The Newer Geography of Jobs

Arpit Gupta

Corporate Back to Work Plans Can Map Urban Recovery

A big agenda I’ve explored the last few years is the impact of remote work on real estate markets and cities. By severing the link between work and home, remote work — whether it is hybrid or fully remote — has the potential to drastically reshape spatial patterns of housing, commuting, and spending. These shocks in where people live, work, and transit then have large knock-on consequences on our urban form designed to accommodate these choices.

A key challenge in this research has been trying to figure out what firms are actually doing with remote work choices, and how persistent these choices are likely to be. There is a lively debate on this front, with different data sources like the surveys from Nick Bloom’s group, the BLS, and the ACS providing slightly different estimates of remote work prevalence and persistence.

There is a new data entrant in this area which I think is potentially quite promising. This data comes from Scoop which has put out a Flex Index — a new company level dataset on stated corporate back to work plans. Covering close to four thousand firms with over million employees, this is a very promising data collection effort to figure out what corporations are actually doing: which is essential to figure out the likely impacts on commercial office and urban areas.

Remote, Hybrid, and in Person

Companies broadly are choosing between three broad categories of back-to-office plans, which in turn have different implications for real estate:

1. Fully on Site: This is the default for jobs that don’t involve a computer, and can’t be easily done remotely. It’s the status quo before the pandemic, and still makes sense for many firms for reasons of culture or inertia. Even with the occasional Friday at home; this choice will entail minimal changes to real estate footprints for firms and the resulting knock-on consequences for cities.
   ‍
   ‍
2. Structured Hybrid: This is the format many companies have shifted towards, allowing workers to spend some days of the week at home and the rest in the office. There several design elements to choose from here; including possibly requiring a minimum number of days, trying to coordinate and ensure a specific number of days in the office, or some blend of those approaches (which are tracked in the Flex Index).
   ‍

Companies are generally able to still reduce office demand under hybrid work through a few different ways. Under “hot desking,” employees don’t get dedicated workstations, but instead reserve space on an as-needed basis. With “neighborhoods” concept; offices are laid out to accommodate a mix of different users (some private space for deep work; some collaboration space) and workers pick and choose where they want to be as in a University library. Other approaches may entail workers rotating into the office, depending on their job function, or even sharing space with other firms to avoid overlap on days of the week.
‍
Hybrid work like this can potentially lend itself to increased urban agglomerations. Worker housing choice is typically limited by reasonable commuting distances, and when workers only have to come in a few days a week, they are much more willing to tolerate extraordinary commutes. These make trips like Philadelphia-NYC, or DC-NYC potentially viable for workers, resulting in de-densification as workers spread out across the broader metropolitan area. In turns this means that workers in a much larger geographic catchment zone are potentially participants in local labor market dynamics. ‍
‍
3. Fully Flexible: This accommodates fully-remote firms, as well as those allowing employees to choose when to come in at all. Naturally, when firms pull the plug completely on office plans; they are going to cut down on their office expenses enormously, saving something like $15,000 per worker on average in New York City. From the urban side, the real big shift that you can expect to happen is workers just leaving the entire metro area entirely, to potentially shift to other metropolitan areas.
‍
Before the pandemic, urban workers were purchasing a bundled good: access to high quality job markets, and also high quality amenities. Fully remote work removes this bundle, allowing workers to pick and choose which one they want. It turns out that many workers were getting those expensive urban amenities as a side benefit of labor market access, and given the same job, they’d rather live in a cheaper area (potentially closer to family and friends for instance).
‍
One of the cool things this Flex Index data allows is an examination of how different firms and regions are making these three office choices. You see that firm size matters a lot: small companies — those with only a few employees — are very likely to choose fully flexible plans allowing for a lot of remote work. This fits in with other survey evidence suggesting that startups, in particular, are going remote-first.

Source: Flex Index (flex.scoopforwork.com) employee surveys and publicly available data on company office requirements (January, 2023). The Flex Index is presented by Scoop (scoopforwork.com)

This makes a lot of sense. Real estate leases are commonly classified as an operating expense; but my NYU Stern colleague Aswath Damodaran has argued that they should really be often be thought of as a financing expense. Lease payments are like interest payments in that they are financial obligations which take priority above other investors. Ultimately, leasing an asset is just an alternative to purchasing the asset and paying for it with interest payments secured by it. As a result, any firm — especially a startup — is basically heavily diluting investors, and especially the founders, by taking on leases. Startups, and financially constrained small companies generally, can therefore greatly extend their runway and operations if they figure out how to make fully remote work stick.

In the long-run, my guess is that we are going to see startups which figure out remote work really spread these practices across the rest of corporate America. The large companies of tomorrow are the high growth startups of today, and so it’s instructive to see there what their work practices are.

For the moment, at least, you see that larger firms are instead more likely to rely on either fully on site requirements, or (especially for large firms) structured hybrid work arrangements. This changes a bit at the very top of the employee distribution, which has more fully flexible jobs than you might expect. This is partially driven by interesting patterns among the health/biotech space. Small biotech firms basically have strong in person demands — lots of wet lab space — but large health insurance and pharma companies seem to have a little more room for fully flexible work plans than you might expect.

The Geography of Remote Work

The other thing we can do with the is plot the geography of remote work plans based on the location of the headquarters. Looking at the fully flexible remote plans, you see they are heavily concentrated on the West Coast: California, of course, but also Washington and Oregon; and even Idaho, Montana, and Colorado. There is another cluster of fully remote firms around the DC area (this would be even more the trend if you added in federal employers, many of which are adopting remote work heavily — 85% of all patent examiners are fully remote, for instance). And other centers of remote work in parts of the Midwest, and the Northeast (NYC and Boston).

