Type to search

The geography of working from home and the implications for the service industry

Breaking News

The geography of working from home and the implications for the service industry


The Covid-19 outbreak has led to an unprecedented rise in the number of jobs done from home. This column discusses the implications of this shift for locally consumed services such as restaurants, hairdressers, and gyms. Using precise data on the location of homes and offices of workers across the UK, it finds that there is large heterogeneity in the impact of working from home on these businesses. While city centres suffered a significant drop in demand for services, suburban neighbourhoods experienced an increase in demand. Policies aimed at helping the service industry should take this diverse impact into account.

For many workers, the Covid-19 pandemic has led to an unprecedented shift in where they work, from the office to the home. This shift has further altered consumption patterns, as working from home for many meant a change in where they eat lunch or have coffee, and where after-work drinks take place, if they happen at all. Pre-pandemic, workers also visited retail shops near their offices, had their haircut between meetings, and visited gyms before work. Hence, by changing the geography of where we work, working from home has consequences for the geography of demand for locally consumed services. How will this unprecedented shift to working from home affect location and employment of businesses offering locally consumed services?

An estimated 37% of jobs can be done from home with large variation across industry and occupation (Dingle and Neiman 2020 Adams-Prassl et al. 2020). The Business Impact of Covid Survey finds that, by the end of 2020, approximately 30% of UK jobs where done from home. Similar numbers are found for the US (Bick et al. 2020). Further, recent YouGov polls find that a significant number of workers and employers plan to continue working from home, at least part of the time, even after the pandemic. 

In our new research, we aim to quantify the magnitude of this geographic relocation of work, coined the ‘zoomshock’, and illustrate the short- and medium-run implications of this relocation for employment in the locally consumed service (LCS) economy (De Fraja et al. 2021). Jobs in the LCS industry, including hospitality, leisure and high-street retail, cannot be done from home. Further, these services must be consumed in the same geographic space in which they are produced; one cannot ger a haircut at a Hackney barbershop while in Guilford. Therefore, the reallocation of work activity has had uneven consequences, increasing LCS demand in some neighbourhoods and reducing it in others. This constitutes another novel channel by which the economic costs of the Covid-19 pandemic vary across the economy.

Measuring the zoomshock

We measure the zoomshock using pre-Covid-19 data on workers’ precise residence and work locations. For a given neighbourhood, the zoomshock is then calculated as the difference between the inflow of workers due to homeworking and the outflow of workers due to homeworking. This measure captures the net change in economic activity for a neighbourhood due to working from home. The sign (positive or negative) and magnitude of this measure will depend on the amount of pre-Covid-19 work that was performed in neighbourhood, the suitability of that work for working from home (following Dingel and Neiman 2020), and the number of residents in the neighbourhood who commute to work pre-Covid-19 and who can work from home. We measure the zoomshock for all neighbourhoods (defined as Middle Super Output Areas) in England, Wales and Scotland. 

This metric not only captures the change in the geography of economic activity, but also reflects a change in work-based demand for local goods and services. For example, neighbourhoods in central London, with high pre-Covid-19 economic activity, are likely to experience a large outflow relative to inflow, as they were previously heavily ‘importing’ workers from the suburbs – they hence experience a large negative zoomshock (Figure 1). However, neighbourhood highstreets located in popular commuter towns, such as St Albans in Hertfordshire, may experience a large inflow of workers relative to the outflow – and hence a positive zoomshock. 

The zoomshock and the locally consumed service industry

Understanding the geographic shift in work activities has important implications for our understanding of both the short and longer-run consequences for the LCS recovery in the wake of the Covid-19 public health crisis. 

First, different neighbourhoods are affected differently by the zoomshock. Office-dense city centres are experiencing considerable declines in output and employment while relatively less-densely populated suburbs are experiencing an increase in productive activities. This fits with anecdotal evidence for the Greater London Authority. As much as  70% of the large workforces in central local authorities, such as the City of London and Westminster, commute to work and can work from home (De Fraja et al. 2020). This has led to an exodus from inner London to the residential neighbourhoods of outer London (Figure 1) and beyond.

