July 16, 2019 Reading Time: 9 minutes

Economists love markets. Markets tell us what to do, in efficient, decentralized ways, through the signals conveyed by prices. Prices tell you how much someone else values the thing you are considering using, even if you don’t know who that person is. But one economist, Ronald H. Coase, had an annoying question, way back in the 1930s. His question was simple: if markets are so great, why are there firms? Later, in 1991, Coase won the Nobel Prize for answering that question: the answer is transactions costs.

The point is that using the market system is expensive. Not prohibitively expensive, or no one would use markets for anything. But expensive. To hear economists tell it in their classrooms, the price mechanism is costless, bordering on magic, telling everyone what to do. If I’m a farmer, my decision on whether to grow wheat or soy will be based on the expected price. That price contains all sorts of information about trends in consumption, weather in other countries, and new regulations, but I don’t need to know any of that. All I need to know is the price.

Outside of our classrooms — and that is where Ronald Coase wanted economists to go; leave your classroom and go out and talk to people! — it doesn’t work that way for many people. Not many of your neighbors come home at the end of a hard day at work and say, “Man, prices were in a bad mood today, yelling. Yelling at me and ordering me around.” Instead, they say that about their boss.

Their boss is a person, a person who works in a firm. A firm is a collection of contracts, some of which are highly specific. But many of those contracts are pretty vague; we call those “jobs.” You go in to “work,” and your boss tells you what to do. It’s possible that your boss is using prices to decide what tasks to assign, but it’s more likely that your boss is getting orders from his boss, who is getting orders from her boss, and so on. Firms are hierarchies, small self-contained networks of command within the larger system of market chaotic and unplanned market processes.

How big are firms? After all, the question can go the other way. We started with “If markets are so great, why are there firms?” But, if firms are so great, why aren’t markets engulfed? Why isn’t there one big firm? A number of countries have tried a version of this, of course. It’s called “socialism,” where market processes are completely replaced by centrally planned, bureaucratic operations. As William Niskanen pointed out, in his famous 1971 book Bureaucracy and Representative Government, some aspects of the workings of firms are much closer to bureaucracy than to markets.

But even short of central planning, why isn’t the most efficient market system actually one big firm? If firms can reduce the transaction costs of chaotic markets, why do we observe markets at all? The answer, as I already hinted, is still transaction costs. One way to think of it is an everyday decision, one that happens so often we don’t think much about it.

It’s the “make or buy” decision. Every firm, on every margin, has to decide whether to purchase that thing or that service from the outside market or arrange contracts to produce it internally. I mean everything. Suppose your firm makes cars. Should it make the engines for the cars, or buy them? There is a large literature on the costs of contracts and the “hold-up” problem with such complicated contracts; we should probably make the engines ourselves. What about the steel and other metals that go into the engines and the rest of the car; should we mine and refine those metals, or should we buy them? How about the bread for the employee cafeteria; should we make the bread for the sandwiches? Should we raise the wheat that goes into the bread?

The point is that the firm has to compare the cost of making the product with the cost of buying the product, at each stage of production. No car company has a production line where a worker attaches two bolts to hold on a fender and then goes on eBay to auction off the partially completed chassis. Instead, the chassis moves smoothly along the production line to the next worker, who connects some wiring and adds insulation. Eventually, the car is offered for sale, and then the price mechanism tells the firm whether it was worth making the car, depending on whether the sale price exceeds the cost of all the workers and suppliers in that nexus of contracts called the firm.

But the signal is not very clear. We don’t get much information about any of the specific steps along the way, and we get no indication about the importance of different components of costs or production decisions. We don’t know if we spent too much on some aspects of production, or if consumers just don’t like the car. It’s hard to know just why a car company is failing, or for that matter why it’s succeeding. That brings us to the animating force of the firm, an entity not present in bureaucracy.

This entity goes by several names; it could be management, or it could be entrepreneurship, or it could be leadership. But the collection of contracts called “the firm” usually contains some contract that involves an attempt to purchase this quality. It’s the boss. And there is a particular structure of ownership and control that aligns incentives in an especially attractive way: make the manager of the firm, the one who gives the orders to the workers, the “residual claimant,” or the owner of all or part of the excess of revenues over costs. Many people call this “residual” profit, but it is not a return to capital; capital is paid off in the nexus of contracts just like the other inputs. No, the owner-manager has an advantage because he or she is most likely to know, and to be able to act on, all the complex local information of time and place that can make a firm a success or failure.

There are many interesting examples of the problem of transaction costs in management; I myself have written about the idea of having no employees at all, just to show that’s a terrible idea. One of the most interesting ideas is the notion that the manager-on-the-spot is really being paid to encourage “team production,” where it’s difficult for anyone, even members of the team, to tell if everyone is pulling their weight.

More than 35 years after Coase posed his question, and gave his answer, two other economists (Armen Alchian and Harold Demsetz) gave a more specific account of the kind of transaction costs that make firms important in markets. They dispute the centrality of Coase’s “order and direction” explanation, and emphasize instead the unique problem of “team production.”

