Chapter 23 Monopoly We will now turn toward an analysis of the polar opposite of the extreme assumption of perfect competition that we have employed thus far.1 Under perfect competition, we have assumed that industries are composed of so many small firms that each firm has no impact on the economic environment in which decisions are made. As a result, we could assume that individual firms in an industry simply take the market price as given as they determine how much to produce in order to maximize profits. In the case of a monopoly, on the other hand, the firm must make a decision not only on how much to produce but also on what price to charge. There is, in the case of monopoly, no “market” to set the price. In this sense, the monopolist has some control over her economic environment (i.e. prices) that the competitive producer lacks. While we will often talk about a “monopoly” as if it was a fixed concept, it is important to keep in mind that monopoly power comes in more and less concentrated doses. Under perfect competition, the demand that a firm faces for its product is perfectly elastic because of the existence of many firms that produce the same product at the market price. Whenever a firm faces a demand curve for its product that is not perfectly elastic, it has some market power. For instance, I might produce a particular soft drink in a largely competitive market for soft drinks, but my soft drink is nevertheless a bit distinctive. In a sense, my soft drink is therefore a separate product with a separate market, but in another sense it is part of a larger market in which other firms produce close but imperfect substitutes. The demand curve for my soft drink may then not be perfectly elastic – giving me some market power, but that power is limited by the fact that there are close substitutes in the larger soft drink market. If, on top of the existence of close substitutes, there is free entry into the soft-drink market, my market power is limited even more. We will treat this type of market in Chapter 26 as one characterized by “monopolistic competition.” In other settings, of course, there is less of an availability of substitutes for a particular firm’s product. If there are market entry barriers that keep potential competitors from producing substitutes, my monopoly power would then be considerably more pronounced, and the demand for my product considerably less elastic. For now, we will simply treat monopolies as firms that face downward sloping demand curves in an environment where barriers to entry keep other firms from entering to produce substitute goods, and we will simply keep in mind that the elasticity of demand 1 This chapter presumes a basic understanding of demand and makes frequent references to the partial equilibrium models of Chapters 14 and 15. It furthermore presumes a basic understanding of cost curves as derived in Chapter 11 and summarized in Section 13A.1 of Chapter 13. 856 Chapter 23. Monopoly for the monopoly’s product is closely connected to just how powerful a monopoly we are dealing with. When we get to Chapter 26, it will become clear that the stark model of monopoly in this chapter is an extreme model that rarely holds fully in the real world, but it gives us a good starting point to talk about market power – just as perfect competition gives us a useful starting point to talk about competition. 23A Pricing Decisions by Monopolist We begin our analysis of monopoly power by analyzing how the profit maximizing condition of marginal revenue being equal to marginal cost translates into optimal firm decision making when a firm faces a downward sloping demand curve. At first, we’ll assume that the firm is restricted in its pricing policy in the sense that it can only set a single price per unit of output – a single price that is charged to every consumer. We then proceed to think about how a monopoly might want to differentiate the price it charges to different consumers – and under what conditions that is possible. Finally, we will talk explicitly about what kinds of barriers to entry might in fact result in real-world monopolies, and how the nature of the barrier to entry might determine the extent to which we think monopoly power is a problem that requires government intervention. Before moving on, however, recall the two ways in which we thought about profit maximization for price taking firms in Chapter 11. We first set up the profit maximization problem under the assumption that the competitive firm takes price as fixed and solved for the profit maximizing production plan by finding the tangency between isoprofit curves with production frontiers. This method no longer holds for monopolists – because the method presumed a fixed price that the price setting firm simply took as given. We then developed a two-step profit maximization method – with the first step focusing solely on the cost side (where firms attempt to minimize cost) and the second step adding revenue considerations (given the price that competitive firms take as given). Since output price plays no role in the cost-minimizing problem where the firm simply asks “what is the least cost way of producing different levels of output”, this step is the same for monopolists. The difference enters in the second step where we compare revenue to cost – with revenue for the monopolist depending on the price that the monopolist chooses (rather than the price that is set by the market). We can therefore use everything we learned about cost curves – marginal costs, average costs, recurring fixed costs, etc. – and will thus focus on step 2 of the two-step profit maximization method in analyzing monopoly decisions. 23A.1 Demand, Marginal Revenue and Profit For competitive producers, price is the same as marginal revenue. Put differently, the competitive producer knows that she can sell any amount of the good she could feasibly produce at the market price, and so the marginal revenue she receives for each good she produces is simply the price set by the interactions of producers and consumers in market equilibrium. She could, of course, choose to sell her goods at a lower price, but that would not be profit maximizing. If, on the other hand, she tries to sell her goods at a price above the market price, consumers will simply shop at a competitor. While the market demand curve in competitive markets is therefore downward sloping, the demand curve for each competitive producer is perfectly elastic at the market price. For a monopolist, however, the market demand is the same as the firm’s demand since the monopolist is the only producer in the market. As a result, the monopolist gets to choose a point on the market demand curve — which involves a simultaneous choice of how much to produce and 23A. Pricing Decisions by Monopolist 857 how much to charge. When a monopolist decides to increase output, she therefore confronts the following trade-off: On the one hand, she gets to sell more goods to consumers, but on the other hand she sells all her goods at a lower price than before. Thus, as a monopolist increases output, her marginal revenue is not equal to the price she charged initially because she will have to lower price in order to sell the additional output. 23A.1.1 Marginal Revenue along a Market Demand Curve Suppose we consider a demand curve first illustrated in Graph 18.3 in Chapter 18 and replicated here as Graph 23.1a. The first unit produced by a monopolist facing such a market demand for her goods can be sold for approximately $400. Thus, the marginal revenue for the first unit of output is approximately $400. Next, suppose the monopolist was currently producing 199 units of the output for $300.50 each. Were this monopolist to produce two additional units of output, she would have to lower her price to $299.50 in order to sell all 201 goods. She would therefore experience an increase in her total revenues of $599 for the 200th and 201st good, but she would simultaneously lose $1 on each of the first 199 goods she is producing. Her marginal revenue from producing two additional units is therefore $400, or approximately $200 for each of the two units. Graph 23.1: Linear Demand and Marginal Revenue Next, suppose that the monopolist was producing 399 units, selling each at $200.50, and suppose she considered producing two additional units. She would then have to lower the price to $199.50 in order to sell the additional two units, earning an additional revenue of $399 on those units but losing $399 on the units she previously produced because she had to lower the price by $1 for each of the 399 units. Thus, her marginal revenue from producing two additional units is 0. The marginal revenue curve for this monopolist is then depicted in panel (b) of Graph 23.1. It begins at the same point as the demand curve because the marginal revenue of the first good is approximately $400. When the monopolist is at approximately point B on the market demand curve, we demonstrated above that her marginal revenue from producing an additional unit is approximately $200, and when the monopolist is at approximately point A on her demand curve, 858 Chapter 23. Monopoly her marginal revenue from producing an additional unit is approximately 0. Connecting these gives us the blue line that shares the intercept of the demand curve but has twice the slope. Exercise 23A.1 What is the marginal revenue of producing an additional good if the producer is at point C on the demand curve in Graph 23.1? 23A.1.2 Price Elasticity of Demand and Revenue Maximization You can already see in Graph 23.1 that marginal revenue is positive when price elasticity is below -1, becomes zero as the price elasticity of demand approaches -1 (and becomes negative when price elasticity lies between -1 and 0). This implies that total revenue for the monopolist increases as she moves down the demand curve until she reaches the midpoint where price elasticity is equal to -1, and total revenue falls if she moves beyond that midpoint into the range of the demand curve where price elasticity is between -1 and 0. As a result, the maximum revenue the monopolist can raise occurs at the midpoint of a linear demand curve where price elasticity is equal to -1. Exercise 23A.2 Where does M R lie when price elasticity falls between -1 and 0? This is closely related to our discussion of consumer spending and price elasticity in Chapter 18. In Graph 18.4, we illustrated that consumer spending rises with an increase in price along the inelastic portion of demand while it falls with an increase in price along the elastic portion of demand. For the monopolist, consumer spending is the same as revenue. Thus, if a monopolist finds herself on the inelastic portion of demand, she knows she can increase revenue by raising the price. If, on the other hand, she finds herself on the elastic portion of demand, she can increase revenue by lowering price. Consumer spending — and thus revenue — is therefore maximized when price elasticity of demand is exactly -1. Exercise 23A.3 Where does a monopolist maximize revenue if she faces a unitary elastic demand curve such as the one in Graph 18.5? 23A.1.3 Profit Maximization for a Monopolist Like all producers, however, monopolists do not maximize revenue — they try to maximize profit which is economic revenue minus economic costs. Thus, in order for us to see what combination of price and quantity a monopolist will choose (assuming she produces at all), we need to know not only marginal revenue but also marginal cost. First, suppose that the marginal cost of producing is zero. In that case, the monopolist’s M C curve is a flat line that lies on the horizontal axis on Graph 23.1b, intersecting the M R curve at 400 units of output. If the monopolist has no variable costs, maximizing revenue and maximizing profit is exactly the same thing — and so the monopolist would simply choose point A on the demand curve where price elasticity is exactly equal to -1. By selling 400 units at $200 each, revenue and profit (not counting recurring fixed costs) is then equal to $80,000. So long as recurring fixed costs are not larger than $80,000, the monopolist would then choose to produce 400 units of output in both the short and the long run. Exercise 23A.4 True or False: If recurring fixed costs are $40,000, then the monopolist will earn $80,000 in short run economic profit and $40,000 in long run economic profit. 23A. Pricing Decisions by Monopolist 859 Next, suppose that the monopolist has the more common U-shaped M C curve depicted in Graph 23.2a. If this monopolist produces a positive quantity, she will choose the quantity xM where M C intersects M R and charge the price pM that allows her to sell everything she is producing. So long as the short run average (variable) cost at xM is less than pM , this implies the monopolist will in fact produce in the short run, and so long as average long run cost (including recurring fixed costs) at the quantity xM lies below pM , she will produce in the long run. Graph 23.2: Profit Maximization for a Monopolist Exercise 23A.5 Suppose M C is equal to $200 for all quantities for a monopolist who faces a market demand curve of the type in Graph 23.1. At what point on the demand curve will she choose to produce? The first thing we can then observe is that, whenever M C is positive, a monopolist will choose to produce on the elastic part of demand. This is because, for any positive M C, the intersection of M C and M R must lie to the left of the intercept of M R with the horizontal axis — which in turn occurs where price elasticity is exactly equal to -1. This should make intuitive sense: We know that, if a monopolist ever finds herself on the inelastic portion of demand, she can raise revenue by increasing price and producing less. If producing costs something, this implies that, whenever a monopolist is on the inelastic portion of demand, she can raise revenue and reduce costs by producing less and charging a higher price. As a result, it makes no sense for a monopolist to produce on the inelastic portion of demand. Exercise 23A.6 Suppose a deep freeze causes the Florida orange crop to be reduced by 50% causing the price for oranges to increase. As a result, we observe that the total revenues of Florida orange growers increases. Could the Florida orange industry be a monopoly? (Hint: The answer is no.) Second, the concept of a “supply curve” that we developed for competitive firms does not make any sense when we talk about monopolists. A supply curve illustrates the relationship between the price set by the market and the quantity of output produced by a profit maximizing firm. But a monopolist does not have a “market” that sets price — the monopolist herself sets the price. Thus, 860 Chapter 23. Monopoly for any given demand curve and any technology that results in cost curves, the monopolist simply picks a supply point. 23A.1.4 Monopoly and Dead Weight Loss Finally, we can see in Graph 23.2 that the profit maximizing monopolist will produce an inefficiently low quantity. In panel (b) of the graph, consumer surplus (assuming no income effects) can be identified as area (a + b) and monopolist surplus (in the short run, or in the absence of recurring fixed costs) as area (c + d). But there are additional units of output that could be produced at a marginal cost below the value consumers place on that output. Such additional output could be produced all the way up to the intersection of M C and demand at output x∗ , and additional surplus of (e) could be produced if a benevolent social planner rather than a monopolist were in charge of production. Thus, area (e) is a deadweight loss which arises because the monopolist strategically restricts output in order to raise price to its profit maximizing level. Notice that the deadweight loss does not arise because the monopolist makes a profit. Even if a social planner forced the monopolist to produce the quantity x∗ and sell it at the appropriate price along the demand curve, the monopolist might make a profit — it just would not be as large as it is when the monopolist raises price to pM and restricts output. Rather, the deadweight loss emerges from the fact that the monopolist is using her power to strategically restrict output in order to raise price. The monopolist’s market power then causes self-interest to come into conflict with the “social good” – at least when the social good is measured in efficiency terms – unless something else interferes and causes the monopolist to produce more. Exercise 23A.7 Suppose that demand is as depicted in Graph 23.1 and M C=0. What is the monopolist’s profit maximizing output level and what is the efficient output level? What if M C=300? Exercise 23A.8 True or False: Depending on the shape of the M C curve, the efficient output level might lie on the elastic or the inelastic portion of the demand curve. Exercise 23A.9 True or False: In the presence of negative production externalities, a monopolist may produce the efficient quantity of output. Exercise 23A.10 True or False: If demand were not equal to marginal willingness to pay (due to the presence of income effects on the consumer side), the deadweight loss area may be larger or smaller but would would nevertheless arise. 