Managerial Decisions for Firms with Market Power

By Thomas, C.R., Maurice, S.C.

Edited by Paul Ducham


Even though we have not set forth a precise way to measure a firm’s market power, you have probably figured out that the amount of MARKET POWER is related to the availability of substitutes. The better the substitutes for the product sold by a firm, the less market power the firm possesses. However, there is no single measurement of market power that is totally acceptable to economists, policymakers, and the courts. Economists have come to rely on several measures of market power. These methods are widely used, frequently in antitrust cases that require objective measurement of market power.

Any of the methods of measuring the market power of a firm will fail to provide an accurate measure of market power if the scope of the market in which the firm competes has not been carefully defined. This section begins by discussing how to determine the proper market definition: identifying the products that compete with one another and the geographic area in which the competition occurs. Then we discuss some measures of market power.

Market Definition

A market definition identifies the producers and products or service types that compete in a particular geographic area, which is just large enough to include all competing sellers. As you can see by this definition of a market, properly defining a market requires considering the Level of Competition in both the product dimension and the geographic dimension of a market. Although the methodology of appropriately defining a market is primarily of interest to firms engaged in federal or state antitrust litigation—specifically cases involving illegal monopolization of a market or the impact of a proposed merger on the merged firm’s market power—managers should know how to properly define the firm’s market in order to measure correctly the firm’s market power. We will now discuss some guidelines for determining the proper product and geographic dimensions of a market.

A properly defined market should include all the products or services that consumers perceive to be substitutes. A manager who fails to identify all the products that consumers see as substitutes for the firm’s product will likely overestimate the firm’s market power. The CEO of Coca-Cola would be foolish to view the company as a monopolist in the production of cola soft drinks and expect it to enjoy substantial market power. No doubt Coca-Cola’s syrup formula is a closely guarded secret, but most soft-drink consumers consider rival brands of soft drinks, as well as a variety of noncarbonated drinks such as iced tea and Gatorade, as reasonable substitutes for Coca-Cola.

The geographic boundaries of a market should be just large enough to include all firms whose presence limits the ability of other firms to raise price without a substantial loss of sales. Two statistics provide guidelines for delineating the geographic dimensions of a market: (1) the percentage of sales to buyers outside the market and (2) the percentage of sales from sellers outside the market. Both percentages will be small if the geographic boundary includes all active buyers and sellers. These two guidelines for determining the geographic dimensions of a market are sometimes referred to as LIFO and LOFI: little in from outside and little out from inside.

As mentioned earlier, economists have developed several measures of market power. We will discuss briefly only a few of the more important measures.

Elasticity of Demand

One approach to measuring how much market power a firm possesses is to measure the elasticity of the firm’s demand curve. A firm’s ability to raise price without suffering a substantial reduction in unit sales is inversely related to the PRICE ELASTICITY OF DEMAND. The less elastic is demand, the smaller the percentage reduction in quantity demanded associated with any particular price increase. The more elastic is demand, the larger the percentage decrease in unit sales associated with a given increase in price. The elasticity of demand is greater (i.e., more elastic) the larger the number of substitutes available for a firm’s product. As demand becomes less elastic, consumers view the product as having fewer good substitutes.

Although a firm’s market power is greater the less elastic its demand, this does not mean a firm with market power chooses to produce on the inelastic portion of its demand. In other words, market power does not imply that a manager produces where |E| < 1; rather, the less elastic is demand, the greater the degree of market power. We will demonstrate in this chapter that a monopolist always chooses to produce and sell on the elastic portion of its demand.

Relation The degree to which a firm possesses market power is inversely related to the elasticity of demand. The less (more) elastic the firm’s demand, the greater (less) its degree of market power. The fewer the number of close substitutes consumers can find for a firm’s product, the smaller the elasticity of demand and the greater the firm’s market power. When demand is perfectly elastic (demand is horizontal), the firm possesses no market power.

The Lerner Index

A closely related method of measuring the degree of market power is to measure the extent to which price deviates from the price that would exist under competition. The Lerner index, named for Abba Lerner, who popularized this measure, is a ratio that measures the proportionate amount by which price exceeds marginal cost:


Cross-Price Elasticity of Demand

An indicator, though not strictly a measure, of market power is the cross-price elasticity of demand. Cross-price elasticity measures the sensitivity of the quantity purchased of one good to a change in the price of another good. It indicates whether two goods are viewed by consumers as substitutes. A large, positive cross-price elasticity means that consumers consider the goods to be readily substitutable. Market power in this case is likely to be weak. If a firm produces a product for which there are no other products with a high (positive) crossprice elasticity, the firm is likely to possess a high degree of market power.

The cross-price elasticity of demand is often used in antitrust cases to help determine whether consumers of a particular firm’s product perceive other products to be substitutes for that product. Using cross-price elasticities, antitrust officials try to determine which products compete with one another. For example, antitrust officials might wish to determine the degree of market power enjoyed by Nike brand athletic shoes. Nike Corporation has spent a great deal of money advertising to establish a prominent position in the market for athletic shoes. To determine which other products compete with Nike, the cross-price elasticity of the quantity demanded of Nike shoes with respect to a change in the price of a rival’s product can be calculated. Using such cross-price elasticities, antitrust officials can determine whether consumers view Nike as having any real competitors in the market for athletic shoes.

Relation If consumers view two goods to be substitutes, the cross-price elasticity of demand (EXY) is positive. The higher the cross-price elasticity, the greater the perceived substitutability and the smaller the degree of market power possessed by the firms producing the two goods.

These are only a few of the measures of market power. The courts in antitrust cases and the Justice Department in merger and acquisition hearings sometimes use a combination of measures, including concentration ratios and share of the market. It is also not always clear just how high a cross elasticity or how low an elasticity constitutes “too much” market power. If you are ever involved in such a hearing, you should be aware of the problems in measuring market power. Illustration 12.1 shows the difficulty of determining what constitutes a market and what determines the amount of market power.


Entry or potential entry of new firms into a market can erode the market power of existing firms by increasing the number of substitutes. Therefore, as a general case, a firm can possess a high degree of market power only when strong barriers to the entry of new firms exist. A strong barrier to entry exists when it is difficult for new firms to enter a market where existing firms are making an economic profit. Strong barriers to entry hinder the introduction of new, substitute products and protect the profits of firms already in the market.

An example of a strong barrier to entry is a cable TV franchise granted by a city government to only one cable company. This fortunate company is protected from other firms’ competing away any economic profits and is close to being a monopoly. Note that we said “close” to being a monopoly because the cable company has some outside competition even though it is the only cable company in town. Possible substitutes, though certainly not perfect ones, might be regular broadcast television, satellite dishes, radio, books and magazines, rental movies, and so on. Thus the firm would be a monopoly if the cable TV market is the relevant market but not if the entertainment market is the relevant market. We should note that in cases in which a government body protects a firm from entry by other firms into a market, it typically regulates the protected firm.

Weak barriers to entry generally exist in most retail markets. Retail stores typically do not have much market power because entry by other firms into the market is easy and there are good substitutes for the products of firms selling in the market. The products are not perfect substitutes, as is the case for perfect competition, because other firms cannot sell identical products or sell in the same location. However, firms can produce close substitutes. Therefore, even though perfect competition would not exist in such markets because products are not perfect substitutes, no firm has much market power since it cannot raise its price much above its rivals’without a substantial loss of sales. In this chapter we will focus on structural entry barriers, which are barriers arising from cost and/or demand conditions in the market.

Economies of Scale

An important barrier to entry is created when the long-run average cost curve of a firm decreases over a wide range of output, relative to the demand for the product. Consequently, a new firm that wishes to enter this type of market must enter on a large scale in order to keep its costs as low as the large-scale firm or firms already operating in the market. The necessity of entering on a large scale is usually not a barrier to entry by itself, but when it is coupled with relatively small product demand, a strong barrier to entry can be created.

Consider an industry where four existing firms each produce about 200,000 units annually to take advantage of substantial economies of scale. At the current price of the product, annual sales are running at about 800,000 units per year. While many entrepreneurs could obtain the financial backing to enter this industry with a largescale plant capable of producing 200,000 units, there is no room in the industry for a fifth large-scale producer without a significant decline in the price of the product. Even though a fifth firm could enter the industry producing perhaps 50,000 units annually, the per-unit production costs would be much higher than competitors’ costs because of the substantial economies of scale. There just isn’t room for a new firm to enter this industry on a scale big enough to enjoy costs as low as those of rivals. In such situations, economies of scale create a barrier to entry.

