Contestability vs. competition - once more.
"Contestability" theory was advanced in 1982 by Baumol, Panzar, and Willig as a new theory of industrial organization (see Baumol 1982; Baumol, Panzar, and Willig 1982). It was candidly offered as an effort to overturn the core concepts of the field, replacing competition as an organizing principle with contestability.
This idea and the Baumol group's claims for it soon stirred a critical rejection, on a number of grounds.(1) But the Baumol group were unyielding on many of the points. They have continued to claim favorable properties of contestability which others have disproven (see Baumol and Willig 1986; Baumol, Panzar, and Willig 1988; Baumol 1990, 1994; and Baumol and Sidak 1994).
Also, the Baumol-group authors sought to apply their theory widely in the policy arena. They testified in a variety of settings in efforts to change antitrust and regulatory policies. In forums ranging from antitrust cases in the U.S. Courts, cases before federal regulatory commissions and many state regulatory commissions, to British, Australian and other policy arenas, they have used their "new and superior" theory to press officials toward deregulation and laissez-faire in actual markets.(2) The contestability idea is also linked with their "sustainability" and "Ramsey pricing" theories in order to defend and justify strategic price discrimination by dominant firms. Their testimony continues to be active, and the contestability idea has been taken up by other witnesses representing dominant firms such as AT&T.
After more than a decade of this variety of advocacy, it is appropriate to assess the Baumol group's effort to reorient the research field of Industrial Organization and its related policies. The stakes are large. The policies deal with the core reality of competitive free-market capitalist systems, and the authors have urged their ideas in many settings and sectors. They call for radically different policies. In particular, they assert that monopoly is an innocent phenomenon, which will have no harmful effects. That is because "perfect contestability" will, they say, nullify the effects of monopoly. Seen in perspective, in fact, the theory is just another item in the recurrent arguments favoring a cutback to minimal antitrust policies.(3)
How does this theory currently stand, both in the research literature and as a policy guide? I will try to answer both questions. In Section II, I will suggest that "contestability," as a theory, has been mainly a detour. As for policies, I will review two case studies (the Santa Fe/Southern Pacific railroad merger and long-distance telephone service), where the theory has been invoked in belittling market power.
A fair conclusion would still be what I suggested in 1984: The theory is internally inconsistent, difficult to relate to reality, and hazardous for policy treatments of market power. The Baumol group chose not to answer my numerous criticisms and have offered no significant empirical support, while promoting their idea as a guide for real cases. Of course, a balanced attention to entry conditions can be a helpful element of careful policies, but that was known before the Baumol group launched the contestability idea.(4)
II. A WORD OF HISTORY: CONTESTABILITY AS A DEVIATION FROM THE MAINSTREAM
Because contestability theory has attempted to displace the mainstream field, I must begin by briefly reviewing the evolution of the mainstream field, so that contestability's place and value can be seen.
A. Evolution of the Mainstream Field, Including Potential Competition
The modern field of Industrial Organization began in the 1880s, amid lively controversy (see Shepherd 1990a, ch. 1, appendix A; Scherer and Ross 1990). Henry C. Adams, Richard T. Ely, John B. Clark, and others debated the various elements of competition and monopoly, correctly focusing on market dominance as the main problem. Clark (1887, 45-61) also advanced the idea of potential competition in 1887, but he then let it lapse until the 1920s.
During 1906-1920 the first wave of antitrust challenges to dominant firms (e.g., Standard Oil, American Tobacco) yielded extensive information about the forms and effects of dominance in important industrial markets. During the 1900-1930 period, the development of utility regulation also led to further practical information about "natural monopolies."
Oligopoly theory flowered in the 1930s, along with an extensive development of empirical evidence about industrial concentration. The 1950s and 1960s further developed quantitative research into industrial concentration and its possible causes, and in 1956 Bain (1956) reintroduced barriers to entry as a significant element of market structure.
Bain's extensive work on entry barriers eventually encountered a sharp irony. He regarded barriers - and established them, with extensive research - as a strong reinforcing factor, which would make concentration even more harmful to efficiency. But later authors (such as the Baumol group) soon reversed Bain's emphasis and values. They focused on the possibility of free entry rather than on barriers, and they used possible free entry to urge that dominance and concentration have little or no effect.
Meanwhile, antitrust policies during the 1950s and 1960s explored the proper limits of efficient merger policies. But the concurrent antitrust challenges to dominance were no more than mild during the 1960s, despite frequent exaggerations about those actions.(5) And utility regulation - especially of transport, electricity, and telephone service - was recognized to have been largely passive and often ineffective.
