Firms, Contracts, and Coordination

If Markets Coordinate, Why Do Firms Exist?

Ronald Coase posed a deceptively simple question in 1937. If the price system can coordinate buyers, sellers, and producers, why do firms exist at all? Why does a restaurant hire employees instead of bargaining separately over every plate washed, every table served, and every onion chopped?²

One answer would be that firms replace market coordination with planning. That is only partly right. A firm buys labor, ingredients, electricity, insurance, and capital in markets. It also sells its output in a market. The firm does not escape the price system.

Coase’s deeper answer was that using a market is itself costly. A buyer must discover what is available, compare offers, negotiate, check quality, enforce the agreement, and renegotiate when the world changes. Sometimes it is less costly to bring related activities inside one organization and direct them through continuing relationships.

The term may sound technical, but the idea is familiar. The time spent comparing contractors is a transaction cost. So is checking whether a repair was completed correctly. A lawyer’s fee for writing an agreement may be a transaction cost. So may the risk that an important promise cannot be enforced.

Transaction costs do not mean that markets fail whenever exchange takes effort. They mean that effort belongs in the comparison. A low sticker price from an unreliable supplier may not be the low-cost option after delays, mistakes, and disputes are counted.

A firm has no natural size. It expands only while bringing the next activity inside saves more market transaction cost than it adds in internal coordination cost. It stops expanding when another layer, location, or product becomes more costly to manage inside than through some outside arrangement.

The firm boundary separates activities organized inside the firm from activities obtained through outside contracts. Moving that boundary does not by itself create market power. A vertically integrated firm may still face strong competition, while a small firm with few local substitutes may possess market power.

There is no reason every firm should stop at the same boundary. A national chain may prepare an ingredient centrally while a neighborhood restaurant buys it locally. A manufacturer may own a critical component plant but purchase ordinary office supplies. A hospital may employ nurses while contracting for laundry service. The answer depends on the transaction, not on a general preference for large or small firms.

Make Or Buy

Consider a coffee shop deciding whether to bake pastries itself or buy them from a nearby bakery. Buying may provide several advantages. The bakery can specialize, spread its ovens over many customers, and develop expertise that one coffee shop cannot match. The shop avoids investing in equipment and can switch suppliers if another bakery offers better quality or price.

Making pastries inside the shop may also have advantages. The owner can change the baking schedule as demand changes, watch quality directly, protect a distinctive recipe, and coordinate the kitchen with the rest of the business. No outside supplier has to be persuaded to respond.

The lowest production cost does not automatically decide the question. Suppose the bakery can produce a pastry for less than the coffee shop can. Buying is attractive, but only if the bakery can deliver the required quality and timing at a sufficiently low total cost. Conversely, a desire for control does not automatically justify making. Internal production may require an expensive oven that sits idle most of the day.

The two arrangements can also be combined. The shop might buy standard pastries while making one signature item. It might sign a longer contract with one bakery, provide the recipe, or promise a minimum order. Real organizations often use such mixed arrangements because the underlying problems differ from one input to another.

Make or buy is therefore not a slogan about outsourcing. It is an application of opportunity cost. The firm compares the best feasible ways of organizing the same activity.

Contracts Cannot Predict Everything

An agreement can specify a price, quantity, delivery time, quality standard, and method of payment. It cannot realistically describe every event that might occur over several years.

What happens if ingredient prices suddenly rise? What if a storm blocks the normal delivery route? What if demand triples during a festival? What if a new safety rule requires a different process? What if one party discovers an improvement that was not imagined when the agreement was signed?

An incomplete contract is an agreement that does not specify an enforceable response to every possible future event. Nearly every real contract is incomplete. This is not necessarily a drafting mistake. Predicting, describing, verifying, and enforcing every possibility may be impossible or far too expensive.

That incompleteness matters most when one party must make an investment that depends heavily on one relationship.

Suppose the bakery can buy a machine that produces pastries in the coffee shop’s unusual shape. The machine will improve quality and reduce waste, but it has little value for the bakery’s other customers. Before buying it, the bakery and coffee shop may both expect the relationship to continue.

After the machine has been purchased, the situation changes. The coffee shop knows that the bakery has few alternative buyers for its special output. It may demand a lower price. The bakery may refuse, but walking away would leave it with a machine of little use.

