Organization Design

Organization Design

Colin F. Camerer, Caltech

Version 7 November

Chapter 3: Firms

What is a business firm? A firm is a legally-defined entity which is given a status like a person, for the purposes of contracting. From an economic point of view, a useful way to think of a company is that is a “nexus of contracts” (in Jensen and Meckling’s elegant terminology) among various stakeholders. The firm’s legal status gives it the legal right to contract with various parties—regulators, litigants, workers, stockholders, and sometimes communities—without having to link every party in the nexus to every other party, which would be a headache that would slow things down and enrich only lawyers.

A thought experiment that illustrates the utility of the nexus-of-contracts view is to imagine that a firm could not legally incorporate. What would happen? For example, suppose Intel was not a legal corporation, but instead a web of private contracts. The managers of Intel would have separate employment contracts with individual workers. Since it takes a lot of capital to guarantee workers pay, and build factories, the managers would take out loans or sell shares in the profit stream of the firm to individual shareholders. Bank accounts would have to be set up in each manager’s name, or joint accounts where some subset of managers have the right to deposit checks and withdraw money. Vendors who sell materials to Intel would have separate deals with managers and would get paid from the checking account.

What if a manager left Intel? There would have to be some mechanism for the remaining managers to take over the departing manager’s obligations. If vendors, workers and customers did not know what that mechanism was, they might be reluctant to accept promises of any single manager.

In this fictional world where “Intel” was not a legal entity, people would spend an enormous amount of time contracting, bargaining, recontracting, writing checks, and so forth. All these costs of transacting are economized by defining a fictional legal entity, called “Intel”.As the economist Dennis Robertson wrote, firms are like “coagulated lumps in a pail of buttermilk”.

The fictional entity’s core activities are controlled by managers. Checks are written from an Intel corporate account; receipts are deposited into that account. If regulators like the FTC think Intel has defrauded customers by claiming its products are better than they really are, customers can sue Intel or the FTC can regulate Intel, the corporate entity. What keeps the managers from just creating a shell company, collecting as much money as they can, and producing crappy products and not paying workers? The corporate entity is overseen by a board of directors who are elected by shareholders and who are, typically, legally liable (to an increasing extent) if the firm harms somebody. The whole system works rather well if buyers are wary (the doctrine of “caveat emptor” in consumer law—buyer beware) and workers and customers are too. Furthermore, having a corporate entity enables the corporation—the web of contracts—to outlive the lives of the managers, creating a repository of reputation and goodwill. Knowing that your Intel contracts will be upheld even if one manager or another leaves—provided the outgoing managers have an incentive (and honor) to choose a good successor-- also creates repeated-game incentives which can enforce good behavior in the face of short-run temptations to cheat.

I. Firm boundaries: Make or buy

The nexus-of-contracts view is vague on an important point—which agents are “in” the firm and which are “outside”. That is, what is the economic boundary of the firm?

This apparently simple question has been widely debated for many decades. Different theories emphasize different aspects of firm integration.

One line of reasoning is that firm boundaries reflect an effort to align ex ante incentives— that is, to create a good arrangement before anybody shows up at work or invests in relationship-specific capital. The “decision rights” and “incentive systems” theories emphasize ex ante incentive alignment.

A different line of reasoning is that organization of firms is designed to govern decisions efficiently ex post— that is, after a deal is struck and an incentive to cheat (moral hazard) springs to life.

Before proceeding, let’s think about some concrete examples of integration. We will talk about the “make or buy” decision—does a firm want to contract with an inside worker for a product or service (“make”), or “buy” from an outsider? Table MOB below gives some examples.

One thing that is notable about the examples above is that both make and buy are often used simultaneously. Most magazines have both staff writers, who are assigned to stories and paid an annual salary, and freelancers who pitch assignments and are paid a kill fee (if the negotiated story is not accepted) plus a per-word fee. Similarly, most franchise chains have both company-owned stores and some stores that are owned by franchisees, who pay an upfront fee for the franchise as well as a percentage of revenues. In the company-owned stores, the managers earn a salary; the franchise-owned stores are often managed by the owner, or by a manager who earns a bonus based on revenue or profit. Similarly, wealthy people who have personal chefs also go out to restaurants on occasion.

