Monopoly: a game economists love to play--badly!
In choosing my topic for this address, I take to heart the words of Lord Acton that James Buchanan quoted at the start of his presidential address back in the early 1960s (Buchanan 1964), which for me remains the quintessential model for all such addresses, "[I]t is not the popular movement, but the traveling of the minds of men who sit in the seat of Adam Smith that is really serious and worthy of all attention" (Acton 1904, p. 212). I have written much over the years for the popular movement, as many of you know. However, for this talk I want to focus on the traveling of our collective minds, as represented by conventional wisdom within the profession. In the end, for all of us who aspire to sit in the seat of Adam Smith, it is in the ongoing battle of fundamental ideas for space in what will eventually become conventional wisdom--that we will likely make our most lasting mark.
My topic for the talk is monopoly. Indeed, this is a topic very familiar to those who sit in Adam Smith's seat. Perhaps the theory of monopoly is so firmly established in microeconomic theory that you might wonder how I could possibly think there is anything novel to say about it. However, I submit that our theory of monopoly is in serious need of repair, especially since it is so widely studied and, because of embedded problems, is so misleading when it comes to understanding the dynamics of a market economy and to guiding antitrust and other government policies.
Admittedly, one of my essential points is not totally novel. It is one that the late and great Joseph Schumpeter made, albeit as an aside, in his classic Capitalism, Socialism and Democracy, first published back in 1942 (the year I was born):
A system--any system, economic or other--that at every given point of time fully utilizes its possibilities to the best advantage may yet in the long run be inferior to a system that does so at no point of time, because the latter's failure to do so may be a condition for the level and speed of long-run performance. (1942, p. 83; emphasis in the original)
In this talk, I will seek to elaborate on Schumpeter's counterintuitive and little appreciated point, adding extensions along the way. My thesis will be threefold:
(i) My weak point is that economists' standard static model of monopoly exaggerates in various ways the harm done by the prevalence of monopolies in a market economy.
(ii) My stronger point is that some monopoly prevalence is absolutely necessary for a well-functioning, dynamic market economy.
(iii) My strongest point is that under some market conditions, monopoly pricing, as represented by the standard model, can be, on balance, beneficial to consumers.
That is to say, an economy made up of a perfectly fluid, perfectly efficient, perfectly competitive market system--the idealized standard of market analysis--will, as Schumpeter recognized, likely be inferior in the long run to a system that is less than perfectly efficient at every point in time. Perfect fluidity of resources is hardly the idealized market state that we, as economists, make it out to be.
You will be pleased to hear, I suspect, that my central points are sufficiently elementary that my address will be short. However, you should know that there are many details to my arguments that are intentionally left out because of the time (and journal space) limits. These details are covered in a book that I recently completed with Dwight Lee, called Monopoly: Market Bane or Boon? (2003).
1. The Conventional Monopoly Model
Within the economics profession, monopoly has a well-entrenched definition. In its purest form it is almost everywhere defined as a market structure in which there is a single producer of a product that has no close substitutes and that is protected by consequential, if not prohibitive, barriers to entry. Accordingly, the (pure) monopolist faces the market demand for its good and is capable of choosing among the various price-quantity combinations on its demand curve with the single-minded goal of maximizing firm profits. A profit-maximizing monopolist will, of course, invariably price above marginal cost.
Economists might quibble over when substitutes are sufficiently close to withdraw the monopoly designation of a market structure, but all seem to agree that monopolists of all stripes necessarily face downward-sloping demand curves, with the elasticity of demand affected by how close the substitutes are and how high the entry barriers are.
Regardless of the fine distinctions that can be drawn, the monopolist's ability to affect, consequentially, total market supply of its product enables it to charge more than the competitive price and to extract monopoly rents. As a consequence, under this construction of monopoly, consumers are necessarily harmed by the monopolist's pricing decisions since consumers lose consumer surplus, not all of which is recaptured by the monopolist in the form of rents, an outcome fully appreciated by the venerable Dr. Smith and many philosophers before him. (1) That is to say, any self-respecting monopolist will give rise to some deadweight loss, the so-called Harberger triangle. In this regard and to this degree (the size of the Harberger triangle), the only good monopoly is one that has been made extinct by the forces of Schumpeter's creative destruction. Few seem to appreciate, as Schumpeter did, the good that can come from the actual prevalence, at all points in time, of monopolies and the deadweight losses (in static terms) they impose. After all, a deadweight loss should be exactly what is implied, a drag on the economy. The irony of monopolies is that the static deadweight losses are the sources of a market economy's dynamism.
What is especially interesting about standard monopoly theory in which the deadweight loss of monopoly is precisely identified is the unstated presumption that the organization of a market has absolutely no impact on the cost of production. That is to say, the market supply curve under perfectly competitive market conditions is identical with the monopolist's marginal cost curve, a position that should give any economist familiar with the principal/agency literature reason to pause. If a (perfectly) competitive market is served by numerous small firms, then surely there is a greater correspondence of the interests of the principals with those of the agents (if the principals are not often the same as the agents) than can be the case if the entire market is served by one overarching megafirm--a monopolist--in which many, if not all, of the former principals will be transformed into agents, owning only a minor fraction of the firm's stock, if they are owners at all.
Surely the internal coordinating costs in the two market structures cannot be the same. Just as surely, a market structured as perfect competition cannot tolerate any (but the minimal) agency costs simply because there exists zero opportunity for the firms to make an economic profit (or to cover anything more than minimal production costs). (2) Any perfectly competitive firm that suffers any (but the minimal) agency costs is a firm that is doomed to be replaced. Agency costs are not only possible but also likely in the same market structured as a monopoly.
As Dwight Lee and I have argued (2000), on the one hand, the principal/agency problems embedded in the replacement of a formerly competitive market with a single firm ensure that production will be constrained, independent of any imposed restriction on production that the monopolist imposes to generate economic profits. On the other hand, unless internal coordination/agency costs are zero, the curb in production under monopoly cannot be as great as the standard model suggests. This is true simply because of the growing coordination/agency costs that are bound to emerge as output is restricted and economic profits emerge and become progressively larger with further restrictions. These escalating costs will, at some point, choke off the curb in output before the standard monopoly output level is reached. As a consequence, scholars pressing the rent-seeking perspective may have overstated the extent of rent-seeking because, if monopoly is seen as a principal/agency-coordination problem, there will be less rent to seek than the standard model presupposes. Some of the rents (which will be less than the standard model describes) will be soaked up by internal (rent-seeking) agents.
