In Praise of the Office: The Limits to Hybrid and Remote Work

By Peter Cappelli & Ranya Nehmeh

10 min read

Why does it matter? Because the remote-work revolution may be the best deal employers have ever offered — and convinced employees to pay for.

In 2005, Google was the church of the office — free gourmet meals, nap pods, bring-your-dog Tuesdays, all of it engineered to keep employees on campus as long as possible. By 2021, Google was offering hybrid schedules, permanent remote options, and four weeks of work-from-anywhere annually. The obvious interpretation: the pandemic proved the office was always a fiction. Peter Cappelli, a Wharton management professor, thinks that's exactly backward. What actually happened, he argues, was a perfect storm of crisis solidarity, unprecedented government stimulus, and employees so grateful to keep their jobs that they worked harder than normal — none of which will repeat. The pandemic didn't run a clean experiment. It ran a disaster drill and called it science. This book is the fine print on every glowing work-from-home survey you've read — and it will change how you think about the deal you've been offered.

The Pandemic Was Not a Remote-Work Experiment — It Was a Crisis With Exceptional Side Effects

The pandemic did not prove that remote work works. It proved that people will do almost anything to keep their jobs during a crisis — and those two findings are easy to confuse.

Consider the methodological trap embedded in every upbeat survey from 2020. The Adecco poll of 8,000 office workers found improved quality of work and well-being. A PwC study found 83% of employers calling the remote experiment a success. But the same period that produced those positive reports also saw 20 million Americans lose their jobs in a single month, with roughly one-third of the labor force unable to work at some point during the pandemic. The people answering those surveys were acutely aware that their employment was a lifeline, not a right. When an employer extends unusual trust and accommodation, employees feel a powerful obligation to reciprocate. The performance boost, to whatever extent it was real, may have had nothing to do with working from home and everything to do with the fear and gratitude that came with it.

Essential workers make this case quietly. Warehouse staff, grocery clerks, and delivery drivers also performed well during the same stretch — and they were in the building. If location were the real variable, their results would look different. They didn't.

The hour data is where the story gets uncomfortable. One careful study of skilled IT workers found that measured output held roughly steady through the remote period — which sounds like a win until you notice that hours worked climbed about 30%. Productivity, defined as output per hour, actually fell. The employees were producing the same amount of work; they were just spending significantly more of their lives doing it. That's not evidence that remote work matches office performance. It's evidence that people worked longer to compensate for something.

Mark Zuckerberg put his finger on the deeper uncertainty when he wondered aloud whether his teams were simply 'drafting off' existing bonds — performing well by drawing on relationships built before the shutdown began. Goldman Sachs CEO David Solomon was blunter, calling the whole arrangement 'an aberration.' The pandemic gave remote-work advocates a mountain of survey data. What it did not give them was a clean comparison — no control group, no baseline performance tracking, no way to separate location from crisis. The fine print says the experiment was never really designed to answer the question everyone is now using it to settle.

Jay Chiat Tried This Already — and Employees Ended Up Storing Files in Their Cars

Jay Chiat ran one of the most celebrated advertising agencies in America, and in the early 1990s he decided to blow up the office entirely. His logic was clean: physical space was wasteful and deadening to creativity. So he handed everyone a laptop and a cell phone, stripped out the private offices and individual desks, and replaced them with open clusters of couches, shared tables, and — this part is not a metaphor — domed Tilt-A-Whirl cars salvaged from amusement parks, where employees could perch and brainstorm. Architecture critics were delighted. Employees were not.

Without assigned spaces, workers quietly declared their own. The best couches got claimed early and defended. Supplies disappeared into personal stockpiles because returning them meant losing them. In the Los Angeles office, the situation reached an almost comic extreme: people started keeping files in their cars, running out to the parking lot mid-meeting to retrieve what they needed. When the agency was eventually sold, the new owners moved with notable speed to install desks, landlines, and private offices. The experiment lasted about as long as it took to figure out it wasn't working.

That was 1993. What's striking is that the same cycle has run at least twice more since then. At the turn of the millennium, forecasters were confident that a quarter of the American workforce would be telecommuting by 2006 — and the Wall Street Journal was already reporting employer pullback before the decade turned, the tight labor market that had made remote work a recruiting tool cooling fast and taking the enthusiasm with it. The pattern was consistent enough that by the mid-2000s, hoteling and virtual work had simply, as one industry postmortem put it, "vanished."

The pandemic arrived and the cycle restarted, this time with better video software and a genuine public-health rationale. Both things are real. But the underlying pattern — genuine excitement, confident rollout, quiet problems, eventual reversal — has shown up before. What changed with COVID-19 is the scale and the visibility, not the fundamental dynamics of why people do and don't work well when you take away the shared physical space. Chiat's employees didn't start storing files in their cars because they lacked willpower. They did it because the design of the work environment stopped matching how work actually happens. That problem didn't get solved between 1993 and 2020. It got a better webcam.

