by any shifts in the labor participation rate and the number of people with one or more jobs. We’re trumpeting the statistic that shows 7.8 million people are still trying to get jobs when 10.5 million have two or more jobs in a time when fewer people are working overall, but we’re getting more and more people who need jobs. You need to have a stellar 250k+ jobs on the labor survery every month for a year and to cover the gap between those seeking employment and the multiple employed. So why do they keep insisting that these numbers are great when the last time the labor participation rate was this low, it was 1978, when a little less than 50% of women were working?
Or how about the bizarre contradiction in the data that a family making the median income has more than enough money to afford the median house (assuming they have a 20% downpayment), where the median price of a house is about 420k$, and the median family income is about 95k$, but the places where you’re most likely to earn that median income are also the places where a house costs far more than thecmedian sale price. For example, the median house sale price in California is about 780k$. And that’s before you consider how many people making that median income happen to have 80k$ saved to put towards the downpayment. People can’t afford houses where they live to make enough money to afford a house? This is OK?
My eyes tell me the situation is awful…
We — by which I do not mean the CEA and others of that ilk — sometimes forget that the labor market is not an abstraction, a statistical artifact. The labor market is actual working class people seeking actual jobs doing actual work at actual locations at particular dates and times for a more-or-less set amount of money, all to reproduce their labor power (“feed their families”). I’m a lot more comfortable imaginatively entering into the labor market than I am with BLS statistics, whether pure-in-heart, adjusted, or gamed. And if you want to talk about the labor market and Covid, that’s the approach to use, because most Covid data simply is not tracked, and what little there is not integrated by the BLS to the slightest degree. Which is highly unfortunate, since Long Covid is a mass-disabling event that will affect the labor market for years to come.
And with that, let us move away on from this statistical amuse bouche and enter the world of Covid.
Long Covid and Labor Market Participation
In the previous round-up, I cited two studies. One estimated the total impact of the Covid pandemic at $14 trillion; the other estimated the impact of Long Covid at $3.7 trillion[4]. Today we have two additional studies on Long Covid, the first on disablity in the United States, the second on the labor participation rate in the EU.
1) “Long COVID Disability Burden in US Adults: YLDs and NIH Funding Relative to Other Conditions” (preprint) [medRxiv]. Data from the US Census Bureau’s Household Pulse Survey (HPS) and the Institute for Health Metrics and Evaluation (IHME) /Global Burden of Disease (GBD) Long COVID Study Group:
Long COVID represents a mass disabling event of significant public health concern. Long COVID is associated with a 21% loss of health – comparable to traumatic brain injury or complete hearing loss. Among US adults, 5.3% reported having Long COVID in October 2023 and 1-in-4 with Long COVID consistently report significant activity limitations from its symptoms. Across the 12-month sample of n=757,580 US adults, 1.5% (n= 10,401) met our case definition of disabling Long COVID. Their estimated frequency of the population equates to 3,801,986 adults with long term symptoms after COVID that significantly limits daily activity.
(I would tend to equate disabled with being out of the labor market entirely.) 3,801,986 is a lot, especially considering that we’ll keep adding to the total, if current trends continue.
2) “Long COVID: A Tentative Assessment of Its Impact on Labour Market Participation & Potential Economic Effects in the EU” (PDF) [The European Commission].
To the best of our knowledge, so far, no study has explicitly addressed the impact of long COVID on the EU labour market. The present paper provides a tentative assessment using available estimates from surveys, clinical follow-up studies and model simulations of the prevalence of long COVID. This tentative approach yields an estimated prevalence of long COVID cases of around 1.7% of the EU population in 2021 and 2.9% in 2022, resulting in a negative impact on labour supply of 0.2-0.3% in 2021 and 0.3-0.5% in 2022. In person-equivalents, this means long COVID is assumed to have reduced labour supply by 364,000 – 663,000 in 2021 and by 621,000 – 1,112,000 in 2022 – combining the effect of lower productivity, higher sick leaves, lower hours, and increased unemployment or inactivity. … Available labour market data suggest a mixed picture when it comes to the impact of long COVID…. Other indicators suggest that in 2022, health-related issues (COVID-19-related or of other origin) contributed more than in previous years to a reduction in labour supply at the external (i.e. transition into inactivity) and the internal margin (i.e. a reduction in hours). Notably, there has been a sizeable increase in sick leave reported by several EU Member States in 2022, with acute COVID, long COVID and seasonal respiratory illnesses all likely to have been significant contributors. There has also been an increase in people reporting disability or longterm illness, people inactive due to illness or disability and in part-time work due to illness or disability. The timing and distribution of these observations (with women being more severely affected) could mean that long COVID is a contributing factor. Overall, the health impact of long COVID warrants careful monitoring.
2.9% * ~450 million (the EU population in 2022) = 13,050,000 people. That’s a lot, comparable to the 16 million (Brookings) to 19 million (CDC) with Long Covid in the United States.
Now let’s turn to the market itself, looking at issues beyond disability.