Of course, this perspective suggests another point rarely noted: Not all reductions in agency costs are welfare enhancing in monopolized markets. This is because a reduction in agency costs can ease the monopolist's task of coordinating a greater curb in production.
Furthermore, we must note that in virtually all discussion of monopoly, the deadweight loss--the monopoly problem or market failure--is attributed exclusively to the monopolist's control over supply, which emerges from the existence of barriers to entry. Few economists seem to realize that the monopoly problem could just as easily be laid squarely in the lap of consumers who, because of the very existence of the deadweight loss, should--in a world of perfect information and zero organizational costs--be able to buy off the monopolist and get the competitive output level. The consumers' problem, however, is that they can't effect a buy-off because of their organizational costs, which are made prohibitive by the rampant free-rider problem. The consumers' free-riding can just as easily and as legitimately be construed as the source for any persistent deadweight loss of monopoly as the monopolist's control over supply, which is predicated on the presumption of zero organizational costs (which suggests a theoretical asymmetry of major proportions in the treatment of the two sides of the market that is bound to distort professional assessment of the relative efficiency of perfectly competitive and monopoly models). Indeed, it would surely be more accurate for all of us who attempt to sit in Adam Smith's seat to acknowledge openly that any inefficiency of monopoly is the result of the interaction of the two problems, the monopolist's control of market supply (absent any organizational costs) and consumers' presumed prohibitive organizational costs and lack of control over market demand.
My recasting of the monopoly problem as a two-sided one may not appear to make much of a difference, but, surprisingly, it does. When the monopoly problem is seen exclusively as a supply/ barrier-to-entry problem, the breakdown of monopoly can occur from only one source, the breakdown or the circumvention of the entry barriers. When monopoly is seen as a two-sided problem, there is a whole new way in which the efficiency of monopolized markets can be improved, by changes in consumers' organizational costs that can be based in improvements in the technology of organizational forms (with new profit-maximizing business forms arising to act as surrogate bargainers for consumers). The lower the organizational costs for consumers (and their surrogates), the smaller the deadweight loss of monopoly, a point that is rarely tendered but is certainly relevant when markets are not as perfect as perfect competition, and also one that suggests that the inefficiency of monopoly should not be taken as equal to the Harberger triangle, as conventionally identified.
2. Implications of Conventional Monopoly Wisdom
Setting aside such refinements in theory, the implications of the conventional textbook analysis of monopoly markets are well known.
* Consumers should be expected to oppose (if they have the requisite incentive to do so, which they don't) any efforts on the part of the monopolist to enhance its protective entry barriers, whatever their source. After all, such efforts could only raise the price that a profit-maximizing monopolist would charge and would lower consumers' surplus value.
* Conversely, following conventional monopoly wisdom, the existence of some identified barrier to entry, whether natural or artificial, necessarily translates into exploitive monopoly power. The higher the barrier, or the greater the costs new entrants must incur to enter the market, the greater the monopoly power, and the greater the market inefficiency.
* The protected firm can be expected to exploit its favored market position just to maximize the wealth of the owners. If the firm's protected position is not exploited, then the firm's stock price will suffer. Savvy investors can be expected to buy out the monopoly owners, hike the firm's price to monopoly levels, and then sell their interest in the monopoly firm at a higher price that reflects the firm's garnered monopoly rents. Hence, monopolies must act like monopolies under competitive markets for corporate control. (3)
* Barring their ability to collude (and consumers are not expected to be able to collude for such purposes, given their collective decision-making costs), (4) we would expect consumers to favor antitrust prosecution against monopolists, but only so long as the costs of prosecution are lower than the added consumer surplus resulting from greater competition, lower prices, and greater industry output. Also, consumers might understandably favor the regulation of monopoly, as long as the regulation were actually used to further the interests of consumers, not the regulated monopoly (which, as so many in this association have stressed throughout their careers, might be a political pipedream).
* The competitors of any producer, which has monopoly power because of its market dominance, should be expected to applaud the dominant producer's tendency to restrict output, a point stressed by Baumol and Ordover (1985) nearly two decades ago. This is because the competitors would experience an increase in their demand with the dominant producer's cutback in production. They should also be expected to oppose antitrust prosecution of the dominant producer simply because even the threat of antitrust prosecution can lead the dominant producer to curb its sales by less than otherwise, which means that competitors would experience a reduction in their sales.
* Finally, price discrimination in standard monopoly analysis is a two-edged sword. Price discrimination (especially the perfect kind) can give rise to an expansion of output and a reduction in the deadweight loss of monopoly. At the same time, price discrimination can redistribute the consumer surplus from consumers to the monopolist. Beyond recapturing the otherwise would-have-been deadweight loss, the price discrimination contributes nothing to consumer welfare. In general, all monopoly profits, brought on by higher than competitive prices, are treated as an unearned grab of the surplus that is rightfully the consumer's. Certainly, given conventional treatment of monopolies, monopoly profits serve no productive role in markets.
3. Breaks with Conventional Wisdom
In our new book, Dwight Lee and I lay out in some detail the problems of conventional monopoly theory. Here, I will summarize only several of our key arguments that we develop at length in our book.