Barriers Created by Government

An obvious entry barrier is government. Licensing and franchises are ways monopolies are created by government decree. For example, licenses are granted to radio and television stations by the Federal Communications Commission (FCC), and only those stations possessing a license are allowed to operate. Governments also grant exclusive franchises for city, county, and state services. For example, local telephone and cable television utilities have a great deal of market power in that they are the only regional producers of the products. By law, no other producer can exist.

Another legal barrier to competition lies in the patent laws. These laws make it possible for a person to apply for and obtain the exclusive right to produce a certain commodity, or to produce a commodity by means of a specified process that provides an absolute Cost Advantage. Despite examples to the contrary, however, holding a patent on a product or production process may not be quite what it seems in many instances. Apatent does not preclude the development of closely related substitute goods or closely allied production processes. International Business Machines (IBM) has the exclusive right to produce its patented computers, but many other computers are available and there is competition in the computer market.

Essential Input Barriers

Historically, an important reason for market power has been the control of rawmaterial supplies. If one firm (or perhaps a few firms) controls all the known supply of a necessary and essential ingredient for a particular product, the firm (or firms) can refuse to sell that ingredient to other firms at a price low enough for them to compete. When no others can produce the product, monopoly results. For many years the Aluminum Company of America (Alcoa) owned almost every source of bauxite, a necessary ingredient in the production of aluminum, in North America. The control of resource supply, coupled with certain patent rights, provided Alcoa with an absolute monopoly in aluminum production. It was only after World War II that the federal courts effectively broke Alcoa’s monopoly in the aluminum industry. There have been other such historical examples, but at the present time there are few cases of firms with considerable market power because of exclusive control of a raw material.

Brand Loyalties

On the demand side, older firms may have, over time, built up the allegiance of their customers. New firms can find this loyalty difficult to overcome. For example, no one knows what the service or repair policy of a new firm may be. The preference of buyers can also be influenced by a long successful advertising campaign; established brands, for instance, allow customers recourse if the product should be defective or fall short of its advertised promises. Although technical economies of scale may be insignificant, new firms might have considerable difficulty establishing a market organization and overcoming buyer preference for the products of older firms. A classic example of how loyalty preserves monopoly power can be found in the concentrated-lemon-juice market. ReaLemon lemon juice successfully developed such strong brand loyalties among consumers that rival brands evidently could not survive in the market. The situation was so serious that the courts forced ReaLemon to license its name to would-be competitors.

The role of advertising as a barrier to entry has long been a source of controversy. Some argue that advertising acts as a barrier to entry by strengthening buyer preferences for the products of established firms. On the other hand, consider the great difficulty of entering an established industry without access to advertising. A good way for an entrenched monopoly to discourage entry would be to get the government to prohibit advertising. The reputation of the old firm would enable it to continue its dominance. Anew firm would have difficulty informing the public about the availability of a new product unless it was able to advertise. Thus advertising may be a way for a new firm to overcome the advantages of established firms. The effect of advertising on entry remains a point of disagreement among economists.

Consumer Lock-In

For some products or services, consumers may find it costly to switch to another brand—either an existing rival’s brand or a new entrant’s brand of product or service. Some of the kinds of switching costs incurred by consumers include things such as installation or initiation fees, search costs to learn about availability and prices of substitutes, and costs of learning how to use a new or different product or service. When high switching costs make previous consumption decisions so costly to alter that rivals do not believe they can induce many, if any, consumers to change their consumption decisions, then a situation known as consumer lock-in results. Consumer lock-in, of course, discourages new firms from entering a profitable market, and thus protects incumbent firms from new competition. High switching costs may arise naturally, or firms may strategically design products and services to have high switching costs in order to create a consumer lock-in barrier to entry.

While consumer lock-in can certainly create a strong barrier to entry, high monopoly profits nonetheless create a strong incentive for potential entrants to find ways to overcome a lock-in barrier. For example, when Microsoft decided to enter the market for household financial software with its Money program, Quicken had already established a virtual monopoly, and satisfied consumers seemed unwilling to incur the costs of switching from Quicken to Money. Microsoft, however, overcame this consumer lock-in barrier by designing its Money program to accept financial data files stored in Quicken’s proprietary format so that switchers would not need to reenter their financial data. Microsoft also employed similar commands for its software and provided specialized help menus for users making the switch from Quicken. Thus, by lowering the switching costs facing consumers, Microsoft overcame a consumer lock-in barrier to entry and successfully ended Quicken’s monopoly.

Network Externalities (or Network Effects)

For most goods, the utility you get from consuming the good does not depend on how many other people consume the good. The value or benefit you receive is the same whether the good is purchased by 10 other people or 10 million other people. In contrast to this “normal” situation, however, there are a few special goods and services for which your utility varies directly with the total number of consumers of the good. In other words, a larger number of consumers buying a product will enhance the value you get from that product.2 Such goods are characterized by Network Externalities or network effects. Some examples of goods or services believed to experience network effects include cellular phones, Internet access services, computer operating systems (such as Microsoft Windows or Apple OS X), e-mail, job search or dating service companies, and online auction Web sites (such as eBay).

There are two possible reasons why network externalities arise. First, network externalities are likely to characterize products and services when the usefulness of the product requires connecting consumers. For example, your utility from having a cell phone increases as the number of other people who have cell phones in the cellular network increases. Most people carry a cell phone to make and receive calls to and from many people, so the value of having a cell phone grows with the size of the network of people you can reach. An online dating service will be more valuable to single individuals looking for a companion if the dating service has a large membership. A large electricity transmission grid may be more valuable to individual homeowners than a small grid because a large grid will be more quickly restored to operation should it fail or be knocked out by a natural disaster of any kind (e.g., hurricane, wild fire, flood, etc.) Second, network externalities arise if complementary goods are important to users of the network good. For example, software applications are important complements to computer operating systems. As the number of Apple computer users increases, software companies will write more software for Apple computers, making ownership of Apple computers more satisfying, and in turn increasing demand for Apple computers. Repair service or troubleshooting can be extremely valuable for many products. Consumers may believe that manufacturers will provide better service or phone support when there are many users of a good. A computer “bug” in Microsoft Word is likely to get fixed much sooner than the same bug would be fixed in Corel’s WordPerfect because the network of users is larger. And, even if Microsoft is slow to fix bugs, a large network of bloggers who use MS Word can offer free “work-arounds” on the Internet.

Our primary interest in network externalities in this textbook concerns barriers to entry. Network effects can make it difficult for new firms to enter markets where incumbent firms have established a large base or network of buyers. Since the value of the good depends on the number of users, it will be difficult to enter and compete as a small-size firm. Buyers want to be part of the large network of consumers held by the established incumbent.

As a general rule, maximizing market share is not equivalent to maximizing profit and the value of the firm. However, we also pointed out that network externalities can present an exception to this rule. In a network industry, a price cut by Firm A, which causes a number of buyers to switch from rival firms to Firm A, can set in motion a self-reinforcing or snowball process. The initial gain in Firm A’s market share adds to its network size, causing even more buyers to switch from rivals, in turn further increasing Firm A’s network size and so on until the market tips all the way to Firm A. In this scenario, the profit-maximizing price is lower than it would be if there were no network externalities. Gaining market share can lead to higher profit under these circumstances.

It follows that when network externalities are significant, an incumbent firm possessing a large number of consumers may enjoy a formidable barrier to entry. Buyers value the incumbent firm’s large network and will be reluctant to switch to an entering firm’s product for which the network of other consumers is too small to be attractive.

Sunk Costs as a General Barrier to Entry

The last structural barrier to entry we wish to discuss—sunk costs—can be viewed as a general type of entry barrier that can include the other entry barriers we have discussed above. You will find it helpful to think of managers as making two decisions. The first decision is whether to enter the market, and the second decision, if the manager chooses to enter, is how to set price and output to maximize profit in the market after entering. When a firm must incur “setup costs” to enter a market, these costs are sunk costs the firm must pay to enter and must be treated as the costs of entry. In other words, the costs of entering a market are the sunk costs incurred by making the decision to enter a market. These entry costs, because they are sunk costs, are not costs of doing business once a firm is an incumbent firm—that is, after a firm has entered. All sunk costs of entry should be ignored because they do not affect profit. However, entry costs can serve as a barrier to entry if they are so high that the manager cannot expect to earn enough future profit in the market to make entry worthwhile.

A simple example will be helpful here. Suppose there is a market in which the incumbent firms are making $1,200 of profit each month in long-run equilibrium. You would like to enter this market because after you enter and a new long-run equilibrium is established your firm and the other firms in the market will each earn $1,000 of profit each month. If there are no sunk costs of entry, then you would certainly enter the market. The decision would be an easy one because you could always exit the market if monthly profit later becomes negative. Alternatively, suppose you must incur a lump sum cost of $50,000 to enter and this cost is sunk—you cannot recover any of the $50,000 if you later decide to exit the market. What should you do now that entry is costly? Let’s keep things simple and ignore the time Value of Money (i.e., assume your discount rate is zero). You would choose to enter this market if you were certain the market would last for at least 50 months. As you can see in this simple example, the higher the sunk costs of entry, the greater market profitability must be in order to make entry worthwhile. Thus, high sunk costs relative to the profitability of entry can serve as a barrier to entry.

As you can now see, the other entry barriers discussed previously may require an entering firm to incur some amount of sunk costs to enter the market. Economies of scale may create a barrier to entry if sunk set-up costs must be incurred to acquire a production facility large enough to give you the same unit costs as incumbent firms. As an example of this, you may need to pay for an Environment impact study before you can build your plant. When access to essential inputs is restricted or when patents, licenses, advertising, switching costs or network externalities exist, a firm that wishes to enter must frequently incur costs to overcome these barriers. If the costs of overcoming a structural barrier are sunk and large relative to the profitability of entering the market, then sunk costs can serve as a barrier to entry.

Despite the existence of barriers to entry, firms can lose and have lost their positions of extensive market power. Even quite strong barriers to entry can be overcome. A monopolist can become complacent in its protected position and allow inefficiencies to enter the production process. This raises the cost, and hence the price, and allows new, more efficient firms to enter the market. Some potential entrants are ingenious enough to find ways to lower cost, or (as noted earlier) get around patent protection, or overcome Brand loyalty to the established firm. Thus barriers to entry cannot completely protect the established firm with great market power.

I L L U S T R AT I O N 1 2 . 1

Is Microsoft a Monopoly?

Some of the most contentious issues in the recent antitrust case against Microsoft stemmed from the question of whether Microsoft had a monopoly in the market for PC operating systems. And, even if Microsoft did have a monopoly in its Windows operating system, did it have sufficient market power to harm consumers? And, even if Microsoft possessed sufficient market power to harm consumers, would consumers benefit by breaking Microsoft up into two smaller companies? Don’t think for a moment that we can answer these questions definitively in a short illustration, or even in a long one for that matter. We can’t. But we can illustrate the rich complexities of these interesting questions by surveying the opinions of a number of economists as reported by various business news publications.

Alan Reynolds (Director of Economic Research at the Hudson Institute)

It was routinely reported that Microsoft’s Windows software “runs on more than 90 percent of the world’s PCs.” This fraction would be worrisome if it meant that Microsoft had captured all but 10 percent of the total market for operating systems. To evaluate the usefulness of the reported market share, we must consider the market definition employed to make the calculation. As we emphasized in the text, a properly defined market should include all the products that consumers perceive to be substitutes. Reynolds argued that the Department of Justice defined the market for operating systems far too narrowly, and so inflated Windows’ share of the operating system market.

The Justice Department defined the market in the Microsoft case to be “single-user computers with Intel microprocessors.” Reynolds noted that this narrow definition of the market in which Microsoft competes excludes such competitors as Apple computers, since they don’t use Intel microprocessors; Sun Microsystems workstations; any operating system used as part of a business network (e.g., Solaris and UNIX); and operating systems used in handheld and subnotebook computers. In short, Reynolds believed the Justice Department stacked the deck against Microsoft by excluding many genuine competitors of Microsoft’s Windows operating system. Reynolds also noted that, in high-tech Industries, dominant firms are normal: Quicken has 80 percent of the home-finance software market, Netscape once had 90 percent of the browser market, and Intel has 76 percent of the microprocessor chip market.

Richard Schmalensee (MIT Economist and Expert Witness for Microsoft)

During his testimony as an expert witness for Microsoft, Schmalensee made a particularly insightful point: Microsoft may indeed have owned most of the market for operating systems, but it did not have a high degree of market power and was not a harmful monopoly. Schmalensee calculated that, if Microsoft was indeed a monopolist wielding great market power because it faced little or no competition, the profitmaximizing price for Windows 98 would have been somewhere between $900 and $2,000. The Justice Department’s attorney expressed his astonishment over this calculation by asking Schmalensee if he thought a price of $2,000 made sense as the profit-maximizing price for Microsoft to actually charge for Windows. “Of course not, because Microsoft faces significant longrun competition. That’s precisely the point.” As we explained in the text, the degree of market power a monopolist possesses depends on the availability of close substitutes. Schmalensee explained that not only did Windows 98 face potential competition from new entrants in the future, it also had to compete with two highly successful and widely available rival products: Windows 3.1 and Windows 95. Perhaps a consumer’s best protection from the alleged Microsoft monopoly was to own an early version of Windows.

Franklin Fisher (MIT Economist and Expert Witness for Department of Justice)

”Microsoft has engaged in anticompetitive conduct that has no compelling economic justification but for its effect of restricting competition,” according to testimony in the case by Franklin Fisher, an expert in antitrust matters pertaining to monopoly practices. The government introduced into evidence numerous internal Microsoft memos and strategy documents. The language in these documents painted a picture of a firm obsessed with beating its rivals in every way possible. In one e-mail circulated among the top executives at Microsoft on the topic of subverting rival Java software language: “Subversion has always been our best tactic . . . Subversion is almost always a better tactic than a frontal assault. It leaves the competition confused; they don’t know what to shoot at anymore.” While the tactics employed by Microsoft to beat its rivals do seem ruthless to us, we suspect the same kind of memos would surface if the trial involved Pfizer, Toyota, Bank of America, or any other profit-maximizing firm.

The Economist (Editorial Opinion in the British Business News Magazine)

In an editorial opinion, The Economist expressed its concern that many of the high-tech markets in the New Economy experience network externalities, which increase the likelihood that a single firm may dominate a market. Once a dominant firm establishes a large, “installed” base of customers who use its brand of high technology, consumers may become locked in, creating a monopoly by blocking the entry of new firms and new technologies. For antitrust enforcement agencies charged with the duty of preventing new monopolies and breaking up old ones, the continual product improvement and falling computer product prices make it difficult to demonstrate that consumers are harmed by “monopoly abuse” in high-tech markets. Consequently, The Economist worried that Microsoft could stifle innovation and inflict serious harm to high-tech consumers and the New Economy.

The Economist, like Franklin Fisher, viewed Microsoft’s business Behavior as evidence of its intent to use its market power to maintain its market dominance. “An amazing trail of e-mails and management papers has depicted a company ready, it seems, to do almost anything to protect its Windows monopoly . . . When, as in the Microsoft case, a monopolist’s conduct seems to be chilling innovation in markets in which the competition is largely defined by innovation, the argument for antitrust intervention is compelling.”

Gary Becker (Nobel Prize–Winning Economist at University of Chicago)

The Department of Justice proposed breaking Microsoft into two independent firms: an operating-system company (Windows) and an applications company (MS Office, Internet Explorer, and other Microsoft applications). DOJ believed a breakup was needed to encourage faster technological innovation. Becker saw two problems with the Justice Department’s arguments. First, economists are not sure that competition fosters greater rates of innovation than monopoly. Becker referred to the original thinking on this issue by Joseph Schumpeter (1883–1950), who believed that monopoly markets experience higher rates of innovation than competitive ones. Monopolies stimulate more technological innovation, according to Schumpeter, because they don’t have to worry about competitors (quickly) imitating their innovations, driving down their profits.

