Working Primarily From Home
Workers with disabilities are more likely than those without disabilities to work primarily from home. As shown in Table 1, an average of 5.7% did so in the 2009–2018 period, compared to 4.6% of workers without disabilities (columns 1 and 2). The rates of home-based work are especially high among workers with mobility impairments (6.8%) or difficulty with activities inside the home (7.1%) or outside the home (7.2%) (column 2). The differences by type of disability are illustrated in Fig. 1. The disability gap is smaller but still highly significant when confined just to employees (3.3% compared to 2.8%, in columns 3 and 4), and expands slightly and remains significant when limited only to the self-employed (23.1% compared to 21.3%, in columns 5 and 6).
The disability gap is maintained across almost all personal characteristics in the breakdowns in Table 2. Workers with disabilities remain more likely than those without disabilities to primarily work at home across gender, race/ethnic, and education categories, and regardless of whether they live alone, live with children under age 18, live with elders age 65 or older, or live in a house with Internet access (columns 1 and 2). The exception is that there are no significant differences by disability status for middle-aged workers (age 35–49 or 50–64), so the disability gap is concentrated among younger and older workers (18–34, and 65 plus). The younger cohort of workers with disabilities may be more challenged in finding employment and more likely to work from home as they look for stable work. In comparison, the older cohort of workers with disabilities may be more likely to have retired from prior jobs and opted for telework as a convenience. Also, it is possible that middle-aged workers with disabilities are more motivated to work on site since they are at the peak of their earnings capacity and do not want to receive a wage penalty for working from home.
The occupational distribution does not explain the disability gap in home-based work. As shown in Table 3, workers with disabilities are significantly more likely to work primarily at home in each of the six major occupations.
These relationships are tested more rigorously in Table 4 using linear probability regressions that control for demographic characteristics and detailed occupation (Probit regressions produced similar results). As can be seen, workers with disabilities were significantly more likely to be working primarily at home, both overall and when confined to employees (columns 1 and 4). Controlling for other variables, workers with vision or hearing impairments are less likely than workers without disabilities to work primarily at home. In contrast, workers with mobility impairments and difficulty outside the home are more likely to do so (columns 2 and 5).
How is home-based work changing over time? Figure 2 shows that workers with disabilities were more likely to work primarily at home over the 2009–2018 period. The home-based work rates increased for both workers with disabilities (5.3% to 6.0% after smoothing) and workers without disabilities (4.0% to 5.2%) over this period, but the disability gap narrowed. Table 4 confirms these trends, showing a positive trend in home-based work for both groups, but a more positive trend for workers without disabilities. Note that while the disability-year interaction is negative, the overall predicted trend for workers with disabilities remains positive when the interaction and main effects are added.
Any Work at Home
While the ACS results clearly indicate that workers with disabilities are more likely to work primarily at home, this may not apply to doing “any work” at home. The results from the ATUS in Table 5 indicate that this is the case. Asked about all their activities during a reference day, 22.7% of the employed respondents with disabilities said they performed any paid work at home during the day, which is not significantly greater than the 21.7% of employed respondents without disabilities who gave the same response.
Broken down by disability type, the only significant disability gap in any home-based work occurred for workers with mobility impairments, of whom 26.4% reported any home-based work compared to 21.7% of workers without disabilities. Broken down by broad occupation, the only significant disability gap occurred in management occupations (columns 1 and 2), but this significant difference disappears when examining employees and self-employed separately (columns 3 to 6). So the overall gap may reflect the increased likelihood of workers with disabilities to be self-employed (where the rate of home-based work is higher in general). The disability differences, both among employees and the self-employed, remain statistically insignificant in regressions that control for demographic characteristics and occupation (not reported but available).
The story is slightly different in analyzing the larger sample in the combined CPS Disability Supplements done in 2012 and 2019. While the simple comparisons in Table 6 indicate that workers with disabilities are more likely to primarily work at home, the regressions in Table 7 (columns 1 to 3) find that this difference is no longer statistically significant when controlling for demographic characteristics and occupation.
In contrast to the ATUS results, however, both simple comparisons and regressions show that workers with disabilities are more likely than those without disabilities to do any work at home (Table 6, columns 1 and 2, and Table 7, columns 4 to 6). The regressions also show that employees with disabilities are more likely to be paid for work at home and have a formal arrangement with their employer (Table 7, columns 7 and 8). Work at home appeared to increase for employees in general from 2012 to 2019 (columns 2, 5, 7, and 8 of Table 7, consistent with the positive time trend from ACS data in Table 4), but the insignificant interactions with disability status do not point to different trends for workers with and without disabilities.