The Microsoft Problem
Somewhat in passing, and at the least incisive level of analysis, I might note that recent use of conventional monopoly theory in the Microsoft antitrust case doesn't appear to be at all consistent with economists' expectations that consumers should oppose monopoly restrictions while producers should favor them. There is no firm in modern economic history that has been more squarely identified by the legal system--the Justice Department, the offices of 19 state attorneys general, and the courts (as well as the European Commission dealing with antitrust matters)--as a ruinous monopoly than has the Microsoft Corporation. (5) This is because both the district court and the appeals court have agreed that, when they considered the case, Microsoft's market share of the operating-system market was no less than 80% and could exceed 95% (Jackson 1999, 2000; U.S. Court of Appeals 2001). (6) Moreover, Microsoft is clearly protected by an insurmountable entry barrier that is new to the digital age, the applications barrier to entry. (7)
The irony of the Microsoft antitrust case, which suggests that something could be wrong with monopoly theory (if the courts are right on declaring Microsoft a monopoly), is that the case was fervently pressed by Microsoft's key competitors (including Oracle, Sun Microsystems, AT&T, and IBM) and, at the same time, generally not supported at all by many consumers of software products. Indeed, in spite of Microsoft's being constantly touted as the world's greatest and most destructive of monopolies each day of the four years that the trial was moving through the courts, a sizable majority of consumers tended to side with Microsoft and were often hostile to the government's side of the case. (8) Something is amiss. Either the courts have misidentified Microsoft as a monopolist, or monopolies don't behave exactly as conventional theory suggests. Then, just recently, the Wall Street Journal, as so many other commentators have done, characterized the "core of the case" as "Microsoft's long-standing strategy of adding new features to Windows as a way of derailing would-be competitors" (Wilke 2003). What is so odd about that statement is that "adding new features" to a product could more easily be construed as the behavior of a fierce competitor, not the predatory monopolist that Microsoft has been made out to be by the Justice Department, the District Court, and a host of their supporting economists.
Perhaps the legal system has been misguided by economists' conventional model of monopoly. Perhaps, as Schumpeter has recognized, monopolies (or firms that are thought to be monopolies) are more competitive, efficiency enhancing, and beneficial to consumers in ways and to a degree that conventional monopoly theory fails to capture.
The Problem of Good Creation
Granted, the Microsoft antitrust case amounts to a legal anecdote, or a single data point that is hardly the strongest, most convincing basis for encouraging a professional reconsideration of basic monopoly theory. Fortunately, there are a number of other transparent problems with standard monopoly theory that do not appear to be widely appreciated. Let me count them.
A major source of professional confusion over the contribution of monopoly lies in an assumption of static theorizing relating to all market structures: The good that is produced and subject to analysis is assumed to exist. Indeed, the good that is produced doesn't even fall like manna from heaven. If it did, more economists might wonder about where the good came from and about the institutional mechanisms that are needed to get to the static stage of analysis in which the good is simply there, on the horizontal axis of our graphs we draw in our classrooms.
Under such a market setting, in which the good appears as if by magic, it is understandable why the division of the consumer surplus between the producer and consumers becomes a debatable one, albeit readily cast aside by economists because judgments on the division necessarily involve interpersonal utility comparisons on which economists, as they readily acknowledge, have no special expertise. There is no good reason--philosophical or economic--for the monopolist to claim any part of the surplus as economic profits or to withhold a portion of the surplus from consumers to obtain their economic profits. Indeed, any economic profit taken by the monopolist is, given that the good is assumed into existence, fully equivalent with economists' notion of rents, that gain that is totally unearned. The monopolist did nothing to earn the profits, other than restrict output. The monopolist certainly did nothing to identify an unmet need or want, create the good that could satisfy the unmet need or want, and then develop the market for the good. All of these problems are solved, as if by magic, by economists assuming that the good exists. Nevertheless, as suggested in our standard monopoly model, consumers and broader society, meaning all those people other than the monopolist's owners--have a greater claim to the consumer surplus than does the monopolist. This is self-evident in all the talk about the deadweight loss, inefficiency, or market failure of the monopolist. Why use such derogatory expressions if there isn't a presumption that the consumers have a right to the lost efficiency? Why do we cite the monopolist's economic profits as rents--meaning unearned--without, at the same time, citing consumers for having received unearned consumer surplus?
Perspectives must necessarily change, however, when we consider that in the real world, goods subject to monopolization, and any resulting deadweight loss, must be conceived and created, along with their markets being developed. Surely the distribution of any consumer surplus can be assessed by accounting for who is responsible for the very existence of the good on the horizontal axis, as well as for the existence and development of the market under study. After all, in the real world, markets for goods are no more subject to falling from the skies than are the goods themselves. Both markets and goods typically emerge in tandem, from a creative/developmental process that is never considered in monopoly analysis.
If consumers are responsible for the development of the good and its market, then we can argue that the consumers would clearly have a greater claim on the resulting consumer surplus than would the monopolist. They would also have a justified beef over the monopolist's restricting production. There would clearly be a deadweight loss, meaning a loss of consumer welfare that could have been captured but is not.
On the other hand, if the monopolist is responsible for the creation of the good and the development of the market for the good, then surely the consumers would not have the same entitlement to the resulting consumer surplus as they do when they are responsible for the development of the good and market. Indeed, consumers could be pleased with the monopolist's claiming a portion of the consumer surplus, as well as be pleased to suffer the deadweight loss--if they understood that, by doing so, they would have some surplus left over, the rarely mentioned Dupuit triangle, the area above the monopoly price line and below the demand curve. (9) Indeed, from this perspective, the would-be monopolist can be seen as an organizational form, the central economic function of which is to overcome consumers' free-rider problems in conceiving and developing goods and services that meet their needs and wants. A monopolist may extract monopoly profits and impose the deadweight-loss triangle, once the good and market exist fully formed, as treated in our textbook monopoly models. However, those losses of consumers' surplus can be viewed as necessary costs the consumers are more than willing to incur to ensure their extraction of a host of Dupuit triangles garnered from an array of goods.
Digressions: Mutually Beneficial Trades and the Perfect Fluidity of Markets
At this point, I need to digress. There are two dimensions to the way we treat monopoly in our classes and textbooks that should be recognized for what they are: intellectual embarrassments. First, we start our courses with demonstrations of how trade can be, and must be, mutually beneficial. The mutual benefits are viewed as essential to the consummation of trades. So many of us then move seamlessly into a later treatment of market structures in which, in the case of perfect competition, trade is hardly mutually beneficial. Consumers get all the true gains--the consumer surplus--while producers reap only full cost recovery. Clearly, something is amiss in our stress on the necessity of mutual gains in all economic dealings and the outcome of perfectly competitive markets that is so one-sided. If trades are to be mutually beneficial, then it would appear that some deadweight loss (as conventionally described in our monopoly models) is a theoretical necessity, in the absence of price discrimination, because there is no way else the economic profits can be extracted.
