Who Works from Home? Education, Job Type and Location All Play a Role

After more people started working from home because of the COVID-19 pandemic, remote flexibility became an important nonmonetary benefit and reason for choosing a job.

How much education do workers who are more likely to work from home have, and what kinds of work do they do?

I analyzed the demographics of remote workers to find out, and also looked at the work situations of their neighbors. I found that living in a county where remote work is more common increases the likelihood that an individual worker will work from home, even after factoring in education and occupation.

I used data from the Current Population Survey (CPS), a nationally representative survey carried out monthly by the U.S. Census Bureau and the U.S. Bureau of Labor Statistics. The survey asks people about their demographics, employment status, education, geography, and more. In October 2022, the survey began asking respondents whether they had worked from home in the past week, and if so, for how many hours. I used data from October 2022 to January 2026Data for October 2025 are missing due to a government shutdown. and studied only civilian adults who worked 10 hours or more in the previous week.Restricting the sample to adults who worked at least 10 hours improves comparability across workers. In that sample, 21.6% of workers reported working from home for some time in the previous week, a percentage roughly in line with those from other studies.

Remote Workers Are Educated and White Collar

The CPS asks respondents their highest level of education. I grouped respondents by five education levels and plotted the percentage who reported working from home at some point in the week prior to responding to the survey, as seen in the graph below.

The figure above shows a wide distribution in work-from-home (WFH) rates. Workers who did not complete high school rarely work from home; about 3% in that group reported doing so. On the other hand, roughly 41% of workers with an advanced degree (master’s, Ph.D., JD, etc.) reported conducting at least some remote work—almost twice the national average. Only workers with at least a bachelor’s degree exhibit work-from-home rates above the national average.

The survey also asks respondents what their primary occupation is. I grouped occupations into 22 broad categories and plotted the rate at which different groups say they worked from home in the previous week. The results are in the graph below.

Again, there is a wide range of work-from-home rates; these reflect how feasible it is to do work tasks from a remote location:

• Less than 2% of workers in the food service category—including cooks, bartenders and servers—reported working from home at some point during the previous week.
• Approximately 65% of workers in the computer and math group, such as computer programmers, database administrators and actuaries, participated in remote work the previous week.
• Occupations generally regarded as white collar (in business and finance, management, and legal groups) worked from home at higher rates than manual laborers in the construction, agriculture and production categories.

Remote Workers Cluster in Counties with Higher Education and Income

Workers who are more likely to work remotely tend to be more highly educated and be in white-collar professions. Where are they most likely to live? To find out, I combined individual data on remote work with county-level data on education and income, namely the percentage of adults in a county with a bachelor’s degree and that county’s median household income in 2022. I computed county-level work-from-home rates and related them to the county characteristics.

The scatterplots below demonstrate a strong, positive relationship between county-level education and household income and the percentage of respondents in that county who reported working from home. The U.S., St. Louis County and New York County (Manhattan in New York City) averages are shown as reference points.

It is perhaps unsurprising that more highly educated and higher-income counties have greater proportions of remote workers, since people with higher levels of education work from home more frequently. But is that the full story? An interesting question is whether the rate of people working from home in a particular county influences how likely it is a resident there will work from home. Put another way, does that person, with a given education level and occupation, face different probabilities of working from home depending on where they live? If every other local business offers their employees remote work flexibility, does that make it more likely your boss will offer work from home as an option?

County Trends Affect Residents’ Likelihood of Working from Home

I applied a statistical technique called probit regression to help answer the question of whether the work-from-home rate of the county you live in affects your likelihood of working from home. Think of probit regression as estimating the relationship between two variables while holding all other variables constant. In this case, a person’s education and occupation were held constant to isolate the effect that living in a WFH-intense county has on an individual’s own chances of working from home. I studied the work-from-home rate in an individual’s county excluding the individual; in other words, looking at how many of the person’s neighbors work from home.

Consider an example with two counties: County A, where 10% of other workers participate in remote work, and County B, where 40% of other workers do. The model predicts that a person working a legal occupation with an advanced degree (a lawyer, for instance) living in County A would have around a 45% chance of working from home. If they lived in County B, they would have a 68% chance of working remotely. (See this example and several others in the graph below.) In all cases, the identical individuals’ probability of working from home increases with higher percentages of remote workers in their county.

An important caveat is that these results imply correlation, not causation. I cannot definitively say that living in County B leads to opportunities to do remote work. Rather, with education and occupation held constant, individuals living in counties with higher work-from-home rates are themselves more likely to work from home. This could be because employers adjust their policies in response to local labor market norms, or because workers with remote working flexibility choose to live in similar counties. Pinning down the cause of the clustering of remote workers would require further research.

Notes

1. Data for October 2025 are missing due to a government shutdown.
2. Restricting the sample to adults who worked at least 10 hours improves comparability across workers.