23A.1.5 Monopoly Rent Seeking Behavior and Dead Weight Loss We have demonstrated that monopolists are able to achieve economic profits if they have indeed secured monopoly power in some way. We have furthermore demonstrated that this economic profit comes at a social cost as the monopolist produces below the socially optimal level in order to raise price above marginal cost – and we have denoted that social cost as deadweight loss. The actual deadweight loss may, however, be larger than what we have derived thus far because firms may engage in socially wasteful activity in order to secure and maintain the monopoly power that gives them the opportunity to generate economic profits. There are a variety of ways in which barriers to entry that lead to monopoly power can arise, and we will say more about this later on in this chapter. One possibility, for instance, is that monopoly power is granted through government intervention – with governments granting to a single firm 23A. Pricing Decisions by Monopolist 861 the exclusive right to produce a certain product. In such circumstances, firms may compete for such government favor – in the process expending resources on lobbying politicians. The maximum amount that a firm would be willing to invest in order to secure a government granted monopoly is then equal to the present discounted value of the future profits the firm can expect to make from exercising its monopoly power. It is therefore conceivable that firms will expend resources equal to their monopoly profits in order to get the monopoly power, and it is similarly conceivable that much of these resources are spent in socially wasteful ways. This is referred to as political “rent seeking” – i.e. the seeking of “rents” or “profits” in the political arena. To the extent to which the resources spent on political rent seeking are socially wasteful, this would add to deadweight loss beyond what we have derived in our graphs thus far. 23A.2 Market Segmentation and Price Discrimination So far, we have assumed that the monopolist is constrained in the sense that she can only charge a single price to all of her customers. This is the case when a monopolist cannot effectively differentiate between consumers and their marginal willingness to pay for her product, or when charging different prices to different consumers is illegal. In this section, we will suppose that charging different prices to different consumers — a practice known as price discrimination, is permitted and that the monopolist can segment the set of consumers into those that are willing to pay relatively more and those that are willing to pay relatively less. Even when a monopolist can segment the market into different types of consumers, however, she must have some way of preventing resale to keep those consumers that purchase the product at a low price from selling to those that are being offered the same product at a higher price. Below, we will illustrate three different ways in which monopolists may price discriminate under different circumstances. We will begin with the case where monopolists can perfectly identify each consumer’s demand and can offer each consumer a particular quantity at a particular overall price for that quantity. This is known as perfect (or “first degree”) price discrimination. Then we will consider a case where the monopolist, while still being able to identify each consumer’s demand perfectly, can offer different per-unit prices to different customers who potentially want to buy multiple units of the good. We will call this imperfect (or “third degree”) price discrimination. Finally, we will consider the case where a monopolist knows that there are different types of consumers with different demands but she does not know what type each particular consumer is. We will see that the monopolist can then construct price/quantity packages that cause customers to “reveal their type”, a practice known as “second degree” price discrimination. 23A.2.1 Perfect (or “First Degree”) Price Discrimination We can begin with another extreme assumption: Suppose that the monopolist knows all of her customers extremely well and can thus perfectly ascertain each consumer’s willingness to pay for her product. For example, suppose that I am an artist that has his own studio and gallery. I am the only one who produces my unique type of art, and I know my customers personally and invite them individually to sip snooty wine while pretentiously gazing at my art. To make the analysis as simple as possible, let’s further suppose that each of my clients will buy a single piece of art from me. (After all, my art is so special that owning a single piece produces complete intoxication as my clients spend all their time simply gazing at their wall to view it.) The demand curve for my art is then composed of many different individuals who each place a certain value on one of my pieces of art. As I produce my art, I can therefore invite first the 862 Chapter 23. Monopoly individual who places the most value on my art — and who therefore sits at the very top of the demand curve that I face. Suppose this individual of impeccable taste places a value of $10,000 on my art. In that case, I will charge that individual exactly $10,000. Next, I invite my second biggest fan who might place a value of only $9,900 on my art. I can then sell a piece of art to this individual for exactly $9,900. My marginal revenue for the first piece was $10,000, and my marginal revenue for the second piece was $9,900. Since I can charge different prices to each of my clients, I can therefore produce a second piece of art without foregoing any profit on the first piece. As a result, the demand curve becomes my marginal revenue curve when I can price discriminate perfectly between all my clients. Graph 23.3 illustrates the behavior by a profit maximizing producer who can perfectly price discriminate in this way. Since demand is equal to M R, this producer simply chooses to produce xM where M C intersects demand. No single price is charged as each consumer is charged exactly what she is willing to pay along the market demand curve. Consumers therefore attain no surplus — and all the surplus, equal to the shaded area, accrues to the monopolist. In the process, the efficient quantity is supplied — with any additional quantity costing more than the level at which it is valued in society. Graph 23.3: Perfect Price Discrimination This form of perfect price discrimination is also referred to as first degree price discrimination. While it leads to an efficient quantity of output, it clearly leaves consumers worse off than the non-price discriminating outcome in the previous section. This is because consumers now attain no consumer surplus while they do attain some consumer surplus (albeit at a lower output level) when there is no price discrimination. Efficiency is, as we know, a statement about the maximum overall surplus and says nothing about whether the distribution of the surplus is desirable. Exercise 23A.11 * We simplified the analysis by assuming that each person will buy only 1 piece of art. How would you extend the idea of perfect price discrimination (resulting in demand being equal to marginal revenue) to the case where consumers bought multiple pieces? (The answer is provided in the next section.) 23A.2.2 Imperfect or “Third Degree” Price Discrimination Perfect price discrimination assumes that a monopolist can not only identify perfectly each type of consumer’s demand but can also charge an amount that is exactly equal to each consumer’s total 23A. Pricing Decisions by Monopolist 863 willingness to pay. In our somewhat artificial example of my art studio, we had assumed that each consumer only demands one piece of art (implicitly assuming that the marginal value of the second piece is zero for each consumer). As a result, perfect price discrimination meant that I simply arrived at an individualized price equal to exactly each consumer’s willingness to pay for one piece of art. More generally, consumers have downward sloping demand curves and thus place value on more than one unit of output. Consider, for instance, two types of consumers whose demands are given as D1 and D2 in panels (a) and (b) of Graph 23.4. Suppose further that the producer faces a constant marginal cost of $10 per unit of output. Under perfect price discrimination, the producer would sell 200 units of the output to type 1 consumers and 100 units of the output to type 2 consumers, and she would charge type 1 consumers the entire shaded blue area in panel (a) and type 2 consumers the entire shaded magenta area in panel (b). Thus, when consumers place value on more than one good, perfect price discrimination implies that the monopolist will not charge a per unit price but rather a single price for all the units sold to a consumer together. Graph 23.4: Imperfect (“Third Degree”) Price Discrimination Exercise 23A.12 We might also think of the perfectly price discriminating firm as charging what is called a “two-part tariff ” which consists of a fixed payment that is independent of the quantity a consumer buys and a per-unit price for each unit purchased. Can you identify in the graph above which portion would be the fixed payment and what would be the per unit price for each of the two consumers? In many situations, this seems rather unrealistic. Instead, it might be that a monopolist who can identify different types of consumers is restricted to charging a per-unit price for the goods — a price that can differ for different types of consumers but remains constant for any amount a particular consumer chooses to purchase. If this is the case, the monopolist can no longer perfectly price discriminate but will rather price discriminate “imperfectly”. Such price discrimination is also known as third degree price discrimination. 864 Chapter 23. Monopoly For our example in Graph 23.4, this would imply that the monopolist determines the marginal revenue curve for each of the two types of consumers and then sets output where the constant M C intersects M R. This leads the monopolist to charge the price p1 to type 1 consumers, with those consumers choosing to consume x1 (in panel (a)). Similarly, a potentially different price p2 would be charged to type 2 consumers who would then consume x2 (in panel (b)). Thus, when monopolists can charge a per unit price that differs across identifiable consumer types, they will restrict output below what it would be under efficient first degree price discrimination. As a result, a deadweight loss will arise under imperfect (or third degree) price discrimination. Exercise 23A.13 In our example of me running my art studio and selling to consumers who place value only on the first piece of art they purchase, is there a difference between first and third degree price discrimination? Explain. (Hint: The answer is no.) While we therefore know that deadweight loss will emerge under third degree price discrimination, it is not clear whether eliminating the ability by the monopolist to price discriminate in this way will lead to greater or less deadweight loss. If such price discrimination were deemed illegal, the monopolist would revert to charging a single price to all consumers – which would entail a lower price for the high demanders and a higher price for the low demanders. Conceivably, this uniform price could be such that low demanders will no longer consume any of the good, thus leading to the effective closing of the market in the low demand consumer sector. The welfare losses sustained by low demanders combined with the reduction in profit for monopolists would then have to be weighed against the welfare gains by high demanders. Depending on the types of demand the different consumers have, the elimination of third degree price discrimination could then lead to either a welfare improvement (if the high demanders gain more than the low demanders and the monopolist lose) or an additional welfare loss (if the low demanders and the monopolist lose more than the high demanders gain). Without knowing the specifics in any particular case of third degree price discrimination, it is therefore not possible to make a uniform efficiency-based policy recommendation on how to treat monopolists who engage in third degree price discrimination. Exercise 23A.14 Why do we not run into similar problems of ambiguity in thinking about the welfare effects of first degree price discrimination? 23A.2.3 Non-Linear Pricing and “Second Degree” Price Discrimination Sometimes there are external signals that a firm can use to infer the type of consumer she is facing. Movie theaters know that students will generally have different demands than adults in the labor force, and they may therefore offer student prices that are different from regular prices (and not available to non-students). This is an example of third degree price discrimination. But in many real world circumstances, firms do not have such external signals and therefore are unsure of what types of consumers they face at any given moment. Put differently, it is often difficult to tell by just looking at someone whether that person is a “high demander” or a “low demander” – even if a firm knows how many high demanders there are relative to low demanders. Even in such cases, however, the monopolist can try to find ways of increasing profit through strategic pricing. But since the monopolist cannot tell what type of consumer she is facing, she has to structure her pricing in such a way as to give the incentive to consumers to self-identify who they are. This involves the setting of a single non-linear price schedule — or offering different quantities of the good at different prices. Such a pricing strategy does not explicitly discriminate between different consumers because all consumers are offered the same price schedule for different 23A. Pricing Decisions by Monopolist 865 quantities of the good. Rather, consumers end up paying different average prices based on their choices once they see the non-linear price schedule posted by the monopolist. Suppose, for instance, that the monopolist knows that she has two types of customers – just as in Graph 23.4 in the previous section. But now she cannot tell in any particular instance which type of consumer has entered her store; all she knows is that there is an equal number of both types of consumers in the economy. In Graph 23.5a, we then illustrate the blue type 1 demand curve D1 and the magenta type 2 demand curve D2 within the same picture and again assume a constant marginal cost of $10 per unit of output. If the monopolist could price discriminate perfectly, she would want to offer 200 units of output to type 1 consumers and charge the entire area under D1 — ($2000 + a + b + c). Similarly, she would want to offer 100 units of the output to type 2 consumers and charge the entire area under D2 — ($1000+a). This would result in no consumer surplus and a surplus for the monopolist of (2a + b + c) assuming there is one consumer of each type. Graph 23.5: “Second Degree” Price Discrimination Exercise 23A.15 Explain how this represents separate “two-part tariffs” for the two consumer types (as defined in exercise 23A.12). When the monopolist cannot tell which consumers are type 1 and which are type 2, she cannot 866 Chapter 23. Monopoly implement this perfect price discrimination (nor can she implement the third degree price discrimination from Graph 23.4). This is because type 1 consumers now have an incentive to simply pretend to be type 2 consumers, purchase 100 units at the price ($1000+a) and get consumer surplus of (b). Were the monopolist to offer the 100 and 200 unit packages at the prices suggested above, she could look ahead and know that no one will pick the 200 unit package, leaving her with surplus of only (2a).2 Exercise 23A.16 Why would the monopolist not be able to offer two per-unit prices as in Graph 23.4? In order to induce type 1 consumers to behave differently than type 2 consumers, the monopolist must therefore come up with a different set of price/quantity packages. For instance, the monopolist might continue to offer 100 units at the price ($1000+a) while reducing the price of 200 units to ($2000+a + c). This would equalize the surplus a type 1 consumer will get under the two packages and would therefore make it optimal for type 1 consumers to pick 200 units. (In fact, the monopolist has to charge a price just under ($2000+a + c) for 200 units in order to insure that type 1 consumers will in fact strictly prefer the 200 unit package over the 100 unit package.) As a result, the monopolist would be able to expect a surplus of (2a + c) which is larger than the surplus of (2a) she could expect under the previous price/quantity combinations. Exercise 23A.17 In exercise 23A.12, we introduced the notion of a “two-part tariff ”. Can you express the pricing suggested above in terms of two-part tariffs? In panel (b) of Graph 23.5, however, we can see that the monopolist can do even better by making the package targeted at type 2 consumers less attractive and thus charging more for the package containing 200 units. Consider, for instance, the scenario under which the monopolist offers a package with 90 units and another with 200 units. Type 2 consumers will be willing to buy the 90 units at a price of ($900+d). But now the monopolist can charge ($2000+d + f + g + h) for the 200 unit package — giving an overall surplus of (2d + f + g + h). The surplus of (2a + c) in panel (a) is the same as a surplus of (2d + 2g + h) in panel (b) — which implies that the monopolist’s surplus has changed by (f − g) as she switched from offering the 100 unit package to offering only a 90 unit package. Area (f ) is larger than area (g) — so profit has increased. But once the monopolist recognizes that she can earn higher profit by reducing the attractiveness of the package targeted at type 2 consumers, she can do even better. In panel (b) of the graph, the vertical magenta distance represents the approximate loss in profit from type 2 consumers if the monopolist decreases the type 2 package by another unit (from 90 to 89) while the vertical blue distance represents the approximate increase in profit from type 1 consumers that can now be charged a higher price for the 200 unit package. The monopolist can increase profit by reducing the type 2 package so long as the vertical magenta distance is shorter than the vertical blue distance. Thus, a forward looking monopolist would reduce the type 2 package to a quantity x∗ (where the two distances are equal to one another). This is represented in panel (c) of the graph. Exercise 23A.18 What price will the profit maximizing monopolist charge for x∗ and for 200 units in panel (c) of Graph 23.5? Exercise 23A.19 * We have assumed in our example that there is an equal number of type 1 and type 2 consumers in the economy. How would our analysis change if the monopolist knew that there were twice as many type 1 consumers as type 2 consumers? 