Posted by: Josh Lehner | April 1, 2021

COVID-19 and Job Polarization

Yesterday the Bureau of Labor Statistics released the 2020 occupational data which allows us to take our first look at how the pandemic has impacted employment through the job polarization lens. Our office has been posting employment by wage tier based on industry average wages every month since COVID hit. However that’s really just been a placeholder because we lack solid monthly estimates for occupational data at the state level. It’s an important distinction because if the manager and accountant at a restaurant stayed employed and it was the chefs and waitresses who were laid off, industry employment masks those changes among high- and low-wage jobs even within hard-hit sectors.

With that said, let’s take a look at the fresh 2020 data. This first chart shows the percentage change in employment by major occupational group from 2019 to 2020. While we knew the in-person service sectors have been hammered with the pandemic, it’s even more apparent looking at the occupational data. Both Food Preparation and Personal Care (think barbershops and nail salons) are down almost 20%. Sales (mostly cashiers) is down nearly 10%. On the upside, note that the increases in Scientists and Agriculture are a thousand or two jobs, which is large in percentage terms, if not in actual numbers and may reflect some reclassifications of individual occupations rather than underlying strong growth.

To be honest the thing that stood out the most to me in the chart above was the relatively “mild” declines in Production jobs. Production occupations are the manufacturing jobs that actually do the manufacturing. While a 5% drop is large, it’s not necessarily so when compared with other recent recessions when these types of jobs have been hammered and never fully return. Overall this has been an encouraging development this cycle, or at least falls under the “better than first feared” category. Our office wrote about the manufacturing outlook in more detail in our December 2020 forecast (see PDF pg 16.)

With that in mind, the second chart compares recessionary losses across the past three recessions here in Oregon. In both the tech and housing busts, middle-wage jobs experienced the largest losses, while low- and high-wage jobs fared better. That’s the definition of job polarization. Now, what’s different about COVID is the low-wage jobs have been hit harder than they ever have before, and middle-wage jobs have experienced a moderate recession. Furthermore, high-wage jobs today are estimated to be at historic highs and have avoided losses overall.

Now, keep in mind a couple data caveats. These are technically May 2020 employment estimates, but given the way the occupational data is compiled and calculated, they are really based on November 2019 and May 2020 underlying information. As such they do not have the full impact of COVID incorporated in them, but are certainly broadly representative of what we know has happened and are the best available data we have. This also means that a year from now when the (May) 2021 data is released, that will be based on information from November 2020 and May 2021, and may better reflect the pandemic-related state of the economy. Furthermore, next year BLS is changing their methodology for these data to better include staffing patterns by occupation from firms who do not always report their information. That may or may not be an additional curve ball to keep an eye on.

Given that we know COVID has hit the economy different than past recessions, with workers in low-wage occupations bearing the brunt of the cycle and comparatively mild middle-wage job losses, I think it’s important to take a step back and look at the big picture.

This third chart looks at employment changes in Oregon in recent decades through the job polarization lens. Overall it is great news that high-wage jobs continue to pace overall economic growth. However the concerning issue continues to be that middle-wage jobs have been primarily down for much of the past 20 years, and even in the best of times, they only really regain their lost ground.

Given the middle-wage job outlook has called for only moderate gains during expansions, one of the more concerning parts to the COVID recession was that it hammered the low-wage jobs. While we’ve discussed this before, a lot of times workers struggled to adjust when they lose their traditional, middle-wage job. While a few are able to land high-wage jobs, the vast majority end up taking a low-wage job, moving away in search of work, or dropping out of the labor force entirely. None of this is a good dynamic. However the concern last year was if there were also no low-wage job opportunities then even the existing meager options would dwindle further.

Thankfully, to date, this does not appear to be the case. Middle-wage jobs are not experiencing severe job losses, and even as low-wage jobs have, federal aid has largely kept workers and households afloat in the past year. Looking forward, our office expects both low- and middle-wage jobs to fully recover, at least overall. Some individual occupations and industries are unlikely to entirely regain all their lost jobs, but the dynamics this cycle are different.

Note that the projections in the chart above take our office’s industry forecasts and maps them to occupations. This pattern of high-wage occupations growing the fastest, followed by low-wage ones and middle-wage growing the slowest is quite similar to the latest 10 year occupational projections from our friends over the Oregon Employment Department.

Finally, don’t let the apparent lack of low-wage job growth in 2021 fool you. It is more an artifact of how annual averages work mathematically when the underlying monthly data is moving so quickly. As an example, here is our industry forecast for leisure and hospitality employment in Oregon. You can see that even as the sector adds jobs throughout 2021, the annual average is basically the same as last year. This is in part because last year’s average included some pre-pandemic data and the decline was so swift and severe that when you average over the full year, that annual figure did not fully reflect all of the pain. On the upside we will be experiencing something similar, but in reverse. Job growth will accelerate but the final 2021 annual averages won’t fully reflect that strength. Just something to keep in mind when talking about annual data.

Bottom Line: We knew that workers in the in-person service industries have been the most impacted by the COVID recession. However until yesterday we didn’t know exactly how those losses broke down between occupations even in the hard-hit industries. This cycle is different both in terms of the nature of the losses, in addition to the federal policy response. Given that incomes are up for most households, and pent-up demand is very real, our office does expect a much faster and more complete economic recovery than we have experienced in recent decades. This means there is likely to be relatively little economic scarring and permanent damage. The occupations and industries that have borne the brunt of the recession will recover in the years ahead. Our office will continue to monitor this data to look for structural changes in the staffing patterns across the economy.


  1. Curious, is there a reason why utilities and government jobs are not included here?

    • Hi Mel, thanks for asking. The data here are occupational and not industry. So for utilities, you would have some jobs in management, office and admin support (office staff), engineering, some of the line workers probably in installation, maintenance, and repair jobs, etc. Overall the big structural changes in the economy in recent decades really are at the occupational level (fewer receptionists, more engineers etc) rather than at the industry/sector level.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s


%d bloggers like this: