In 2008, I wrote a short article on an old question. The question was first asked, at least in concise form, by Ronald Coase in his 1937 paper “The Nature of the Firm.” Economists generally focus on the power of the price mechanism to organize highly decentralized yet efficient cooperation. That is, prices tell us what to do, because prices embody all the information about the relative scarcity and the opportunity costs of resources.
But workers don’t come home at the end of a long day and say, “What a day! Prices were in a really bad mood today, and I really had to work hard to please them.” No, workers say that about their bosses, not prices. Most people, most of the time, are working for a boss, in an institution called the “firm.”
Coase noticed that firms exist precisely to suppress the market mechanism based on prices. The reason, according to Coase, is what he called “transaction costs,” or costs of the use of the market. If I am working on a production line, and attach a flange to an assembly using a bolt, I don’t then look for a buyer of the unfinished widget. Instead, the production line moves the widget to the next worker on the line; she attaches a pulley, or a wire, or whatever the boss has ordered.
It’s true that prices are still operating, in the background. The costs of the parts, the wages of the workers, and the prices at which widgets are ultimately sold to consumers combine to signal a profit, or loss, on the operation. But the individual parts of operation are not separately priced. The boss has to decide whether each of the many workers is carrying his or her weight, and adding to or subtracting from the profitability of the firm.
In such a setting, the boss also must try to figure out if there is enough cloud storage or computers for the workers. This equipment is purchased on the market, so again there are price signals operating. But the boss does not sell the computer time to the staff. He gives it away. Is it worth it, in terms of somehow increasing profits or raising the company’s share value? The boss has to guess. The same is true for the accounting group, and housekeeping.
All those bosses, in all those different firms, make wildly varied guesses about how many people to hire, and what to tell them to do. Some of the firms are profitable, some are not. No one knows what specific choices led to increased profits, or to bankruptcy. But all firms face similar “make or buy” choices: should we make this ourselves, or buy it in the market? The company may well decide to buy the machines and hire workers to make the metal housing for the product they sell; the company probably does not make its own wire or fuses; and the company almost certainly does not own the wheat farm to obtain the flour used to make the bread for sandwiches in the employee cafeteria. In short, all firms, without exception, face the “make or buy” choice, and they make different decisions depending on the level of transaction costs, at the margin, for each activity.
You might imagine that one day, in one firm, one manager, perhaps on a whim, outsources the computer services or janitorial services or the legal-advice department. This outsourcing is not a physical transfer to India or Ireland; I just mean that the boss ends the internal, firm-based contractual relationship to employees and instead acquires the services using the market. The boss buys services or inputs on the spot market, after taking bids from several companies. The advantage is obvious: these bids reflect direct competition with other companies providing similar services, companies forced by the direct glare of the market to provide excellent service at low cost.
By looking at the different prices in the bids offered in this competition, the boss learns something: what the cost of providing the service is. Until now, because the work was done in-house, he didn’t really know. It’s hard to fire employees, particularly since most employees are smart enough to work hard enough to get acceptable performance reviews. It’s possible, though it’s not obvious, that the boss also has a hard time motivating the in-house staff, because watching each employee is expensive and tiresome. But it’s easy to fire contracted employees, because you just sign a new contract with a competitor. Why not let the market system do your motivation work? Let’s suppose that our boss sees the company’s profits rise dramatically after outsourcing legal and janitorial services, and that the stock price goes up 18 percent in six months. Life is good, for the boss.
So, one day the boss has this crazy thought. He asks himself a question that has never occurred to him before: why have any employees at all? He never really liked those people anyway, always asking for direction and needing resources to do their jobs. For that matter, why have a building? Why not just sit at home, wearing his jammies and bunny slippers, sipping a nice cup of tea, and outsource everything? He can write contracts to buy parts, he can pay workers to assemble the parts, and he can use shipping companies to box and transport the product.
The boss fires all his employees, and takes bids on all the design, parts manufacture, assembly, and shipping that those people used to do. The next morning, he sleeps late, has a big breakfast, and makes himself some espresso (he grinds his own Esmerelda Geisha beans, which he buys online). Then he puts his bunny-slippered feet up on his desk at home. Firing up his computer (he uses Gmail, because he cancelled the proprietary web service for the company), he notices he has 78 new messages. He checks his cell phone, and sees he has missed 23 calls and 14 texts. What the heck?
Ronald Coase was right, after all. There are “transaction costs” of using markets, and those include not just finding the right good or service and negotiating a price, but also monitoring the agreement for delivery and compliance, and enforcement of contracts or agreements that are breached. Transactions are not self-enforcing. It turns out that organizing all these new contract employees, and coordinating all those different transactions to work together in time and space, is a really hard job.
The manager shucks the bunnies, puts on his suit, and hurries over to his primary contract supplier of inputs. “What are you doing? We have a contract!”
The supplier says, “Who are you again?” The boss gives his name, frantically. Not one of the other contracts can be fulfilled if he can’t get these parts.