2 In Chapter 24, we will introduce the idea of a “sequential game” in which some players move first. We could then say that the monopolist plays such a sequential game with consumers – setting her pricing schedule in stage 1 knowing that consumers will optimize in stage 2. 23A. Pricing Decisions by Monopolist 867 Exercise 23A.20 In Chapter 22, we analyzed situations in which there is asymmetric information between consumers and producers (as in the insurance market). Can you see how the problems faced by an insurance company that does not know the risk-types of its consumers are similar to the problem faced by the monopolist who is trying to second-degree price discriminate? The above example is just one of many that might arise for a monopolist who seeks to offer different per-unit prices to different customers whose type she cannot identify. We will see further examples later on. In the real world, the “packages” offered to different types of consumers may also vary in ways that are related to not just quantity but also quality. For instance, in the airline industry, fares for the same flights are often priced quite differently for business travelers and leisure travelers, with business travelers facing fewer restrictions on when and how they can change their tickets. If these topics are of interest, you should consider taking a course in industrial organization. 23A.3 Barriers to Entry and Remedies for Inefficient Monopoly Behavior So far, we have simply assumed that a particular firm has a monopoly in the market for good x. But how does a firm get such monopoly status in the first place? And how does it hold onto it? We began to discuss this a bit in our brief section on political rent seeking and its implications for deadweight loss. In this section, we will try to dig a bit deeper and point out more explicitly that there must exist some barrier to entry of new firms in order for a monopoly to be able to earn long run positive profits. Such a barrier might emerge simply from the technological nature of production, from different types of legal barriers to entry that we introduced when thinking about political rent seeking or through other channels. 23A.3.1 Technological Barriers to Entry and Natural Monopolies In our discussion of perfectly competitive firms, we never considered the case of a firm that has increasing returns to scale for all output quantities. Rather, we focused on firms that may have increasing returns to scale in their production process for low levels of output but eventually face decreasing returns to scale as output increases. It is because of this assumption that M C and AC curves eventually sloped up. But, while we argued in Chapter 11 that the logic of scarcity requires that marginal product of each input eventually diminishes, there is no particular reason that the production process itself cannot have increasing returns to scale. Exercise 23A.21 Review the logic of how a production process can have diminishing marginal product of all inputs while still exhibiting increasing returns to scale. Now suppose the production process for good x always has increasing returns to scale. This implies, as we illustrated in Graph 12.9, that the M C curve is always downward sloping and always lies below AC, which further implies that any price taking firm will either produce nothing at a particular price or will produce an infinite quantity of the good. But, in a world of scarcity, consumers will not demand an infinite quantity of the good at a positive price — which implies that the assumption of price taking behavior on the part of the firm is not reasonable under increasing returns to scale. It is for this reason that no competitive industry can have firms whose production process always has increasing returns to scale. Similar logic applies when a production process has a large initial or a significant recurring fixed cost together with a constant marginal cost, a case which is illustrated in Graph 23.6a. This can arise in many different contexts. For instance, a large investment in research and development may 868 Chapter 23. Monopoly be required prior to the production of a vaccine, but, once the research is complete, the vaccine can be produced easily at constant M C. Or a utility company might have to invest a large amount in laying electricity lines within a city in order to then be able to provide electricity to everyone at a constant M C. Or a software company might work for years to produce a piece of software that can then be offered at virtually no marginal cost by having customers download it from the internet. Graph 23.6: A Natural Monopoly A natural monopoly is then defined as a firm that faces an AC curve that declines at all output quantities. This declining AC curve can be due to increasing returns to scale everywhere or due to the presence of a recurring fixed cost with constant marginal cost. In either case, we cannot identify a “supply curve” that is equal to the M C curve above AC because M C never lies above AC. It is therefore “natural” for a single firm to emerge as a monopoly. Exercise 23A.22 Can you see in Graph 23.6a that a price taking firm facing a downward sloping AC curve would produce either no output or an infinite amount of the output depending on what the price is? Exercise 23A.23 Suppose the technology is such that AC is U-shaped but the upward sloping part of the Ushape happens at an output level that is high relative to market demand. Can the same “natural monopoly” situation arise? Panels (b) and (c) of Graph 23.6 then add demand and M R curves to the cost curves from panel (a). In panel (b), demand is relatively “high”, and the usual profit maximizing single price pM read off the demand curve at quantity xM where M C and M R intersect results in a positive profit for the monopoly firm. In panel (c), on the other hand, demand is relatively “low” — causing the monopoly to make a loss if it simply produced where M R intersects M C. In order for a firm facing the situation in panel (c) to make a positive profit, it would therefore have to price output differently — thus employing one of the price discrimination strategies discussed in the previous section. In the absence of being able to identify different consumer types, this implies that, in order to produce, the firm would have to engage in a form of pricing that involves more than just a single per-unit price. The most common such strategy for natural monopolists (in the absence of price regulation) is to charge a fixed fee plus a per unit price, which we referred to as a 23A. Pricing Decisions by Monopolist 869 “two part tariff” in exercise 23A.12. In the case of utility companies, for instance, there might be a fixed service fee per month plus a price per unit of electricity consumed. Because the technological constraints are such that multiple firms in such industries would entail higher per unit costs, governments have often favored regulation of natural monopolies over alternative policies to address the deadweight loss from monopoly pricing. Such regulation typically focuses on pricing policies that guarantee a “fair market return” for the natural monopolist while moving production closer to the socially optimal level. Given that the fixed cost is a sunk cost once the monopolist is operating, efficiency would require output where M C crosses the demand curve. But because AC lies above M C, forcing the natural monopolist to price the output at M C would imply negative profits for the monopolist (when profit is defined as long run profit that includes recurring fixed costs). Exercise 23A.24 In a graph similar to Graph 23.6b, illustrate the negative profit that arises when the monopolist is forced to price at M C. Exercise 23A.25 Suppose the fixed cost is a one-time fixed entry cost that is sufficiently large to result in a picture like panel (c). True or False: If the government pays the fixed cost for the firm, it will not have to regulate the firm in order to make sure the firm makes a profit – but he monopoly outcome will be inefficient. For instance, suppose the monopolist faces high recurring fixed costs. Then regulators who attempt to achieve efficient output levels in natural monopolies might aim to set price at M C and allow monopolists to charge an additional “fixed fee” that each customer has to pay independent of the level of consumption. For instance, an electricity provider might charge a fixed “hook up” fee for connecting a household to the service and then a per unit price for each unit of electricity consumed, or a phone company might charge a fixed monthly fee plus a per minute charge for phone calls made. The fixed fees can then be set in such a way as to make the natural monopoly profitable even though the per-unit prices do not cover any of the fixed costs. Exercise 23A.26 Is this an example of a two-part tariff ? Does it result in efficiency? While it is easy to see how this type of regulation works in principle, in practice the regulator unfortunately does not have all the required information to implement the optimal two-part pricing. In particular, the regulator does not typically know the cost functions of the natural monopoly, and the natural monopolist has every incentive to inflate her costs to the regulator in order to obtain higher fixed fees and higher per-unit prices. There are examples in the real world of natural monopolists devising clever schemes involving fake billing from secondary firms in order to show higher costs than they actually incur, and it is not always easy for regulators to identify such falsifications of cost records. The monopolist furthermore has no particular incentive to find innovative ways of lowering costs through technological innovations even if she is perfectly honest in how she reports the costs she actually incurs. For some of these reasons, more recent policy approaches have made an effort to introduce competition into some industries that face these cost curves by having the government pay the fixed costs that cause AC curves to be downward sloping. In the utility industry, for instance, the government could lay (and maintain) the electricity lines to all the houses in a city and then allow any utility company to use these lines in order to “ship” electricity to individual houses. It is much like the government laying a system of roads that different trucking companies can use to deliver goods. With the fixed costs paid by the government, individual electricity suppliers then have only variable costs — and thus flat or upward sloping M C curves. It then becomes once 870 Chapter 23. Monopoly again possible for many different electricity providers to compete for households, with households choosing a provider based on quality of service and price. Exercise 23A.27 Suppose that instead a private company is charged with laying all the infrastructure and then charges competing electricity firms to use the electrical grid. How might this raise a different set of efficiency issues related to monopoly pricing? Would these issues still arise if the government auctioned off the right to build an electricity grid to a single private company? 23A.3.2 Legal Barriers to Entry While monopoly power can certainly arise from technological barriers that prevent several firms from operating simultaneously, it may alternatively arise from legal barriers. Such legal barriers might arise because of general patent and copyright laws that grant the exclusive right to produce particular products (for a certain number of years) to those firms that were awarded the patent or copyright. The motivation behind such laws is not to encourage the formation of monopolies but rather to provide incentives for innovations by ensuring that innovators can profit from their activities for some period. We will discuss the role of patents and copyrights in more detail in Chapter 26. Patent and copyright laws are not, however, the only legal barriers to entry. As we have seen, free entry (in the absence of technological barriers) tends to drive economic profits to zero. Thus, if a firm can successfully lobby the government to protect it from competitors, it will invest resources to accomplish this if the required resources are smaller than the present discounted value of the monopoly profits the firm can expect to earn if legal barriers to entry were erected. As we have already mentioned, to the extent to which such lobbying involves socially wasteful activities, the deadweight loss from government created monopolies may therefore exceed the loss due to the decline in production that results under monopoly profit maximization. Monopoly power has been granted by governments to a variety of firms throughout history. In the 15th and 16th centuries, for instance, the British Crown awarded exclusive rights to shipping companies to establish trade routes in the West Indies and other parts of the world. More recently, airlines routes were regulated in a similar manner, with airlines being assigned exclusive rights to certain routes within the US (prior to airline deregulation). The same was true until the 1970s in the trucking industry and the phone industry. Today, the US Post Office continues to hold the exclusive right to deliver first class mail, although the government now permits carriers like UPS and FedEx to deliver express packages and large ground packages. In each of these cases, you should be able to see how the firm that attained the exclusive rights to serve a particular market benefits from the governments’ entry barriers – and how it might have a vested interest in engaging in socially wasteful lobbying activities in order to retain its monopoly power. 23A.3.3 Restraining Monopoly Power While governments have, as we have mentioned, been prime culprits – for better or worse – of granting monopoly power to certain firms, the increasing awareness of potential social losses from the exercise of monopoly power has also led to government policies aimed at restraining monopolies. The question of when and under what circumstances government intervention is desirable is a complicated one. The tendency of monopolies to limit output in order to raise price has the clear deadweight loss implications that we have discussed. At the same time, patent-protection of innovation may have led to the emergence of products that might otherwise never have seen the light 23A. Pricing Decisions by Monopolist 871 of day – implying the creation of social surplus despite the fact that, at any given moment, more surplus could be gained by forcing monopolies to produce more. (We will have more to say about this in Chapter 26.) And the existence of increasing returns to scale in certain industries implies that natural monopolies may lower per unit costs even as they attempt to use their monopoly status to raise price above marginal cost. We will show in end-of-chapter exercise 23.9 that some of the potential remedies that one might think of applying to monopolies are either ineffective or counterproductive. These include per unit taxes and profit taxes. We have already discussed (in our treatment of regulation of natural monopolies) that attempts to directly regulate the pricing of monopoly goods run into informational constraints because regulators typically do not know the real costs of firms and because such regulation would give little incentive for cost innovations by monopolies. This does not imply that regulation in some circumstances is not the appropriate policy, but it does imply that regulation is no panacea in all cases. In some instances, governments have forced the break-up of monopolies (as in the case of large oil companies many years ago or large phone companies more recently), and in other cases they have found ways of addressing the root causes of natural monopolies by disconnecting the fixed cost infrastructure from the marginal cost provision of services. And in other cases, governments have actively blocked mergers of large companies that might have resulted in excessive monopoly power. Finally, there has been an increasing trend toward deregulation of industries where regulation itself (such as in the airline industry) created monopolies to begin with. If these topics seem interesting to you, you might consider taking a course on antitrust economics or law and economics. In many circumstances, however, the most effective tool for restraining monopoly power has little to do with direct government