The problem begins even if no one ever makes the demand. If the bakery expects to become vulnerable after investing, it may not buy the machine in the first place. A valuable investment is lost because the parties cannot fully protect the future bargain.

Economists call this the hold-up problem. It is not simply a story about one party being dishonest. The investment changes the parties’ alternatives, and the original contract may not cover every later decision.

Several responses are possible. A longer contract can specify prices or adjustment rules. Reputation can make opportunistic behavior costly by reducing future business. A performance bond or other pledge can make a promise more credible. The parties might share ownership of the special asset. The coffee shop might buy the bakery or bring the specialized production inside.

Each response has a cost. A long contract can become rigid. A pledge ties up resources. Shared ownership can create disputes about control. Integration creates internal monitoring and management problems. The goal is not to eliminate every hazard. It is to choose an arrangement whose remaining problems are less costly.

Ownership Means Deciding When The Agreement Is Silent

When a contract gives no answer, someone still must decide what happens. Ownership helps determine who has that authority.

Suppose the special pastry machine can be modified for a new product that neither party imagined. If the bakery owns it, the bakery normally controls that use unless the contract says otherwise. If the coffee shop owns it, the shop has more authority over the decision. Economists call these residual control rights: rights to make decisions that were not assigned in advance.³

Ownership does not mean that the owner personally makes every daily choice. Owners hire managers, delegate authority, obey laws, honor contracts, and depend on workers with better local knowledge. Ownership matters because it supplies a starting rule when the agreement runs out.

This helps explain vertical integration, in which a firm brings more stages of a supply chain under common ownership. A manufacturer might buy an important supplier. A media company might own both content and distribution. A restaurant chain might operate its own food-processing facility.

Integration can protect specific investments and make adaptation easier. It can also weaken the incentives of people who no longer own the results of their decisions. The market bargain has been replaced by an internal management problem, not by perfection.

What Happens Inside A Firm?

Bringing an activity inside does not make everyone care equally about the result. Chapter 12 introduced the principal-agent problem: a person making a decision may not bear all of its gains and losses, while the person bearing those consequences may find the decision costly to observe.

Many firms also depend on team production. A restaurant meal may reflect the combined work of cooks, servers, cleaners, purchasing staff, and managers. When the dining room runs smoothly, it can be difficult to measure exactly how much each person added. When service is poor, it may be equally difficult to identify one cause.

Monitoring can help, but monitoring takes time and can focus attention on whatever is easiest to count. Sales targets, customer ratings, cameras, checklists, and performance reviews may reveal useful information. They can also be gamed or miss important work.

A residual claimant receives what remains after the organization’s other obligations have been paid. An owner who keeps the remaining profit has a strong reason to watch costs, improve service, and respond to local information. That does not mean every employee should become the full owner. Ownership also means bearing risk and supplying capital, and team output cannot always be divided cleanly.

Armen Alchian and Harold Demsetz asked the next question directly: Who will monitor the monitor? If the monitor receives the same pay whether the team performs well or poorly, the monitoring job creates its own incentive problem. Hiring another monitor merely moves the question up one level.

Their answer was to make the monitor the residual claimant. Better monitoring can raise the team’s output and lower waste, which makes the residual larger. Poor monitoring makes it smaller. The monitor’s own reward therefore depends on doing the monitoring job well.⁶

This helps explain why the for-profit firm is such a common way of organizing economic activity. Profit is not merely money left over after production. The claim to profit gives an owner a strong reason to monitor, improve, and reorganize the team. Large corporations complicate the story because owners and managers are often different people, but they do not make the question disappear.

Why Was A McDonald’s In Russia Different?

A McDonald’s restaurant is a McDonald’s restaurant, right? Customers expect the familiar menu, signs, and service. But producing that familiar meal can require very different organization in different economies.

When the first Moscow restaurant opened in 1990, Russia was beginning its transition from central planning toward a market economy. McDonald’s restaurants elsewhere could take many things for granted: specialized food suppliers, reliable delivery, familiar quality standards, and contracts with private firms. Those arrangements did not yet exist in Russia in the same form.

McDonald’s therefore built its own processing and distribution complex and closely organized farms and other suppliers. It brought more activity inside its own system because the outside market could not yet provide everything it needed reliably.⁸

As capable private suppliers developed, McDonald’s began buying more from outside firms. The restaurant customers saw may have looked much the same, but the organization behind the counter changed.