Examining the reasons workers offered for working at home, Table 6 shows that workers with disabilities were less likely to cite reduced commuting and coordination of work and family needs, and more likely to cite illness, disability, or health reasons.
Therefore, the ATUS and CPS results give a somewhat mixed message on disability differences in working at home. While the ATUS results do not show significant differences by disability status in doing any work at home (either in simple comparisons or regressions), the CPS regression results point to a significantly higher likelihood of work at home for both employees and self-employed workers, and a higher likelihood of paid work and a formal arrangement for employees with disabilities compared to those without disabilities. We put a somewhat heavier weight on the CPS results, given the larger sample size and the additional measures related to work at home. Also, as noted in the data section, there is at least a five-month lag between the disability measure and the work at home measure for the ATUS data, making the disability measure less reliable in the ATUS data.
Wages of Home-Based Workers
How does working at home affect the wages of workers with and without disabilities? Table 8 reports on wage regressions with standard controls that include detailed occupation. As shown at the bottom of the table, the disability pay gap is slightly larger among homeworkers (-13.5%) than among non-homeworkers (− 10.4%) (column 1), with similar gaps when looking just at employees or self-employed workers (columns 2 and 3). While the differences in these disability gaps between homeworkers and non-homeworkers are small, they are statistically significant.Footnote 2
Along with comparing workers with and without disabilities, another interesting comparison is between homeworkers and non-homeworkers. Table 8 shows that employees without disabilities earn slightly more on average (3.7%) if they work primarily at home, while the difference for employees with disabilities is not significant. In contrast, self-employed workers earn less by working at home, with home work gaps of 8.1% among workers without disabilities and 8.6% among those with disabilities (each of these gaps is statistically significant by itself, but the difference between these gaps is not significant).
The principal implication from these results is that the pay gaps faced by workers with disabilities are not substantially altered by working at home: they continue to earn less than workers without disabilities when working at home. Therefore, the potential benefit of increased home-based work for people with disabilities would appear to lie in greater employment opportunities, not increased pay.
Potential for Jobs to be Done at Home
The analysis of current patterns in home-based work begs the question of what jobs can potentially be done at home. The expansion of computer and information technologies enables more and more jobs to be done remotely. In a 2020 study, Dingel and Neiman studied which jobs could potentially be done entirely at home, using data on occupational characteristics from the federal O*Net database . They assign a potential to each occupational category, the category designated with a two-digit number, and then present a score for each category reflecting how many jobs within that category could be done entirely at home. Those scores are replicated in column 1 of Table 9. The actual number of workers working primarily at home in the 2014–2018 period (using ACS) is shown in column 2. The numbers of actual home-based workers in each category (column 2) are uniformly lower than the number of potential home-based workers (column 1), illustrating substantial room for expansion of home-based work in each occupational category. There is a strong positive correlation of 0.59 between the two measures, as shown at the bottom of Table 9, providing some validation for Dingel and Neiman’s measures of which occupations are more amenable to work at home.
How much potential is there for workers with disabilities to work primarily at home? Column 3 of Table 9 shows the number of workers with disabilities in each of the two-digit occupational categories. People with disabilities appear to have less potential for expanding telework in their current jobs: the distribution of workers with disabilities has a strong negative correlation of -0.68 with the potential for jobs to be done entirely at home, as shown at the bottom of Table 9. This negative relationship is also demonstrated by estimating the percent of jobs that could be done entirely at home, using current occupational distributions of workers with and without disabilities. Based on their current occupational distributions, Table 9 shows that close to two in five (39.5%) of the jobs held by workers without disabilities could be done entirely at home, compared to only one-third (34.0%) of the jobs held by workers with disabilities.
In sum, while workers with disabilities are more likely to be working from home currently, their current jobs offer less potential for substantially expanding the extent to which they could work primarily at home. This reflects their disproportionately greater likelihood of being in service and blue-collar occupations that are less amenable to home-based work (e.g., buildings and grounds cleaning and maintenance, where workers with disabilities have the greatest representation). It is also possible that the nature and severity of their disabilities are more likely to limit them to those occupations that cannot be done from home.