Table MOB: Some examples of make-or-buy decisions

Exchange partners / Buy from the market / Make internally
A-B / B has residual decision rights, bears income risk / A has residual decision rights, pays B a fixed wage and so A bears income risk
Landowner-sharecropper / Rent land (fee + % crop yield) / Hire sharecropper (pay regular wage)
Restaurant-grower / Buy herbs from farmer / Own herb gardens
Distributor-movie theater / Consign movie to theater (theater pays fee + %gross) / Own movie theater
Chain-franchises / Franchisee (pays upfront fee + % revenue) / Company-owned store (manager gets fixed wage)
Magazine-writer / Freelance journalist (paid “kill fee” + $/word) / Staff writer (paid regular wage)
Client-chef / Eat at a restaurant (pay $ per meal) / Hire personal chef (pay regular wage)
Company-lawyers / Outside counsel (lawyers bill by the hour) / Inside counsel (lawyers paid regular wage)
Grower-picker / Spot picker paid by the hour, bushel, or pound / Picker paid regular wage
Company-consultants / Management consultant (pay per hour, or total job fee) / Internal planning staff (pay regular wage)
Movie star-concierge / Rent as bundle with hotel (pay per night) / Staff assistant (pay regular wage)

The fact that make and buy often coexist is really important. It means that choosing whether to make or buy is not a matter of choosing the obviously optimal organizational form. Instead, firms choose a make-or-buy mixture. The fact that a mixture is optimal might reflect several factors. Optimality of a mixture could reflect individual differences—some fast-food restaurant managers are highly motivated by profit and like personal control, so they scrape together capital to buy a franchise, while others like a regular salary and don’t mind being told what to do by a corporate boss. Optimality of a mixture might also come about because firms use what is learned from one kind of store to inform them about how to manage the others. For example, suppose you have a fast-food chain and aren’t really sure how much a single store should earn. The company-owned stores create a benchmark for the franchisees and vice versa. If franchisees are doing better, that provides information that company-owned managers are not doing a great job.

Besides mixtures of made-and-bought goods and services, there are interesting hybrid organizational forms in the economy. Often a broad firm boundary exists around a bunch of economic activities, and coexists with a different arrangement in which the economic activities are controlled by separate people or smaller firms inside the boundary.

An example is general contractors who work on houses. If you have work done on your house, you probably hire a general contractor who has access to large pools of workers with specialized skills (floors, roofs, carpentry, painting…). Most contractors do not actually employ works who are paid a fixed wage and called up when they’re needed. Instead, the contractor hires workers as needed on a spot market. So when you hire a contractor, you are hiring a person who hires other people.

Another example is shopping malls. A mall is a physically linked group of stores which are privately-run. The mall is usually managed by a mall owner who makes sure the rent is paid, and the mixture of stores is balanced and attractive. For example, at a typical mall there is a shoe store, a women’s boutique, a gourmet deli, and so on. Then you walk inside Bloomingdale’s, which is a store in the mall. Inside Bloomingdale’s there is a shoe store, a women’s boutique, a gourmet deli…In a sense, Bloomingdale’s itself is a mall with in a mall…except that the mall stores are privately owned and run, and the departments in Bloomingdale’s are staffed by employees who earn a fixed wage and (typically) don’t get a bonus if business is good.

We take for granted that firms efficiently integrate different activities, or else those activities are more efficiently put under a single-firm umbrella. Some striking examples show how different organizational forms might work. For example, there is a seaside area in Hong Kong where people go to eat wonderful seafood. Instead of going to a traditional restaurant, you first go to an outside fish market and haggle with the fishmonger about exactly what fish to buy. They weigh the fish on a crude balance scale and put it in a plastic bag. Then you decide which “restaurant” will cook the fish. The fishmonger walks you over to your restaurant or hands you the fish to carry yourself. From the moment you sit down, you have a typical restaurant experience—except that you picked out the fish at a wholesale price a hundred yards away. The restaurant bill usually includes the cost of the fish, but the fishmonger is paid separately for the fish by the restaurant.

Why does this “unbundled” restaurant flourish only in Hong Kong (and maybe elsewhere, but not in the US)? Nobody knows for sure. It might be a social convention (like 50-50 sharing contracts among corn farmers in Illinois). Part of the explanation is probably that Hong Kongese are very fussy about their fish, and want the fish very fresh. Unbundling the restaurant experience creates competition among fish-sellers, which reduces price and increases quality, which fussy customers appreciate. The separation is also permitted by the simple fact that it is easy to buy a single fish in one place and throw it in a wok in another place nearby. Buying a whole cow in order to eat a steak wouldn’t work so well.

The unbundled restaurant raises a question: Why don’t the fish-sellers get together, and just create a wholesale market that sells directly to the restaurants (which is the way most wholesale fish businesses work)? Part of the explanation might be show—the Hong Kong area is known by tourists who are amused by buying a fish in one place and eating it in another. Or it might be driven by fussy consumers, who want more control over the exact price and quality (at the wholesale level) of the fish they buy.