Two men jointly lift heavy cargo into trucks. Solely by observing the total weight loaded per day, it is impossible to determine each person’s marginal productivity. With team production it is difficult, solely by observing total output, to either define or determine each individual’s contribution to this output of the cooperating inputs. The output is yielded by a team, by definition, and it is not a sum of separable outputs of each of its members.” (Alchian and Demsetz, 1972, p. 779; emphasis in original)

They claim that the two key features of the firm both have to do with “metering.” First, team production requires that someone with specialized knowledge observe the activities of workers, since observing output alone tells you nothing about individual effort or skill. Second, this same monitor who specializes in metering inputs also devises production-specific reward and punishment mechanisms that bring individual and group incentives into line. To give the monitor the proper incentives to do this job effectively, the monitor may also be the “residual claimant,” or the person who receives the excess of revenues over costs from the team’s activities.

The Alchian- and Demsetz conception of the firm adds an important element to the problem of division of labor, most particularly in team-production activities. Having a monitor to meter effort, to assign punishment and reward, and to be directly responsible for the result sharply reduces the transactions costs of production. It is not immediately obvious that adding, and paying extra for, a monitor actually reduces costs. But it’s true!

An even clearer example of the “pulling your weight” idea comes from Steven N.S. Cheung, a very gifted economist of the institutional–transaction costs school. I’ll tell a simpler version of Cheung’s account (if you want the full version, the article is worth your time).

Imagine a group of workers pulling a barge upstream. There is a footpath beside the river, and the coolies struggle to pull the large, heavy cargo boat against the current. Walking alongside the coolies is the boss, a man with a whip. He looks at the calf muscles of the coolies, and if he sees them bunch up with strain, he does nothing. If he sees slack calf muscles, he whips the back of the offender.

Now, assume that the coolies are voluntarily employed as barge haulers, and can leave if they want. There are other, kinder freight barge–hauling “firms,” or teams of workers out there. If you don’t like the system with a boss who makes sure everyone is working, you can always find a friendlier team with no whips, and no monitoring.

Of course, the problem with team production is that everyone makes more money if everyone pulls their hardest. But if there are 10 men, plus a boss, I’ll make almost as much money if I hardly pull at all — while grunting to pretend I am pulling — if everyone else pulls. Conversely, if the other 9 don’t pull hard and I do honestly strain for the good of the team, I capture only 1/10 of the extra value I create. Either way, without monitoring you get inferior effort. So it’s certainly possible that a worker might volunteer to be coerced, and sign up with a firm, a set of connected contracts to form a team, with voluntary acceptance of harsh punishment for shirking.

You likely already see the difficulty, though. Who will monitor the monitor? If all of us are busy pulling, we can’t check to see if our teammates are also honestly expanding effort; that’s the point, the reason we need a boss. But if the boss plays favorites, letting his buddies shirk, or if the boss is a sadist, and gets pleasure intrinsically from wielding the whip, that doesn’t solve the shirking problem about effort and it creates a new shirking problem for monitoring.

A clever way to solve the problem is to extend team production, in effect expanding the firm, to “make” monitoring internally rather than to “buy” it by hiring a whip boss. Just take turns. Each of us gets to rest, and wield the whip, for one hour per day. If someone gets a little too happy about whipping, they can expect to be punished in turn. This is an example of “worker democracy,” in a sense, because even the monitoring of team production is handled by the team itself.

That’s why I like the barge-hauling example so much. You could imagine a pure market in these services. A mob of workers press together, all clamoring for a job today hauling a barge. The barge captain walks up and says, “You, and you, and you … big guy in the back!” Different barge captains pay different amounts, and eventually the crews are all filled. Everyone picks up their rope, and away they go.

None of the crews have worked together before, and they don’t expect to work together again. They slog along, no one pulling very hard. But all of them are angry, because they are barely moving, and they won’t make any money this way. The captain will only pay them when they finish the trip, and that could take forever.

But then a firm, or something like a firm, emerges in response to market pressures. One set of pullers contract as a group, and provide a monitor. They might at first rotate the job of monitor, and empower the monitor to use a whip on slackers. Or they might hire a small sharp-eyed fellow to specialize as the monitor, since he can’t pull well but can run back and forth to check on those calf muscles. Hiring the firm will be much cheaper for the barge owner than hiring 10 workers piecemeal, both because the initial contracting costs are lower but also because the firm is also paying the monitoring costs.

In no time, this group, or firm, will have more business than it can handle. They can charge less, and still make more money, because they can finish far more trips per day. Other worker groups will either be forced to form competing firms, or find another job. From the barge captain’s perspective, the owner of the boat who just wants to contract out the “pull me upstream” part of the thing, the transactions cost of signing up one group, instead of 10 or 12 individuals, is much less. The reputation of the firm as reliable may even allow them to raise their price a little, as barge captains in a hurry seek them out because of their higher quality. What this means is that all the explanations that have been offered, from transactions cost to team production to monitoring shirking, can operate at once.

And that’s really the key point here. Firms do not really operate outside the market, but rather are a market-driven process themselves. Certainly there is a command-and-control system of direction within the firm, but that doesn’t mean that market forces are not operating. 

As so often happens, what looks like a market failure (transactions costs, collective-action problem, etc.) is nothing of the kind. The creation of firms as a way of decreasing costs, increasing output, and securing increased profits is itself an important market process. 

Michael Munger

Michael Munger

Michael Munger is a Professor of Political Science, Economics, and Public Policy at Duke University and Senior Fellow of the American Institute for Economic Research.

His degrees are from Davidson College, Washingon University in St. Louis, and Washington University.

Munger’s research interests include regulation, political institutions, and political economy.

Books by Michael Munger

Get notified of new articles from Michael Munger and AIER.