The pandemic has now forced millions of people to work from home, typically those with higher incomes and in higher-income economies. Surveys suggest that, to date, employers have been positively surprised by the efficacy of this new way of working.
The Covid-19 pandemic brought the abrupt closure of offices, factories and stores. Huge numbers of employees have been working from home. An even larger number of workers, with jobs that are deemed essential or that cannot be done at home, have either faced an elevated economic risk of layoffs or an elevated health risk involved in leaving home each day.
The ability to work from home contributes to the unequal impact of Covid-19 across individuals, communities and countries. It plays a key role in macroeconomic analysis designed to guide policy-makers. And it may shape the contours of a transformed workplace even after the public health crisis abates.
Here, we review what was known about working from home prior to the pandemic, and discuss how early assessments of the ability to work from home shaped efforts to understand the incidence and transmission of the damage brought by Covid-19. We discuss what we have learned during the pandemic from actual data on remote work, and we conclude with a view about what this might mean for the working environment in the long run.
Who worked from home before the pandemic?
Prior to the pandemic, working from home was relatively uncommon. In the UK, only about 5% of the workforce reported working mainly from home in 2019 (Office for National Statistics, 2020). Just over a quarter of UK workers had ever worked at home. In the United States, pre-pandemic surveys show that although roughly one-quarter of workers did at least some of their work at home, the median worker believed this was not feasible for almost 95% of their tasks (Mas and Pallais, 2020).
As the pandemic first started its global spread, economies went into lockdown, with workers commonly asked to stay home. A natural question in thinking about the economic incidence and propagation of the virus was: ‘who can work from home?’ Of course, knowing who worked from home before the pandemic does not reveal who could work from home when forced to do so by the pandemic.
Researchers have often employed occupational characteristics to predict what might happen to workers in response to a variety of economic shifts. For example, economists used the US Dictionary of Occupational Titles to identify routine tasks that could be computerised (Autor et al, 2003).
Economists have also estimated which US jobs were most susceptible to US firms shifting their tasks to foreign workers, a possibility brought about by the telecommunications revolution of the late 1990s. The first investigations of ‘offshorability’ used surveys administered by a US programme called O*NET to classify each occupation’s susceptibility (Blinder, 2009).
Who could work from home during the pandemic?
Researchers used O*NET survey questions describing jobs to estimate that 37% of US jobs could plausibly be done entirely at home, as would be required under the strictest social distancing measures (Dingel and Neiman, 2020).
For example, if surveys reveal that an occupation requires operating vehicles or equipment, or primarily involves walking or running, that occupation would be classified as not able to be done entirely at home. Many of the jobs classified as theoretically able to be done at home – including, for example, most teachers and some doctors – had not in fact been previously done at home to any significant extent.
The share of jobs that researchers estimate could be performed entirely at home varies substantially across industries and places. In Boston, which has a large share of workers in research and technology, nearly 45% of jobs can be done from home; whereas in Las Vegas, with its huge entertainment and hospitality industry, only 30% of jobs can be done from home (Dingel and Neiman, 2020).
Perhaps the most striking pattern is the variation by individual income: higher-paying jobs are far more likely to be possibly performed from home in both the United States (see Figure 1) and the UK (Costa-Dias et al, 2020).
Figure 1: Jobs that can be done at home typically earn higher wages
Source: Dingel and Neiman (2020)
A number of economists have corroborated the finding that the ability to work from home is not evenly distributed across the population (Avdiu and Nayyar, 2020; Mongey et al, 2020). Jobs that are difficult to perform under social distancing policies are more often held by workers who are less likely to hold a college degree, to have a large amount of liquid wealth, to be white or to have been born in the United States (Mongey et al, 2020; Bell and Blanchflower, 2020)
Economists’ initial classifications of the feasibility of working from home for each occupation were based on the characteristics of jobs in the United States. These measures were combined with data from the International Labour Organization on the occupational mix of 86 countries to assess the extent to which work can be done at home across the globe (Dingel and Neiman, 2020). Consistent with the variation within the United States, a far lower share of jobs can be done at home in lower-income countries (see Figure 2).
Figure 2: Share of jobs that can be done at home, by GDP (PPP) per capita
Source: Dingel and Neiman (2020)
One might worry about applying a US-based classification of jobs to countries with very different economic circumstances. But recent research applies a similar job characteristics method to surveys administered in developing countries and also finds that the ability to work from home is rare in those economies (Saltiel, 2020).
An important caveat to these findings is that farming in low-income countries, while infeasible inside one’s home, might be done while safely outside on the farmer’s own property (Gottleib et al, 2020). Similar analyses have been performed in a number of developed economies (Barbieri et al, 2020; Boeri et al, 2020; Stratton, 2020).
Related question: Who can work from home and how does it affect their productivity?
In addition to improving our understanding of which workers will be most affected by Covid-19 and resulting stay-at-home orders, the share of jobs that can be performed at home at the industry and aggregate level has been an important input into quantitative analysis aiming to draw out the macroeconomic implications of the pandemic and to guide policy-making (Jones et al, 2020; Kaplan et al, 2020; Rampini 2020).
These shares also play important roles in economic analysis that thinks through how damage caused by the shock in one sector of the economy may in turn affect other sectors (Baqaee and Farhi, 2020; Bonadio et al, 2020).