Becker also argued that the Department of Justice has not provided any quantitative evidence that the dominant position held by Microsoft in operating systems had slowed technical progress in the computer- Internet industry:

The government and its experts cite a few potential innovations that were supposedly discouraged by Microsoft’s aggressive behavior. Even if these examples are valid, the government does not consider whether there have been other innovations stimulated by a large market for new software applications made possible by the dominant Windows platform.

Over the last 40 years, enormous technological progress has occurred in the computer-Internet industry. That progress, Becker pointed out, did not slow down as Microsoft built its powerful position in operating systems during the last 20 years of this period. Maybe Microsoft’s rivals who complained in court were hoping the Justice Department would protect them from competition rather than promote competition?

As we told you at the beginning of this illustration, we wish we could give you the answer to all of these questions, but we can’t. Indeed, the answers proved difficult for all involved in this case. Eventually, the trial judge, Judge Thomas Penfield Jackson, found Microsoft guilty of illegal monopolization and ordered Microsoft to be split into two firms. On appeal, the U.S. Appeals Court reversed the breakup order and removed Judge Jackson from the case. In November 2001, Microsoft and the Justice Department reached a settlement on penalties and remedies that received final approval in November 2002 by the new judge in the case, Judge Colleen Kollar-Kotelly. Clearly, the question of illegal monopolization proved to be quite challenging for all concerned. You should try to reach your own conclusion and discuss your reasoning with classmates and your professor. This case will likely be debated for many years.


We will now analyze the profit-maximizing decision of firms that are pure monopolies. Keep in mind that the fundamentals of this monopoly decision apply to a large extent to all firms with market power. The manager of a monopoly treats the market demand curve as the firm’s demand curve. As was the case for perfect competition, we assume that the manager wishes to maximize profit. Thus the manager of a monopoly firm chooses the point on the market demand curve that maximizes the profit of the firm. While the manager of a monopoly does, in fact, determine the price of the good, price cannot be chosen independent of output. The manager must choose price and output combinations that lie on the market demand curve.

In Figure 12.1, for example, if the manager wishes to charge a price of $14 per unit, the monopoly firm can sell (consumers will buy) only 900 units of the product. Alternatively, if the manager decides to sell 900 units, the highest price that can be charged for this output is $14. So, while the monopolist can choose both price and output, the two choices are not independent of one another.

In practice, some monopolists choose price and let market demand determine how many units will be sold, whereas other monopolists choose the level of output to produce and then sell that output at the highest price market demand allows. Consider your electric utility company. Electric utilities set the price of a unit of electricity, say, 10 cents per kilowatt-hour, and then stand ready to supply as many kilowatt-hours as consumers wish to buy at that price. You can be sure that your electric company has estimated its demand function and knows approximately how much electricity will be demanded at various prices.

Alternatively, an automobile manufacturer might decide to produce 300,000 cars of a particular model in a given year. The manufacturer sells these cars at the highest possible price given the existing market demand. Again, you can be sure that the automobile manufacturer has estimated the demand for its cars and knows approximately the average price at which each car can be sold.

Given the demand curve facing a monopolist, choosing price to maximize profit is equivalent to choosing output to maximize profit. To be consistent with our discussion of profit maximization under perfect competition, we will view the monopolist as choosing output to maximize profit.

The basic principle of profit maximization—profit is maximized by producing and selling the output at which marginal cost equals marginal revenue—is the same for the monopoly as for the competitive firm. Amanager can increase profit by expanding output as long as the marginal revenue from the expansion exceeds the marginal cost of expanding output. A manager would reduce output if marginal revenue is less than marginal cost. The fundamental difference for a monopolist is that marginal revenue is not equal to price.

Principle A monopolist chooses the point on the market demand curve that maximizes profit. If marginal revenue exceeds marginal cost, a profit-maximizing monopolist increases output. If marginal revenue is less than marginal cost, the monopolist does not produce these additional units.

Demand and Marginal Revenue for a Monopolist

A monopoly, facing a downward-sloping demand, must lower the price in order to sell more. As shown in Figure 12.1, marginal revenue is less than price for every unit sold except the first. You will recall that marginal revenue is the change in the firm’s total revenue from an additional unit of sales; symbolically, MR= ΔTR/ΔQ. In Figure 12.1, if the firm sells 900 units at $14 each, you can see from the marginal revenue curve that the marginal or additional revenue from selling the 900th unit is $8. This means that reducing the price just enough to increase sales from 899 to 900 adds $8 to the firm’s revenue, rather than the $14 price at which the 900th unit is sold. The reason is that in order to sell the 900th unit, the firm must reduce the price on the 899 units it could have sold at the slightly higher price.

Although we set forth a technical analysis of the relation between MR and P, we can perhaps give you a bit more understanding of why MR is less than P with a hypothetical example. Suppose you manage a small appliance store that has been selling 20 radios a day at $50 apiece. You want to increase your sales of radios, so one day you reduce the price to $49. Sure enough, you sell 21 radios that day at the reduced price. So you sold one more at $49. You check the cash register and compare the receipts with those from previous days. You had been receiving $1,000 (= $50 * 20). Now you see that you have taken in $1029 (= $49 * 21) from selling radios. Your revenue increased by $29, but what happened to the $49 at which the additional radio was sold? Did someone steal $20 from the register? What happened was that, in order to sell the 21st radio, you had to take a $1 price reduction on the 20 you could have sold at $50. This $1 price reduction accounts for the “missing” $20.

Figure 12.1 illustrates the relation between demand and marginal revenue for a linear demand curve. When demand is linear, marginal revenue is twice as steep as demand and consequently lies halfway between demand and the vertical axis. When MR is positive, between 0 and 1,500 units, demand is elastic. When MR is negative, above 1,500 units, demand is inelastic. When MR equals 0, at 1,500 units, demand is unitary elastic.

Relation The market demand curve is the demand curve for the monopolist. Because the monopolist must lower price in order to sell additional units of output, marginal revenue is less than price for all but the first unit of output sold. When marginal revenue is positive (negative), demand is elastic (inelastic). For a linear market demand, the monopolist’s marginal revenue is also linear, with the same vertical intercept as demand, and is twice as steep.

Maximizing Profit at Southwest Leather Designs: An Example

Southwest Leather Designs specializes in the production of fashionable leather belts for women. Southwest’s original designs are sometimes imitated by rival leather goods manufacturers, but the Southwest logo is a registered trademark that affords the company some protection from outright counterfeiting of its products. Consequently, Southwest Leather enjoys a degree of market power that would not be present if imitators could make identical copies of its belts, trademark and all.

Table 12.1 presents the demand and cost conditions faced by the manager of Southwest Leather Designs. Columns 1 and 2 give the demand schedule for 1,000 through 9,000 units of output (leather belts) in discrete intervals of 1,000. Column 3 shows the associated total revenue schedule (price times quantity). The total cost of producing each level of output is given in column 4. The manager computes profit or loss from producing and selling each level of output by subtracting total cost from total revenue. Profit is presented in column 7. Examination of the profit column indicates that the maximum profit ($56,020) occurs when Southwest Leather Designs sells 6,000 belts at a price of $18.92.

The manager of Southwest Leather Designs can reach the same conclusion using the marginal revenue–marginal cost approach. Marginal revenue and marginal cost are shown, respectively, in columns 5 and 6. The marginal revenue from selling additional leather belts exceeds the marginal cost of producing the additional belts until 6,000 units are sold. After 6,000 units the marginal revenue for each of the next 1,000 belts is $5.48 per belt while the marginal cost for each of the next 1,000 belts is $6.25 per belt. Clearly, increasing output and sales from 6,000 to 7,000 belts would lower profit. Thus profit must increase until 6,000 units are produced; then profit decreases thereafter. This is the same solution that was obtained by subtracting total cost from total revenue: An output of 6,000 belts maximizes profit.

The example in Table 12.1 is shown graphically in Figure 12.2. Since marginal revenue and marginal cost are per-unit changes in revenue and cost over discrete changes in output of 1,000 units, we plot these values in the middle of the 1,000- unit interval. For example, marginal revenue for the first 1,000 units sold is $35 per unit for each of these 1,000 units. We plot this value of marginal revenue ($35) at 500 units of output. We do this at all levels of output for both marginal revenue and marginal cost.