The supplier checks. “Oh, yeah, I see the order. Sorry; we’re running behind. We’ll have the stuff to you by the end of the week, maybe the following Monday.”
Within three days, the boss is fired, because the company is plunging into bankruptcy. Stock prices fall by 75 percent; the newly hired boss makes a desperate and only partly successful attempt to rehire the old employees, the ones who knew how everything actually worked, back to their old jobs.
2019: The Return of the Bunny Slippers
As I said, the story above is the one I told, at greater length, in 2008. What about today? Have things changed?
Yes, they have, though it’s hard to say just how much. If Coase were alive today, he might ask a different question: why do we own, when we could rent? But the answer — because it’s always the answer, to every important question in economics! — would be the same: transaction costs. If I need two holes in this wall, right now, what’s the cheapest and most convenient way for me to get that done? Owning a drill, which I also have to pay to store. But if there were a fast and cheap way to get a rental drill delivered, and a way to ensure trust in that transaction, I wouldn’t own a drill, or for that matter any other tools. Long-term employment contracts are the answer to a problem: how do I get this job done on the production line, while ensuring that all sorts of other complex tasks can also be performed if I need those done instead? It is complicated and expensive to hire a worker who has to schedule the job and then travel to the work site. Employees are the answer.
Until they aren’t, because the new sharing economy has reduced the transaction costs of “renting.” As I discussed in my 2018 book, Tomorrow 3.0, there are millions of new firms in the U.S. that have been created in the last 10 years. But 80 percent of those “firms” have zero employees. That is, the firm is just the owner, who works under market contracts rather than working as an employee. Some days, at least, those bosses likely wear bunny slippers and stay home, managing their firm from a laptop on the porch.
And far more of those who are employees are temporary or contract workers. If this movement continues, the very notion of a “firm” may be blurred. A group of people, each with specialized skills and a ratings-based reputation on LinkedIn, would be hired for a project. At the project’s completion, the group would break up, only to re-form anew in kaleidoscopically different combinations of workers and projects.
Hollywood films, for example, were once made by the major “studios” such as Metro-Goldwyn-Mayer or 20th Century Fox. These “studios” now are distributors, and movies are made by “gig” workers, hired for the duration of the shooting of the film. There are about 150 different disciplines involved in making a movie, all those job titles you see at the end of the credits, including the gaffer, the key grip, and the appealingly titled “best boy.”
If you make a movie, you go to LinkedIn and choose one of each of these workers. On the first day of shooting, the team works well together because the refinements of division of labor in the industry are clear and well-organized. After the film is completed, the gig is over.
Adam Davidson describes this “Hollywood Model” of gig employment in a 2015 New York Times article:
Recently I visited a movie set. It was the first day of production, and I arrived just as the sun was coming up, but already, around 150 people were busy setting up that day’s shot in an abandoned office building. Crew members were laying electric cables and hanging lights.… Carpenters were putting the finishing touches on a convincing prop elevator — I pushed the call button and waited, until I finally realized it was a fake. [I tried] to figure out something that mystified me as the day went on: Why was this process so smooth? The team had never worked together before, and the scenes they were shooting that day required many different complex tasks to happen in harmony: lighting, makeup, hair, costumes, sets, props, acting. And yet there was no transition time; everybody worked together seamlessly, instantly. The set designer told me about the shade of off-white that he chose for the walls, how it supported the feel of the scene. The costume designer had agonized over precisely which sandals the lead actor should wear. They told me all this, but they didn’t need to tell one another. They just got to work, and somehow it all fit together.…
A project is identified; a team is assembled; it works together for precisely as long as is needed to complete the task; then the team disbands. This short-term, project-based business structure is an alternative to the corporate model, in which capital is spent up front to build a business, which then hires workers for long-term, open-ended jobs that can last for years, even a lifetime.
In a way, none of this is surprising: Specialization and the gains from division of labor are limited by the extent of the cooperation horizon. If the transaction costs of extreme specialization can be reduced, the scope for innovation is expanded. Hollywood has full-time professional “bug wranglers” to manage scenes with bugs as “actors.” No studio might be able to afford a full-time bug wrangler, so under the old system the job did not exist. But in the gig economy, with a thousand films all being made simultaneously by an ephemeral but highly organized team of the sort Adam Davidson describes, there may be enough work to keep several bug wranglers busy.
There are some clear advantages to the gig economy, for both workers who have specialized skills and for companies trying to remain small and nimble. Coase said that firms only exist if they are lower-transaction-costs means of organizing production. But that formulation is consistent with the existence of “firms” that have exactly one employee, who is also the CEO. Firms may be able to rent capital equipment and labor for very short periods, increasing the productivity of the workers for the period that they are employed and dramatically reducing the fixed costs of the firm. In the limit, firms themselves might simply become individuals or small teams that hire out for specific projects. Workers in this system would be private contractors, not “employees” in the traditional sense.
And everywhere, for everyone, there may be a lot more time for bunny slippers and excellent espresso.