Using new data from the 2021 Business Response Survey, a large, nationally representative survey of U.S. private sector businesses, this article presents unique estimates of telework patterns observed during the coronavirus disease 2019 (COVID-19) pandemic. We find that, between July and September 2021, 13 percent of all U.S. private sector jobs involved teleworking full time and 9 percent involved teleworking some of the time. Telework was less common in establishments that increased base wages during the pandemic. The share of establishments that increased telework was larger among establishments that started offering flexible work hours or compressed work schedules after the pandemic hit. Telework was also associated with reductions in workplace square footage and relocation. Within each industry sector, low-paying establishments had a smaller share of jobs that involved telework.
Businesses made many adjustments to their operations and policies during the coronavirus disease 2019 (COVID-19) pandemic. One key adjustment was shifting some employees from working in person at offices and stores to teleworking (also known as working from home or remote work). By limiting in-person contact among employees, telework allowed businesses to maintain operations and workers to remain employed.
Many researchers have demonstrated the importance of telework to the U.S. economy during the COVID-19 pandemic. Since early 2020, when the pandemic hit, telework has expanded sharply.1 Jose Maria Barrero, Nicholas Bloom, and Steven J. Davis estimate that telework accounted for about 50 percent of paid work hours between April and December 2020, compared with 5 percent before the pandemic.2 At the beginning of the pandemic, the extent of job loss was much lower for workers who were able to telework. Matthew Dey, Harley Frazis, Mark A. Loewenstein, and Hugette Sun found that, between February and April 2020, employment fell 8 percent among workers in occupations in which telework is feasible, compared with 21 percent among workers in occupations in which telework is not feasible.3
In this article, we examine telework during the pandemic by using data from the 2021 Business Response Survey to the Coronavirus Pandemic (BRS), which was collected between July and September 2021. The BRS is a special U.S. Bureau of Labor Statistics survey that describes how U.S. businesses have changed their operations and employment since the pandemic hit. We show how telework is related to other employer pandemic responses, including offering flexible work schedules, changing the physical size of workplaces, and relocating. We also show how telework varies by establishment characteristics, such as industry, average wages paid, and geography.
According to an earlier BRS survey conducted between July and September 2020, 31 percent of establishments increased telework offered to employees because of the pandemic.4 Notably, there was substantial variation by establishment size and industry. Large establishments (those with 500 or more employees) were more than twice as likely to have increased telework than were smaller establishments.5 In the sectors of educational services, finance and insurance, information, and management of companies, more than 50 percent of establishments increased telework.6 By contrast, in both agriculture and accommodation and food services, less than 10 percent of establishments did so.7 Many questions remained after the 2020 BRS: How many jobs involved telework? What portion of jobs involved teleworking full time? Did all jobs within an establishment involve teleworking similar amounts of time? One goal of the 2021 BRS was to provide answers to these questions.
Other researchers have estimated the extent of telework during the pandemic. Using data from the Current Population Survey, Matthew Dey, Harley Frazis, David S. Piccone Jr., and Mark A. Loewenstein found that, between October and December 2020, 22 percent of workers were teleworking because of the pandemic.8 The Federal Reserve Bank of Atlanta estimated that, as of January 2021, 15 percent of private sector jobs involved teleworking full time and 36 percent involved teleworking at least 1 day per week.9 Beyond giving estimates for the share of jobs that involve telework, we show how telework is associated with establishment characteristics and establishment responses to the pandemic.
Survey questions on telework
There were 83,032 respondents to the 2021 BRS, which was conducted from July to September 2021 and collected information from a representative sample of private sector establishments in the United States.10 One part of the survey asked a sequence of three questions about the telework behavior of employees at each establishment.11 Specifically, it asked respondents to report the percentage of their employees currently teleworking all the time, some of the time (but not all), and rarely or never, allowing us to distinguish among three groups of workers. The survey instructions directed respondents to ensure that the three responses summed to approximately 100 percent, so that each employee would be counted in one of the three groups.
For each group, the survey instrument provided a drop-down menu of 13 potential responses (all percentages): 0, 5, 10, 20, 30, 40, 50, 60, 70, 80, 90, 95, and 100. Some collected data were edited to improve their accuracy. First, because some respondents appear to have misunderstood the sequence of questions, more than 20 percent of those who provided a response for all three groups reported 0 percent for each group. Upon further analysis and a brief followup survey (described in 2021 BRS technical notes), it was determined that a very large percentage of these respondents intended to report that 100 percent of their employees rarely or never teleworked.12 For the purposes of the present analysis, these respondents are treated as reporting that 100 percent of their employees rarely or never telework.
Another issue is that some respondents gave implausible responses for the three groups—responses totaling a number either too low or too high to be usable. We constructed an acceptable range for this total and dropped any respondents falling outside that range. Because the respondents had to use a drop-down menu with limited response options, some of them might have had a valid set of answers with a total as low as 80 (actual percentages: 29–29–42; rounding down would give a 20–20–40 response) or as high as 120 (actual percentages: 33–33–34; rounding up would give a 40–40–40 response). If a respondent provided a response for at least one but not all three groups, we imputed 0 for the missing responses.
Because the answers of some respondents did not sum to exactly 100, a final step in the data editing process involved normalizing all responses so that their total equals 100. For instance, a response of 40–40–40 would be multiplied by the adjustment factor of 100/120, which would yield an imputed response of 33–33–33.
The final sample of usable responses contains data for 72,324 establishments.
For each establishment, there are three responses to the telework questions:
- TeleworkFT = percentage of jobs that involve teleworking full time
- Teleworksome = percentage of jobs that involve teleworking some of the time (but not all)
- Teleworknone = percentage of jobs that involve teleworking rarely or never
Our estimates are weighted averages across all respondents with a valid response to the telework questions and a valid response to any additional question used in a tabulation. The weights used to calculate these averages are employment weights, constructed as described in the 2021 BRS technical notes.13 To estimate, for example, the percentage of jobs that involve teleworking full time, we use the formula
where is the employment weight for establishment , which is based on the sector, state, and size class of establishment . In the calculation of , the employment for every combination of sector, state, and size class in the fourth quarter of 2020 in the Quarterly Census of Employment and Wages (the sample frame for the BRS) is divided evenly among all valid respondents in the same cell.
To estimate the percentage of jobs that involve teleworking at least some of the time, we combine the estimates of the percentage of jobs that involve teleworking full time and the percentage of jobs that involve teleworking some of the time:
Lastly, we provide estimates for a related question asking whether telework increased for some or all employees since the start of the pandemic. These estimates, which represent the percentage of establishments that responded “yes,” are weighted averages calculated with a formula similar to equation (1). In this calculation, however, we use establishment weights instead of employment weights. Our sample includes respondents with valid answers to this question, the sequence of three telework questions, and any additional question used in a tabulation.
Telework and other employer responses
Overall, 33 percent of establishments increased telework for some or all employees during the pandemic.14 (See table 1.) At the time of the 2021 BRS, 13 percent of jobs involved teleworking full time. More broadly, 22 percent of jobs involved teleworking at least some of the time (that is, some of the time or full time), with the remaining 78 percent involving teleworking rarely or never.