actions and more to do with the fact that, when a monopoly does exercise its power to create profit, there is a powerful incentive for entrepreneurs to find new ways to challenge that monopoly power. A firm may, for instance, have captured a large portion of the market, perhaps for no other reason than being first and making early, strategically smart decisions (as in the case of Microsoft and its Windows operating system). There is no doubt that such firms will use their monopoly power to their advantage, but they may also be more cognizant of the threat of competitors (that may find ways of producing substitutes) than our simple static models of monopoly behavior predict. The more a firm exercises its monopoly power, the greater is the incentive for others to find ways of producing such substitutes, and a forward looking monopolist should take that into account when setting current prices, as we will see in upcoming chapters. Sometimes barriers to entry that may seem rock-solid at one time can fall quickly with new technological innovations, as, for instance, with the sudden emergence of cell phone technology, internet calling and cable provision of telephone service that are challenging traditional phone companies. In such environments, governments can play in important role in insuring that existing firms (such as traditional phone companies) do not successfully erect barriers of entry through legislation or regulation (by prohibiting, for instance, cable companies or internet providers from providing telephone service). Just as there exists a powerful incentive for innovators to find ways of breaking barriers to entry by existing firms, there is a similarly powerful incentive on the part of existing firms to find other ways of shoring up these barriers to entry in order to preserve market power. Exercise 23A.28 In the 1970s when OPEC countries raised world prices for oil substantially by exercising their market power, the Saudi oil minister is said to have warned them: “Remember, the Stone Age did not end because we ran out of stones.” Explain what he meant and how his words relate to constraints that monopolies face. 872 23B Chapter 23. Monopoly The Mathematics of Monopoly From a mathematical point of view, monopolies engage in the same optimization problem that competitive firms undertake except that monopolies have additional choice variables. Both types of firm face some cost function that emerges from the cost minimization problem and tells them the total cost c(x) of producing any quantity x. We should note at the outset that for much of our development below we will assume that dc(x)/dx = c – i.e. the firm faces a constant marginal cost. This simplifies some of the analysis in convenient ways, and we will explore different marginal cost schedules in some of the end-of-chapter exercises. Exercise 23B.1 Explain why the cost minimization problem in the firm’s duality picture is identical for firms regardless of whether they are monopolies or perfect competitors. A monopoly that is restricted to charging a single per-unit price then solves the problem max π = px − c(x) subject to p ≤ p(x), x,p (23.1) where the price the monopolist charges when trying to sell the quantity x cannot be greater than the price for that quantity given by the inverse demand function p(x). The perfect competitor’s problem could be written in exactly the same way, except that for the perfect competitor the inverse demand function is simply p(x) = p∗ , where p∗ is the market price. Thus, price ceases to be a choice variable when price is set by the competitive market, but it is a choice variable for a monopolist who faces a downward sloping demand curve. Since the monopolist will set price as high as she can while still selling all the goods she produces, the inequality in equation (23.1) will bind – i.e. p = p(x). The monopolist’s problem can therefore be re-written as max π = p(x)x − c(x). x (23.2) Note that by choosing the optimal quantity xM , the monopolist implicitly chooses the optimal price pM = p(xM ) once we have substituted the constraint into the objective function of the optimization problem. And because of the resulting one-to-one mapping from quantity to price, the monopolist’s problem could alternatively be written as max π = px(p) − c(x(p)) p (23.3) where x(p) is the market demand function (as opposed to the inverse market demand function p(x) in the previous problem.) Whether we view the monopolist as choosing quantity as in equation (23.1) or price as in equation (23.3), the same monopoly quantity and price will emerge. When a monopolist is not restricted to charging a single per-unit price, she has additional decisions to make as we have seen in our discussion of price discrimination in Section A. The exact nature of that choice problem then depends on what the firm knows and what pricing strategies are available to the firm. If the firm can identify consumer types prior to consumption choices by consumers, first and third degree price discrimination become possible (assuming re-sale can be prevented), and if the firm only knows the distribution of consumer types in the population, second degree price discrimination becomes possible. Different forms of such discrimination are furthermore restricted by the types of pricing schedules that firms are permitted to post, as we will see a little later in the chapter. Fundamentally, however, the firm is still just maximizing profit by making production choices and potentially by engaging in strategic price differentiation. 23B. The Mathematics of Monopoly 23B.1 873 Demand, Marginal Revenue and Profit Suppose that the market demand facing a monopolist is of the form x(p) = A − αp, (23.4) which gives rise to an inverse market demand 1 A − x. (23.5) α α For consistency, we will use this market demand specification repeatedly, both in this chapter as well as in the following chapters that deal with other market structures within which firms might operate. p(x) = 23B.1.1 Marginal Revenue and Price Elasticity For the monopolist, total revenue is then equal to price times output, where price is determined by the inverse market demand curve; i.e. 1 A 1 A − x x = x − x2 . (23.6) T R = p(x)x = α α α α In Section A, we argued verbally that the marginal revenue curve for a monopolist has the same intercept as the inverse demand curve but twice the slope. This is easily verified mathematically, with marginal revenue simply the derivative of T R with respect to output dT R A 2 = − x. (23.7) dx α α More generally, we can write the inverse demand function as p(x) and total revenue as T R = p(x)x. Using this expression, we can differentiate T R with respect to x to get MR = dp x. (23.8) dx Now suppose we multiply the second term in equation (23.8) by (p(x)/p(x)). Then we can write the expression for M R as dp x . (23.9) M R = p(x) 1 + dx p(x) M R = p(x) + Recall that the price elasticity of demand for an inverse demand function p(x) is given by ǫD = (dx/dp)(p(x)/x), which is just the inverse of the second term in parenthesis in the above equation. Thus, we can write the expression for M R as 1 M R = p(x) 1 + . (23.10) ǫD Suppose, for instance, that we are currently at the mid-point of a linear demand curve (such as the one in Graph 23.1a) where the price elasticity of demand is equal to -1. Equation (23.10) then tells us that marginal revenue at that point is equal to 0, precisely as we derived in panel (b) of Graph 23.1. 874 Chapter 23. Monopoly Exercise 23B.2 Use equation (23.10) to verify the vertical intercept of the marginal revenue curve in Graph 23.1b. 23B.1.2 Revenue Maximization In order to maximize total revenue T R, the monopolist would simply set M R equal to zero. Using equation (23.10) for M R, it follows immediately that revenue is maximized when ǫD = −1. With the linear demand specified in equation (23.4), this implies an output level of A/2. Exercise 23B.3 Set up a revenue maximization problem for the firm. Then verify that this is indeed the revenue maximizing output level and that, at that output, ǫD = −1. 23B.1.3 Profit Maximization The monopolist’s profit maximization problem differs from revenue maximization in that costs are taken into account. This problem, already introduced at the beginning of this section, can be written as max π = p(x)x − c(x), x (23.11) where c(x) is the total cost function (that is derived from the production function as described in our producer theory chapters earlier in the text).3 Taking first order conditions, we get M R = p(x) + dc(x) dp x= = M C. dx dx (23.12) Exercise 23B.4 Can you use equation (23.10) to now prove that, so long as M C > 0, the monopolist will produce where ǫD < −1? For instance, suppose that market demand is linear as specified in equation (23.4) and c(x) = cx. Then our M R = M C condition implies 2 A − =c α α (23.13) which further implies a monopoly output xM and price pM of xM = A − αc A + αc and pM = . 2 2α (23.14) Exercise 23B.5 Illustrate that profit maximization approaches revenue maximization as M C = c approaches zero. Exercise 23B.6 Verify for the example of our linear demand curve and constant marginal cost c that it does not matter whether the firm maximizes profit by choosing x or p (as in the problems defined in equations (23.1) and (23.3) above). 3 Recall that the cost function is really a function of output x as well as input prices. We are suppressing the input price notation since input markets are not a focus for us here. 23B. The Mathematics of Monopoly 23B.1.4 875 Constant Elasticity Demand and Monopoly Markups Another way to write the optimal monopoly price emerges from substituting our elasticity-based expression for M R from equation (23.10) into the M C = M R condition of equation (23.12); i.e. 1 = M C. p 1+ ǫD (23.15) p − MC −1 . = p ǫD (23.16) Re-arranging terms, we then get The difference between price and M C – i.e. (p − M C) – is called the monopoly markup because it represents how much the monopolist “marks its price up” above marginal cost where we would expect competitive firms to produce. The left hand side of equation (23.16) is called the monopoly markup ratio – which is simply the markup relative to the price charged by the monopolist. (The markup ratio is also called the Lerner Index.) Since the price elasticity term ǫD is negative, this cancels the negative sign on the right hand side and makes the mark-up itself positive. Suppose, then, that instead of facing a linear demand curve for which price elasticity differs at each point, a monopolist faces a constant-elasticity demand curve of the form x = αp−ǫ for which the price elasticity of demand is −ǫ everywhere. Equation (23.16) then tells us that the monopolist’s markup ratio is inversely proportional to the price elasticity of demand. This implies that the markup ratio (and the markup itself) approaches zero as the price elasticity of demand approaches minus infinity. That certainly makes intuitive sense: as the price elasticity of demand approaches minus infinity, the monopolist faces a demand curve that increasingly looks like the demand curve faced by a perfect competitor. When working with the family of constant elasticity demand curves, the price elasticity of demand is therefore a nice measure of the degree of monopoly power that the firm actually has. 23B.2 Price Discrimination when Consumer Types are Observed In Section A of the chapter, we differentiated between three different types of price discrimination that monopolists might employ depending on what they know about their consumers and the degree to which the monopolist can prevent consumers from undermining the price discrimination. In cases where monopolists can identify demand by each consumer, the firm can perfectly (or firstdegree) price discriminate and capture the consumers’ entire surplus – as long as something prevents consumers from selling the goods to each other. When monopolists are restricted to charging perunit prices but are not restricted to charging the same per unit price to all consumers (whose demand they can again identify), we illustrated how they can employ third degree price discrimination – again assuming that consumers cannot engage in re-sale. Finally, if monopolists know that different consumers have different demands but cannot identify which consumer is which type, we saw that the firm can second-degree price discriminate by designing (non-linear) price/quantity combinations that cause consumers to self select into packages based on their type. We will begin in this section with the mathematically easier cases of first and third degree price discrimination where we assume that firms observe consumer types prior to setting pricing policies. 876 23B.2.1 Chapter 23. Monopoly Perfect or First Degree Price Discrimination As we illustrated in Section A, first degree price discrimination implies that the firm will charge the consumer her marginal willingness to pay for each of the goods she purchases. Suppose that a monopolist faces a constant marginal cost M C and let pc = M C represent the per unit price we would expect under perfect competition. For a particular consumer n, let CS n represent the consumer surplus n would receive under competitive pricing, with the consumer choosing to consume where pc crosses her demand curve Dn . One way to think of perfect price discrimination is to think of the monopolist as continuing to charge a per-unit price of pc but supplementing this with a fixed fee that the consumer has to pay before she can purchase anything at all. Notice that this fixed fee is a sunk cost for the consumer once it is paid – and therefore has no impact on the quantity the consumer will purchase once the fee is paid. The only question for the consumer is then whether she wants to pay the fixed fee in order to be able to purchase from the monopolist. Since she expects a consumer surplus of CS n when she faces a per-unit price of pc in the absence of a fixed fee, she will be willing to pay any fixed fee that is less than or equal to CS n . The monopolist can therefore set a two-part tariff, with the overall payment P n charged to consumer n equal to P n (x) = CS n + pc x. (23.17) Under this two-part tariff, the monopolist has set a price policy for consumer n that will leave the consumer with no surplus but results in the efficient level of consumption by consumer n. The fixed portion of the price policy is different for each type of consumer, which implies the monopolist must know each consumer’s type in order to implement the first-degree price discrimination if consumers have different demands. Exercise 23B.7 Illustrate graphically the two different parts of the two-part tariff in equation (23.17). 23B.2.2 Third Degree Price Discrimination Suppose now that the monopolist is selling to two different distinct markets but is limited to charging per-unit prices in each market (and thus cannot implement a two-part tariff of the type in equation (23.17)). With knowledge of the two inverse demand functions p1 (x) and p2 (x) for the two markets, the monopolist will then try to maximize her profit across the two markets by choosing how much to produce in each market (and thus also how much to charge in each market); i.e. the monopolist will solve the problem max π = p1 (x1 )x1 + p2 (x2 )x2 − c(x1 + x2 ), x1 ,x2 (23.18) where c is the firm’s total cost function. Taking first order conditions, we get ∂π = p1 (x1 ) + ∂x1 ∂π = p2 (x2 ) + ∂x2 which can simply be rewritten as dp1 1 x − dx1 dp2 2 x − dx2 dc = 0, dx dc = 0, dx (23.19) 23B. The Mathematics of Monopoly 877 M R1 = M C = M R2 , (23.20) where M Ri is the marginal revenue function derived from the ith market’s inverse demand function. Since we know from equation (23.10) how to write M R functions in price elasticity terms, we can write this as p 1 1+ 1 ǫD 1 1 = MC = p 1 + ǫD 2 2 (23.21) which simply extends equation (23.15) to two separate markets, with the “mark-up” in each market reflecting the price elasticity in each market. This then implies (ǫD2 + 1)ǫD1 p1 = . p2 (ǫD1 + 1)ǫD2 (23.22) Put into words, regardless of what the M C of production is, the price charged in one market relative to that charged in the other market depends only on the price elasticities of demand in the two markets when M C is constant. Suppose, for instance, that a monopoly faces constant marginal cost equal to c and that the demand functions in two different markets are x1 = A − αp and x2 = B − βp. These demand functions give rise to inverse demand functions p1 = B − x2 A − x1 and p2 = , α β (23.23) and the first order conditions requiring marginal revenue to be equal to marginal cost in both markets imply x1 = A − αc B − βc and x2 = 2 2 (23.24) p1 = A + αc B + βc and p2 = . 2α 2β (23.25) and Exercise 23B.8 Verify that equation (23.22) holds for this example. (Be sure to evaluate elasticities at the optimal output levels.) Exercise 23B.9 * True or False: The higher priced market under (third degree) price discrimination is more price inelastic. As we noted in our Section A discussion of third degree price discrimination, the welfare effect of eliminating such discrimination is ambiguous and requires an analysis of the gains by low elasticity consumers relative to the losses by high elasticity consumers (and the monopolist). 