Evidence on working during the pandemic
What do we know about the actual work experience several months into the pandemic? As traditional data sources collect information on remote work with some lag, economists have turned to non-traditional sources to produce contemporaneous measures. Two recent studies use online surveys to produce real-time measures of working from home for the United States. These studies estimate that between 35% and 50% of individuals were working at home (Bick et al, 2020; Brynjolfsson et al, 2020).
While the latter figure exceeds the earlier estimates of the share of US jobs that can be performed entirely at home, these numbers are not necessarily comparable: in the short run, many employees may perform parts of their jobs at home rather than being wholly furloughed or laid off. In addition, those who can work from home may be overrepresented in online surveys.
Recent tabulations of US job losses by industry derived from payroll-processing data illustrate that the feasibility of working from home is only one factor that influences employment outcomes (Cajner et al, 2020). As of late April, the industries with the largest employment declines were restaurants, scenic and sightseeing transport, and amusement and recreation, with employment declines of 50% or more. These are industries in which a small share of jobs can be performed at home.
Related question: Which firms and industries have been most affected by Covid-19?
At the same time, the largest employment increases were in heavy and civil engineering construction, non-store retail, and couriers and messengers, which are also industries in which a small share of jobs can be performed at home. The pandemic-induced necessity of constructing hospitals and delivering groceries to residences caused employment in these industries to grow more than 10%. Those who are working from home during the pandemic rely on many essential services that cannot done remotely.
Real-time surveys have also been employed outside the United States. In a survey of UK firms, 37% of employees were reported to be working from home in both April and May 2020 (Decision Maker Panel, 2020). Another survey-based study compared the extent of job losses as of early April in the UK, the United States and Germany (Adams-Prassl et al, 2020). The study concludes that workers less able to do tasks at home were more likely to experience reduced hours, lower earnings and job losses.
Related question: What are the effects of coronavirus on the UK and US labour markets?
What is the future of remote work?
Even if the pandemic is short-lived, it may trigger permanent shifts in occupational practices. Economic history offers many examples in which temporary shocks had persistent effects on production. For example, the US civil war reduced cotton supplies to Great Britain for only a few years, but this caused British cities that specialised in manufacturing cotton textiles to have permanently smaller populations (Hanlon, 2017).
More recently, it has been documented that the decline of middle-skill occupations focused on routine tasks over the last four decades has occurred almost entirely during economic downturns (Jaimovich and Siu, 2020). The post-recession recoveries have been ‘jobless’ in that the routine occupations never came back. These patterns suggest that some cities or some office occupations may not necessarily return to their pre-pandemic levels.
Will in-person work concentrated in large cities be less prevalent even after the public health crisis abates? The answer is not at all obvious. After all, videoconferencing, document-sharing and other technologies facilitating remote work during the pandemic have been available for years, yet employment in skill-intensive service industries has been increasingly concentrated in large, expensive cities rather than becoming geographically dispersed (Kolko, 2010; Eckert et al, 2020).
The current pandemic could make remote work more prevalent in the long run through three different mechanisms:
- First, firms and workers may realise that they had been underestimating the potential of working from home.
- Second, firms and workers may make costly investments to enable remote work during the pandemic that then cause working from home to be more attractive even after the pandemic (Baldwin, 2020).
- Third, a permanently elevated risk of infectious diseases would make working from home more attractive relative to pre-pandemic arrangements.
We are too early in the pandemic-driven trial of working from home to forecast confidently the future of work, but recent surveys of businesses suggest that some shifts may stick. A survey of 1,500 hiring managers found that more than half reported working remotely has gone better than expected, while only one in ten said it has gone worse than expected (Ozimek, 2020).
Revised beliefs may make working from home more prevalent after the pandemic. A survey of more than 4,000 executives and small-business owners found about one-quarter had purchased new digital tools to assist employees working from home and a large majority expect that tool usage to be permanent after Covid-19 (Azurite Consulting, 2020). Such sunk investments could permanently shift some jobs from offices to homes.
Finally, it is worth remembering the astonishing speed with which the work-from-home transition occurred. No one would characterise the setting in which remote work was adopted on this scale – with no time to plan and in the face of worldwide health and economic crises – as the ideal laboratory for testing a new style of work. It seems likely that firms and workers will get better at it over time.
Where can I find out more?
How many jobs can be done at home? In the first study of this question during the pandemic, Jonathan Dingel and Brent Neiman estimate that 37% of US jobs could be plausibly done entirely at home. A webinar presentation and discussion of the research is available in the Centre for Economic Policy Research series on Covid Economics: Vetted and Real-Time Papers.
US and UK labour markets before and during the Covid-19 crash: David Bell and David Blanchflower provide an early overview of the effect of the shock in the UK and the United States and analysis of the labour market situation at the beginning of the crisis.
Getting people back into work: Monica Costa Dias, Christine Farquharson, Rachel Griffith, Robert Joyce and Peter Levell discuss some key economic issues that the UK government needs to face when thinking about how best to get people back into work
Inequality in the impact of the coronavirus shock: evidence from real time surveys: Abi Adams-Prassl, Teodora Boneva, Marta Golin and Christopher Rauh have a project to survey people in the UK, Germany and the United States providing real-time information on labour market effects of the crisis.
Covid, hysteresis, and the future of work: VoxEU editor-in-chief Richard Baldwin argues that Covid-19 has changed the future of work via four shocks: massive job losses, massive digital transformations, massive debt burdens, and massive costs of socially distanced office space.
The future of remote work: A report by Adam Ozimek, chief economist at Upwork.
Impact of Covid-19 on business decision making, spending and recovery: Azurite Consulting survey of 3,400 Americans and 1,000 international respondents