(Bloomberg View) —It’s been more than 10 years since I last worked in an office on any regular basis, and remote work has worked for me and my employers. And yet, when I read that Apple and IBM were moving away from telecommuting and toward more traditional office-time requirements, my first thought was: “What took them so long?”

Don’t get me wrong; remote work has real benefits. I shave two hours of commuting off of every workday, time that I can instead spend getting work done.  Early in my telecommuting career, in fact, I had the following conversation with a manager who wanted me to spend more time at the office.

“I’ll be happy to. But I’m already working more than 12 hours a day, so my commute is going to have to come out of my work output, not my personal time.” (Pause) “What do you want me to do?”

“Enjoy your home office.”

These benefits are obvious. And thus, as far back as the science-fiction stories of the 1950s, people have been predicting that telecommunications would one day take the place of face time and cubicles. Yet these expectations have been steadily disappointed by reality. It turns out that some kinds of information travel very well by wire, but others get lost in transmission. It also turns out that those kinds of information are often vital to a company’s work.

To understand why, it may help to go back to the theory of the firm, and a question that economists have struggled with: Why do companies exist? Why don’t we all act as free agents, bidding our services out in the marketplace, rather than binding ourselves into subordinate relationships with larger entities?

There are a lot of answers to that question, but one of the biggest ones, provided by the eminent economist Ronald Coase, is “transaction costs.” Paying a lawyer to write you a contract is a transaction cost. So is the time you spend finding someone to contract with. If the transactions costs are too high, then deals cannot be profitably done.

Firms are often a good way to solve the problem of transactions costs. Because everyone involved is there for the indefinite future, they don’t have the same trust problems that come from doing one-off deals, and managers don’t have to keep going to the trouble of finding labor and negotiating every time they want something done.

Firms have inefficiencies, too, of course, because they have to manage all that labor (and often, to keep more around than they may need at any given moment). But they are so good at reducing transactions costs that they are still, in many cases, more efficient than simply bidding every single service on the open market.

But one of their most effective means of reducing transaction cost is that most elusive of business-journal ideals: corporate culture. When job performance is difficult to specify in minute detail — as one would for a contractor — corporate culture is what ends up determining how hard your employees work, how far they will go out of their way to help out a co-worker in trouble, what lengths they will go to in order to satisfy customers.

This culture cannot be transmitted by writing it all down in a manual somewhere, or exhortatory speeches by managers; it is transmitted in a thousand little interactions that show more than they tell. This is exactly the sort of information that gets lost if your employee’s interaction with the firm consists largely of daily video chats.

Then there’s the problem of transmitting other kinds of information. It’s easy enough to send a document or a spreadsheet from headquarters to a remote worker and back. The real obstacle is how to transfer the stuff that you don’t put into those confidential documents, a million little bits of knowledge about the markets you compete in, the firm’s challenges, the changes going on in management. Call it “sub-information”: the stuff people don’t even really know they know.

Electronics are a remarkably effective barrier to this sort of information. Humans are evolved for face-to-face interaction, and something about electronics stiffens us, turns us more formal and less social, even when we are still sitting in the same room. And when we’re miles apart? Forget it. We transmit our conscious knowledge, but leave out all the little things that only come through face to face, in casual conversation, unplanned and unintended, but nonetheless, the lifeblood of a firm.

When I lived in New York and was immersed in the financial capital of the world, I used to make a lot of confident and wrong predictions about politics, while tearing my hair out in despair at the stupid things Washingtonians said about finance. Then I moved to Washington, and realized I was rapidly losing my savvy about the financial industry, while gaining new understanding about politics.

That’s not because I’m constantly circulating through Georgetown cocktail parties, or taking important meetings with high-level political figures, which I mostly don’t. No, I gain this sub-knowledge at dinner parties with mid-level civil servants, from other journalists in the quiet moments before think-tank panels begin, from little asides in sit-down interviews about something else. I can type anywhere, but the job that I do, the way that I do it, can be done from only one city on this earth.

And so with firms: Each big company is a sort of little city unto itself. If the city tries to scatter itself to the four winds, the traffic stops, and the city starts to die. No wonder these big companies are starting to recall their residents.

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.