The lesson is larger than McDonald’s. Telling a firm to “use the market” assumes that reliable suppliers and the institutions needed for exchange already exist. Sometimes a firm must first help create the conditions that make buying possible.

How Might Technology Change The Firms You Work For?

A firm’s boundary changes when the cost of coordinating through markets changes relative to the cost of coordinating inside the firm. Artificial intelligence may change both.

AI may make it easier to find an outside specialist, compare offers, draft an agreement, and check routine work. A firm might then hire fewer permanent specialists and purchase more services when needed. But AI may also make it easier to share knowledge, coordinate teams, and monitor work across many locations. That could allow a larger firm to manage more activity inside. There is no reason the effect must run in only one direction.

Smart contracts may lower the cost of carrying out some agreements. A smart contract is computer code on a blockchain that performs a preset action when specified conditions are met.⁹ For example, payment might be released automatically after a digitally verified delivery.

Code can make a promise automatic only after someone has decided exactly what counts as keeping the promise.

But the lesson about incomplete contracts still applies. Code works best when the condition is clear and can be verified electronically. It cannot by itself decide whether a late delivery was reasonable, whether unusual quality is good enough, or what the parties intended in a situation they never imagined. New technology may automate part of an agreement without eliminating judgment, trust, ownership, or law.

The firms you work for may therefore look different from those of the past. Some may consist of a small permanent team connected to outside specialists. Others may use AI to coordinate large groups spread across many locations. Many will combine both approaches. The useful prediction is not that every firm will become larger or smaller. It is that firm boundaries will change as the costs of coordinating change.

The Coordination Problem Never Disappears

Return to the pencil and the iPhone. Their production requires the plans of many people to fit together even though no one person possesses all the knowledge or directs the entire economy.

This book began with scarcity and opportunity cost. Scarcity forces choices. Specialization and trade create gains, but they also make people more dependent on one another. Prices communicate information and guide adjustment. Property rights help determine who may decide and who bears the consequences. Contracts support promises. Firms coordinate teams when repeated market exchange is too costly. Incentives shape how people respond, while information problems, market power, and transaction costs limit what any arrangement can accomplish.

The same lesson applies to public policy. A rule does not erase scarcity or prevent people from adjusting. Every policy changes incentives, creates trade-offs, and may produce unintended consequences. Finding a flaw in the current arrangement tells us to look for alternatives. It does not tell us that any particular alternative will work better. The question is whether a feasible alternative would coordinate people and resources better.

Economics therefore gives us more than a collection of graphs. It gives us a disciplined way to ask who chooses, what information they possess, which incentives they face, who bears the cost, and how the rules shape cooperation.

The wonder is not that coordination sometimes fails. It is that millions of people with different knowledge and different goals so often find ways to cooperate. Understanding the institutions that make that possible is one of the central tasks of economics.

Chapter Study Map

Chapter Summary

1. Complex production requires several coordination tools. Prices are central, but contracts, firms, property rights, standards, reputation, and law also help separate plans fit together.
2. Markets, contracts, and firms are different ways of coordinating specialized activity.
3. Transaction costs include finding partners, negotiating, monitoring, enforcement, and adaptation.
4. Make or buy compares total costs under realistic alternatives, not production cost alone.
5. Contracts are incomplete because specifying and enforcing every future event is impossible or too costly.
6. Relationship-specific investment can make an investor vulnerable after the investment is sunk, discouraging useful investment.
7. Longer contracts, reputation, pledges, shared ownership, and integration can reduce hold-up risk but create new costs.
8. Ownership assigns authority over decisions a contract did not specify.
9. Team production creates monitoring problems because individual contributions may be difficult to observe.
10. Making a monitor the residual claimant ties the monitor’s reward to the team’s remaining profit and helps answer who monitors the monitor.
11. Vertical integration can improve control and adaptation but creates internal agency and bureaucracy.
12. McDonald’s Russia illustrates how a transition economy may lack suppliers and market institutions that firms elsewhere take for granted.
13. AI can lower market transaction costs and internal coordination costs, so it does not imply one direction for firm size.
14. Smart contracts can carry out preset actions when clear digital conditions are met, but they cannot settle every unforeseen dispute.
15. Good economic analysis compares feasible institutions and the costs each one solves and creates.

