Nicholas Bloom, Stanford University economist and cofounder of WFH Research—the preeminent group that’s been tracking remote work attitudes and statistics across dozens of countries—is the authority on the matter, but even he doesn’t know. He guessed that countries with the highest density of knowledge workers—often high earners and remote-capable—would have the highest share of remote work. 

But when he set out on his latest round of research, he discovered he was wrong. In his latest working paper, titled “Working From Home Around the Globe,” Bloom and his team found an even odder indicator of remote work occurrence: national language, which plays a bigger role than national income. And countries speaking one language in particular, English, lead the charge. 

Rates of remote work are above average in English-speaking countries. While the average employee across the 34 countries WFH Research observed works remotely 0.9 days per week, that figure jumps to 1.4 days in the U.S., the U.K., Canada, Australia, and New Zealand.  

Workers in each of these countries work from home in greater numbers than in other similarly wealthy countries where English isn’t spoken, such as Germany (one remote day per week) and South Korea (0.4 days). But as to why this is the case, Bloom and his fellow researchers are perplexed. While they’re continuing their investigation, he told Fortune on Monday, he does have a few conjectures, from infrastructure to apartment size. 

Why (we think) the English-speaking world leads the WFH charge

Bloom’s best understanding: The unique management and culture in the U.S. and wealthy English-speaking nations fosters a strong, remote-capable environment. 

More to the point, he said, English-speaking countries skew wealthy. As prior research has found, wealthy countries and their wealthy inhabitants tend to work jobs more amenable to flexibility. But then again, he pointed out, Japan is wealthy, too, and remote work is fairly rare there. “So that’s not all of it,” he said. 

Another potential factor is infrastructure. Residents of English-speaking countries often live in larger apartments or houses, which makes remote work doubly appealing compared with cities with less space per person. And while that could be a major factor, Bloom noted that the U.K. has “a pretty high population density, similar to Northern Europe,” yet a much greater incidence of WFH. Also hard to overlook is New York City, the financial capital of the world, where apartments are notoriously cramped, yet office buildings remain half-empty. 

Then comes the question of lockdowns. Many Asian countries emerged from the total lockdown phase of COVID quickly, and thus spent fewer months toying with different remote work arrangements. As a result, returning to the office in full force is less of a surprise. On the other hand, many English-speaking countries had more intense lockdown experiences, which pushed more people into adopting WFH. This may have been signified by more home office investments or improved WFH management practices by employers. Again, an asterisk: China, Bloom points out, had aggressive lockdowns “and not much WFH now.” 

That brings Bloom back to management and culture, which he ultimately thinks is the deciding factor. Countries with ample remote work tend to have strong performance management cultures and decentralized practices, he said. This is critical for a successful remote work arrangement, which requires trust and communication in equal measure. Indeed, Bloom’s research from back in 2012 found that better managed firms have more decentralization, or “no boss to watch you all the time.”

For better or worse, the U.S. tends to be a harbinger of workplace trends, wrote Pilita Clark of the Financial Times. English-speaking countries tend to mimic the way U.S. companies conduct business quickest, and non-English-speaking countries follow along a bit later down the line. That’s why Bloom expects most of them to reach levels similar to where the U.S. is now, despite their paltry WFH figures. Per data Kastle Systems provided exclusively to Fortune, the average office in the top 10 major U.S. metro areas was 49.2% full last week. 

For bosses, that looks half empty, but workers who are insistent on maintaining their flexibility would argue it’s half full.