Three years into the pandemic, life is so close to back to normal that Ticketmaster is failing to process the wild demand for Taylor Swift tickets and e-commerce shopping has come crashing back to earth as people leave their homes to return to brick-and-mortar stores.

But the remote workers refuse to give in. How very privileged of them. 

All these years into the pandemic, remote work is still inaccessible to large parts of the workforce, according to a new working paper from the London School of Economics and Political Science. All the media discussion about the future of the office (including this one) applies almost exclusively to a “metropolitan elite,” the paper says.

The study found that while remote working has performed well enough for large companies and white-collar employees based in cities, it is still struggling to catch on at smaller companies and in poorer regions of the world.

“Before declaring that the office and our cities are dead, policy makers should be careful about seeing the world solely through the eyes of the metropolitan elite, whether they’re based in London, Milan or Washington. Not everyone is working from home, or can work from home, even if they want to,” Riccardo Crescenzi, an economist at LSE and one of the paper’s co-authors, said in a statement.

The authors analyzed work-from-home data dating back to well before the pandemic in Italy, which is among the few developed countries to have information on the remote status of every single employee in its economy, according to the study. The authors then compared real work-from-home data with existing measures of how many jobs could potentially be done remotely. 

The study combined the results with territorial economic data in Italy, and concluded that while many jobs could in theory be done remotely, in practice many are not. And the reason is that policymakers have yet to make remote work, well, work for everyone.

“The value of these findings is that they show that existing literature and existing commentators have overemphasized the importance of working from home,” Crescenzi told Fortune. “The rhetoric around work from home…is sending the message that there is no problem and that this technology will be an inclusive solution, while on the contrary, our results suggest that this is not necessarily the case.”

Who can really work remotely?

Questions have been raised since the beginning of the pandemic on the potential inequalities that come with working from home, questions that have yet to be fully resolved.

At the height of the pandemic in 2020, a paper by the International Monetary Fund found that remote work was much easier to implement in certain countries and economic sectors than others. The paper found that nearly 100 million workers in 35 advanced and emerging countries were vulnerable to losing jobs they couldn’t do remotely, with young, low-income, and female workers most likely to be penalized.

In the U.S., the rise of remote work also caused racial and ethnic divides to solidify. When lockdown orders were first announced in March 2020, Black and Hispanic workers were significantly less likely to be able to work remotely, according to a study by the Economic Policy Institute, a think tank. Students from low-income backgrounds were also disadvantaged by virtual learning models, with many unable to afford the necessary equipment to attend school digitally.

But two years later, many of these issues have yet to be resolved, and have migrated over to small companies with fewer resources or are located in poorer areas and are unable to support working remotely.

“We’re still facing the exact same issues because nothing has been done,” Crescenzi said. 

He added that the remote work conversation has so far focused on those of us who “live in cities and work in highly knowledge-intensive sectors,” but the dominating rhetoric on the benefits of remote work risks exacerbating inequalities from the early days of the pandemic.

Remote work divides

The LSE study found that small companies with fewer resources are significantly less likely to allow working from home relative to large corporations. Less than 1% of companies with fewer than 10 employees allowed working from home, according to the study, compared to 70% of large firms with more than 250 employees.

The study found that in Lazio and Lombardy—two of Italy’s wealthiest regions that are respectively home to Rome and Milan—more than 21% of employees were able to work from home when forced to do so. But Italy’s economy is notoriously territorially divided, as the vast majority of industry and wealth are located in the north, while in many of Italy’s southern regions—which last year was responsible for only around 20% of the country’s economic output—fewer than 2% of workers were able to work remotely, according to the researchers.

The study found that even when accounting for economic differences between Italy’s north and south, the gap between how many people could conceivably work from home and those who actually did was “staggering.” LSE’s Crescenzi also said that existing data on working from home tends to be skewed, since surveys are mostly conducted through online forms that remote workers are more likely to answer.

Crescenzi and the other researchers suggested that remote working remains a privilege for white-collar workers based in cities and employed by large companies, while working from home is unlikely to permeate through to most of the remaining workforce.

In the U.S., similar results were reported in a September study by the Economic Innovation Group, an advocacy organization. The study found that factors including occupation, education levels, and how expensive the local housing market is can determine the popularity of remote work in a specific region. In the U.S., remote work remains concentrated in large cities and coastal areas, while people are not as likely to be working from home in less-developed areas, the study found.

And another July report by WFH Research, which has been surveying remote work attitudes monthly since the start of the pandemic, even called working from home a “big-city phenomenon” that is significantly more common in large cities in the U.S. than small towns or rural areas.

The unequal distribution of remote work carries some significant risks for policymakers, Crescenzi said, as it could funnel investments away from crucial infrastructure that non-remote workers need to keep doing their jobs.

“It’s something that might lead to underinvestment in public transport, for example,” he said, adding that lower ridership on public transit could lead to reduced service and make it more expensive in terms of both time and money for non-remote workers to get to work.

“This is an example of how the assumption that the office space is dead and is not important for anyone anymore, runs the risk of making the cost for those who have to work from the office even higher, in favor of the narrative,” Crescenzi said.