The Most-Cited Pro-WFH Study Is Quietly a Warning

The Stanford call center study is the one advocates reach for first. Researchers ran a genuine randomized experiment at a large Chinese call center — half the workers sent home, half kept in the office — and found that home workers were 13% more productive. That number travels well. It's clean, it comes from a real experiment, and it gets cited in countless op-eds as proof that remote work delivers. But the same research team kept watching, and what they found afterward doesn't make the highlight reel.

When the company later hired workers directly into remote roles — people who had never sat in that office — their performance ran 18% below their on-site counterparts. The original productivity bump belonged specifically to workers who already knew the job, the culture, and their colleagues. Strip that context away and the number flips. The remote workers in that study were also 12% less likely to advance than the people who stayed on-site. The productivity gain was real. It went to the employer. The career penalty was also real. It went to the employee.

A nationally representative study tracking American workers across more than a decade found that telecommuting didn't reduce the hours people spent working in the office. It extended total working hours instead. Employees weren't substituting home time for commute time. They were adding home time on top of everything else.

Remote Work Saves Your Employer Money. The Career Costs Are Yours to Keep.

The benefits of remote work split cleanly along organizational lines — and the split rarely gets named. Employers pocket the savings. Employees absorb the costs.

The arithmetic is straightforward once you look at it. When a worker shifts permanently to remote, the employer recovers a desk, a parking space, a portion of the lease. The US Patent Office calculated that its work-from-anywhere program saved $120 million in a single year. That money went to the balance sheet, not to the employees who made it possible. Meanwhile, a second call center study found that workers would accept about 8% lower pay in exchange for the right to work from home. Employers noticed. The compensation discount is already being built into offers, quietly converting a flexibility benefit into a wage subsidy flowing in one direction.

The career side of the ledger tells the same story. Managers consistently attribute stronger qualities to the employees they can see — not because they are more rigorous evaluators in person, but because visibility does actual cognitive work. The result shows up in promotion data: UK government figures from before the pandemic found remote employees were 40% less likely to advance than their office-based peers. Remote workers understand this, which is why research finds them compensating with extra effort — accepting undesirable assignments, attending meetings at inconvenient hours, volunteering for work nobody asked them to do. They are paying a hidden tax on their own time just to stay legible to management, and they still fall behind.

The WeWork CEO, whose business depends on companies renting offices, made the structural logic explicit when he called remote employees the least engaged workers in any organization. The Washingtonian's editor was blunter still, writing in a widely read essay that it is simply easier to lay off people you do not see regularly. She published that while her own staff was working remotely. They walked out for a day. Both statements were ungraceful. Neither was wrong about the underlying reality.

The honest frame is not whether remote work delivers productivity gains. Some of the time, for some roles, it does. The frame is: who keeps those gains? The employer gets the rent savings, the wage discount, and the extended hours. The employee gets flexibility — and hands back visibility, advancement, and leverage without always realizing the transaction is happening.

The Wyoming Test: Why Your Zip Code Shouldn't Determine Your Salary — But Will

Imagine your company asks a consultant to redesign its supply chain. She does excellent work. At billing time, your finance team proposes paying her less because she lives in Tulsa rather than New York. The hourly rate was always a market price for her expertise — her mortgage has nothing to do with it.

That logic applies to your own employees, too. Yet Facebook, Twitter, and Stripe all announced during the pandemic that remote workers who relocated to cheaper areas would see their pay adjusted downward to match local cost-of-living levels. Stripe's version was almost elegant in its construction: the company offered a $20,000 bonus to anyone willing to leave San Francisco, paired with a permanent 10% pay cut. Do the arithmetic and the bonus runs out in roughly two years — after which the employee is simply earning less, indefinitely, for the same work.

The incoherence runs deeper than the math. Programmers in Silicon Valley aren't paid generously because their rent is high. They're paid generously because the skills they have are scarce and the market competes for them. The causal arrow points the other way: those cities are expensive because high earners have spent decades bidding up housing. Cutting pay when someone moves to Sacramento doesn't reflect a change in what their skills are worth — it captures the savings for the employer. That logic also only runs in one direction. No technology CEO working remotely from a ranch in Wyoming has been asked to accept a salary adjustment to reflect Teton County housing prices. When the policy would apply to executives, it doesn't seem to come up. That asymmetry reveals what the policy actually is: not a neutral accounting of market conditions, but a mechanism to extract value from employees who have already shown, by accepting the deal, that they'll pay a price to live somewhere they prefer.