Figure 1 Zoomshock for the Greater London Authority

The geography of working from home and the implications for the service industry 1

Note: This figure maps the change in economic activity (based on number of jobs performed) by middle super output area for the Greater London area. Blue areas indicate an increase in activity, red areas indicate a decrease in activities.
Source: De Fraja et al. (2021). 

Second, by moving workers from the office to the home, the zoomshock reallocates a significant portion of the demand for delis, cafés and hairdressers, from urban neighbourhoods with many of these services to suburban neighbourhoods with relatively few (Figure 2). We find that, on average, neighbourhoods experiencing a positive zoomshock have 687 employees in the LCS industry; neighbourhoods experiencing a negative zoomshock have 2,139 employees in the LCS industry. This is important for short-to-medium run employment in the LCS industry. Even when workers can move freely from one place to another to follow demand, capital will be much slower to move. Existing coffee shops, bakeries, and hairdressers in suburban neighbourhoods will quickly hit capacity constraints. Further, demand is dispersed in the suburbs compared to dense urban centres, increasing the cost of supplying pre-Covid-19 levels of aggregate demand. Even with increased working from home, some suburbs won’t have the critical mass needed for certain parts of LCS which depend upon a large catchment (e.g. gyms). Notice that the described threats to LCS employment result purely from the zoomshock, and our discussion ignores any other additional threats that Covid-19 has created. 

Figure 2 Zoomshock and local service employment

The geography of working from home and the implications for the service industry 2

Note: This figure regresses, for each MSOA, the % change in employment due to homeworking against the log-employment in the local service industry. Binned into 100 evenly sized groups. Right-hand-side figure includes only MSOAs in the Greater London Area, left-hand-side figure includes all MSOAs outside the Greater London Area.
Source: De Fraja et al. (2021).   

Third, when comparing different local authorities (or other work areas), the aggregate LCS employment consequences of the zoomshock will be larger in some areas than in others. Local authorities in which the neighbourhood-level zoomshock is concentrated in a few neighbourhoods are at greater risk of experiencing a large employment loss in the LCS industry. This follows from the supply constraints mentioned above: a zoomshock that leads to a large increase in demand for one neighbourhood will strain the capacity of existing LCS businesses more than a zoomshock that leads to a small increase in demand across many neighbourhoods.     

Concluding remarks

Working from home has allowed many workers to continue their employment with minimal interruption while following public health guidelines during the Covid-19 pandemic. Moreover, for many it has led to substantial savings in time and money spent commuting. However, this redistribution of workers has several important implications for how Covid-19 impacts the locally consumed service economy, and which policies should be chosen to aid the economic recovery of these local businesses. The allocation of scarce resources to promote the economic recovery should reflect the heterogeneity of the zoomshock: in some neighbourhoods local services will benefit from the increase in homeworking, while in other neighbourhoods local services will suffer.  

How much of our work will continue to be done from home after the Covid-19 pandemic is of critical importance for the recovery policy. Our city centres are unlikely to return to pre-Covid-19 levels of local service demand if workers do not return to pre-Covid-19 levels of office use. Even a widespread move to working from home one day a week would be a substantial shock. In this case, policies which support the survival of all local businesses will be in vain. Rather, business should be encouraged to move to neighbourhoods for which the zoomshock has been positive.     


Adams-Prassl, A, T Boneva, M Golin and C Rauh (2020), “Work Tasks that Can Be Done from Home: Evidence on Variation within & Across Occupations and Industries”, CEPR Discussion Paper No 14901. 

Barrot, J, B Grassi and J Sauvagnat (2020), “Sectoral Effects of Social Distancing”, HEC Paris Research Paper No FIN-2020-1371.

Bick, A, A Blandin and K Mertens (2020), “Work from home after the COVID-19 Outbreak”, Federal Reserve Bank of Dallas Working Paper 2017.

De Fraja, G, J Matheson and J Rockey (2021), “Zoomshock: The geography and local labour market consequences of working from home”, Covid Economics 64: 1–41. 

Dingel, J and B Neimen (2020), “How many jobs can be done at home?”, Journal of Public Economics 189.

Leave a Comment

Your email address will not be published. Required fields are marked *