With all this empirical, theoretical and policy diversity, the field of industrial organization had by 1970 developed a healthy variety and balance. The elements of market structure, including market shares, concentration and entry conditions, were being explored, along with economies of scale and other possible determinants of structure and performance (Shepherd 1990a; Scherer and Ross 1990).
Yet in the 1970s there arose a drumfire of claims by conservative lawyers and economists that the 1960s had brought extreme antitrust policies toward concentration. Also, the Bell System came under antitrust challenge in 1974, which ultimately led in 1984 to its divestiture. AT&T employed the economists that later became known as the Baumol group, at least partly in order to formulate ideas which were used to resist the antitrust suit and later to resist entry into its markets. In the special setting of the 1970s and 1980s, the Baumol group could claim that its ideas were merely professional expressions seeking to reach a proper balance, rather than to advance dubious claims.
B. Origins and Content of Contestability Theory
The idea and analytical forms of contestability theory were created by Baumol and various younger colleagues (including Panzar, Willig, and Elizabeth E. Bailey). Their employer (the American Telephone & Telegraph Company, then also known as the Bell System), was the owner of the massive and complex core telephone system in the U.S. Contestability theory - particularly as it was framed by the Baumol group - tended directly to advance the interests of this company. Born in this setting of forensic, policy-related work, the theory has been used with unusual vigor by these same economists and others in order to advance dominant-firm interests in a variety of other regulatory and antitrust cases.
"Contestability" theory was first presented in 1982 by Baumol (1982) and then by Baumol, Panzar, and Willig (1982) later in the year, with an unusual degree of self-congratulation and promotion. The idea of perfect contestability, they said, "will transform the field and render it far more applicable to the real world" and "be extraordinarily helpful in the design of public policy" (Baumol, Panzar, and Willig 1982, xiii, xxii). "We hope to provide a unifying framework for a pure theory of industrial organization where none was available before" (p. 3).
Because these entry concepts were more general, they said, they rendered obsolete the mainstream concepts of competition. In the future, economists wishing to judge whether a market contained any market power would need to consider only the condition of entry - whether the market was perfectly contestable (that is, it had what might be called "super-free" entry). If entry were super-free, then all dimensions of an efficient competitive outcome would always be achieved, as by the wave of a magic wand.
Even if the market contained a pure monopolist, or a dominant firm, or a group of tight oligopolists which held most of a highly concentrated market, those degrees of monopoly would now be irrelevant. Under the conditions of contestability, all of these market-dominating firms would be compelled by the fear of new entry to conform precisely to the pure competitive outcome, holding their prices strictly down to costs. Contestability would nullify all market power and yield optimality.
Moreover, all market shares that actually emerged and remained in such a market would be justified by superior costs. No excess market shares (that is, market shares larger than the levels that were dictated by economies of scale) would occur. Accordingly, the contestability doctrine reinforced the "efficient-structure" hypothesis, which asserts that whatever market structure actually exists is also the normative best (Shepherd 1988a). Any seeming market dominance would reflect only the existence of superior efficiency, as well as creating no market power.
The theory relieved all market dominance, even pure monopoly, of the need to defend itself. Not even hard-line Chicagoans had been willing to declare pure monopoly to be virtuous, rather than probably harmful.
It seemed to be a miraculous theory, and it turned out like many supposed miracles to involve certain illusions. The idea soon met the usual fate of new ideas; it was debated, whittled down to reality, and has eventually taken a modest place in the pure theory of markets. A series of skeptical articles quickly established that the contestability approach was more like a oddity than a general theory (see Schwartz 1986). The assumptions underlying it were shown to be not only impossibly restrictive but also mutually inconsistent (Shepherd 1984). The theory was shown to be not robust when there were small deviations from its strict assumptions (Schwartz 1986).
By 1984 Bailey and Baumol (1984) conceded that the airline industry did not fit the theory, even though that industry had been radically deregulated (and monopoly-creating mergers were later permitted) on the claim that it was a leading example of perfect contestability. By 1986 Baumol and Willig (1986) further admitted that contestability was mainly a source of interesting theoretical insights. The main empirical future they foresaw for the theory was in trying out the theory in experimental game situations. The Baumol group, and other advocates of the idea, have been unable to identify any significant real markets that could be said to be reasonably "contestable," and even those instances departed from some of the assumptions.
In 1984 I noted in some detail that the theory was internally inconsistent in two of its fundamental assumptions. The theory is based on three highly restrictive assumptions:
1. Entry is free, without limit. When a new firm is induced to enter by the incumbent's small excess in price, the entrant can do far more than gain a foothold, as conventional free entry theory envisages. The entrant can instantly duplicate and entirely replace any existing firm, even a complete monopolist. There are no costs or significant lags in entry, and the entrant can match all dimensions of size, technology, costs, product variety, brand loyalties, and other advantages of all existing firms.