Fraction of firms with Fully Flexible work plans in state. Source: Flex Index (flex.scoopforwork.com) employee surveys and publicly available data on company office requirements (January, 2023). The Flex Index is presented by Scoop (scoopforwork.com)

Now, a natural question is whether this regional pattern just reflects some of the industry or size factors I just mentioned. But if you control for those variables, and plot the areas that have more or less fully flexible work plans, you actually get a very similar picture below. That suggests there are some durable geographic factors, above and beyond firm composition, that seem to drive remote working choices.

State's relative propensity for Fully Flexible work plans, controlling for local firm industry and size. Source: Flex Index (flex.scoopforwork.com) employee surveys and publicly available data on company office requirements (January, 2023). The Flex Index is presented by Scoop (scoopforwork.com)
‍
Then we have the states that have a lot of structured hybrid work. This graph is more a story of the Sunbelt — the booming cities in the Southeast, Florida, Texas, and other Midwestern states in general.

Fraction of firms with Structured Hybrid work plans by state. Source: Flex Index (flex.scoopforwork.com) employee surveys and publicly available data on company office requirements (January, 2023). The Flex Index is presented by Scoop (scoopforwork.com)

Finally, you have the places where fully in person work is sticking around the most. Here some traditional Northeastern cities in the NYC area, as well as Chicago and DC seem to pop up. Many of these are financial firms, some of which have gone back to more of an in-person presence than before.

Fraction of firms with fully in-person work plans by state. Source: Flex Index (flex.scoopforwork.com) employee surveys and publicly available data on company office requirements (January, 2023). The Flex Index is presented by Scoop (scoopforwork.com)

Connecting these dots, I think you can see some broader patterns for the future of different office and real estate markets across the country:

• West Coast Cities are facing huge threats. They are heavily involved with tech firms and startups that have adopted fully remote work the most; and also just generally seem to have local cultures which are shifting towards more fully flexible work. That’s true for places like San Francisco, but it’s really a trend you see broadly across the region: including Los Angeles, Portland, Seattle, and Denver. These areas are just going through a radical transformation in which firms are going fully remote, office demand is dropping substantially, and workers are free to move wherever they like. ‍

*You see the impact of that in migration data for the Bay Area. One thing to keep in mind is that urban areas in the US have actually always had a lot of net domestic migration, they are just continually replenished by net foreign migration and natural births. This general pattern changed briefly in the post-GFC window, when you saw some net domestic US migration into cities. That was partially driven by “push” factors — the post financial crisis shutdown in housing production — as well as new “pull” factors reflecting high job growth in urban clusters, as well as new amenities (this is when “hipsters” became a thing; speakeasies and espresso joints really took off).

This is precisely the era that Enrico Moretti hails in the New Geography of Jobs — but it actually stops being true soon after the book’s publication around 2015, as Jim Russell has noted. After that point, cities like San Francisco (but there is a similar trend in places like New York) started to see more domestic net outmigration, but it was still roughly balanced by foreign immigrants, even with the lower post-Trump numbers. This pattern drastically changed in the pandemic — you have the combination of a shutdown in net foreign migration, further secular decreases in births, and historically unprecedented increases in net domestic migration. ‍

Some of the key factors here include the NIMBY policies of high cost urban metros, which didn’t allow for enough housing to be built, meaning that workers in these cities are giving up their high wages in rents. The amenity value of these cities started to suffer, with pandemic shocks and crime increases. And finally you have pull factors from the rest of the country, which has been steadily building more housing and getting the sorts of fancy cocktail and coffee establishments to meet those worker needs. That led to this exodus from those cities.
‍

• The Continued Rise — And Sprawl — of the Sunbelt. The flip side here is the presence of cities in places like Texas or Georgia which have not gone fully remote, but are instead sticking to fully in-person or structured hybrid work plans. But remote work is not leaving these areas untouched either. Because many workers have structured hybrid schedules, they will likely be able to commute in only 2-3 days a week to their offices in Dallas or Miami. In turn, this implies that urban agglomerations might stretch out even further in these areas.

This is important — because it provides a new lease on life for Sunbelt metros. In the old days, an increase in house prices in Houston would just mean the construction of another ring of roads around the city. However, lately we’ve seen house prices go up even in Texas and other metros — you start reaching the limits of car commuting at the periphery of these metros, and would ordinarily have to get more dense. But if you’re working hybrid, maybe some of those far exurbs start to look more viable, and you can get even more real estate activity further out into the country. There will still be an office demand shock in these markets, but one that might turn the corner given enough in-migration and suburban sprawl. It’s not a future urbanists are going to be excited about, but one that offers many Americans the comforts of suburban living and frequent-ish in person work (even in these markets, office demand is falling.)
‍

• Uncertain Future for Legacy East-Coast Metros. That leaves other East Coast metros like New York, Chicago, and DC in a difficult position. On the one hand, like many West Coast cities they are looking at shocks to office and urban demand coming from fully flexible remote work. However, they are generally buffered a bit by the presence of fully in-person work demands, particularly in those certain key industries like finance. There’s some hybrid work as well, which is not as common in the West, which is good for New York and Chicago suburbs.

Ultimately though, we’re just going to have to wait and see what happens! It should be an eventful next decade. The better we keep improving our data and understanding of remote work; the more research we are ultimately going to see on its broad social impacts.