878 Chapter 23. Monopoly 2-Part Tariff Restrictions for Different Forms of Price Discrimination None 1st Degree 3rd Degree 2-part tariff 2nd Degree F1 =0 =0 F2 =0 =0 = F1 1 p p2 = p1 = p1 Table 23.1: F n = type n’s fixed charge; pn = type n’s per-unit price 23B.3 Discrimination when Consumer Types are Not Observable First and third degree price discrimination are relatively straightforward since firms are assumed to know the types of consumers they face. When they do not know the consumer types but are only aware of the fraction of the population that falls into each category, the monopolist’s problem becomes more difficult and involves more strategic considerations. In particular, since the monopolist has no external signal about the consumer types she is facing, she must design her pricing policy in such a way that consumers themselves choose to reveal what type they are through the types of purchases they make. As you may have noticed already in Section A, all the various ways of thinking about monopoly pricing involve the firm choosing two-part tariffs of the form P n (x) = F n + pn x for n = 1, 2. (23.26) In other words, we can express each of the pricing strategies as separate two-part tariffs aimed at the two types of consumers. The difference in all these strategies is that in some cases we are restricting fixed charges F n to be zero and in some cases we are restricting the monopolist to only a single pricing schedule. Table 23.1 illustrates this for the forms of price discrimination we have treated above and those we are about to discuss below. For instance, we began the chapter in Section 23B.1 with a monopolist who was restricted to charging a single per-unit price to all consumers, effectively assuming F 1 = F 2 = 0 and p2 = p1 as in the first column of the table. Under first degree price discrimination, on the other hand, we make no restrictions on the fixed and perunit prices that the monopolist can use. Under third degree price discrimination, no fixed fees are permitted (i.e. F 1 = F 2 = 0) but no restrictions are placed on the per-unit prices the monopolist can charge. Below we will revisit the case where no restrictions are placed on fixed fees or perunit prices (as in first degree price discrimination) but under the informational constraint that the firm cannot observe consumer type prior to consumers making their purchasing decisions. This is second degree price discrimination, represented in the last column of Table 23.1. But we will build up to this full second-degree price discrimination by first considering the case where a firm does not observe consumer type and is restricted to posting a single two-part tariff (rather than separate two-part tariffs aimed at different consumer types). This is represented in the second-to-last column in Table 23.1. To simplify the analysis to its essentials, we will also allow a single preference parameter to differentiate the different consumer types in this section.4 In particular, suppose that consumer n has tastes for the monopoly good x that can be represented by the utility function 4 Previously, we have allowed different consumer types to differ in both the intercept and the slope of their demand curves. 23B. The Mathematics of Monopoly 879 U n = θn u(x) − P (x), (23.27) 5 where P (x) is the total charge for consuming quantity x. Differences in consumer tastes are then captured by differences in the value of θn . Note that this is not the typical type of utility function we have worked with given that it is defined over only a single good. However, as we demonstrate in a short appendix, this type of “reduced form” utility function can be justified as arising from preferences that are separable (between other consumption and the good x) when the overall spending on the good x represents only a small portion of the consumer’s income. In fact, we demonstrate in the appendix that we can assume identical underlying (separable) preferences where consumers differ only in their income, and that the differences in the value of θn in the reduced form utility function above are then simply related to underlying differences in consumer income. 23B.3.1 Second Degree Price Discrimination with a single Two-Part Tariff As already mentioned, we begin our consideration of second degree price discrimination with a restricted version that we did not discuss in Section A of the chapter: a version in which the monopolist is limited to using a single two-part tariff for both consumer types (rather than different two-part tariffs aimed at the two different types). If the monopolist is so constrained, P (x) has to take the form P (x) = F + px, (23.28) where F is the fixed charge and p the per-unit price – with neither being superscripted by n (since the same price schedule applies to both types). Maximizing consumer utility from equation (23.27) given the two part tariff from equation (23.28) entails the simple optimization problem max θn u(x) − F − px x (23.29) and gives us the first order condition du(x) = p. (23.30) dx The analysis becomes particularly clean if we assume the following functional form for u(x): θn 1 − (1 − x)2 (23.31) 2 which has a first derivative with respect to x that is just (1 − x). Plugging this into equation (23.30) and solving for x, we then get the consumer’s demand function as u(x) = xn (p) = θn − p . θn (23.32) Notice that we therefore have specified underlying preferences in such a way as to once again have linear demand curves of the form x(p) = A − αp where A = 1 and α = 1/θn . 5 This exposition draws on similar exposition in Tirole, J. (2001), The Theory of Industrial Organization, Cambridge, MA: The MIT Press. For the interested student, this text is an excellent reference for matters related to market power, but it is quite advanced. 880 Chapter 23. Monopoly Exercise 23B.10 Intuitively, why does the fixed charge F from the two-part tariff not show up in the demand function? Exercise 23B.11 Derive the price charged to consumer n by a third-degree price discriminating monopolist with constant marginal cost c. In Graph 23.7 below, we depict the inverse demand curve for this demand function and illustrate the consumer surplus triangle CS n that, for a particular per-unit price p with F = 0, is of size CS n (p) = (θn − p)2 (θn − p)xn (p) . = 2 2θn (23.33) Graph 23.7: Consumer n’s Inverse Demand Curve Now suppose that a monopolist faces two types of consumers, type 1 and 2 with preference parameters θ1 and θ2 respectively and with θ1 < θ2 . Suppose further that the monopolist knows that a fraction γ < 1 of the consumers are of type 1, with the remaining fraction (1 − γ) made up of consumers of type 2. Finally, suppose the monopolist faces a constant marginal cost of c. Whatever per-unit price the monopolist chooses, she then has to respect the constraint that the lower demand consumer 1 will not choose to consume any of the good if the fixed charge F is set above CS 1 (p) = (θ1 − p)2 /2θ1 .6 Thus, for a given per-unit price p, the monopolist’s optimal fixed charge is CS 1 (p). Knowing this, the monopolist needs to determine the optimal per-unit charge in the two-part tariff. One way to think of this is as a process in which the monopoly maximizes the expected profit from each encounter with a consumer – knowing the fractions of the consumer pool that fall into one type or the other. This expected profit takes the form E(π) = CS 1 (p) + γ(p − c)x1 (p) + (1 − γ)(p − c)x2 (p). (23.34) The CS 1 (p) term is simply the fixed charge that we have concluded the firm will set in its two part tariff, a charge that will be paid by both types of consumers. Thus, the firm receives that for sure each time a customer shows up. With probability γ, the firm faces a consumer of 6 This constraint is often referred to as the individual rationality constraint. 23B. The Mathematics of Monopoly 881 type 1 who will purchase x1 (p) at price p. When multiplied by the difference between price p and marginal cost c, we get the expected additional profit from facing this type of consumer. Similarly, with probability (1 − γ) the firm will face a consumer of type 2 and with it an additional profit of (p − c)x2 (p). Substituting in for what we derived for CS 1 (p), x1 (p) and x2 (p), and rearranging terms, the expected profit can then be expressed as (1 − γ) γ (θ1 − p)2 p (23.35) + (p − c) 1 − + E(π) = 2θ1 θ1 θ2 Exercise 23B.12 Verify that this equation is correct. The only choice variable for the monopolist in this expected profit equation is p. Thus, maximizing the expected profit subject to the implicit constraint that only a two-part tariff can be employed is simply maximizing E(π) by choosing p. Solving the first order condition from this maximization problem for p, we get the optimal per-unit price p∗ p∗ = Exercise 23B.13 ** c(γθ2 + (1 − γ)θ1 ) 2(γθ2 + (1 − γ)θ1 ) − θ2 (23.36) Verify that this equation is correct. In panel (a) of Graph 23.8, the line CS 1 (p∗ ) + p∗ x represents the two part tariff P (x) that indicates, for any quantity x, the total price charged to consumers. What makes this a two-part tariff is that the line has a vertical intercept – which puts in place a fixed cost to the consumer for purchasing from the firm. Were the line to go through the origin, we would have a simple per-unit price. Graph 23.8: Second Degree Price Discrimination with Two-Part Tariffs In panel (b) of the graph, we illustrate the shape of indifference curves for the two types of consumers, with the blue indifference curves representing type 1 and the magenta indifference curves representing type 2. Consumers prefer to have more of x and less of P – and thus become better off as they move toward indifference curves to the south-east of the graph. 882 Chapter 23. Monopoly Exercise 23B.14 Are these preferences convex? Exercise 23B.15 Note that each set of blue and magenta indifference curves cross once, with the magenta indifference curve having a steeper slope at that point than the blue indifference curve. Can you give an intuitive explanation for this? Finally, in panel (c) of the graph, we put indifference curves and the two-part tariff-induced constraint into a single graph to illustrate the consumers’ optimal choices, with type 1 consumers optimizing at point A and type 2 consumers optimizing at point B. Note that the optimal blue indifference curve for type 1 crosses the origin, which implies that type 1 consumers are as well off at point A as they are at point (0, 0) where they consume no x and pay no price. Put differently, consumers of type 1 attain zero consumer surplus at point A under the two-part tariff that has been set by the firm. Exercise 23B.16 Given what you know of how the firm constructed the two-part tariff, can you give an intuitive explanation for this? While the firm that is implementing the two-part tariff does not know what type of consumer she faces prior to a consumption decision, the graph illustrates that the two-part tariff allows the firm to know what type of consumer she faced after the decision has been made. Put differently and in the language of Chapter 22, the firm has induced a separating equilibrium, with the consumer types signaling their type through their consumption choices. 23B.3.2 Second Degree Price Discrimination more Generally In our definition of second degree price discrimination in Section A, we did not limit the monopolist to using a single two-part tariff but allowed her to create price/quantity packages that in effect allowed her to charge different fixed fees and different per-unit prices. In order to reconcile our treatment here with the graphs we drew in Section A, particularly Graph 23.5, we can again consider the problem using demand curves rather than indifference curves. Panel (a) of Graph 23.9 then illustrates the inverse demand curves for type 1 (blue) and type 2 (magenta) as well as the per-unit price p∗ in the single two-part tariff that we just derived. Exercise 23B.17 Explain why, for the preferences we have been working with, the two demand curves have the same horizontal intercept. Since we know the monopolist sets the fixed charge in the two-part tariff equal to the consumer surplus type 1 would get under only the per-unit price, the shaded blue area is equal to the fixed charge F . This implies zero consumer surplus for type 1 consumers and consumer surplus equal to the magenta area for type 2 consumers. We began our exploration of second degree price discrimination in Section A by proposing that the firm set a per-unit price at M C = c and then charge the highest possible fixed fees to each consumer type such that the consumers would in fact choose different bundles. We replicate this in panel (b) of Graph 23.9 for the demand curves we are working with, taking the liberty of drawing these in a particular way so as to minimize the number of areas we have to keep track of. Here, the firm sets its per-unit price at M C and then charges a fee F 1 = (a + b + c) to type 1 (thereby capturing all of type 1’s consumer surplus) and a fee F 2 = (a + b + c + d) to type 2 consumers. The expected profit from a consumer of unknown type is then (a + b + c + (1 − γ)d) under this pricing policy, while it is (a + b + (1 − γ)(c + e)) under the single two-part tariff we calculated in the previous section. 23B. The Mathematics of Monopoly 883 Graph 23.9: Two-Part Tariff illustrated with Demand Curves Exercise 23B.18 Why is F 2 = (a + b + c + d) the highest possible fixed fee the firm can charge to type 2 consumers given that it sets per unit prices at M C and charges type 1 F 1 = (a + b + c)? Exercise 23B.19 Why is the expected profit from the single two-part tariff (a + b + (1 − γ)(c + e))? It is then easy to see in this example that charging the proposed different fixed fees might in fact result in more profit for the monopolist. Suppose, for instance, that γ = 0.5. Then the expected profit from a given consumer of unknown type under different fixed fees and marginal cost pricing is (a + b + c + 0.5d) while it is (a + b + 0.5(c + e)) under the single two-part tariff with per-unit price p∗ . Since areas c and e are equal to each other, the profit from the two-part tariff can also be written as (a + b + c), which is lower than the profit from charging two different fixed fees and pricing at marginal cost. Exercise 23B.20 Can you think of alternative scenarios under which the single two-part tariff yields more profit? But then we also illustrated in Graph 23.5 that, when allowed to design fixed fee and per unit pricing packages that differ in both dimensions, the monopolist can do better by raising the per-unit price on the low demand consumer and thus increasing the fixed fee for the high demand consumer. Complete freedom in designing pricing when faced with different consumer types then results in high demand consumers purchasing the socially optimal quantity but paying a higher fixed fee, and the lower demand consumers purchasing sub-optimal quantities and paying a lower fixed fee. A potentially optimal level of second degree price discrimination (analogous to what we derived in Section A) is pictured once again in Graph 23.10. It can be viewed as consisting of two separate two-part tariffs, with consumers free to choose which one to select. The two-part tariff targeted at low-demand consumers consists of a per unit price p accompanied by a fixed fee equal to that 884 Chapter 23. Monopoly Graph 23.10: Optimal Second Degree Price Discrimination using Two-Part Tariffs consumer type’s consumer surplus CS 1 (p) under the per unit price p. Under this two-part tariff, type 1 consumers will choose x1 (p) and pay a total tariff P 1 = CS 1 (p) + px1 (p). (23.37) which is equal to the shaded blue area in the graph plus the rectangle cx1 (p) underneath the shaded blue area. The tariff aimed at high demand consumers, on the other hand, consists of a per-unit price c equal to marginal cost and the highest possible fixed fee that will keep type 2 consumers from taking the two-part tariff aimed at type 1 consumers. This will result in type 2 consumers purchasing the quantity x2 (c), leaving them with consumer surplus equal to the shaded blue, green and magenta areas in the absence of a fixed fee. Since type 2 consumers can obtain consumer surplus equal to the shaded green area by accepting the two-part tariff aimed at low demand consumers, the most that the firm can then charge in a fixed fee is equal to the shaded blue plus the shaded magenta areas. The resulting two-part tariff P 2 aimed at type 2 consumers is then given by 2 P = ( 1 CS (p) + (p − c)x1 (p) + " p2 (x1 (p)) − c 2 #) x2 (c) − x1 (p) + cx2 (c), (23.38) where the first bracketed term represents the shaded blue area and the second bracketed term represents the shaded magenta area – which together compose the fixed fee charged to type 2 consumers. This implies that the firm can expect profit of π 1 (p) = CS 1 (p) + (p − c)x1 (p) from type 1 consumers and (23.39) 23B. The Mathematics of Monopoly 885 π 2 (p) = CS 1 (p) + (p − c)x1 (p) + " p2 (x1 (p) − c 2 # x2 (c) − x1 (p) . (23.40) from type 2 consumers. The expected profit from encountering a consumer of unknown type is then E(π) = γπ 1 (p) + (1 − γ)π 2 (p) or 1 1 E(π) = CS (p) + (p − c)x (p) + (1 − γ) " p2 (x1 (p) − c 2 # x2 (c) − x1 (p) (23.41) The only variable in the expression for E(π) that is under the control of the monopolist is the price p – because the setting of p determines the fixed charges that can be levied on the two types of consumers and we already know that the per-unit price for type 2 consumers is c. Thus, the monopolist’s problem is to choose p to maximize E(π) and then to define the two-part tariffs for the two consumer types accordingly. For the preferences we have used in this section, we can substitute in for the various functions in E(π) and write the firm’s problem as (θ2 − c) (θ1 − p) (θ1 − p)2 (θ1 − p) (1 − γ) θ2 p + (p − c) + −c − max E(π) = p 2θ1 θ1 2 θ1 θ2 θ1 (23.42) With a bit of careful math, the first order condition for this maximization problem can then be solved for p to yield p= θ1 γ θ1 − (1 − γ)θ2 c (23.43) from which the two-part tariffs P 1 and P 2 can be derived. We have then derived full second-degree price discrimination in the form of two separate twopart tariffs, with different fixed fees and different per-unit prices targeted at the two consumer types in such a way as to get each consumer type to utilize the two-part tariff intended for her while maximizing the monopolist’s profit (conditional on the monopolist not being able to a priori identify the consumer types). There is one final caveat for the monopolist who is contemplating this pricing policy: If there are sufficiently many high demand consumers (i.e. if γ is sufficiently low) or if the high demanders have sufficiently greater demand than low demanders (i.e. θ2 is sufficiently above θ1 ), it may be better for the monopolist to write off the type 1 market and simply set a single two-part tariff intended to extract the most possible surplus from type 2 consumers. You can see this clearly in Graph 23.10. Suppose, for instance, that γ = 0.5 – implying an equal number of type 1 and type 2 consumers. By choosing second degree price discrimination, the monopolist chooses to forego capturing the shaded green area in type 2’s consumer surplus in exchange for instead getting the shaded blue area of type 1’s consumer surplus. The alternative is for the firm to capture the green area of type 2’s surplus and not offer anything that type 1 consumers would choose – thus foregoing the shaded blue area. Note that, in our graph, the green area is larger than the blue area. Thus, with γ = 0.5, the monopolist is better off engaging in first degree price discrimination with respect to type 2 consumers (and not sell to type 1 consumers) than to engage in second degree price discrimination. 