II. Theories of the firm: Holdup and decision rights

Gibbons (2004) sorts through theories of the firm in a very useful way. He notes that some theories emphasize how organizing anticipates and helps resolve ex post problems arising from costly rent-seeking over quasi-rents created by specific assets.

Other theories emphasize ex ante allocations of decision rights, or systems of decision rights (incentive systems). Let’s discuss each in turn.

Ex post: Asset specificity, holdup, and integration

The earliest theories of the firm were pioneered by Williamson (1971,1975). Williamson emphasized “holdup” problems created by a conjunction of factors. Williamson emphasized three factors:

• The inability to completely specify a contract, which said who would do what in different foreseeable contingencies;
• “Relationship-specific” investments (or just “specific assets”)—once these investments are made, there are “quasi-rents” that will be competed over by both parties;
• “Opportunism”, or “self-interest seeking with guile”, which means a behavioral tendency for both sides to think they deserve the lion’s share of the quasi-rents created by specific assets and ruthlessly grab whatever they can.

To motivate what Williamson was after, think of a place you might regularly have breakfast (we’ll take Hugo’s in West LA as a personal example). There are many other places to have breakfast, but you like Hugo’s and go there often. You get to know the menu, the people who work there, when Hugo’s is most crowded, and so forth. Similarly, many people eat at Hugo’s—if you quit going there they would hardly notice. This is the typical competitive situation described in economics books— one breakfast place and one customer don’t make that much difference to either side.

Now suppose you, and Hugo’s, moved at the same time to a very small town. Suppose there is only one restaurant—Hugo’s—and you eat there all the time, and are one of a few regular customers. Moving to a small place creates a specific asset—if you don’t go to Hugo’s the closest substitute is not as cheap or good (it’s pizza delivery, fast food, or cooking at home); and if you don’t go to Hugo’s their business suffers a lot. Now there is a “quasi-rent” which is valuable to both sides. So how are the gains from this quasi-rent shared?

Enter opportunism. If both sides are best friends, there might be no problem. But if both sides seek to get the most for themselves, and think they deserve more, then the value of the quasi-rent is up for grabs.[1] Hugo’s might raise its prices; since you have nowhere else to go, you’ll pay the extra amount. Or you might demand some special deal as a valued customer, pressuring Hugo’s should cut the price to keep you as a customer. This process of making demands on the other party, after a specific asset or quasi-rent is created, is called “holdup”.[2]

The possibility of holdup creates two different kinds of potential inefficiency. One is that both players incur resources haggling over their share of the quasi-rent, after it is created. Another is that players anticipate such haggling, and they will be reluctant to invest in a way that creates a holdup problem. (In our example, even if it would be efficient for the staff from Hugo’s to move to a remote location to serve a single client, they may be reluctant to do so because of the ex-post struggle over quasi-rents that would occur.)[3]

Williamson’s point is that firms anticipate the costs of holdup, or the reluctance to invest. The solution is integration into a single firm. If the firm can reduce inefficiencies from haggling over the quasi-rent, or can force optimal investment in the face of potential holdup, then the firm has a competitive advantage.

While Williamson explained the benefits of integration (i.e., reduction of costs associated with holdup), he did not clearly articulate the costs of integration. Later theories (and also Simon’s 1951 account of employment relations) more carefully balanced the two.

Empirical studies of the importance of asset-specificity have focused on make-or-buy decisions about different kinds of parts within a company. The typical study uses questionnaires to asset both how specific the design of the parts are [ADD TYPICAL QUESTION HERE], how site-specific the parts are (is the part needed by a geographically-separate division of the company), and how “complex” the part is. Some parts are very carefully designed to fit a firm’s specifications—increasing the holdup potential. Parts which are site-specific are more vulnerable to contracting problems, since it is harder for firms and suppliers to physically sit down and negotiate over what they need. Some parts are very complex, which creates more dimensions of quality (and often, speed) over which the firm and supplier can haggle.

Most of these studies confirm that parts which are more firm-specific, site-specific and complex are more likely to be “made” from internal divisions of a firm, rather than “bought” from independent suppliers. For example, Masten (1984)[4] studied parts used in an aerospace contractor building satellites and other complex pieces of equipment. Table MASTEN shows the probability that the company bought the part from one its own divisions, depending on whether the part was designed specifically and complex. Complex, design-specific parts are made internally 92% of the time. Parts that are neither design-specific nor complex are made internally only 2% of the time.