Blog: Is Work-From-Home an Existential Crisis for Office Space?
By William Wheaton
Founding Consultant, CBRE-EA
The press speaks almost daily about the empty office space in America’s CBDs and whether this is a temporary response to Covid19 or a more permanent demise in the need for downtown offices.[i] In fact, “Telecommuting” was actually a term first coined in the 1970s, which then gained popularity in the 1990s. Furthermore, well prior to the onset of the Pandemic, scholars began actively studying the WFH transition, suggesting that the US labor market might finally be experiencing a structural change in the nature of work.[ii] In this paper, I would like to jump into this debate and apply some long-established and well recognized models of urban land use or “Urban Form “ to the question of whether the demand for office space, and in particular central office space could be in structural decline. I will speak purely theoretically, leaving until the end, the empirical question of how well theory lines up with the far-to-sparse data on WFH.
The now almost 100-year-old “Monocentric” model of cities has long provided the back bone for much economic analysis of transportation and urban land use. That said, in the last 25 years a growing group of urban economists have generalized this model, allowing firms to endogenously select their location, rather than simply assuming that all economic activity takes place at the single “center’ or CBD. This location choice by firms is based on three relationships.[iii]
a). Firms should be able to pay workers lower wages when their choice of location enables workers to commute less. There is strong empirical support for this principle: observationally equivalent workers are in fact paid more for making long CBD commutes relative to much shorter trips to a more local suburban office.
b). Firms can justify the higher costs of centralized agglomeration when such clustering improves productivity through so-called “Marshal” spillovers between the clustered firms. The most likely source for these spillovers – one that receives some empirical support – is found in the operation of local labor markets. Workers are able to further specialize their skills and move between firms (improving productivity) when such moves require no residential relocation. When all firms are in just one location – the CBD – this is obviously true. Firms in turn are able to more quickly fill vacancies when they are located in the midst of a plentiful labor supply. Locally “thick” labor markets offer advantages to both sides.
c). In the end Urban Form emerges as a hypothesized tradeoff between these two considerations. It can range between being “monocentric” or having all economic activity totally concentrated at a central site (when spillovers are large and commuting easy) to being far more dispersed or “polycentric” (when spillovers are weak and commuting difficult).
In the current context, we can think of WFH employment (enabled by technology) as an extreme form of polycentric dispersal – one that involves no commute by workers, and large reductions for firms in the cost to house their workers in office space. It is unclear whether some share of the commute savings – initially accruing to workers – might eventually be split with employers in the form of lower wages. Similarly, the savings to firms in office space might eventually be shared with workers in the form of higher wages. Applying Nash Bargaining principles these two savings combined might be divided evenly between workers and firms. Without getting into the details, the two “savings” together could well add up to 15-20% of current wages.
In terms of Urban Form, the end result is that the physical concentration of working in the CBD could be greatly diminished. Employment per se would be relocated to worker residences. Highways would flow freely again. Whatever physical presence by workers is still required by office firms would likely occur in the most convenient location for periodic gatherings – probably still the CBD – but with a much reduced usage of space.
But what about the advantages of physical proximity in terms of the operation of the labor market? Ironically, these should totally dissolve because firms would employ few if any workers at physical sites – only virtually at home. When employed at home, workers could search for productivity-enhancing job changes with any firm – regardless of its physical location since work is remote. Similarly, firms can fill vacancies with any worker, anywhere, all online! Marshallian spillovers come from physical proximity and when this proximity ceases to be important, and vanishes, so do the spillovers. What is the justification for CBD agglomerations if all workers are remote?
Now for some caveats. First, many types of work are not suited to WFH. Work where employees have to operate machinery, have to engage in physical labor, provide services in person are clearly not subject to these arguments. Surveys suggest that only a third of all jobs have the potential for WFH in the sense that the majority of their tasks can be accomplished away from a “place of work” [iv]. Second, for those industries and occupations where WFH is applicable, its adoption should be most widespread in locations where office rents are high, and the journey to those buildings is long and arduous. So for office space, buildings in the CBD seem far more at risk than those in satellite, suburban locations. Finally, if we ignore spillover productivity and focus purely on individual output, it still remains to be seen if workers in a WFH environment are more or less as privately productive as they are when at a place of work.
What is missing from the discussion at this point is a data at the firm, establishment or building level about the recent status of their workers. This would allow us to see how these transitions are varying not just by occupation or industry, but also by location. To date, we have to rely on surveys of workers, most of which do not provide information on either place of residence, or place of work. Without this verification, we cannot tell for sure as to whether WFH will have concentrated impacts in certain, largely CBD based industries, or whether it will be more widespread. This is the data that will allow us to determine if office space in general, after more than a century of rapid growth, is in fact facing its existential moment in history.
[iii] McMillen, Singell. “Work Location, Residence Location and the Intraurban Wage Gradient.” Journal of Urban Economics 32 (1992): 195-213.
White, “Location Choice and Commuting Behavior in Cities with Decentralized Employment.” Journal of Urban Economics 24 (1988): 129-152.
Timothy, Wheaton, “Intraurban Wage Variation, Employment Location and Commuting Times“, Journal of Urban Economics, 50, 2 (2001), pp 338-366.
Wheaton, “Commuting, Congestion and Employment Dispersal in Cities with Mixed Land Use”, Journal of Urban Economics, 55,3, (2004), pp 417-438.
Fallick, Fleischman, Rebitzer, “Job Hopping in Silicon Valley; micro foundations of a High Tech cluster”, Review of Economics and Statistics, 88, 3 (2006), pp 472-481.
Helseley, Sullivan. “Urban Subcenter Formation.” Regional Science and Urban Economics 21, 2 (July 1991).
Redding, Rossi-Hansberg. “Quantitative spatial economics”. Annual Review of Economics 9, 21 (2017)
[iv] Brick, Blandin, Mertens, “Work from Home After the COVID-19 Outbreak”, Federal Reserve Bank of Dallas, WP-2017, (2020),
Adams-Prassl, Boneva, Golin, Rauh, ”Work that can be done from Home: Evidence on Variation with and across Occupations and Industries”, IZA paper 13374 (2020).
Dingel, Neiman, “How Many Jobs Can be Done at Home?” Working Paper 26948. National Bureau of Economic Research. (2020).