886 Chapter 23. Monopoly F1 F2 p1 p2 x1 x2 CS 1 CS 2 E(π) TS Different Forms of Monopoly Price Discrimination None 1st Degree 3rd Degree 2-part tariff 2nd Degree $0 $28.13 $0 $23.63 $12.50 $0 $52.08 $0 $23.63 $33.33 $72.50 $25.00 $62.50 $31.25 $50.00 $72.50 $25.00 $87.50 $31.25 $25.00 0.2750 0.7500 0.3750 0.6875 0.5000 0.5167 0.8333 0.4167 0.7917 0.8333 3.7813 0 7.0313 0 0 20.0208 0 13.0208 23.3724 18.7500 18.8021 40.1042 20.0521 28.2552 29.1667 30.7031 40.1042 30.0781 39.9414 38.5417 Table 23.2: θ1 = 100, θ2 = 150, γ = 0.5, c = 25 Exercise 23B.21 If the monopolist is restricted to offering a single two-part tariff (rather than two separate tariffs intended for the two consumer types), is she more or less likely to forego second degree price discrimination in favor of first degree price discrimination with respect to the high demand type? 23B.3.3 Comparing Different Monopoly Pricing: An Example We noted at the beginning of our discussion of second-degree price discrimination that we can think of each of the pricing strategies we have covered as different personalized two-part tariffs of the form P n (x) = F n + pn x. Under some strategies we assume fixed charges F n to be zero; under others we require them to be equal for the different consumer types (as summarized in Table 23.1). This then gives us a convenient way of comparing the different forms of price discrimination. Table 23.2 undertakes this comparison for a particular example in which θ1 = 100, θ2 = 150, γ = 0.5 and the marginal cost c = 25. The first column begins by presenting the outcome of monopoly behavior when no price discrimination takes place, with the next two columns presenting the outcome for first and third degree price discrimination where the firm knows each consumer’s type and the final two columns presenting the outcome when the firm does not know each consumer’s type and is at first restricted to using a single two-part tariff and then permitted to employ separate two-part tariffs aimed at the two consumer types. In each case, we begin with the fixed fees and the per-unit prices charged to the two consumer types and then report the consumption levels, consumer surpluses and the firm’s expected profit per consumer. The final row of the table then sums the consumer surpluses and the firm’s profit to arrive at the total surplus. We know from our work above that first degree price discrimination results in full efficiency, with the entire surplus accruing to the firm. It is therefore not surprising to see that the firm’s profit and the total surplus are the largest under first degree price discrimination, nor is it surprising that this is the least preferred outcome for consumers whose entire surplus is taken in fixed fees by the monopolist. It should also not be surprising that the firm’s profit is the lowest when it is not permitted to engage in any price discrimination. After all, we can see from Table 23.1 that the firm is most restricted in its pricing policy in that case, with no possibility of charging a fixed fee and no possibility of differentiating between the consumer types in terms of the per-unit price charged. These restrictions are lifted partially under third degree price discrimination, resulting in higher 23B. The Mathematics of Monopoly 887 firm profit, and fully lifted under first degree price discrimination. It is therefore natural to expect the firm’s profit from third degree price discrimination to fall in between the no-discrimination and full (first degree) discrimination scenarios. In the case where firm’s can discriminate but do not know the consumer types (represented in the last two columns), it is again not surprising that the firm makes more profit than it dues in the no-discrimination case, nor should it be surprising that firm profit is higher when the firm can charge two separate two-part tariffs (in the last column) than when it is restricted to a single two-part tariff (in the second-to-last column). The only case that is theoretically ambiguous with respect to firm profit is the comparison between third degree price discrimination and the two forms of second degree price discrimination in the last two columns. For our particular example, it turns out that both forms of second degree price discrimination result in greater profit than third degree price discrimination, but for other examples the reverse could be true. Exercise 23B.22 From looking at Table 23.1, it seems that the firm is unambiguously less restricted in its pricing under second degree price discrimination than under third degree price discrimination. So how could it theoretically be the case that profit is higher under third degree price discrimination? We can summarize these implications in two sets of equations, with π(None) ≤ π(2-part tariff) ≤ π(2nd Degree) ≤ π(1st degree) (23.44) comparing profit under the second degree price discrimination scenarios to the extremes of no discrimination and perfect discrimination, and with π(None) ≤ π(3rd Degree) ≤ π(1st degree) (23.45) comparing third degree price discrimination to these same extremes. Exercise 23B.23 * Can you think of a scenario under which all the inequalities turn to equalities in the two equations above? (Hint: Think of goods for which consumers demand only 1 unit.) Turning from profit to consumer surplus, we can derive the following implications for the low demand consumers: 0 = CS 1 (1st Degree) = CS 1 (2-part tariff) = CS 1 (2nd Degree) ≤ CS 1 (None) ≤ CS 1 (3rd Degree). (23.46) Exercise 23B.24 Can you give an intuitive explanation for why this has to hold? For the high demand consumers, however, the implications for consumer surplus are not nearly as unambiguous. We can definitively conclude that 0 = CS 2 (1st Degree) ≤ CS 2 (3rd Degree) ≤ CS 2 (None) (23.47) CS 2 (1st Degree) ≤ CS 2 (2nd Degree) ≤ CS 2 (2-part tariff), (23.48) and but we again cannot be certain about how consumer surplus for the high demand type under no and third degree price discrimination compares to consumer surplus under the two forms of second 888 Chapter 23. Monopoly degree price discrimination. In our example, third degree price discrimination happens to be worse for high demand consumers than either of the forms of second degree price discrimination but no discrimination is better than second degree price discrimination. The theoretical ambiguities with respect to profit and consumer surplus of high demand consumers then create theoretical uncertainty about the overall efficiency (or total surplus) under different monopoly behavior. The only conclusions that hold regardless of the types of demand are that total surplus is largest under first degree price discrimination. For instance, by simply changing γ in our example from 0.5 to 0.4, the ranking of total surplus changes from one in which second degree price discrimination is more efficient than no discrimination which is more efficient than third degree price discrimination (as illustrated in Table 23.2) to one where no discrimination is more efficient than third degree price discrimination which is more efficient than second degree price discrimination. It is therefore important from an efficiency-focused policy perspective to know as much as possible about underlying demands before intervening in monopoly pricing behavior. Furthermore, it may be that policy makers are less concerned about monopoly profit and more concerned about consumer welfare, in which case overall surplus is not the relevant outcome to consider. Exercise 23B.25 Can you think of any definitive policy implications if the goal of policy is to maximize consumer welfare (with no regard to firm profit)? Exercise 23B.26 Explain all the zeros in Table 23.2. Exercise 23B.27 In Table 23.1 we note that there are no restrictions on per-unit prices for the two consumer types under either first or second degree price discrimination, with firms being able to tell consumer types apart in the former case but not the latter. Yet in Table 23.2, the firm appears to be charging exactly the same per unit prices to the two consumers under first degree price discrimination when it can tell the consumers apart and different per-unit prices under second-degree price discrimination when the firm cannot tell the consumer types apart. Explain this intuitively. 23B.4 Barriers to Entry and Natural Monopoly In Section A, we concluded with a discussion of barriers to entry that create monopolies and particularly focused on the case of natural monopolies that are characterized by downward sloping average cost curves. The mathematical treatment of such monopolies is relatively straightforward, and we therefore leave its development to end-of-chapter exercise 23.8. We will also return to the role of barriers to entry in creating market power in Chapters 25 and 26. Conclusion In this chapter, we have begun exploring market power by focusing on the extreme case in which a single firm controls the entire market for a particular good and thus faces the market demand curve rather than the perfectly elastic demand curve that arises for a firm’s output under perfect competition. We noted at the beginning that “market power” is a relative concept that is closely linked to the price elasticity of demand that the firm is facing, with infinite price elasticity representing the extreme case of no market power. We then illustrated how monopolies can take advantage of market power to increase profit, whether it is by charging a single per-unit price to all consumers or by price discriminating in various ways that depend on which pricing strategies are available to the 23B. The Mathematics of Monopoly 889 firm, whether it is possible to prevent re-sale and how much information regarding consumer types the firm has. Unless a firm is able to perfectly price discriminate, we concluded that monopoly behavior results in deadweight loss because monopolies will strategically restrict output in order to raise price. This deadweight loss might be even higher in cases where firms engage in socially wasteful activities in order to attain or maintain monopoly power. At the same time, we noted that our models probably over-predict the size of deadweight losses in many circumstances in which a single firm might in fact control the market for a particular good but in which its monopoly power is disciplined by fear of the possible entry of future competitors. In the case of government-induced monopolies, however, our models may under-estimate the deadweight loss if monopolists expend resources to lobby for government protection. The emergence of deadweight loss from the existence of market power raises the possibility that government intervention in markets characterized by market power might result in efficiency enhancements. But whether such intervention is possible and will in fact lead to increased efficiency depends on the precise nature of the monopoly and the information available to policy makers. In some cases, monopolies might exist for good reasons – such as in the case of natural monopolies that have cost curves which make the presence of multiple firms in the market inherently inefficient. Government intervention in such cases might require information about cost curves that is not readily available to regulators, with the added problem that firms have an explicit incentive to misrepresent their true costs and a possible incentive to not innovate if regulation simply guarantees a “fair market return.” At the same time, we discussed market-based interventions, such as the public provision of the fixed cost infrastructure that might open up the possibility of multi-firm competition along the infrastructure that would otherwise result in a natural monopoly. Often, monopolies exist because governments create market power. Governments might, as we will see more clearly in Chapter 26, offer market power in the form of copyrights and patents in order to provide powerful incentives for innovations that might otherwise not occur, and the surplus from such innovation may well outweigh the deadweight losses from underproduction that arises due to the granted market power. At the same time, governments might grant market power as a result of lobbying efforts by firms that seek profit, thus bestowing “concentrated benefits” on owners of the firm while creating “diffuse costs” that nevertheless exceed the benefits. In such circumstances, efficiency and consumer welfare would clearly be enhanced by the removal of such market power. Finally, when faced with a monopoly exercising its market power through price discrimination, we found that it is not always obvious whether the mere tempering of price discrimination through government intervention will necessarily increase social welfare. In such circumstances, much depends on the underlying specifics of the case. As a result, courts that are asked to adjudicate in anti-trust law suits that challenge monopoly pricing will typically need to take great care to understand the specifics of the case at hand. Our focus in this chapter has been exclusively on the ways in which monopolists can use pricing to exercise market power and generate profit. There are, however, a variety of other ways in which monopolies might exercise market power. These include differentiating the quality of its output across different consumer types and strategically bundling different goods so as to extend monopoly power from one market to another. An entire course can easily be taught on such topics, and probably is taught in your department under the heading of antitrust economics or industrial organization??. If this chapter has been interesting to you, you might want to consider taking such a course in your future studies. We will proceed in Chapters 25 and 26 by investigating market structures that lie in between the extremes of perfect competition and monopoly. Before doing so, however, we need to develop 890 Chapter 23. Monopoly some concepts that assist economists in thinking about strategic behavior – concepts that come under the heading of game theory. It turns out that we have implicitly begun to use some of these concepts in this chapter as we thought through the strategic choices made by monopolists under different pricing strategies (as we illustrate in end-of-chapter exercise 24.11 in the next chapter). We will now formalize these and other concepts – and then return to the topic of market power and its impact on efficiency in a wider array of settings. Appendix: Deriving a “Reduced Form” Utility Function from Separable Preferences In Section 23B.3 we introduced what we called a “reduced form” utility function representing preferences for the monopoly good x that took the form U n = θn u(x) − P (x) (23.49) where θn became our preference parameter that distinguished consumer types and P (x) was the total charge to the consumer for consuming the quantity x of the monopoly good. We indicated at the time that this way of representing preferences for a single good can be derived from a more typical utility function over x and a composite good y. We furthermore indicated that one can assume that consumers in fact have identical underlying preferences and that the parameter θn is simply a measure of consumer income, with consumer demands therefore differing solely because of underlying income differences. We will now illustrate this more fully. Suppose that consumers have underlying preferences that can be represented by the utility function U(x, y) = u(x) + v(y). (23.50) If spending on the monopoly good x represents a relatively small fraction of the consumer’s income I, we can approximate this utility function by writing it as U (x, I) ≈ u(x) + v(I) − P (x) dv(I) . dI (23.51) When we then solve the optimization problem maxx U (x, I), the terms v(I) plays no role in the first order conditions, leaving only the portion (u(x)−P (x)dv(I)/dI) as relevant for the optimization problem. We can then define θ = 1/(dv(I)/dI) and multiply this relevant portion of the utility function by θ to get e θ) = θu(x) − P (x) with θ = U(x, 1 . dv(I)/dI (23.52) The term θ is then simply the inverse of the “marginal utility of income”. It is common to assume that marginal utility of income declines in income – i.e. dv(I)dI < 0. Since θ is the inverse of marginal utility of income, this implies that θ is increasing in incomce – i.e. dθ/dI > 0. Suppose, then, that we have two consumers with identical preferences that can be represented by the separable utility function in equation (23.50) but their incomes are I1 < I2 . Then we can represent their preferences for purposes of determining demand for the monopoly good x by the equation 23B. The Mathematics of Monopoly U (x) = θn u(x) − P (x) with θ1 < θ2 . 