In Figure 12.2, marginal revenue equals marginal cost at 6,000 units of output, which, as you saw from the table, is the profit-maximizing level of output. The demand curve shows that the price that Southwest Leather Designs will charge for the 6,000 belts is $18.92.

We turn now from a specific numerical example of profit maximization for a monopolist to a more general graphical analysis of a monopolist in the short run. In this case, we will assume for analytical convenience that output and price are continuously divisible

Short-Run Equilibrium: Profit Maximization or Loss Minimization

A monopolist, just as a perfect competitor, attains maximum profit by producing and selling the rate of output for which the positive difference between total revenue and total cost is greatest; or it attains a minimum loss by producing the rate of output for which the negative difference between total revenue and total cost is least. As long as total revenue covers total avoidable cost, this condition occurs when marginal revenue equals marginal cost. As was the case for the perfectly competitive firm, when price is less than average variable cost, the manager shuts down production in the short run. We will first discuss profit maximization and then loss minimization.

The position of short-run equilibrium is easily described graphically. Figure 12.3 shows the relevant cost and revenue curves for a monopolist. Because AVC and AFC are not necessary for exposition, they are omitted. Note that demand is the downward-sloping market demand curve. Marginal revenue is also downwardsloping and lies below the demand curve everywhere except at the vertical intercept. Figure 12.3 shows a situation in which price exceeds average total cost, and thus the monopolist earns an economic profit.

The monopolist maximizes profit by producing 200 units of output where MR = SMC. From the demand curve, the monopolist can (and will) charge $7 per unit. Total revenue is $1,400 (= $7 * 200), or the area of the rectangle 0ABE. The average total cost of producing 200 units of output is $5. Total cost of producing 200 units is $1,000 (= $5 * 200), or the area of the rectangle 0DCE. Economic profit is TR minus TC, $400 (= $1,400 - $1,000), or the shaded area ABCD. Since price is greater than average total cost at the equilibrium output of 200 units, the monopolist earns an economic profit. This need not be the case, however.

People often have the idea that monopoly firms can always make a profit; if the firm is making losses, it can simply raise price until it makes a profit. It is, however, a misconception that all monopolies are ensured a profit. Figure 12.4 illustrates a monopolist that makes losses in the short run. Marginal cost equals marginal revenue at 50 units of output, which, from the demand curve, can be sold for $75 each. Total revenue, then, is $3,750 (= $75 * 50), or the area 0DCE. Since average total cost is $80 per unit, total cost is $4,000 (= $80 * 50), or the area 0ABE. Since total cost exceeds total revenue, the firm experiences a loss of $250 (= $4,000 - $3,750), which is the shaded area ABCD.

Note that in Figure 12.4 the monopolist would produce rather than shut down in the short run since total revenue (area 0DCE) exceeds the total variable cost of $3,250 (= $65 * 50), or area 0GFE. After all variable costs have been covered, there is still some revenue, $500 (area GDCF), left over to apply to fixed cost. Since total fixed cost in this example is $750 (= $15 * 50), or area ABFG, the firm loses less by producing 50 units than by shutting down. If the monopolist shuts down, it would, of course, lose its entire fixed cost of $750.

If demand decreases so that it lies below AVC at every level of output and the monopolist could not cover all its variable cost at any price, the firm would shut down and lose only fixed cost. This is exactly the same shutdown rule as that of the perfect competitor.

We should note that a monopolist would never choose a situation in which it was producing and selling an output on the inelastic portion of its demand. When demand is inelastic, marginal revenue is negative (see Table 6.4, page 225). Since marginal cost is always positive, it must equal marginal revenue when the latter is also positive. Thus the monopolist will always be on the elastic portion of demand.

In the short run, the primary difference between a monopoly and a perfect competitor lies in the slope of the demand curve. Either may earn a pure profit; either may incur a loss.

Principle In the short run, the manager of a monopoly firm will choose to produce the output where MR = SMC, rather than shut down, as long as total revenue at least covers the firm’s total avoidable cost, which is the firm’s total variable cost (TR ≥ TVC). The price for that output is given by the demand curve. If total revenue cannot cover total avoidable cost, that is, if total revenue is less than total variable cost (or, equivalently, P < AVC), the manager will shut down and produce nothing, losing an amount equal to total fixed cost.

Long-Run Equilibrium

A monopoly exists if there is only one firm in the market. Among other things, this statement implies that entry into the market is closed. Thus, if a monopolist earns an economic profit in the short run, no new producer can enter the market in the hope of sharing whatever profit potential exists. Therefore, economic profit is not eliminated in the long run, as was the case under perfect competition. The monopolist, however, will make adjustments in plant size as demand conditions warrant, in order to maximize profit in the long run.

Clearly, in the long run, a monopolist would choose the plant size designed to produce the quantity at which long-run marginal cost equals marginal revenue. Profit would be equal to the product of output times the difference between price and long-run average cost:

π = P * Q - LAC * Q = Q (P - LAC)

New entrants cannot come into the industry and compete away profits—entry will not shift the demand curve facing the monopolist.

Demand conditions may change for reasons other than the entry of new firms, and any such change in demand and marginal revenue causes a change in the optimal level of output in both the short run and the long run. Suppose demand does change, due perhaps to a change in consumer income. In the short run, the manager will adjust output to the level where the new marginal revenue curve intersects the short-run marginal cost curve (or it will shut down if P < AVC). This short-run adjustment in output is accomplished without the benefit of being able to adjust the size of the plant to its optimal size. The plant size that minimizes the cost of production varies with the level of output. Hence, in the long run, the manager would adjust plant size to the level that minimizes the cost of producing the optimal level of output. If there is no plant size for which long-run average cost is less than price, the monopolist would not operate in the long run and would exit the industry.

Principle The manager of a monopoly firm maximizes profit in the long run by choosing to produce the level of output where marginal revenue equals long-run marginal cost (MR = LMC). unless price is less than long-run average cost (P < LAC), in which case the firm exits the industry. In the long run, the manager will adjust plant size to the optimal level; that is, the optimal plant is the one with the short-run average cost curve tangent to the long-run average cost at the profitmaximizing output level.

This principle is illustrated in Figure 12.5. The level of output that maximizes profit in the long run is 350 units, the point at which MR = LMC. In the long run, the manager adjusts plant size so that 350 units are produced at the lowest possible total cost. In Figure 12.5, the optimal plant size is the one with shortrun average total cost and marginal cost curves labeled ATC1 and SMC1, respectively. Thus the average cost of producing 350 units is $50 per unit. The manager will sell the 350 units at a price of $55 to maximize profit. Long-run profit is $1,750 [= Q * (P - LAC) = 350 * ($55 - $50)], or the area ABCD. By the now-familiar argument, this is the maximum profit possible under the given revenue and cost conditions.

I L L U S T R AT I O N 1 2 . 2

Quasi-Fixed Costs and Pricing Decisions by Stainless Steel Makers

The CEO of Universal Stainless & Alloy Products Inc., a capital-intensive manufacturer of specialty steel products that possesses some degree of market power (specialty stainless steel is not a homogeneous commodity) raised its prices twice over a 30-day period, according to a recent article in The Wall Street Journal. According to the CEO, the price hikes were in direct response to falling demand for stainless steel products. Can raising price be the optimal response to falling demand? The short answer: No. In this Illustration we will examine the confusion experienced by Universal and several other stainless steel makers when they raised their prices to counter falling demand for stainless steel product.

The source of confusion for the stainless steel CEOs concerns their apparent misunderstanding of the role fixed costs play in decision making, both fixed and quasi-fixed costs in the stainless steel industry. “Regular” fixed costs do not matter for decision-making purposes. Managers should never make production or pricing decisions for the purpose of spreading fixed costs over more units of output. Furthermore, the firm’s decision to shut down in the short run (or to exit in the long run) is decided by comparing revenue to avoidable costs: When total revenue covers total avoidable costs, keep producing; if not, shut down (or exit in the long run). Fixed costs play no role in the short-run shutdown decision or the pricing and production decisions because they must be paid even if output is zero. In contrast, any quasi-fixed costs a firm might incur do matter, but only for making the shutdown decision in the short run. Since quasi-fixed costs can be avoided if output is zero, the firm’s avoidable costs include both variable costs and quasi-fixed costs. Thus, a stainless steel firm that employs quasi-fixed inputs will produce in the short run only if total revenue covers total avoidable cost, which is the sum of variable costs of production plus the quasi-fixed costs. As you can see, the existence of quasi-fixed costs increases the amount of revenue (i.e., raises the shutdown price) that must be reached in order to continue producing rather than shut down in the short run.

