Does Telework Harm Older Workers? A First Look at Technological Change Spurred by COVID-19
The good, the bad, and the unknown.
1) If you want to take a look at my code, please see this link.
2) If you want to take a look at some event-study regressions I have run that better control for within-person effects, occupation and industry trends, etc., please see this link.
I was curious to see what’s happening to those lucky enough to telework amid the pandemic—whom the technological shock, actually, shocked.
Starting in March 2020, an unprecedented number of people took to working from their living room couches, their kitchen tables, and (for those of us who prefer a horizontal lifestyle) their beds. Amid a pandemic, this saved jobs and lives. But, from what I’ve seen, economists haven’t yet dug into the ramifications of such a vast and abrupt technological change.
The speed of work-from-home technology adoption between the months of January 2020 and April 2020 presumably is unmatched by any previous technological change in history. Much of the technology may seem benign to those of us who grew up in the virtual age. How different is joining a zoom call than FaceTiming on a cellphone? Is Slack all that different from group messaging? But the abrupt and widespread adoption of new technologies may open up new chasms of discrimination and inequality. This is important now and for the future: Writers have probed the possibility that telework will go on forever, (or at least will become increasingly popular over time).
So, what’s happening now?
I used data collected in the Current Population Survey (CPS) to look. First, I categorized occupation groups by whether or not they had above median or below median take-up of teleworking technology during the pandemic (as measured by the proportion of workers in that occupation who worked from home due to the pandemic). Basically: Did your people in your job convert to teleworking? Right away, you see stark differences in unemployment and employment rates. The figure below shows that for occupations in which more people used teleworking technology, employment dropped by significantly less than in occupations in which fewer people used teleworking technology.
This pattern suggests the obvious: Being able to work from home during the pandemic helped. Jobs in which workers were able to move to remote work saved people’s lives and better preserved their incomes. If writers, bankers, and us economists (very top of the pile for teleworking) didn’t get hit with a plague in 2020 and, instead, in 1980—without this technology—maybe that would be different.
But worth thinking about, as is en vogue in contemporary economic studies of technological change, are the ways in which new technologies introduce new tasks and skills to the mix of those needed to efficiently and effectively complete a job, and perhaps automate away some tasks that once needed a human to complete. With the change in needed skills, will we see the recent abrupt shift impact workers differently according to age? Is it harder for older workers to adapt to the new landscape of on-the-job tasks? How will more veteran workers cope with the Sisyphusian tasks of unmuting one’s self or re-setting the router?
Taking a look at unemployment incidence in the CPS by age and likelihood of teleworking (predicted using one’s occupation), we see yes, it seems so! Unemployment rates are much lower for those in high-teleworking occupations. But also in high-teleworking occupations, there is a clear increase in the likelihood of facing unemployment as a worker ages. The positive relationship between age and unemployment is not seen in professions that don’t have a high take-up of teleworking technologies.
While the chart below does not prove definitively that technology directly harms more seasoned workers, it certainly suggests that as the skills and tasks needed for daily work change, our wiser peers may find it harder to maintain job security.
Going one step further, into the world of event study regressions, I look at how employment status is changing over time for workers older than age 60 who experience a “layoff” at some point in the sample (which I define as moving from employed in one month to unemployed and looking for work in the next survey month). With a bunch of controls to account for time trends and individual-level means, I find that starting in April 2020, workers over age 60 seem to be leaving employment (and the labor market entirely) at higher rates if they report being in an occupation that took up teleworking at higher rates during the pandemic. Each point plotted in the chart represents the effect of more teleworking in one’s occupation on the probability that they are either employed (top panel) or not in the labor-force (NILF, bottom panel). We see a clear downward trend in employment and an upward trend in NILF status starting in April 2020. Importantly, these charts are showing the impact of telework technology on this group (60+ year olds who experienced a layoff at any point in the sample period).
This isn’t an argument against technological change. But it’s important to understand the wrinkles. How a good change (working from home to save lives during a pandemic) can find the cracks in our system and do harm (in this case, reducing job security for older workers). Studying this technological change will help us identify these issues, and bolster the gaps. Understanding the inequitable impact of technological change is key to making sure that two steps forward doesn’t come with one step back.
P.S. Here’s some quick data on a related issue.
What can we learn about the incidence of technological change on those individuals who could not continue their jobs via teleworking. In particular, how does the patriarchal and racist foundations of our social schema shape the impact of technological change. Here’s some quick data I saw while looking at data for teleworkers.
By income bracket (and subsequently by race, gender, and nationality), teleworking possibilities vary dramatically. For example, those in managerial work saw rates of teleworking well over 50% between May 2020 and November 2020. In contrast, for those in food services, teleworking rates hovered around 10%. We know that occupational sorting results in womxn, Black and Latinx folk bearing the brunt of this work. No doubt there will be much to learn moving forward about the ways in which technological change wrenches open already existent inequities.
P.P.S. Economists were literally ranked highest in terms of fraction of workers reporting telework at ~95%.