Quantity v. Quality in the Labor Market
Covid as a “mass disabling event” was coined by WaPo editor Francis Steadman in 2022 (and super-wierdly framed as an issue of identity politics for the newly disabled). The central issue is that reinfection increases Long Covid risk, and so the size of the disabled mass will keep increasing over time (given the Biden Administration’s maleficent policy of mass infection without mitigation). From the Sloan-Kettering Institute:
A May 26, 2023 op-ed in the Boston Globe by Wes Ely, MD, MPH — an ICU physician and the Chair of the Department of Medicine at Vanderbilt University Medical Center — discussed the risks humanity is taking with being repeatedly infected with COVID-19, describing what he sees as playing “disability roulette” and that the COVID-19 pandemic has shifted from an emergent infection with significant morbidity and mortality to an “ongoing mass disability event”.
While society yawns, impatient to move on from the COVID-19 pandemic, Americans still play disability roulette. About 1 in 10 of the 110,000 people who catch COVID this week in the United States, many for a second or third time, will be left lastingly ill. Even some vaccinated people; even some young, previously healthy people, after only mild cases.
No longer a mass death event, COVID-19 is an ongoing mass disability event. Every seven days, 25,000 more people join the 10 million in our country suffering memory loss, heart problems, dizziness, extreme fatigue, and more owing to the virus. Globally an estimated 65 million people have this new chronic health condition.
From the University of Nebraska Medical Center:
Since it’s been estimated that over 80% of Americans have been infected with COVID-19 at least once, concern about reinfection is valid. Indeed, a person can get COVID-19 once, twice, three times or more.
A new study analyzed medical records from the Department of Veterans Affairs of nearly 41,000 people who suffered COVID-19 reinfection. For those who had COVID-19 two times or more, the data appeared to show two times increased risk of long COVID and chronic fatigue.
This handy chart from Sloan Kettering gives a visual represenation of the odds of getting Long Covid with repeated infections:
Covid as a mass-disabling event will affect both the quantity and the quality of the labor force. First, quantity. In Germany:
In 2023, sick leave in Germany once again exceeded the record level of 2022, pushing the German economy into recession. This is reported by the “Rheinische Post” with reference to a study by the Association of Research-Based Pharmaceutical Companies (VFA): “Significant work absences led to considerable production losses – without the above-average sick days, the German economy would have grown by almost 0.5 percent,” according to the study, which has not yet been published. As it was, however, the economy shrank by 0.3 percent. “If sick leave had not been so high again, an additional 26 billion euros would have been generated in 2023. Instead of a mild recession, there would have been an increase of just under half a percent in 2023,” write authors Claus Michelsen and Simon Junker, according to the report.
Yes, we must infer the increased sick leave was caused by the Covid pandemic; the causes of sick leave weren’t in scope for the study. Not so in the UK:
As per Long Covid Support’s briefing ahead of tomorrows meeting, Covid-19 continues to have a profound effect on chronic disability in the UK which could be addressed with prevention, treatment, and workplace flexibility: Almost 2M people in the UK had long covid (ONS, 2023)…. Over half of patients with Long Covid reduced their paid work hours, including 17% no longer in paid work. NHS trusts in England lost more than 1.8M working days to Long Covid absences March 2020-Sept 2021.
Covid as a mass disabling event also affects the quality of the work force. In our previous post, we gave several examples of cognitive dysfunction, especially executive function. Studies like those come out so often they’re almost a genre. Here are a few new ones:
1) “Attentional impairment and altered brain activity in healthcare workers after mild COVID-19” [Brain Imaging and Behavior]:
As a core component of cognitive and behavioral processes, attentional function plays a key role in basic and higher functions and has a large impact on daily life and work…. As a high-risk group, healthcare workers (HCWs) are continuously exposed to SARS-COV-2 infection and its consequences during clinical work. Faced with the complex social environment and high workload of pandemic prevention, we believe that even in HCWs with mild SARS-COV-2 infection, impairment of attentional function persists and affects later clinical work. Therefore, we recruited HCWs with mild SARS-COV-2 infection to explore in-depth, attentional network impairments, using ANT, to understand the changes in patients’ levels of attentional function impairments comprehensively…. Our study found varying degrees of reduced efficiency in both the attentional orienting and executive control networks in the patient group. The corresponding neuropsychological background test results also indicated the presence of impairments in general attention and executive function.”
Hence, an HCW who’s had mild Covid is more likely to mix up your pills or misread your chart. Or, extrapolating, forget to bolt on the door to an aircraft. And for the patients–
2) “Factors associated with cognitive impairment in patients with persisting sequelae of Covid-19” [The American Journal of Medicine]:
Patients with PASC [Long Covid] are almost 4 times more likely to evidence cognitive dysfunction compared to normal controls. Forty-four percent of patients with PASC demonstrated cognitive deficits about 7 months from infection.
Hence, truck drivers running red lights, carpenters no long being able to sketch plans, aircraft controllers losing track of their airplanes, and so on.