The WSJ article quotes several of the CEOs as they explain their reasoning for raising their specialty steel prices. One stainless steel executive explains his price hike as follows:

Unlike (price) increases announced in recent years, this is obviously not driven by increasing global demand, but rather by fixed costs being proportioned across significantly lower demand.

In other words, decreasing demand caused a drop in the quantity of stainless sold, and this reduction in output caused an increase in average fixed costs since there were fewer tons of stainless steel over which to spread or “proportion” the fixed costs. The figure below shows the demand for a stainless steel maker that possesses some degree of market power and earns positive profit before falling demand causes a loss. The original demand and marginal revenue conditions, denoted DA and MRA, require price to be set at point a on demand in order to maximize profit prior to the decrease in demand (SMC = MRA at Q*A). (Notice that point a lies slightly above ATC, creating a positive profit for the steel firm.) When demand decreases to DB, which also causes marginal revenue to shift to MRB, the new optimal price is at point b on DB (SMC = MRB at Q*B ). As you can see, moving from point a to point b must reduce total revenue because both price and quantity fall. Even if the firm kept price unchanged after the decline in demand (see point f ), total revenue would still decline. In either case, at point f or point b, profit is now negative since price is lower than average total cost.


What should the stainless steel maker do once demand decreases? At least one mistake, raising the price of stainless steel, was made, and perhaps a second mistake was made: continuing to produce stainless steel rather than shutting down operations. Let’s first consider the price hike. Raising price after a decrease in demand (i.e., setting price somewhere above point f along DB), does several undesirable things. Raising price above point f on DB instead of lowering price to point b only adds to the loss of sales when demand for stainless falls. The global recession caused demand for steel to shift leftward, which is something management can do nothing about, but as you can see clearly in the figure, management’s decision to raise prices made matters even worse. Not only will there be fewer tons of steel to spread fixed costs over—a completely irrelevant concern in any case—but stainless steel demand is elastic here and raising price also reduces total revenue! Specifically, by choosing, incorrectly, to set price higher than point f instead of at point b, some total revenue is needlessly sacrificed since demand is elastic at point b (i.e., MR is positive).

On the second issue of shutting down stainless steel production, the CEO must distinguish between fixed cost and quasi-fixed cost, because stainless steel production involves substantial quasi-fixed costs. At Universal’s facility an enormous blower serves as a pre-heater by blasting 2,300 degree Fahrenheit hot air at the giant ladle holding the molten steel. The preheater must continue running even when no steel is being produced in order to prevent a melt-down of the refractory bricks that heat the ladle and to avoid damage caused by starting and stopping the motors. The cost of running the giant pre-heater is the same whether Universal makes one batch or a dozen batches of steel.

It is clear from this description of stainless steel production that running the huge blower represents a quasi-fixed cost: A lump amount of “blowing" is needed for the first ton and this amount is constant as more tons are produced. And, in contrast to “regular” fixed costs, the quasi-fixed cost of running the blower is avoidable if the plant shuts down. Unless the now lower total revenue can cover all the variable costs plus the quasi-fixed costs of running the pre-heater, the stainless steel firm should shut down operation in the short run and lose only its total (unavoidable) fixed cost. According to the article, the stainless steel makers are continuing to produce in the short run, so we will assume that total revenue is indeed sufficient to cover all avoidable costs. We show this situation in our figure. As you can verify at point b, the new price, P*B, is greater than the sum of average variable cost (AVC) plus average quasi-fixed cost (AQFC) at point c on the curve labeled “AVC + AQFC ”.

Now we can explain why raising price (incorrectly) might also cause the firm to make the wrong shutdown decision. Based on the figure we have constructed, the optimal price exceeds average avoidable cost (i.e., point b lies above point c), so total revenue from selling Q*B tons of steel at price P*B will indeed cover total avoidable cost. However, since the CEO made the wrong pricing decision by setting price somewhere above point f on DB, we cannot be sure that total revenue at the wrong price will be sufficient to warrant continued production in the short run. While the WSJ article does not tell us whether or not the firm’s revenue is covering all of its avoidable costs, you can see from this Illustration, nonetheless, how making the wrong pricing decision could cause the firm needlessly to shut down in the short run and lose more than the minimum possible loss.

We can summarize this rather complicated Illustration with two general points. First, as a general rule, when the demand for your product falls, raising price is not the optimal response if profit-maximization is your aim. We cannot help but wonder why the CEO at Universal Stainless & Alloy Products made this decision twice. Perhaps he raised price again after the first price hike failed to raise revenue or reduce average fixed costs. Second, making an incorrect pricing decision by setting price higher than optimal will result in lower total revenue, which could lead the manager mistakenly to shut down.



Thus far we have analyzed monopoly profit maximization in terms of the output decision. As was the case for competition, the manager can also maximize profit by choosing the optimal level of input usage. Choosing the optimal level of input usage results in exactly the same output, price, and profit level as choosing the optimal level of output would. We now discuss the monopoly firm’s input decision assuming that there is only one variable input.

The analytical principles underlying the input decision for the manager of a monopoly are the same as those for managers of perfectly competitive firms. But since price does not equal marginal revenue for a monopoly, P * MP is not the correct measure of the marginal revenue product (MRP)—the increase in revenue attributable to hiring an additional unit of the variable input. Suppose a monopolist employs an additional unit of labor, which causes output to increase by the amount of the marginal product of labor. To sell this larger output, the manager must reduce the price of the good. Each additional unit adds marginal revenue (MR) to total revenue. Thus the additional unit of labor adds to total revenue an amount equal to marginal revenue times the marginal product of labor:


For example, suppose hiring the 10th unit of labor increases output by 20 units (MP = 20). To sell these 20 additional units of output, the monopolist must lower price. Further suppose that marginal revenue is $5 per additional unit. Thus the additional revenue attributable to hiring the 10th unit of labor is the $5 additional revenue received on each of the 20 additional units of output produced and sold, or $100 (= $5 * 20). The marginal revenue product of the 10th unit of labor is $100.

Recall that in the case of perfect competition, marginal revenue product is measured by multiplying price (= MR) by the marginal product of labor. Also recall that MRP for a perfect competitor declines because marginal product declines. For a monopolist, marginal revenue product declines with increases in input usage not only because marginal product declines but also because marginal revenue declines as output is increased.

Figure 12.6 shows the positive portion of MRP below ARP, which is the relevant portion of the MRP curve for a monopolist employing labor as its only variable input. Just as for a perfectly competitive firm, a monopolist shuts down and hires no labor when the wage rate exceeds average revenue product (w > ARP) at the level of input usage where MRP = w. Suppose the wage rate is $45. To maximize profit, the manager should hire 400 units of labor at a wage rate of $45. To see why this is the optimal level of labor usage, suppose the manager hires only 300 units of labor. Hiring the 301st unit of labor adds slightly less than $58 to total revenue while adding only $45 to total cost. Clearly, hiring the 301st unit increases profit, in this case, $13 (= $58 - $45). The manager should continue to hire additional units of labor until MRP = w1 = $45 at point A in Figure 12.6. If the manager mistakenly hired more than 400 units, say, 500 units of labor, the additional revenue from hiring the last unit of labor ($30 for the 500th unit) is less than the additional cost, $45, and profit falls if the 500th worker is hired. Getting rid of the 500th worker lowers cost by $45 but revenue falls by only $30; thus, reducing labor by 1 unit increases profit by $15. And each additional 1 unit reduction in labor similarly increases profit until labor usage is reduced down to the 400th worker.

If the wage rate falls to $30 per unit (shown by the horizontal line w2), the manager should hire 500 units of labor (point B) to maximize monopoly profit. Similarly, at a wage of $58, the manager would hire 300 workers (point C). Thus you can see that, over the relevant range, the MRP curve is the monopolist’s demand curve for a single variable input.