Second, perfect competition is perfectly efficient, but only because of some very extreme assumptions. I understand that many economists who sit in Adam Smith's seat acknowledge the unlikelihood of perfect efficiency because of supposed market imperfections such as the real-world cost of gathering information on prices (if nothing else) that is assumed away in the perfectly competitive market. However, I am not so sure that economists widely recognize the extent of the monumental imperfection built into the perfectly competitive model by the assumption of perfect fluidity of resources among markets (grounded in the explicit assumption of zero entry and exit costs). In such a world, the potential for economic gains above costs are zero. The underlying irony of the model is that there would be no incentive for any producer to get off his or her proverbial (economic) dime and do anything different (and better). Everyone could rightfully assume that any effort to get off the dime and do something different (and better) would be to no good end because any net gains would evaporate as soon as the potential for the gains is realized.
Granted, we might escape the paradox of perfect competition--that economic profits get competed away when no one has an incentive to move--by assuming it away and falling back on what amounts to a cop-out: "This is static analysis." However, once that position is taken, then Schumpeter's essential insight regarding ever-present efficiency in markets leading to a second-best world is elevated because he is talking about the consequences of markets over time, in a world that is subject to change and potential improvement, and human welfare is not restricted to the efficiency limits that can be achieved within the confines of static analysis, in which the goods are given and never change, and in which the potential for economic evolutionary development to higher states of potential welfare (beyond the state that can be achieved by an efficient allocation of known resources among known goods) is never imagined. In such a world, impediments to the fluidity of resources, which can form the foundation of market power and deadweight losses, can be welfare enhancing, contrary to what is so easily, and mistakenly, deduced from standard microstatic modeling of markets.
There are at least two reasons that resource impediments can improve welfare: First, they can provide positive incentives for people to get off their proverbial economic dimes and do something different and better. Second, impediments to the movement of resources can offer firms some assurance that they can recover all costs of production, including the costs associated with
* conceiving of goods that satisfy needs and wants,
* creating the products, and
* developing the markets for the goods that are created.
In a world of perfect fluidity of resources, firms that incur such costs can have no hope of recovering them. This is because once a good is shown to be profitable in markets that have already been developed by someone at some cost, competitors will enter, competing the price down to the costs involved in the actual production of the goods. Such a fluid state of resource movement will allow the firms who dare to create the new and improved goods and develop their markets to recover only the costs associated with production, not their product and market development costs.
Surely, as basic reproduction costs shrink and as product and market development costs escalate, the ultimate efficiency of perfect competition must become progressively more suspect. With the emergence of digital goods, or those products and services that can be reproduced with 1s and 0s (or with electrons), the threat to innovation and economic development in perfectly fluid markets for goods in which the marginal cost of production is close to zero (if not zero) has become self-evident.
Patent and copyright laws are legally imposed impediments to the fluidity of resource flows, which highlights a central point that is, to my knowledge, rarely mentioned in discussions of the standard monopolized market--namely, monopolies, at least to some degree and prevalence, can be a positive force for economic development, in spite of any deadweight loss and economic profit that necessarily emerge, by the legally imposed market protections against entry.
If the government has any legitimate concern with Microsoft's influence in the economy, especially over the innovativeness of software and related markets, in my considered opinion, which I have argued at length (McKenzie 2000a), it has little or nothing to do with Microsoft acting as a monopolist, that is, one that seeks to restrict sales and push up its prices and profits. Rather, it has everything to do with the company's potential competitiveness that stems from its ubiquity and its nearly zero cost of entering software markets adjacent to Windows--and hence, the fluidity with which Microsoft can move into adjacent markets. Microsoft can allow a variety of applications to be developed by outsiders and then cherry pick among the profitable applications, incorporating only those applications in Windows that show promise of being profitable. It can then take over the newly developed markets, as it did in the case of the browser market, without incurring all the upfront costs of developing the actual product category and the market for particular products that fill an identified product category. Here, the Microsoft threat stands on par not with OPEC, a cartel of firms that has sought to restrict market supply of a given product, but with all the digital pirates around the world who can sit back and do nothing more than run copying machines, forcing the prices of creative products down to the marginal cost of digital copying, which can be negligible. (10)
However, there are at least two good reasons caution is in order in appraising the critics' claims that Microsoft's competitive aggressiveness in entering adjacent software markets has chilled innovation. First, Microsoft may have ultimately chilled Netscape's interest in developing its browser (beyond Netscape's initial efforts in the late 1990s to fend off Microsoft's entry into the browser market). After all, Netscape's share of the browser market has largely dried up. However, Microsoft has no doubt spurred innovation in other software and Internet markets precisely because its inclusion of Internet Explorer in Windows made web browsing immediately available to practically every computer user. In other words, the Internet may not have developed as rapidly as it did, and to the extent it did, had Microsoft not bundled Internet Explorer with Windows.
Second, the comparison of Microsoft with software pirates is not completely fair. Pirates face rampant prisoner's dilemma problems, meaning that individual pirates can contribute to the destruction of innovators' incentives to produce original works but have no incentive to curb their own piracy activity, simply because they cannot individually have a meaningful impact on the overall level of piracy and, thus, innovation. Microsoft is in a radically different market position, given its market dominance. It needs to curb its impact on innovation if for no other reason than it can feed on the stream of innovations in adjacent software markets. It has an incentive to curb its invasion of adjacent markets, if not stimulate the development of adjacent software products. This isn't to say that there is no Microsoft problem for potential competitors, but it is to say that the problem is not of the same magnitude as the problem of piracy is likely to be. Moreover, efforts to break up Microsoft into competing operating-system companies (as the Justice Department proposed [Klein 1998], and which the District Court Judge Thomas Penfield Jackson accepted in his ruling [Jackson 2000]) could have the exact opposite effect of that intended, namely, a greater chill on innovation in adjacent software markets because the competing software firms would have less incentive to invade adjacent software markets.