Key Terms

• Transaction cost: A cost of finding a partner, reaching an agreement, checking performance, enforcing promises, or adapting the agreement.
• Make or buy: The choice between producing an input inside a firm and purchasing it through the market.
• Incomplete contract: An agreement that does not provide an enforceable answer for every possible future event.
• Relationship-specific investment: An investment whose value depends heavily on continuing a particular exchange.
• Hold-up problem: The risk that changed bargaining positions after a specific investment will discourage that investment.
• Residual control right: Authority to decide how an asset is used when a contract does not specify the decision.
• Vertical integration: Common ownership of more than one stage of a supply chain.
• Team production: Production in which several people’s efforts combine into an output and individual contributions are difficult to measure.
• Residual claimant: A person or organization receiving what remains after other obligations are paid.
• Franchise: An arrangement allowing a local operator to use a brand and business system subject to contractual rules and payments.
• Smart contract: Computer code on a blockchain that performs a preset action when specified conditions are met.

Review Questions

1. What question did Coase ask about the existence of firms?
2. What are transaction costs? Give four examples.
3. Why does the existence of firms not mean that prices have stopped coordinating economic activity?
4. What does a make-or-buy decision compare?
5. Why might a supplier with a lower production cost still not be the lower-cost choice?
6. What is an incomplete contract, and why are real contracts incomplete?
7. What makes an investment relationship specific?
8. How can hold-up risk discourage a useful investment before any dispute occurs?
9. Name three possible responses to hold-up and one cost of each.
10. What is a residual control right?
11. Why does ownership not mean that an owner personally makes every decision?
12. What is vertical integration?
13. Why does integration change rather than eliminate transaction and incentive problems?
14. What is team production, and why can it make individual effort difficult to measure?
15. What is a residual claimant?
16. How does making the monitor a residual claimant help answer “Who will monitor the monitor?”
17. How does a franchisee’s incentive differ from that of a salaried manager?
18. Why does a franchisor continue to impose rules and monitor quality?
19. What does the McDonald’s Russia case show about changing firm boundaries?
20. Give one way AI can favor outsourcing and one way it can favor internal organization.
21. What is a smart contract?
22. Why do smart contracts not eliminate incomplete contracts?
23. What does it mean to compare institutions rather than ideals?

Economic Reasoning Questions

1. A coffee shop can bake muffins for $1.70 each or buy them for $1.45. The supplier sometimes misses the morning delivery, causing an expected loss of $0.35 per muffin ordered. Which option has the lower measured total cost? What other information would you want?
2. A supplier is considering a machine useful mainly to one buyer. Explain how the investment changes the supplier’s outside options and why the supplier may invest too little.
3. Compare a five-year fixed-price contract with vertical integration as responses to the specialized-machine problem. Identify one advantage and one new cost of each.
4. A company owns a machine, but its employment and supply contracts do not say whether it may be used for a newly discovered product. Explain why ownership matters.
5. A restaurant chain is deciding whether to franchise a distant location. Identify the local-information, monitoring, risk-bearing, and brand-quality considerations.
6. A company brings delivery service inside after repeated late shipments. Explain the problem integration may solve and two internal costs it may create.
7. A region develops reliable private suppliers, standard quality certification, and faster contract enforcement. Predict how these changes could affect firm boundaries.
8. An AI tool reduces the cost of finding freelance designers and checking routine work. Trace the likely effect on outsourcing.
9. A different AI tool makes it easier to share expertise and coordinate workers across many locations. Trace the possible effect on internal organization.
10. A smart contract releases payment after a delivery scanner records that a shipment arrived. What part of the exchange can the code handle? What problems might still require human judgment?

Optional Research And Discussion Questions

1. Choose a familiar business and identify one input it makes, one it buys, and one hybrid arrangement it uses. Explain each boundary using production cost, transaction cost, information, and control.
2. Investigate one franchise system. What does the local operator own or control? What does the franchisor control? How are revenue, investment, and risk divided?
3. Choose one workplace use of AI. Identify one coordination cost it lowers and explain whether it might move work inside a firm, outside the firm, or in both directions.

Source Notes