We now show that a monopolist would never choose a level of variable input usage at which the average revenue product is less than the marginal revenue product (ARP < MRP). If, at the level of input usage where MRP = w,



w > PQ/L


wL > PQ

which implies that total variable cost exceeds total revenue, and the profitmaximizing monopolist would hire 0 units of the variable input and shut down.

Principle When producing with a single variable input, a monopolist will maximize profit by employing that amount of the input for which marginal revenue product (MRP) equals the price of the input when input price is given. Consequently, the MRP curve, over the relevant range, is the monopolist’s demand curve for the variable input when only one variable input is employed. The relevant range of the MRP curve is the downward-sloping, positive portion of MRP for which ARP > MRP.

As for the case of a competitive firm, the manager of a firm with market power that employs two or more variable inputs maximizes profits by choosing input levels so that the marginal revenue product equals the input price for all inputs simultaneously.

Recall that, for a price-taking firm, the profit-maximizing condition that the marginal revenue product of labor equals the wage rate (MRP = w) is equivalent to the profit-maximizing condition that product price equals marginal cost (P = SMC). By “equivalent” we mean that regardless of whether the manager chooses Q or L to maximize profit, the resulting levels of output, labor usage, and profit are identical. We will now demonstrate that, for a monopolist, the profit-maximizing condition MRP = w is equivalent to the profit-maximizing condition MR = SMC.

Suppose the manager of a monopoly firm chooses the level of output to maximize profit. The optimal output for the monopolist is where


Recall that

SMC = w/MP

where MP is the marginal product of labor and w is its price. Substituting this equation for marginal cost, the profit-maximizing condition MR = SMC can be expressed as

MR = w/MP


MR * MP = w

MRP = w

Thus you can see that the two profit-maximizing rules are equivalent: MR = MC implies MRP = w, and vice versa.

Relation For a monopolist, the profit-maximizing condition that the marginal revenue product of the variable input must equal the price of the input (MRP = w) is equivalent to the profit-maximizing condition that marginal revenue must equal marginal cost (MR = MC). Thus, regardless of whether the manager chooses Q or L to maximize profit, the resulting levels of input usage, output, price, and profit are the same in either case.



As we pointed out at the beginning of this chapter, the general model of monopoly is useful in the analysis of firm behavior in other types of markets in which firms have some degree of market power but are not pure monopolies. Firms in such markets, facing downward-sloping demands, attempt to maximize profit in the same way a monopoly does: by setting MR = MC. In these intermediate markets, between firms with the most market power (monopoly) and firms with the least (perfect competition), certain complications arise for the profit-maximizing decision. In this section, we analyze intermediate market structure in which firms have the least market power of all firms that are not perfect competitors: monopolistic competition.

Monopolistically competitive markets are characterized by (1) a large number of relatively small firms; (2) products that are similar to, but somewhat different from, one another; and (3) unrestricted entry and exit of firms into and out of the market. The only difference between monopolistic competition and perfect competition is that under monopolistic competition firms produce a differentiated product. The major difference between monopolistic competition and monopoly is that under monopolistic competition firms can easily enter into and exit out of the market. Thus, as the name implies, monopolistic competition has characteristics of both monopoly and perfect competition.

Product differentiation under monopolistic competition prevents a firm’s demand from becoming horizontal. Real or perceived differences between goods, though slight, will make them less than perfect substitutes. For example, gasoline stations in a particular city are good, but not perfect, substitutes for one another. Your car would run on gasoline from any gasoline station, but stations differ in location, and people’s tastes differ: Some people prefer BP, some prefer ExxonMobil, some prefer the service at Joe’s, others prefer Julie’s service. And the differentiating characteristics go on and on. The most important point is that although the products are similar, they are differentiated, causing each firm to have a small amount of market power.

We will first set forth the theory of monopolistic competition in its original form, as developed by Edward Chamberlin in the 1930s.4 Because each firm in the market sells a slightly differentiated product, it faces a downward-sloping demand curve, which is relatively elastic but not horizontal. Any firm could raise its price slightly without losing all its sales, or it could lower its price slightly without gaining the entire market. Under the original set of assumptions employed by Chamberlin, each firm’s output is so small relative to the total sales in the market that the firm believes that its price and output decisions will go unnoticed by other firms in the market. It therefore acts independently.

As you will see, the theory of monopolistic competition is essentially a long-run theory; in the short run, there is virtually no difference between monopolistic competition and monopoly. In the long run, because of unrestricted entry into the market, the theory of monopolistic competition closely resembles the theory of perfect competition.

Short-Run Equilibrium

With the given demand, marginal revenue, and cost curves, a monopolistic competitor maximizes profit or minimizes loss by equating marginal revenue and marginal cost. Figure 12.7 illustrates the short-run, profit-maximizing equilibrium for a firm in a monopolistically competitive market. Profit is maximized by producing an output of Q and selling at price P.

In the situation illustrated, the firm will earn an economic profit, shown as the shaded area PABC. However, as was the case for perfect competition and monopoly, in the short run the firm could operate with a loss, if the demand curves lies below ATC but above AVC. If the demand curve falls below AVC, the firm would shut down.

In its original form, there appears to be little competition in monopolistic competition as far as the short run is concerned. Indeed Figure 12.7 is identical to one illustrating short-run equilibrium for a monopoly. In the long run, however, a monopoly cannot be maintained if there is unrestricted entry into the market. If firms are earning economic profit in the short run, other firms will enter and produce the product, and they will continue to enter until all economic profits are eliminated.

Long-Run Equilibrium

While the short-run equilibrium for a firm under monopolistic competition is similar to that under monopoly, the long-run equilibrium is more closely related to the equilibrium position under competition. Because of unrestricted entry, all economic profit must be eliminated in the long run, which occurs at an output at which price equals long-run average cost. This occurs when the firm’s demand is tangent to long-run average cost. The only difference between this equilibrium and that for perfect competition is that, for a firm in a monopolistically competitive market, the tangency cannot occur at minimum average cost. Since the demand curve facing the firm is downward-sloping under monopolistic competition, the point of tangency must be on the downward-sloping range of long-run average cost. Thus the long-run equilibrium output under monopolistic competition is less than that forthcoming under perfect competition in the long run.

This long-run result is shown in Figure 12.8. LAC and LMC are the long-run average and marginal cost curves for a typical monopolistically competitive firm. Suppose that the original demand curve is given by Dm. In this case the firm would be making substantial economic profits because demand lies above LAC over a wide range of output, and if this firm is making profits, potential new entrants would expect that other firms in the market are also earning economic profits. These profits would then attract new firms into the market. While the new firms would not sell exactly the same products as existing firms, their products would be very similar. So as new firms enter, the number of substitutes would increase and the demand facing the typical firm would shift backward and probably become more elastic (though not perfectly elastic). Entry will continue as long as there is some economic profit being earned. Thus entry causes each firm’s demand curve to shift backward until a demand curve such as D in


Relation Long-run equilibrium in a monopolistically competitive market is attained when the demand curve for each producer is tangent to the long-run average cost curve. Unrestricted entry and exit lead to this equilibrium. At the equilibrium output, price equals long-run average cost and marginal revenue equals long-run marginal cost.

In closing our discussion of monopolistic competition, we briefly mention two points. First, according to the original model as set forth here, firms act independently when making decisions, ignoring the actions of other firms in the market. In reality firms may not act independently when faced with competition from closely related firms, possibly because of proximity; in fact, they may exhibit a great deal of interdependence and intense personal rivalry. This change in assumptions will not alter the long-run, zero-profit conclusions of the theory, however.

Short of getting the government to prevent entry, there is nothing firms in a monopolistically competitive market can do about having their profits competed away. Even if the firms were to conspire to fix a price, new firms would enter. Each firm would find its demand decreased and its sales reduced until price equaled average cost and economic profits were zero, although possibly at a higher price than would occur in the absence of the price-fixing agreement.

Second, we have emphasized that, under monopolistic competition, profits are competed away in the long run. This is correct in general. But we do not mean to imply that there is no opportunity for astute managers to postpone this situation to the future by innovative decision making. Firms selling in monopolistically competitive markets can and do advertise and change product quality in an effort to lengthen the time period over which they earn economic profit. Those managers who are successful in their Marketing Strategy can sometimes earn profit for a long time. Some firms can reduce their cost. However, successful strategies can be imitated by competitors’ selling a product that is rather similar. Therefore, under monopolistic competition there is always a strong tendency for economic profit to be eliminated in the long run, no matter what strategies managers undertake.