4. Failures and Monopoly Profits
In the static model world in which goods are given, normal profits need cover no more than opportunity cost, but little in the way of risk costs. This is because so much of the risk of real-world markets has been precluded by the assumption that the goods are given. They don't have to be developed. Their markets are also given, which means that the markets (and their demands) do not have to be developed. The prospects of failure have also been severely restricted, given that the goods that are produced are indeed goods, which means they have been shown to provide value, given that the demand for the good is drawn into our blackboard graphs.
The world is much different when the goods and their markets are subject to creation and development. In such a world, product failure can be rampant. Successful products can be quickly duplicated, with the result that producers of many products cannot hope to recover their upfront product and market development costs. If there are real losses in such real-world markets, then economic profits--monopoly market positions with attendant monopoly profits--must be not only a prospect but also a reality, at least somewhere in the economy. Economic profits must be a prospect, or else producers of portfolios of products could not hope to offset their expected losses on some products and earn a normal profit over all the products in their portfolios. Hence, I submit to you that Schumpeter was right: The prospects of monopoly positions--and their attendant inefficiencies--are absolutely necessary for a dynamic, growth-oriented, welfare-enhancing market economy.
And if the prospects of monopoly are to carry the requisite level of incentives, then real, working monopolies in the economy--ones that actually curb production to allow monopoly profits to exist--must be a reality. Otherwise, the prospect of monopoly would be so many empty words, devoid of any chance that economic behavior would be affected.
I concede that with some twists and turns in the arguments, an economist could retort: "Well, a single firm intent on devising a portfolio of products--for example, toys--need not be a real monopoly in the way you suggest. After all, the array of products are produced under one organizational roof and the firm, on balance, behaves in accord with the competitive model, earning only normal profits." However, such a retort would miss my broader point. We can back away from an individual firm and consider an investor who invests in an array of firms, knowing full well that one or more of her (or his) investments will be failures. To expect to earn, on balance, nothing more than a normal rate of return on her investments, she must figure that one or more of her investments will acquire a monopoly position to some degree and, accordingly, will earn monopoly profits. This is to say that the reality of some level and prevalence of monopolies can actually improve the facility and efficiency with which resources are deployed across an economy. It is also to say that if the potential for monopoly market positions were totally wiped out by (perfectly) efficient antitrust prosecution, the dynamism of the economy could be undercut simply because investors would then be left with investment opportunities that yield normal profits or less, with their actual average profits earned across an array of investments being less than the much heralded minimum normal profits.
My stronger point can now be stated: When viewed outside the economist's box of static analysis for a given product, even a pure, deadweight-loss-producing monopoly can be welfare enhancing, not so much by what it does in its own market but by its indirect impact across markets. In its own market, it can restrict sales and raise its price and profits. In doing so, it can give rise to the deadweight loss that we cover in our lectures. However, my point is that the net impact of even a pure monopoly cannot be judged and should not be judged the way it almost always is judged in our lectures and textbooks--that is, exclusively by its impact within the limits of its own market boundaries. After all, entrepreneurs throughout the economy will take note of the pure monopolist's economic profits. The fact that those economic profits exist and persist can motivate a host of entrepreneurs to try to acquire the same kind of market positions. Many can be expected to fail in their efforts to acquire monopoly positions, earning perhaps no more than normal profits (or even less). The important point is that even a bad monopolist can give rise to offsetting, if not more than offsetting, welfare gains in the economy (to Dupuit triangles at least, if not the entire consumer-surplus triangle) that I submit are rarely, if ever, mentioned in our blackboard discussions of monopoly harms. (11)
5. The Welfare-Enhancing Force of Product Price
One unheralded source of the perceived harm of monopoly emerges from the dictum that we all teach our beginning economics students: The price of the product affects the quantity demanded, not demand. Market forces other than price affect demand, resulting in an increase or decrease in the underlying value, and potential consumer surplus, of the good. I appreciate the professional insistence on the conveyed rule about price because students easily confuse and misuse demand and quantity demanded. However, we professors should understand that our classroom rules need not always be carried over to our own appraisals of the relative merits of different market structures in which price can indirectly, if not directly, affect demand and the underlying market valuations of the units consumed, which suggests that a monopolist's price increase does more than force the emergence of the Harberger triangle; it can give rise to an increase in the size of the Dupuit triangle. In this address, I will take up four illustrative cases.
The Client Effect
The standard rule regarding how price doesn't affect demand is predicated on an assumption of all consumers being homogeneous. That is hardly reflective of real-world markets in which people who would be willing and able to buy a good only at low prices are often radically different (for example, in terms of income, culture, dress, and tastes for complementary goods) from consumers who would be willing and able to buy the good at higher prices. As Dwight Lee and I have argued in this Association's journal (1998), when price is raised, not only does the level of consumption change, the actual combination of consumers changes for many products. In the case of hotels, the Four Seasons Hotels may raise their prices above the level charged by Motel 6s not only because Four Seasons Hotels provide chocolates on their guests' pillows (and other amenities throughout their hotels) but also because they seek to ration out of the market potentially undesirable guests who would detract from the value of the hotel experience for their targeted guests. This client effect translates not only into greater profits for the hotel chain but also greater demands and larger Dupuit triangles for the guests who are served. Surely any expansion of the Dupuit triangles for guests must be set in contrast to the emergence of any Harberger triangles for the guests who are not served.
Monopoly Profits as a Bond
Those familiar with principal/agency theory have often pointed to how businesses need to develop hostages or bonds that they, the businesses, can offer their customers as assurances that agreed-upon standards of contracted service will be provided. The hostages or bonds are said to make the contracts self-enforcing, which can lower transaction costs, improving the overall efficiency of markets. Organizational theorists have pointed to how brand names and unnecessarily luxurious building facades can provide the needed assurances to buyers: If a firm that has invested in the brand names or luxurious facades fails to meet agreed-upon standards for the products that are delivered, its reputation can suffer, nullifying in part or in total the firm's investment in its brand name and building facade.