891 (23.53) Thus, low demand consumers will be those with less income than high demand consumers. This then implies that, for instance, under full second degree price discrimination, lower income consumers purchase the monopoly good at a higher per-unit price but are charged a lower fixed fee than high demand consumers. End of Chapter Exercises 23.1 Suppose that the demand curve for a product x provided by a monopolist is given by p = 90 − x and suppose further that the monopolist’s marginal cost curve is given by M C = x. A: In this part, we will focus on a graphical analysis – which we ask you to revisit with some simple math in part B. (It is not essential that you have done Section B of the chapter in order to do (a) through (d) of part B of this question.) (a) Draw a graph with the demand and marginal cost curves. (b) Assuming that the monopolist can only charge a single per-unit price for x, where does the marginal revenue curve lie in your graph? (c) Illustrate the monopolist’s profit maximizing “supply point”. (d) In the absence of any recurring fixed costs, what area in your graph represents the monopolist’s profit. (There are actually two areas that can be used to represent profit – can you find both?) (e) Assuming that the demand curve is also the marginal wilingness to pay curve, illustrate consumer surplus and deadweight loss. (f) Suppose that the monopolist has recurring fixed costs of an amount that causes her actual profit to be zero. Where in your graph would the average cost curve lie? In particular, how does this average cost curve relate to the demand curve? (g) In a new graph, illustrate again the demand, M R and M C curves. Then illustrate the monopolist’s average cost curve assuming the recurring fixed costs are half of what they were in part (f). (h) In your graph, illustrate where profit lies. True or False: Recurring fixed costs only determine whether a monopolist produces – not how much she produces. B: Consider again the demand curve and M C curve as specified at the beginning of this exercise. (a) Derive the equation for the marginal revenue curve. (b) What is the profit maximizing output level xM ? What is the profit maximizing price pM (assuming that the monopolist can only charge a single per-unit price to all consumers)? (c) In the absence of recurring fixed costs, what is the monopolist’s profit? (d) What is consumer surplus and deadweight loss (assuming that demand is equal to marginal willingness to pay). (e) What is the cost function if recurring fixed costs are sufficiently high to cause the monopolist’s profit to be zero? (f) Use this cost function to set up the monopolist’s optimization problem and verify your answers to (b). (g) Does the average cost curve relate to the demand curve as you concluded in part A(f)? (h) How does the profit maximization problem change if the recurring fixed costs are half of what we assumed in part (e)? Does the solution to the problem change? 23.2 Everyday and Business Application: Diamonds are a Girl’s Best Friend: Historically, most of the diamond mines in the world have been controlled by a few companies and governments. Through clever marketing by diamond producers, many consumers have furthermore become convinced that “diamonds are a girl’s best friend” because “diamonds are forever.” In fact, the claim is that the only way to show true love is to give a diamond engagement ring that costs the equivalent of 3 months of salary. (We will refer to this throughout the exercise as “the claim.”) A: For purposes of this question, assume that diamonds are only used for engagement rings, that there is no secondary market for engagement rings and that the diamond industry acts as a single monopoly. 892 Chapter 23. Monopoly (a) Let x be the size of diamonds (in karats). Draw a demand curve for x (with the price per karat on the vertical axis) – and make the shape of this demand curve roughly consistent with the claim at the beginning of the question. (b) If this claim is true, what is the price elasticity of demand for diamonds? (c) What price per karat would be consistent with the diamond monopoly maximizing its revenues (assuming the claim accurately characterizes demand)? (d) What price is consistent with profit maximization? (e) How large would the diamonds in engagement rings be if the marketing campaign to convince us of the claim at the beginning of the question was fully successful and if the diamond industry really has monopoly power? (f) True or False: By observing the actual size of diamonds in engagement rings, we can conclude that either the market campaign has not yet fully succeeded or the diamond industry is not really a monopoly. B: Suppose that demand for diamond size is x = (A/p)(1/(1−β)) . (a) What value must β take in order for the claim to be correct? (b) How much revenue will the diamond monopoly earn if the claim holds? Does this depend on what price it sets? (c) Derive the marginal revenue function (assuming the claim holds). Assuming M C > 0, does M R every cross M C? (d) If M C = 0, how large a diamond size per engagement ring is consistent with profit maximization (assuming the claim holds)? (e) Suppose the diamond monopoly has recurring fixed costs that are sufficiently high to cause it’s profits to be zero. If marginal costs were zero, what would be the relationship between the demand curve and the average cost curve? (f) Suppose β = 0.5 and M C = x. What is the profit maximizing diamond size now? (g) What if instead β = −1? 23.3 Business and Policy Application: Monopoly Pricing in Health Insurance Markets: In Chapter 22, we worked with models in which high and low cost customers compete for insurance. Consider the level x of health insurance that consumers might choose to buy, with higher levels of x indicating more comprehensive insurance coverage. A: Suppose that there are relatively unhealthy type 1 consumers and relatively healthy type 2 consumers. The marginal cost of providing additional insurance coverage is then M C 1 and M C 2 , with M C 1 > M C 2 . Unless otherwise stated, assume that d1 = d2 – i.e. the individual demand curves for x are the same for the two types. Also, suppose that the number of type 1 and type 2 consumers is the same, and some portion of each demand curve lies above M C 1 . (a) Begin by drawing a graph with the individual demands for the two types, d1 and d2 , as well as the marginal costs. Indicate the efficient levels of health insurance x1∗ and x2∗ for the two types. (b) Suppose the monopolist cannot tell consumers apart and can only charge a single price to both types. What price will it be and what level of insurance will each type purchase? (c) How does your answer change if the monopolist can first-degree price discriminate? (d) What if he can third-degree price discriminate? (e) Suppose you worked for the Justice Department’s anti-trust division and you only cared about efficiency. Would you prosecute a first-degree price discriminating monopolist in the health insurance market? What if you cared only about consumer welfare? (f) In the text we suggested that it is generally not possible without knowing the specifics of a case whether third degree price discrimination is more or less efficient than no price discrimination by a monopolist. For the specifics in this case, can you tell whether type 1 consumers are better off without this pricediscrimination? What about consumer type 2? (g) Would it improve average consumer surplus to prohibit the monopolist from third-degree price discriminating? Would it be more efficient? B: Suppose next that we normalize the units of health insurance coverage such that the demand function is xn (p) = (θn − p)/θn for type n. You can interpret x = 0 as no insurance and x = 1 as full insurance. Let θ1 = 20 and θ2 = 10 for the two types of consumers, and let M C 1 = 8 and M C 2 = 6. 23B. The Mathematics of Monopoly 893 (a) Determine the efficient level of insurance for each consumer type. (b) If a monopolist cannot tell who is what type and can only charge a single per-unit price for insurance, what will she do assuming there are γ type 1 consumers and (1 − γ) type 2 consumers, with γ < 0.5? (Hint: Define the monopolist’s expected profit and maximize it.) (c) What would the monopoly price be if γ = 0? What if γ = 2/7? What is the highest that γ can be and still result in type 2 consumers buying insurance? (d) Suppose that the monopolist first-degree price discriminates. How much insurance will each consumer type purchase? How much will each type pay for her coverage? (e) How do your answers to (d) change if the monopolist third-degree price discriminates? (f) Let the payment that individual n makes to the monopolist be given by P n = F n + pn xn . Express your answers to (c), (d) and (e) in terms of F 1 , F 2 , p1 and p2 . (g) Suppose γ = 0.5 – i.e. half of the population is type 1 and half is type 2. Can you rank the three scenarios in (c), (d) and (e) from most efficient to least efficient? (h) Can you rank them in terms of their impact on consumer welfare for each type? What about in terms of population weighted average consumer welfare? * Business and Policy Application: Second-Degree Price Discrimination in Health Insurance Markets: In exercise 23.3, we analyzed the case of a monopoly health insurance provider. We now extend the analysis to second-degree price discrimination, with x again denoting the degree of health insurance coverage. A: Consider the same set-up as in part A of exercise 23.3 and assume there is an equal number of type 1 and type 2 consumers. (a) Begin again by drawing a graph with the individual demands for the two types, d1 and d2 , as well as the marginal costs. Indicate the efficient levels of health insurance x1∗ and x2∗ for the two types. 23.4 (b) Under second degree price discrimination, the monopolist does not know who is what type. What two packages of insurance level x and price P (that can have a per-unit price plus a fixed charge) will the monopolist offer? (Hint: You can assume that, if consumers are indifferent between two packages, they each buy the one intended for them.) (c) Is the outcome efficient? Are consumers likely to prefer it to other monopoly pricing strategies? (d) Suppose next that the demand from type 1 consumers is greater than the demand from type 2 consumers, with d1 intersecting M C 1 to the right of where d2 intersects M C 2 . Would anything fundamental change for a first-degree or third-degree price discriminating monopolist? (e) Illustrate how a second-degree price discriminating monopolist would now structure the two health insurance packages to maximize profit. Might relatively healthy individuals no longer be offered health insurance? (f) True or False: Under second-degree price discrimination, the most likely to not buy any health insurance are the relatively healthy and the relatively young. B: ** Consider again the set-up in part B of exercise 23.3. Suppose that a fraction γ of the population is of type 1, with the remainder (1 − γ) of type 2. In analyzing second degree price discrimination, let the total payment P n made by type n be in the form of a two-part tariff pn = F n + pn xn . (a) Begin by assuming that the monopolist will set p2 = p and p1 = M C 1 = 8. Express the level of insurance x2 for type 2 consumers as a function of p. Then express consumer surplus for type 2 consumers as a function of p and denote it CS 2 (p). (b) Why would a second-degree price discriminating monopolist set F 2 equal to CS 2 (p) once she has figured out what p should be? What would the payment P 2 (p) made by type 2 consumers to the monopolist be under p and F 2 (p)? (c) Suppose M C 2 < p < M C 1 . For p in that range, what is the largest possible F 1 that the monopolist can charge to type 1 consumers if she sets p1 = M C 1 = 8. (Hint: Draw the graph with the two demand curves – and then ask how much consumer surplus type 1 consumers could get by simply pretending to be type 2 consumers and accepting the package designed for type 2 consumers.) (d) Suppose instead that M C 1 < p < 10. What would now be the largest possible F 1 that is consistent with type 1 consumers not buying the type 2 insurance (assuming still that p1 = M C 1 = 8? (Hint: Use another graph as you did in the previous part to determine the answer.) 894 Chapter 23. Monopoly (e) Given that the fraction of type 1 consumers is γ (and the fraction of type 2 consumers is (1 − γ)), what is the expected profit E(π(p)) per customer from setting p2 = p when M C 2 < p < M C 1 ? What if M C 1 < p < 10? (f) For both cases – i.e. for M C 2 < p < M C 1 and when M C 1 < p < 10 – set up the optimization problem the second degree price discriminating monopolist solves to determine p. Then solve for p in terms of γ. (Hint: You should get the same answer for both cases.) (g) Determine the value for p when γ = 0. Does your answer make intuitive sense? What about when γ = 0.1, when γ = 0.2, when γ = 0.25? True or False: As the fraction of type 1 consumers increases, health insurance coverage for type 2 consumers falls. (h) At what value for γ will type 2 consumers no longer buy insurance? If we interpret the difference in types as a difference in incomes (as outlined in the appendix), can you determine which form of price discrimination is best for low income consumers? 23.5 Business and Policy Application: Labor Unions Exercising Market Power: Federal anti-trust laws prohibit many forms of collusion in price setting between firms. Labor unions, however, are exempt from anti-trust laws and are allowed to use market power to raise wages for their members. A: Consider a competitive industry in which workers have organized into a union that is now renegotiating the wages of its members with all the firms in the industry. (a) To keep the exercise reasonably simple, suppose that each firm produces output by relying solely on labor input. How does each firm’s labor demand curve emerge from its desire to maximize profit? Illustrate a single firm’s labor demand curve (with the number of workers on the horizontal axis). (Note: Since these are competitive firms, this part has nothing to do with market power.) (b) On a graph next to the one you just drew, illustrate the labor demand and supply curves for the industry as a whole prior to unionization. (c) Label the competitive wage w ∗ and use it to indicate in your first graph how many workers an individual firm hired before unionization. (d) * Suppose that the union that is negotiating with the firm in your graph is exercising its market power with an aim toward maximizing the overall gain for its members. Suppose further that the union is sufficiently strong to be able to dictate an outcome. Explain how the union would go about choosing the wage in this firm and the size of its membership that will be employed by this firm. (Hint: The union here is assumed to have monopoly power – and the marginal cost of a member is that member’s competitive wage w ∗ .) (e) If all firms in the industry are becoming unionized, what impact will this have on employment in this industry? Illustrate this in your market graph. (f) Suppose that those workers not chosen to be part of the union migrate to a non-unionized industry. What will be the impact on wages in the non-unionized sector? B: * Suppose that each firm in the industry has the same technology described by the production function f (ℓ) = Aℓα with α < 1, and suppose that there is some fixed cost to operating in this industry. (a) Derive the labor demand curve for each firm. (b) Suppose that the competitive wage for workers of the skill level in this industry is w ∗ . Define the optimization problem that the labor union must solve if it wants to arrive at its optimal membership size and the optimal wage according to the objective defined in A(d). (It may be more straightforward to set this up as a maximization problem with w rather than ℓ as the choice variable.) (c) Solve for the union wage w U that emerges if the union is able to use its market power to dictate the wage. What happens to employment in the firm? (d) Can you verify your answer by instead finding M R and M C from the perspective of the union – and then setting these equal to one another? (e) Given the fixed cost to operating in the industry, would you expect the number of firms in the industry to go up or down? 23.6 * Business and Policy Application: Monopsony: A Single Buyer in the Labor Market: The text treated extensively the case where market power is concentrated on the supply side – but it could equally well be concentrated on the demand side. When a buyer has such market power, he is called a monopsonist. Suppose, for instance, the labor market in a modest-sized town is dominated by a single employer (like a large factory or a major university). In such a setting, the dominant employer has the power to influence the wage just like a typical monopolist has the power to influence output prices. 