Thus far in our discussion of profit-maximizing decisions, we have assumed that the manager has only one plant in which to produce the firm’s product. Many firms, however, produce output in more than one plant. We will now show you how firms should allocate production among multiple production facilities. Even though our discussion here focuses on a firm with market power, the rule that we develop here applies to all firms, regardless of the degree of market power possessed.

When firms produce in more than one plant, it is likely that the various plants will have different cost conditions. The problem facing the firm is how to allocate the firm’s desired level of total production among these plants so that the total cost is minimized.

For simplicity, suppose there are only two plants, A and B, producing the desired total output level (QT) of 450 units, but at different marginal costs such that


where plant A produces 160 units (QA) and plant B produces 290 units (QB). In this situation, the manager should transfer output from the higher-cost plant A to the lower-cost plant B. As long as the marginal cost of producing in plant B is lower, total cost of producing QT units can be lowered by transferring production. For example, suppose MCA equals $25 (for the 160th unit at plant A) and MCB equals $10 (for the 290th unit at plant B). One unit of output taken away from plant A lowers the firm’s total cost by $25. Making up the lost unit by producing it at plant B increases total cost only by $10, and 450 units are still produced. As you can see, however, the total cost of producing 450 units falls by $15. The firm would continue taking output away from plant A and increasing the output of plant B, thus lowering total cost, until MCA = MCB. This equality would result because MCA falls as the output of plant A decreases and MCB rises as the output of plant B increases. Thus we can conclude that marginal costs must be equal for both plants in order to minimize the total cost of producing 450 units.

Principle For a firm that produces using two plants, A and B, with marginal costs MCA and MCB, respectively, the total cost of producing any given level of total output QT (= QA + QB) is minimized when the manager allocates production between the two plants so that the marginal costs are equal: MCA = MCB.

The total output decision is easily determined. The horizontal summation of all plants’ marginal cost curves is the firm’s total marginal cost curve. This total marginal cost curve is equated to marginal revenue in order to determine the profitmaximizing output and price. This output is divided among the plants so that the marginal cost is equal for all plants.

The two-plant case is illustrated in Figure 12.11. Demand facing the firm is D, and marginal revenue is MR. The marginal cost curves for plants A and B are, respectively, MCA and MCB. The total marginal cost curve for the firm is the horizontal summation of MCA and MCB, labeled MCT. Profit is maximized at that output level where MCT equals marginal revenue, at an output of 175 units and a price of $45. Marginal cost at this output is $20. Equalization of marginal cost requires that plant A produce 50 units and plant B produce 125 units, which of course sums to 175 since MCT is the horizontal summation of MCA and MCB. This allocation equalizes marginal cost and consequently minimizes the total cost of producing 175 units.

To further illustrate the principle of optimally allocating output in a multiplant situation, we turn now to a numerical illustration. As you will see, the algebra is somewhat more complex than it is for the single-plant case, but the principle is the same: The manager maximizes profit by producing the output level for which marginal revenue equals marginal cost.

Multiplant Production at Mercantile Enterprises

Mercantile Enterprises—a firm with some degree of market power—produces its product in two plants. Hence, when making production decisions, the manager of Mercantile must decide not only how much to produce but also how to allocate the desired production between the two plants.

The production engineering department at Mercantile was able to provide the manager with simple, linear estimates of the incremental (marginal) cost functions for the two plants:

MCA = 28 + 0.04QA and MCB = 16 + 0.02QB

Note that the estimated marginal cost function for plant A (a plant built in 1998) is higher for every output than that for plant B (a plant built in 2005); plant B is more efficient.

The equation for the total marginal cost function (the horizontal sum of MCA and MCB) can be derived algebraically using the following procedure. First, solve for both inverse marginal cost functions:

QA = 25MCA - 700


QB = 50MCB - 800

Next, QT (= QA + QB) is found by summing the two inverse marginal cost functions. Recall, however, that the horizontal summing process requires that MCA = MCB = MCT for all levels of total output QT. Thus it follows that

QA = 25MCT - 700


QB = 50MCT - 800

Summing the two inverse marginal cost functions results in the inverse total marginal cost function:

QT = QA + QB = 75MCT - 1,500

which, after taking the inverse to express marginal cost once again as a function of output, results in the total marginal cost function:

MCT = 20 + 0.0133QT

The marginal cost functions for plants A and B and the associated total marginal cost function are shown in Panel Aof Figure 12.12. The process of horizontal summation can be seen by noting that when MC = $40, QA = 300 units (point A), QB = 1,200 units (point B), and QT = QA + QB = 1,500 units (point C). Thus, if 1,500 units are to be produced, the manager should allocate production so that 300 units are produced in plant Aand 1,200 units are produced in plant B. This allocation of production between the two plants minimizes the total cost of producing a total of 1,500 units.

Note that when QT is less than 600 units, plant A is shut down and only plant B is operated. Until Mercantile increases total production to 600 units or more (point K), the marginal cost of producing any output at all in plant A is greater than the marginal cost of producing additional units in plant B. For output levels in the 0 to 600-unit range, MCB is the relevant total marginal cost curve since QA = 0. For total output levels greater than 600 units, Mercantile Enterprises will operate both plants and MCT is the total marginal cost function.

Suppose that the estimated demand curve for Mercantile’s output is

QT = 5,000 - 100P

The inverse demand function is

P = 50 - 0.01QT

and marginal revenue is

MR = 50 - 0.02QT

Equating marginal revenue and total marginal cost,

50 - 0.02QT = 20 + 0.0133QT

 and solving for QT, the profit-maximizing level of output for Mercantile Enterprises is Q*T = 900. At this output level, marginal revenue and total marginal cost are both $32 at point E in Panel B of Figure 12.12. To minimize the cost of producing 900 units, the production of the 900 units should be allocated between plants A and B so that the marginal cost of the last unit produced in either plant is $32:

MCA = 28 + 0.04QA = 32 and MCB = 16 + 0.02QB = 32

Hence, for plant A, Q*A = 100, so 100 units will be produced in plant A. For plant B, Q*B = 800, so 800 units will be produced in plant B.

Now suppose that forecasted demand decreases and a new forecast of the demand for Mercantile’s output is

QT = 4,000 - 100P

Given that the corresponding marginal revenue function is

MR = 40 - 0.02QT

the firm’s profit-maximizing output (where MR = MCT) declines to 600 units. At this output, marginal revenue and marginal cost are both $28. Equating MCA and MCB to $28, the manager found that for plant A, Q*A = 0, and for plant B, Q*B = 600. With the new (lower) forecast of demand, plant A will be shut down and all the output will be produced in plant B. As you can verify, if demand declines further, Mercantile would still produce, using only plant B. So for output levels of 600 or fewer units, the total marginal cost function is MCB.

In effect, the total marginal cost function has a “kink” at point K in the figure. The kink at point K represents the total output level below which the high-cost plant is shut down. Akink occurs when marginal cost in the low-cost plant equals the minimum level of marginal cost in the high-cost plant, thereby making it optimal to begin producing with an additional plant. [The low-cost (high-cost) plant is the plant with lowest (highest) marginal cost at Q = 0.] The output at which the kink occurs is found by setting marginal cost in the low-cost plant equal to the minimum value of marginal cost in the high-cost plant:

MCB = 28 = 16 + 0.02Q

so the high-cost plant begins operating when Q exceeds 600 units.

The preceding discussion and example show how a manager should allocate production between two plants to minimize the cost of producing the level of output that maximizes profit. The principle of equating marginal costs applies in exactly the same fashion to the case of three or more plants: Marginal cost is the same in all plants that produce. The only complication arises in the derivation of total marginal cost.

Once the total marginal cost function is derived, either by summing the individual plants’ marginal cost curves graphically or by solving algebraically, the manager uses the total marginal cost function to find the profit-maximizing level of total output.

Principle A manager who has n plants that can produce output will maximize profit when the firm produces the level of total output and allocates that output among the n plants so that

MR = MCT = MC1 = . . . = MCn