Economic profit achieved by monopolies can similarly serve as a firm's hostage or bond. If the monopoly firm performs poorly or abrogates its contracts with its buyers, it can lose its monopoly profits. Hence, the monopoly can add value in the form of assurance to its products and can increase demand and the Dupuit triangle. Moreover, the economic profit can represent the deep pockets attorneys need to pursue suits against a firm that fails to live up to its contracts. Hence, the economic profit that emerges as a monopoly raises its price can do more than curb production and give rise to deadweight loss; it can increase the value of the units that are purchased. Once again, the Dupuit triangle can be expanded, perhaps with this expansion more than offsetting the emergence of any deadweight loss that could be recovered if the competitive output level were sold at the competitive price. Note that perfect competitors have zero opportunity to offer any economic profit as a hostage or bond, but then the good in a perfectly competitive market is obviously very simple, not likely requiring self-enforcing contract assurances, given that virtually anyone can produce it (by the assumptions of numerous producers in the market and zero entry costs for anyone who wants to enter the market).
Monopoly Prices as Signals of Consumer Value
One of the unrecognized problems with prices established in perfectly competitive markets, as stressed by Stanford University economist Paul Romer (1994) in his contributions to the "neo-Schumpeterian growth theory," is that such prices reflect goods' marginal values only and, in turn, are equated only with producers' marginal costs of production. The disincentive producers have to innovate under such circumstances is particularly acute in markets with high fixed costs (relative to marginal costs), as noted, but the incentive problems on the innovation front of perfectly competitive prices do not go away when fixed costs are relative low. This is because the prices do not and cannot reflect the full inframarginal values of the goods that can be created. (12) In static analysis, this observation has no consequences, because the goods are given, meaning they do not have to be created, but the observation can be critically important when the goods must first be created (and/or improved). With prices not reflecting the goods' full market value, entrepreneurs have an impaired incentive to innovate, that is, create the goods that can be then subjected to static analysis. While monopoly prices might give rise to the static deadweight loss, they also allow entrepreneurs to capture more of created goods' full market value, which can mean more goods than otherwise being created. Put another way, perfectly competitive prices can imply static Pareto efficiency but can also result in a form of dynamic inefficiency, or an underallocation of resources to the creation of new products and improvements in old products. If created, such goods could enhance consumer welfare over time by far more than welfare is undercut by monopoly prices at any point in time, given the array of products that are then available.
As Romer (1994) suggests, static competitive analysis assumes away one of the more important problems any economy faces, the appropriate allocation of resources to the extraordinarily complex problem of selecting those goods that will in fact be created from a virtual infinite array of goods that could be created. In highly competitive markets, the high-fixed-cost problem will cause some goods that should be created (because their total value exceeds all costs) to go uncreated altogether, with development resources (mis)allocated toward the (over)creation of goods that do not confront the high-fixed-cost problem. Competitive prices that allow for recovery of fixed costs can still lead to an underallocation of resources in the creation of the most highly valued new goods. And the problem of fixed costs, which are unavoidable when considering the creation of goods, is obviously of greater importance than economists have been willing to admit:
Evidence that fixed costs are important comes from the observation that many services and goods are simply not available at any price in many parts of the world. If there were no fixed costs, one should find that all possible goods, services, and production processes and types of exchange are available to firms located everywhere in the world. (Romer 1994, pp. 24-25) (13)
One of the totally unappreciated benefits of monopoly prices is that they allow for at least a partial correction of this dynamic resource allocation problem over time.
In conventional presentations (and outside the rent-seeking literature), the monopoly model is developed as if the monopoly has no history involving the process by which the firm achieved its monopoly position, and as if any current economic profit is unearned because of any lost tie to the monopolist's past pricing strategy. However, in the case of network goods, or those goods the value of which is determined by how many buyers/users there are, a firm's current pricing strategy has direct and indirect tie-ins to its past pricing strategy. That is to say, a firm's monopoly profits could have been earned by how it developed and priced its product throughout its history.
In the case of the operating-system market in the initial stages of its development, a firm such as Microsoft has two incentives to hold the price of Windows down. The first is customary: to sell more copies of Windows currently. The second is to spur the development of the Windows network. By holding its price down, Microsoft sells more copies of Windows, which adds to the value of Windows for each buyer, because each buyer can then be compatible with more users. In addition, because there are more Windows users, more application developers will be more inclined to write programs for Windows, which in turn can increase the value and demand for Windows. With the prospect of the market tipping in favor of the projected to-be-dominant operating system and with the prospects of the users of the to-be-dominant operating system being locked in, which affords the dominant operating system a measure of monopoly power, a firm such as Microsoft has all the more incentive to lower its initial price, just so it can be the dominant producer with the market power to extract monopoly profits. Indeed, the prospects of monopoly profits in the future can spur Microsoft (and other operating-system firms) to aggressively price its products, charging near-zero and even below-zero prices. Such an initial pricing strategy can result in initial losses, but such losses can be construed as a part of the firms' initial investments to acquire the later monopoly profits. Indeed, the later monopoly profits (construed to be such because the then monopolist is charging above marginal cost) may be nothing more than payback for the initial losses.
If evaluated at the point at which the firm has acquired its market dominance, it may appear that the monopoly (Microsoft) is imposing a deadweight loss on the operating-system market. Why? It is not then selling at the competitive output level at which marginal cost equals price. Indeed, there may be no dispute that the dominant firm is restricting output at the point at which it has acquired market dominance. At the same time, the dominant firm could have easily increased human welfare by its pricing strategy orchestrated throughout the time its market dominance developed. This is because the dominant firm's initial low prices could have not only expanded the size of the network but also increased the speed of the network's development, a result that could have benefits of its own, that is, benefits realized independently of the size of the network. Indeed, the greater the projected economic profits in the future, we might surmise, the more eager to-be-dominant firms would be willing to speed up the development of the network by its initial prices.
6. Locked-In Consumers
It's easy to think of consumers who are locked in to the purchase of a given network good--for example, Windows--as a source of monopoly power and, hence, welfare loss for consumers. In a network market, that is not necessarily the case. The reason is simple: If the network is broken up, the consumers can lose value from having fewer consumers of the network good. In the case of Microsoft, consumers can lose value from having fewer computer users with whom they are compatible. Plus, a breakup of the Windows network can lead to fewer applications being available for Windows.
