Some of the top economists in academia and even Wall Street are saying that actually, maybe that’s too harsh. Thomas Philippon, an economist at New York University, argued in a widely read 2022 paper that the modern workforce has been dragging its feet, so to speak, on a slow but continuous basis since the Industrial Revolution. It comes down to a concept called “total factor productivity,” which describes the growth a company or nation can attain without adding to their labor force or capital. In other words, working smarter, not harder. 

A recent Goldman Sachs report, written by a team led by chief economist Jan Hatzius, throws support to Philippon’s claim, pointing out that productivity growth has either stalled or dropped over the past five years. No, it’s not because of remote work, or rapid executive turnover, or any other modern plight. “Trend productivity growth simply tends to fall over time,” the analysts wrote. 

Rather than productivity growing exponentially thanks to newer and more cutting-edge technologies, things actually tend to gradually even off. Per Goldman, the slow drip model rather than the exponential model “helps to explain some but not all of the underperformance of productivity growth over the last 15 years relative to the long-term historical average.” This is the ultimate big, if true, because Philippon looked at hundreds of years of data, going back to 1890, shortly after the Industrial Revolution.

All of this being said, will bosses change their tune on the merits of in-person work? 

The productivity paradox is about way more than where you work—or with what tech

Philippon argues that contrary to the decades-old assumption that total factor productivity (TFP) grows exponentially as human ingenuity expands, it actually grows linearly. He literally looked at data going back hundreds of years, using British TFP data from 1600 to 1914 and American data of TFP from 1890 to the present. If true, this is both good and bad news.

The good news is that the sputtering growth rate of TFP, instead of the warning of plummeting innovation that most economists see, is more like business as usual. As long as annual increments of TFP remain high, slow growth rates are no reason for executives to panic, or cast blame on any modern trend, like work flexibility or job hopping. The bad news, of course, is that the economy has been trickling along for literally hundreds of years, absent a few big bangs.

Goldman analysts tend to agree. And, they add: Beyond just remaining level, workers are actually slowing down. “The historical average rate of TFP growth is not the right baseline expectation for future TFP growth—it is too optimistic,” they write. “Instead, we should expect TFP growth to slow over time.” That, they say, helps partly explain the past 15 years of underperforming productivity growth. 

Goldman Sachs

Generative AI could actually throw a wrench into Philippon’s argument, they write, noting that the productivity gains stemming from AI adoption “could more than offset the underlying growth slowdown over the next 10-20 years.” Still, Hatzius cautions that this is “possible but very premature.” 

Many bosses who are throwing their weight—and dollars—behind AI adoption in their workplaces would certainly hope that’s true. Doubly so if they believe workers are simply working less hard. Earlier this year, per EY-Parthenon research and Bureau of Labor Statistics data, the U.S. underwent five consecutive quarters of year-over-year productivity declines for the first time since 1948. Gregory Daco, chief economist for EY-Parthenon, told Fortune at the time that the productivity drop stemmed partly from remote work—but not entirely. “From our clients across sectors, we hear similar stories of reduced productivity because of the new work environment,” he said. 

But people were also working longer hours (which means higher labor utilization) while barely putting out more work product, which shakes out to reduced productivity overall.

“When you have an environment in which output is outpacing labor growth, that’s an environment of stronger productivity,” Daco said. “When you have the opposite, when output growth is sluggish but labor growth is strong, you have a weak productivity environment.” 

What if generative AI is the real deal?

Goldman’s Hatzius gives himself some wiggle room by noting a flaw in Philippon’s argument: That just because TFP does not grow exponentially, that doesn’t mean it has to grow linearly instead. He agrees with Philippon that “new transformative technologies with wide-ranging effects on the economy like electricity” have come along from time and time and ushered in new periods of faster productivity growth.

Another reason Daco offered for productivity dropping substantial turnover and churn; as a lingering impact of the Great Resignation and quiet quitting, many workers have taken to leaving their jobs after short spells, feeling no sense of loyalty. “Because people were job-hopping so regularly, there wasn’t really a chance to bring them up to the speed, or productivity, that a former worker would’ve had,” Daco said. 

Nonetheless, “the whole idea of remote work and flexible work is to allow people to be more productive.” To that end, a recent report from think tank Economist Impact found that flexible work actually isn’t to blame for workers losing focus—it’s actually a lack of employee choice and “weak infrastructure for a thriving workplace.”  

So even if a hybrid arrangement stands the test of time, the workers who gravitate towards the office are likely to reap the most rewards. “We’ll probably see more weight towards three to four days in the office, rather than one or two, if the labor market slows,” Daco said in May. 

That’s probably unwelcome news to workers who have long maintained that they’re just as productive—if not more productive—at home. But Philippon’s findings, bolstered by Goldman Sachs’s note, should provide a solid source of comfort, and a strong rebuttal to a pro-office boss: Humans have been becoming gradually less productive since Eli Whitney debuted the cotton gin—why, even with rapidly advancing AI, would that change now?

This headline has been modified to remove the implication that Jan Hatzius was critiquing CEOs’ objections to remote work on the grounds of declining productivity. The text is unchanged.