23B. The Mathematics of Monopoly 895 A: Suppose that there is a single employer for some type of labor, and to simplify the analysis, suppose that the employer only uses labor in production. Assume throughout that the firm has to pay the same wage to all workers. (a) Begin by drawing linear labor demand and supply curves (assuming upward sloping labor supply). Indicate the wage w ∗ that would be set if this were a competitive market and the efficient amount of labor ℓ∗ that would be employed. (b) Explain how we can interpret the labor demand curve as a marginal revenue curve for the firm. (Hint: Remember that the labor demand curve is the marginal revenue product curve.) (c) How much does the first unit of labor cost? Where would you find the cost of hiring a second unit of labor if the firm could pay the second unit of labor more than the first? (d) We are assuming that the firm has to pay all its workers the same wage – i.e. it cannot wage discriminate. Does that imply that the marginal cost of hiring the second unit of labor is greater or less than it was in part (c)? (e) How does the monopsony power of this firm in the labor market create a divergence between labor supply and the firm’s marginal cost of labor – just as the monopoly power of a firm causes a divergence between the output demand curve and the firm’s marginal revenue curve? (f) Profit is maximized where M R = M C. Illustrate in your graph where marginal revenue crosses marginal cost. Will the firm hire more or fewer workers than a competitive market would (if it had the same demand for labor as the monopsonist here)? (g) After a monopolist decides how much to produce, he prices the output at the highest possible level at which all the product can be sold. Similarly, after a monopsonist decides how much to buy, he will pay the lowest possible price that will permit him to buy this quantity. Can you illustrate in your graph the wage w M that our dominant firm will pay workers? (h) Suppose the government sets a minimum wage of w ∗ (as defined in (a)). Will this be efficiency enhancing? (i) We gave the example of a modest-sized town with a dominant employer as a motivation for thinking about monoposonist firms in the labor market. As it becomes easier to move across cities, do you think it is more or less likely that the monopsony behavior we have identified is of significance in the real world? (j) Labor unions allow workers to create market power on the supply side of the labor market. Is there a potential efficiency case for the existence of labor unions in the presence of monopsony power by firms in the labor market? Would increased mobility of workers across cities strengthen or weaken this efficiency argument? B: Suppose that the firm’s production function is given by f (ℓ) = Aℓα (with α < 1) and the labor supply curve is given by ws (ℓ) = βℓ. (a) What is the efficient labor employment level ℓ∗ ? (Hint: You should first calculate the marginal revenue product curve.) (b) At what wage w ∗ would this efficient labor supply occur? (c) Define the firm’s profit maximization problem – keeping in mind that the wage the firm must pay depends on ℓ. (d) Take the first order condition of the profit maximization problem. Can you interpret this in terms of marginal revenue and marginal cost? (e) How much labor ℓM does the monopsonist firm hire – and how does it compare to ℓ∗ ? (f) What wage w M does the firm pay – and how does it compare to w ∗ ? (g) Consider the more general case of a monopsonist firm with production function f (ℓ) facing a labor supply curve of w(ℓ). Derive the M R = M C condition (which is the same as the condition that the marginal revenue product equals M C) from the profit maximization problem. (h) Can you write the M C side of the equation in terms of the wage elasticity of labor supply? (i) True or False: As the wage elasticity of labor supply increases, the monopsonist’s decision approaches what we would expect under perfect competition. 23.7 Business and Policy Application: Taxing Monopoly Output: Under perfect competition, we found that the economic incidence of a tax – i.e. who ends up paying a tax – had nothing to do with statutory incidence – i.e. who the law said should pay the tax. 896 Chapter 23. Monopoly A: Suppose the government wants to tax the good x which is exclusively produced by a monopoly with upward sloping marginal cost. (a) Begin by drawing the demand, marginal revenue and marginal cost curves. On your graph, indicate the profit maximizing supply point (xM , pM ) chosen by the monopolist in the absence of any taxes. (b) Suppose the government imposes a per-unit tax of t on the production of x – thus raising the marginal cost by t. Illustrate how this changes the profit maximizing supply point for the monopolist. (c) What happens to the price paid by consumers? What happens to the price that monopolists get to keep (given that they have to pay the tax)? (d) Draw a new graph as in (a). Now suppose that the government instead imposes a per-unit tax t on consumption. Which curves in your graph are affected by this? (e) In your graph, illustrate the new marginal revenue curve – and the impact of the consumption tax for the monopolist’s profit maximizing output level. (f) What happens to the price paid by consumers (including the tax)? What happens to the price received by monopolists? (g) In terms of who pays the tax, does it matter which way the government imposes the per-unit tax on x? (h) By how much does deadweight loss increase as a result of the tax? (Assume that demand is equal to marginal willingness to pay.) (i) Why can’t monopolists just use their market power to pass the entire tax onto the consumers? B: Suppose the monopoly has marginal costs M C = x and faces the demand curve p = 90 − x as in exercise 23.1. (a) If you have not already done so, calculate the profit maximizing supply point (xM , pM ) in the absence of a tax. (b) Suppose the government introduces the tax described in A(b). What is the new profit maximizing output level? How much will monopolists charge? (c) Suppose the government instead imposed the tax described in A(d). Set up the monopolist’s profit maximization problem and solve it. (d) Compare your answers to (b) and (c). Is the economic incidence of the tax affected by the statutory incidence? (e) What fraction of the tax do monopolists pass onto consumers when monopolists are statutorily taxed? What fraction of the tax do consumers pass onto monopolists when consumers are statutorily taxed? 23.8 Business and Policy Application: Two Natural Monopolies: Microsoft versus Utility Companies: We suggested in the text that there may be technological reasons for the barriers to entry required for the existence of a monopoly. In this exercise, we consider two examples. A: Microsoft and your local utilities company have one thing in common: They both have high fixed costs and low variable costs. In the case of Microsoft, the fixed cost involves producing software which, once produced, can be reproduced cheaply. In the case of your local utility company, the fixed cost involves maintaining the infrastructure that distributes electricity to homes, with the actual delivery of that electricity costing relatively little if the infrastructure is in good shape. (a) Let’s begin with Microsoft. Draw a graph with low constant marginal costs and a downward sloping demand curve. Add Microsoft’s marginal revenue curve and indicate which point on the demand curve Microsoft will choose (assuming, until later chapters, that it is not worried about potential competitors). Then draw a second and similar graph for your local utilities company. (b) There is one stark difference between Microsoft and your local utilities company: Microsoft has not asked the government for help to allow it to operate but has instead been under strict scrutiny by governments around the world for potential abuse of its market power. Utility companies, on the other hand, have often asked for government aid in regulating prices in such a way that the companies can earn a reasonable profit. What is missing from your two graphs that can explain this difference? (c) Put into words the “problem” in the two cases from a government’s perspective (assuming the government cares about efficiency)? (d) In the case of Microsoft, how can the granting of a copyright on the software explain the existence of “the problem”? How much is Microsoft willing to pay for this copyright? 23B. The Mathematics of Monopoly 897 (e) Now consider the “problem” in the utilities industry. How would setting a two-part tariff allow the utilities company to produce at zero profit? If properly structured, might its output level be efficient? (f) Explain how the alternative of having the government lay and maintain the infrastructure on which electricity is delivered could address the same “problem”. (g) What would be the analogous government intervention in the software industry – and why might you think that this was not a very good idea there? (Hint: Think about innovation.) Could you think of a way to offer a similar criticism regarding the proposal of having the government provide the infrastructure for electricity delivery? B: We did not develop the basic mathematics of natural monopolies in the text and therefore use the remainder of this exercise to do so. Suppose demand for x is characterized by the demand curve p(x) = A − αx. Suppose further that x is produced by a monopolist whose cost function is c(x) = B + βx. (a) Derive the monopolist’s profit maximizing supply point – i.e. the price and quantity (pM , xM ) under the implicit assumption of no price discrimination. (b) At the output level xM , what is the average cost paid by the monopolist? (c) How high can fixed costs be and still permit the monopolist to make non-negative profit by choosing the supply point you calculated in (a)? (d) How much is Microsoft willing to pay its lawyers to get copyright protection? (e) Suppose both Microsoft and your local utility company share the same demand function. They also share the same cost function except for the fixed cost B. Given our description of the “problem” faced by Microsoft versus your utility company, whose B is higher? (f) Suppose B for the utility company is such that it cannot make a profit by behaving as you derived in (a) and suppose there are N households. Suggest a two-part tariff that will allow the utility company to earn a zero profit while getting it to produce the efficient amount of electricity. (g) Suppose the government were to build and maintain the infrastructure needed to deliver electricity to people’s homes. It furthermore allows any electricity firm to use the infrastructure for a fee δ (per unit of electricity that is shipped). Can the electricity industry be competitive in this case? What has to be true about the fee for using the infrastructure in order for this industry to produce the efficient level of electricity? 23.9 Policy Application: Some Possible “Remedies” to the Monopoly Problem: At least when our focus is on efficiency, the core problem with monopolies emanates from the monopolist’s strategic under-production of output – not from the fact that monopolists make profits. But policy prescriptions to deal with monopolies are often based on the presumption that the problem is that monopolies make excessive profits. A: Suppose the monopoly has marginal costs M C = x and faces the demand curve p = 90 − x as in exercise 23.1. Unless otherwise stated, assume there are no recurring fixed costs. In each of the policy proposals below, indicate the impact the policy would have on consumer welfare and deadweight loss. (a) The government imposes a 50% tax on all economic profits. (b) The government imposes a per-unit tax t on x. (In problem 23.7, you should have concluded that it does not matter whether the tax is levied on production or consumption.) (c) The government sets a price ceiling equal to the intersection of M C and demand. (Hint: How does this change the marginal revenue curve?) (d) The government subsidizes production of the monopoly good by s per unit. (e) The government allows firms to engage in first-degree price discrimination. (f) Which of the analyses above might change if the firm also has recurring fixed costs. (g) True or False: In the presence of distortions from market power, price distorting policies can be efficient. B: Suppose demand and marginal costs are as specified in part A. Unless otherwise stated, assume no recurring fixed costs. (a) Determine the monopolist’s optimal supply point (assuming no price discrimination). Does it change when the government imposes a 50% tax on economic profits? (b) Suppose the government imposes a $6 per unit tax on the production of x. Solve for the new profit maximizing supply point. (c) Is there a price ceiling at which the monopolist will produce the efficient output level? 898 Chapter 23. Monopoly (d) For what range of recurring fixed costs would the monopolist produce prior to the introduction of the policies in (a), (b) and (c) but not after their introduction? (e) What is the profit maximizing output level if the monopolist can perfectly price discriminate? (f) How high a per-unit subsidy would the government have to introduce in order for the monopolist to produce the efficient output level? (g) For what range of recurring fixed costs does the monopolist not produce in the absence of a subsidy from part (f) but produces in the presence of the subsidy? If recurring fixed costs are in this range, will the monopolist produce the efficient quantity under the subsidy? 23.10 Policy Application: Pollution and Monopolies: In Chapter 21, we discussed the externality from pollutionproducing industries within a competitive market. A: Suppose now that the polluting firm is a monopolist. (a) Begin by illustrating a linear (downward sloping) demand curve and an upward sloping M C curve for the monopolist. Indicate the efficient level of production in the absence of any externalities. (b) Draw the marginal revenue curve and illustrate the monopolist’s profit maximizing “supply point”. (c) Suppose that the monopolist pollutes in the process of producing, with the social marginal cost curve SM C therefore lying above the monopolist’s marginal cost curve. Does this change anything in terms of the monopolist’s profit maximizing decision? (d) Illustrate a SM C curve with sufficient pollution costs such that the monopoly’s output choice becomes efficient. (e) True or False: In the presence of negative production externalities, the per-unit tax that would cause the monopolist to behave efficiently might be positive or negative (i.e. it might take the form of a tax or a subsidy). (f) Suppose that the production externality were positive instead of negative. True or False: In this case, the monopolist’s output level will be inefficiently low. B: Suppose a monopolist faces the cost function c(x) = βx2 , but production of each unit of x causes pollution damage B. (a) What is the marginal cost function for the monopolist? What is the social marginal cost function? (b) Suppose demand curve is equal to p(x) = A − αx. Determine the monopolist’s output level xM (assuming no price discrimination). (c) What is the monopoly price? (d) For what level of B is the monopolist’s output choice efficient? 23.11 Policy Application: Regulating Market Power in the Commons: In exercises 21.9 and 21.10, we investigated the case of many firms emitting pollution into a lake. We assumed the only impact of this pollution was to raise the marginal costs for all firms that produce on the lake. A: Revisit part A(g) of exercise 21.10. (a) How does a merging of all firms around the lake (into one single firm) solve the externality problem regardless of how large the pollution externality is? (b) Suppose you are an anti-trust regulator who cares about efficiency. You are asked to review the proposal that all the firms around this lake merge into a single firm. What would you decide if you found that, despite being the only firm that produces output x on this lake, there are still plenty of other producers of x such that the output market remains competitive. (c) Suppose instead that, by merging all the firms on the lake, the newly emerged firm will have obtained a monopoly in the output market for x. How would you now think about whether this merger is a good idea? (d) How would your answers to (c) and (d) change if the externality emitted by firms on the lake lowered rather than raised everyone’s marginal costs? B: Suppose, as in exercise 21.9 and 21.10, that each of the many firms around the lake has a cost function c(x) = βx2 + δX where x is the firm’s output level and X is the total output by all firms around the lake. (a) In exercise 21.10B(a), we discussed how a social planner’s cost function for each firm would differ from that of each individual firm. Review this logic. How does this apply when all the firms merge into a single company that owns all the production facilities around the lake? (b) Will the single company make decisions different from that of the social planner in exercise 21.10? What does your answer depend on?