As Dwight Lee and I (2001) have argued, consumers of a network can have a demand for being locked in and a demand for the dominant producer to make monopoly profits. Consumers can also have an economic interest in the dominant producer aggressively fending off potential competitors by aggressively adding features (browsers, for example) and by engaging in predatory pricing. The very fact that a network producer demonstrates it is willing to defend its network aggressively means that more people can be expected to want to join the network. In Microsoft's case, predatory pricing and nonpricing marketing strategies can increase the array of applications available to Windows users and increase consumer welfare, in spite of any deadweight loss economists might identify at the time Microsoft exerts its dominant position and charges monopoly prices. Such monopoly prices can lead to more consumer welfare left for consumers (the Dupuit triangle) than would have been left had Microsoft not aggressively tried to become the monopolist that it is today.
7. Concluding Comments
I recognize that my arguments are only partially developed. I hope you will find them more fully developed in the book Dwight Lee and I have completed. My effort here has been directed mainly at undermining the facile conclusion readily drawn from standard treatments of the perfectly competitive models--namely, that perfect fluidity in the movement of resources across market is necessarily some economic ideal and that any market structure that is dependent on rigidities in the flow of resources always detracts in some way and to some extent (as identified by the familiar Harberger triangle) from human welfare. Maybe that is the case in the narrow context of our static models in which the goods are given. But the world most of us live in, and want to live in, is one in which goods are not given. They, and the markets for them, are subject to creation and improvement. In this larger real-world context, Joseph Schumpeter had a point that is as ironic as it is worth remembering and teaching: that a system filled with static inefficiencies at every point in time can be superior, in the long run (up to a point) to a system perfectly efficient at every point in time.
(1) Adam Smith noted that
A monopoly granted either to an individual or to a trading company has the same effect as a secret in trade or manufactures. The monopolist, by keeping the market constantly understocked by never fully selling the effectual demand, sells its commodities much above the natural price, and raise their emoluments, whether they consist of wages or profit, greatly above their natural rate. (1904, p. 61)
(2) Perhaps to be completely fair to conventional theory of the firm, there may be some agency costs embedded in the cost structure of even perfectly competitive firms. However, my essential point stands that under perfect competition agency costs can be no higher than the absolute minimum attainable by the very best control mechanism, which suggests that any other market structure must have higher agency costs if for no other reason than the threat of removal by real or would-be competitors is less severe.
(3) If they could (at little or no cost), consumers would counter the monopolist's market power by colluding with the intent of taking one or some combination of the following collective actions: (i) buying off the monopolist, which would involve consumers paying the monopolist to expand output toward the competitive level (with the deadweight-loss triangle ensuring that the payment could result in a net gain to both consumers and the monopolist); (ii) setting a maximum price consumers would pay the monopolist equal to the competitive level, which would induce the monopolist to expand production to the competitive output level; or (iii) agreeing to suppress resale markets for the monopoly product in order that the monopolist might price discriminate, leaving consumers with more output and, perhaps, greater consumer surplus.
(4) Conventionally, the power of a monopolist to extract monopoly rents is a consequence of barriers to entry. However, following Olson (1965), a monopolist's power to extract rents is the result of the inability of large groups of consumers to band together to pursue their common interest of expanding output and consumer surplus. From this perspective, the power of monopoly could be said to emanate from consumers' collective decision-making costs.
(5) For the details of how standard monopoly theory has been at the foundation of the legal case of Microsoft, consider the Justice Department's original complaint (Klein 1998).
(6) As Reynolds (1999) and I (2000a) suggest, Microsoft's market share is subject to considerable debate.
(7) The appeals court unanimously upheld "the District Court's finding of monopoly power in its entirety" (U.S. Court of Appeals for the District of Columbia Circuit 2001, p. 15), drawing that conclusion mainly on this determination:
That barrier--the "applications barrier to entry"--stems from two characteristics of the software market: (1) most consumers prefer operating systems for which a large number of applications have already been written; and (2) most developers prefer to write for operating systems that already have a substantial consumer base.... This "chicken-and-egg" situation ensures that applications will continue to be written for the already dominant Windows, which in turn ensures that consumers will continue to prefer it over other operating systems. (2001, p. 20)
Again, as is the case with so much of the government's case and courts' rulings against Microsoft, the applications barrier to entry has been misunderstood and grossly exaggerated (McKenzie 2000b).
(8) On the other hand, survey after survey throughout the three years of court proceedings found that a substantial number of computer users--upwards of three-fourths--have favored the Microsoft side in the ongoing legal debate and would have liked nothing better than for the Justice Department and the 19 state attorneys general pursuing the case to leave Microsoft alone. For example, Americans for Technology Leadership (Van Lohuizen 2001), an advocacy trade association for a number of technology firms, but mainly financed by Microsoft, found just after the appeals court handed down its decision in late June 2001 that
* 84% of the 500 registered voters polled believed Microsoft had benefited consumers, whereas only 8% believed that Microsoft had harmed consumers;
* 74% disapproved of preventing the shipment of Windows XP, which would, at the time of the poll, be shipped with additional integrated applications, and 57% of the respondents felt "strongly" about this point;
* 68% felt that Microsoft's competitors, not consumers, were benefiting most from the case; and
* 78% said that Microsoft's competitors should focus more time innovating and less time litigating.
Moreover, 77% of those polled had a "very favorable" or "somewhat favorable" impression of Microsoft, whereas only 12% had a "somewhat unfavorable" or "very unfavorable" impression of Microsoft. Those ratings compare very favorably with the ratings received by AOL-Time Warner (58%/21%), Sun Microsystems (31%/3%), and Oracle (25%/4%). Only IBM among the four alternative firms covered in the survey had a better rating (80%/5%).
(9) The Dupuit mangle was named for Jules Dupuit, who, in an 1844 article, was the first to recognize its importance.