The Flexibility Deal Has a Trapdoor: From Remote Employee to Contractor

What happens to a remote employee who stays remote long enough? The answer is already visible — not as a prediction but as a documented trajectory.

Start with how organizations respond when trust erodes at a distance. Without the social fabric of shared space, managers default to the only substitute they can see: output metrics, check-in calls, and eventually software that watches the clock. Teleperformance, a French company managing 380,000 workers across 34 countries, announced in early 2021 that it would conduct random webcam scans of remote employees. Workers who needed a break had to enter a formal 'break mode' and document why. Eating during a shift was prohibited. Cappelli doesn't present this as a scandal. He presents it as a destination — the predictable endpoint of a low-trust remote arrangement that was never redesigned for the distance. Monitoring defeats the flexibility that made remote work attractive, breeds resentment, and severs the sense of mutual obligation that holds employment together. Once that's gone, what remains is a relationship that looks increasingly like contracting: specified deliverables, explicit terms, no expectation of investment from either side.

Permanently remote workers are already structurally closer to contractors than their in-office peers realize. They receive fewer promotions, face higher layoff risk, and operate outside the informal flows of information and sponsorship. Dropbox recognized this clearly enough to reject the two-tier hybrid model entirely. The company's reasoning, as the author reads it, was that permanently remote employees would inevitably become second-class citizens inside a split organization — so rather than create that divide, Dropbox went fully virtual. Most organizations won't make that call. They'll keep some people in the office and some out, and the ones out will find their employment gradually hollows: fewer opportunities, more monitoring, and eventually a conversation about whether their role might be better structured as a contract engagement.

The flexibility deal was always this: you get to choose where you sit, and in exchange, you become easier to reprice, easier to overlook, and easier to cut. Teleperformance just posted the destination.

The Office Was Never Perfect — But You Should Know What It Was Actually Doing

The office was a control mechanism — but calling it that is like calling a classroom a detention room. Technically accurate, completely misleading about what was actually happening inside.

Organizational culture is the set of unwritten instructions that tells employees how to behave when no rule covers the situation and no manager is watching. You learn it the way you learn an accent: by immersion, by watching people around you, by absorbing what gets rewarded and what gets quietly punished. Strip away the physical presence and that transmission largely stops. A fully remote new hire is, in functional terms, closer to a contractor than an employee — you can hand them a document about company values, but you can't put them in a room where those values are being modeled.

The British Chartered Institute of Internal Auditors polled its members and found something specific worth paying attention to: fraud detection depends on informal proximity-based channels — the kind where someone notices, in passing, that a colleague is acting strangely, spending oddly, avoiding certain conversations. Those signals don't travel over Slack. Neither does the low-level social awareness that tells you something is off before you can articulate why. The warning isn't nostalgia for cubicles. It's about the particular work the office was quietly doing that video calls don't replicate automatically.

The pandemic did produce one genuine management improvement worth naming — and it's at serious risk of disappearing. Supervisors, unable to use visibility as a proxy for performance, were pushed in some organizations to check in weekly with remote employees: a simple practice, asking how work was actually going, that turned out to matter enormously for engagement and catching problems early. It worked. But when offices reopen, managers tend to assume that seeing people means knowing how things are going for them. That assumption was always wrong. The pandemic just made it temporarily impossible to hide behind.

The honest assessment isn't that the office was good and remote work is bad. The office was doing several jobs at once — transmitting culture, generating informal learning, creating social accountability — and most organizations never had to think carefully about any of that because proximity handled it automatically. Now that proximity is optional, someone has to replace those functions deliberately. Skip that work, and you don't get freedom. You get the Teleperformance webcam.

What Teleperformance Should Make You Worry About

The path of least resistance is hard to resist. The real-estate savings were easy to take. The culture investment was easy to skip. And when you skip it long enough, you end up needing a way to know whether 380,000 people are working — and you reach for the only tool you've actually built, which is surveillance. Teleperformance's webcam scans weren't designed by sadists. They were designed by managers who had never built anything better. That's the logic the book leaves you with: monitoring isn't a personality defect, it's what fills the vacuum when trust infrastructure doesn't exist.

The question worth carrying forward isn't whether you prefer the office or your kitchen table. It's whether the organization you're part of has done the harder work. And the harder work has a pretty unglamorous face: it's the manager who blocks off thirty minutes every week to ask each direct report what's unclear, what's stuck, and what they're not saying in the group chat. That weekly check-in — the one discussed in Section 7 — isn't a perk or a retention tactic. It's deliberate culture transmission, which is just a fancy way of saying someone decided to make the invisible visible on purpose, before a webcam scan had to do it for them.