“Hence” is doing a lot of work in both places. Perhaps we could wait a decade or so for NIH to blow another billion dollars on a useless study. Or we could make reasonable inferences. Remember: Covid is a mass-disabling event. So those two studies affect the entire population, of which those who enter the labor market are a part.
Effect of Changes in the Labor Market
Here’s one[5] knock-on effect of changes that Covid brought about in worker behavior.
1) Commercial Real Estate (CRE). Workers, in the aggregate, changed their preferred location to work. From the Boston Fed, “Lease Expirations and CRE Property Performance“:
The pandemic-induced shift to remote work appears to have led to a large and persistent decline in the demand for office space, especially in central business districts (CBDs). On the other hand, the deterioration in commercial real estate (CRE) loan performance has been relatively modest to date, as long-term leases have shielded commercial-property owners from the effects of diminished demand for space to a fair degree. To shed light on how these properties will perform in the longer term as more leases expire, this paper analyzes how lease expirations have affected property performance historically and investigates how these patterns have changed so far for leases that have expired since the COVID-19 outbreak.
Conclusion
Speaking of the imagination, there’s historical precedent for pandemic-driven labor market upheavals:
A pandemic kills off a chunk of the population, especially the more vulnerable working class. The labor force depletes. The labor of the remaining people who are willing to work is suddenly worth higher wages. People in power pre-pandemic want to deny that labor is suddenly worth more than it was before.
It’s the story of the American labor market in 2021 — and the onset of the “Black Death,” or bubonic plague, in 14th-century Europe.
With over 38 million people leaving their jobs in 2021, the coronavirus pandemic has spawned what some are calling the Great Resignation. Research shows that people want to pursue more fulfilling careers, even without other jobs lined up. Workers seem tired of low-paying, dangerous professions, and want to avoid increased exposure to the pandemic.
And they definitely want more pay.
More bleakly, swaths of the workforce are simply also dying off: line cooks, warehouse employees, and agricultural workers were especially at high risk of death in 2020, according to a study from the University of California, San Francisco.
A new book by English historian Dan Jones makes clear that labor shortages have long followed pandemics, with social unrest not far behind. His “Powers and Thrones,” which looks at roughly 1,000 years of medieval history, includes a discussion of the Wat Tyler rebellion, which he argues was really the working class leveraging its power as a result of the labor shortage that followed the Black Death, or bubonic plague.
It wasn’t just an English phenomenon, either. Thousands of people died in the rebellions that took place throughout Europe during that age, but something else happened — the economy changed forever.
When a pandemic arrived on this scene, the continent eventually lost roughly one in every two people. This was a human tragedy, but it also put labor at a premium. Wages soared as landowners struggled to make sure their crops didn’t die from a lack of harvesters. The sudden decline in the population also meant a decline in land rental prices. Land became dirt cheap, and landlords were desperate for tenants.
More than 800,000 people have died from COVID in the US over 2020 and 2021, and a similar thing is happening on a much smaller scale. Bureau of Labor Statistics data shows that 3 million people are currently missing from the American workforce, while wages have increased significantly for the first time in decades, and set to go even higher in 2022.
According to Jones, the same kind of shift in worker power 600 years ago caused wealthy landowners to panic. They petitioned their rulers to help save them from financial ruin.
In England, King Edward III enacted legislation that made it illegal for workers to claim wages above pre-pandemic rates: the Ordinance and Statute of Labourers. Workers were prescribed wage ceilings depending on their industry, such as masonry or mowing. The ordinance also made it a legal requirement for every able-bodied person under 60 years old to work.
Under the next king, Richard II, these ordinances, along with a slew of new, higher taxes finally triggered what is known as the Peasants’ Revolt, the first great popular rebellion in English history.
Of course, there wasn’t any such thing as “Long Black Death” back in the day; people were either dead, or not. I find it hard to imagine organizing a popular rebellion with a mass of disabled people. Nevertheless….
NOTES
[1] I have toyed with the slogan that “The only market is the labor market,” because, surely, all markets depend on the jobs that human beings do?
[2] From this morning’s Links, container import volumes are up too. So something’s going on out there….
[3] Also from CNN: “The rate at which people are getting recruited for surveys that are used in many of BLS’ monthly reports including employment, Consumer Price Index and Job Openings and Labor Turnover are down sharply from before the pandemic.” Curiously parallel to the problems that election pollsters are having.
[4] In the previous round-up, I did not cite to this article from Nature, though I should have:
The oncoming burden of long COVID faced by patients, health-care providers, governments and economies is so large as to be unfathomable, which is possibly why minimal high-level planning is currently allocated to it. If 10% of acute infections lead to persistent symptoms, it could be predicted that ~400 million individuals globally are in need of support for long COVID.
[5] I wanted to look at productivity and consumption, too, but dang. My sources weren’t good enough. It shouldn’t be hard to show, however, that cognitive dysfunction will have effects on productivity, quality of work, customer satisfaction, and so forth. As for consumption, surely at some point some significant fraction of the population will latch on to the idea that closed, crowded, close-contact spaces are dangerous? Maybe next time.