(10) Of course, Microsoft may not have actually pirated (illegally copied) any of its software, such as Internet Explorer. It did not need to do so. All it needed was to develop a similar product that was sufficiently different to avoid being accused of violating anyone's copyright. The point is that it didn't have to incur the costs associated with identifying the need, developing the product category that would satisfy that need, and developing the market for the product category. It can avoid many of the risk costs associated with developing an array of products that are believed a priori to meet a need or want, many of which prove to be failures.
(11) Of course, many entrepreneurs will seek to become monopolistic in a tried and true way, going to government for the required market protections. I am certainly aware of this problem, which is one good reason for restricting government's ability to establish and protect monopolies. Having said that, however, my line of argument does not preclude government-established and -protected monopolies from, on balance, being welfare enhancing.
(12) In Romer's words, "In competitive markets, prices work at the margin. If good Z already existed, then prices that are equal to marginal cost give the right signals about how much of Z to use in this production process. But these prices do not attach the correct overall value to the associated bundle of goods, and cannot be used as a guide in the decision about whether or not to incur a cost and invent good Z" (1994, p. 16).
(13) Romer (1994, p. 25) then points out that direct evidence of the importance of fixed cost is available from Teece (1977). To assess Romer's point with greater clarity, consider the extreme case of all costs being upfront costs, incurred by the development of a software program. The marginal cost is zero, which would imply that in a perfectly competitive market the efficiency-maximizing price should be zero. However, with a prospective zero price, the good would not be created and made available by rational entrepreneurs. A monopoly price would be a better signal of the inframarginal value of the program, thus increasing the likelihood that the good would be conceived, created, and made available to consumers. Under competitive pricing, consumers would get no consumer surplus. Under monopoly pricing, they would at least get the Dupuit triangle.
Acton, Lord. 1904. Letters of Lord Acton to Mary Gladstone. Edited by Herbert Paul. London, UK: George Allen.
Baumol, W. J., and J. Ordover. 1985. Use of antitrust to subvert competition. Journal of Law and Economics 28:247-65.
Buchanan, J. M. 1964. What should economists do? Southern Economic Journal 30:213-22. Reprinted in What should economists do? by James M. Buchanan. 1979. Indianapolis, IN: Liberty Press, pp. 17-38.
Jackson, T. P. 2000. "First District Court, civil action no. 98-1232, IB, Conclusions of Law, May 5." United States v. Microsoft Corporation. Accessed 8 January 2004. Available http://www.usdoj.gov/atr/cases/f4400/4469.htm.
Jackson, Thomas Penfield. 1999. U.S. District Court, Finding of Fact, United States v. Microsoft Corporation, Civil Action 98-1232, November 5, 1999. Accessed 30 January 2004. Available http://www.usdoj.gov/atr/cases/f3800/msjudgex.htm.
Klein, J. I., et al. 1998. United States v. Microsoft. Complaint, First District Court, civil action no. 98-1232, May 20.
Lee, Dwight R., and Richard B. McKenzie. 1998. How the client effect moderates price competition. Southern Economic Journal 64:741-52.
Lee, Dwight R., and Richard B. McKenzie. 2001. A case for letting a firm take advantage of network effects and 'locked-in' customers. Hastings Law Journal 52:795-812.
McKenzie, Richard B. 2000a. Trust on trial: How the Microsoft case is transforming the rules of competition. Boston, MA: Perseus Books.
McKenzie, Richard B. 2000b. Microsoft's 'applications barrier to entry': The case of the missing 70,000 programs. In Policy briefing. Washington, DC: Cato Institute.
McKenzie, Richard B. 2004. The legal fit of "The Microsoft Problem": Antitrust or copyright? University of California, Irvine, Graduate School of Management Working Paper.
McKenzie, Richard B., and Dwight R. Lee. 2000. Monopoly as a coordination problem. In Public choice essays in honor of a maverick scholar: Gordon Tullock. Boston, MA: Kluwer Academic Publishers, pp. 125-36.
McKenzie, Richard B., and Dwight R. Lee. 2004. Monopoly: Market bane or boon? Unpublished book, Graduate School of Management, University of California, Irvine.
Olson, Mancur. 1965. The logic of collective action: Public goods and the theory of groups. Cambridge, UK: Cambridge University Press.
Reynolds, A. 1999. U.S. vs. Microsoft: The monopoly myth. The Wall Street Journal, 9 April, p. A12.
Romer, Paul M. 1994. New goods, old theory and the welfare costs of trade restrictions. Journal of Economic Development 43:5-38.
Schumpeter, J. A. 1942. Capitalism, socialism and democracy. New York: Harper.
Smith, Adam. 1904. An Inquiry into the Nature and Causes of the Wealth of Nations. London, UK: Methuen and Co., Ltd. Edited by Edwin Carman. Library of Economics and Liberty. Accessed 14 January 2003. Available at hap://www.econlib.org/ library/Smith/smWN0.html.
Teece, David. 1977. Technology transfers by multinational firms. Economic Journal 87:242-61.
U.S. Court of Appeals for the District of Columbia Circuit. 2001. United States of America v. Microsoft Corporation, no. 005212 (consolidated with no. 00-5213). Accessed 13 November 2003. Available at http://www.microsoft.com/presspass/ trial/appeals/06-28opinion.asp.
Van Lohuizen, Jan. "Voter/Consumer Research." Americans for Technology Leadership. Accessed on 3 July 2001. Available http://www.techleadership.org/LatestNews.asp?FormMode=Resease&lD=34. Accessed 8 July 2001. Available hap:// techleadership.org/toplines.pdf.
Wilke, John R. 2003. Judges question U.S. settlement with Microsoft. Wall Street Journal, 5 November, p. A2, A10.
Richard B. McKenzie is the Walter B. Gerken Professor of Enterprise and Society in the Graduate School of Management at the University of California, Irvine. This paper was prepared as his presidential address at the annual meeting of the Southern Economics Association, held in San Antonio, Texas, November 22, 2003.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||2003 Presidential Address|
|Author:||McKenzie, Richard B.|
|Publication:||Southern Economic Journal|
|Date:||Apr 1, 2004|
|Next Article:||Beneficial inequality in the provision of municipal services: why rich neighborhoods should get plowed first.|