People who put in at least some of their work week from home are more likely to come from households that earn $100,000 or more, a data analysis from the Rockefeller Institute of Government released this week found.

The data comes as many workers continue to work remotely since the start of the COVID-19 pandemic and as policymakers, elected officials and business leaders grapple with what is likely a long-term shift in how people do their jobs.

Upper income earners staying home at least part of the work week could cement changes in commuting, how people spend their money when they do so and where they live.

And it comes as New York City Mayor Eric Adams and Gov. Kathy Hochul have sought to bring more people back to offices and onto trains and buses to office spaces in urban downtowns.

“The research in this brief puts real numbers behind assertions that remote and hybrid work are persisting as we enter the pandemic recovery era,” said Laura Schultz, executive director of research at the Rockefeller Institute of Government. “Rates of telework remain considerably elevated over their pre-pandemic levels and this is especially true among higher income earners and those industries where job responsibilities do not need to be completed in-person.”

Researchers found just over 5% of Americans reported some amount of remote work prior to the start of the pandemic. Now, as more offices and job sites have re-opened, nearly 30% of Americans worked remotely in August.

But the types of workers who do so varies: 55% of households that earn $100,000 or more reported working remotely at least once in a two-week period that ended in August. Meanwhile, 28% of households that earn between $50,000 and $99,999 reported telework as an option during that time period.

And for households that make between $25,000 and $49,999, the telework number drops to 14%.

The survey covered 15 of the largest metropolitan areas. Not surprisingly, many jobs in farming, transportation, retail and construction are being down in person, the survey found.