2. Entry is absolute. The entrant can establish itself before an existing firm makes any price response or even threatens to do so. That is because the incumbent regards the entrant as too small to matter.
3. Entry is perfectly reversible. Exit is perfectly free, at no sacrifice of any cost. Sunk cost is strictly zero. The newcomer can get into the market (even to replace the monopolist entirely) and then get out, without incurring any sunk cost whatever.
The impressive deductive results hold only when these pure conditions exist. But they were soon recognized in the literature as being too strict and not observable in any markets - other than perhaps in markets that were already effectively competitive. In those extremely competitive markets, a newcomer might indeed fully replace another small competitor. Yet competition would already be completely effective, so the contestability conditions would be irrelevant to social performance.
The mutual inconsistency is between assumptions 1 and 2. Number 1 assumes that entry is total; the newcomer can oust the incumbent completely and immediately. But number 2 assumes that the incumbent regards the entrant as only negligibly small, with no significant market share. That is necessary for the Bertrand-Nash outcome, for which the incumbent must not bother to react at all.
The two assumptions are diametric opposites. Entry cannot be both trivial and total. If entry is trivial, it has no force. If it is total (or large), the incumbent will respond, with special force where it faces total elimination. That inconsistency is fatal to the claim that the theory rigorously provides normative conclusions about competitive outcomes and efficiency. Although I stressed this flaw in 1984, the Baumol group have offered no answer to it. To my knowledge, in fact, they have not acknowledged the existence of this fatal inconsistency in any discussions.
A further weakness of the theory is to misplace reality, by claiming that potential competition has primacy over actual competition. Contestability theorists assume that entry has infinite force, superseding all rivals or conditions inside the market. That is contrary to reasonable theory and to the overwhelming weight of business experience.
Moreover, entry barriers are a more formidable and widespread problem than the Baumol group has recognized. I have assembled (Shepherd 1990a, 1990b) some 14 sources of entry barriers which the literature has identified. They derive both from "exogenous" causes (that is, basic conditions such as technology) and "endogenous" conditions (that is, voluntary actions taken by the incumbent firms so as to make entry harder). Perhaps the central and most general cause is the customer base held by an incumbent dominant firm. That base must be won away by the entrant, but the incumbent often can defend it fiercely. Many customers are also inert or habitual, rather than quick to switch.
The Baumol group also offered no persuasive rebuttal to the showing (Schwartz 1986) that the theory is not robust. That lack of robustness negates the claim that the theory is more general than the long-established theory of competition. Competition theory has been known for many decades to be robust. And we will see below that Baumol et al. commit a simple error when they pose perfect competition as the only version of effective competition.
Therefore many leading economists soon recognized that contestability was a fragile and limited theory, which adds little to the already rich and growing literature on entry conditions. And they saw contestability's practical applicability to be slight, since no significant real markets that were contestable could be found.
The Baumol group would admit in informal settings that the analysis was meant merely as an "exercise," not pretending to deal with real sectors. The diametrically conflicting assumptions were needed only to make the analysis "tractable" and to give "insights."
Yet the Baumol group have also continued to claim that the theory is powerful, general, and superior to concepts based on competition and monopoly. They and others have testified vigorously in many real antitrust and regulatory cases, invoking the theory. Moreover, the theory has become something of a fixture in textbooks of the "new Industrial Organization theory," and it is routinely taught in many leading departments as part of the field of Industrial Organization (see, e.g., Tirole 1991).
III. RETHINKING THE THEORY
The use of contestability theory to guide policies actually rests on a long and highly specific sequence of ideas about competition and public policies. I will need to review that sequence, in order to place the contestability idea in a correct context for appraisal. Recently Baumol has coauthored a book (Baumol and Sidak 1994, esp. ch. 3)(6) which restates his current assessment of the theory's value. Fortunately, Baumol (the senior author) and Sidak clearly present these ideas, so that by retracing their steps in the following pages I can present the concepts in a clear sequence. I will number the ideas in the sequence, for clarity.
A. The Sole Criterion of Good Economic Performance and the Public Interest is Pareto Static Efficiency
Baumol focuses exclusively on economic efficiency as the objective for good economic performance and the public interest (pp. 25-27). That criterion reduces to static allocative efficiency, which provides a maximization of static consumer and producer surplus. The maximizing of consumer and producer surplus is therefore Baumol's exclusive criterion to guide public policy.
Unfortunately, the efficiency criterion excludes the process of innovation, which may be the most important single element of economic performance. Although static allocative efficiency may be formulated precisely in formal theoretical terms and it is surely a significant goal, it has also been recognized as a narrow analytical matter. Even the concepts of consumer and producer surplus themselves are highly controversial, and there have been few successful efforts actually to measure them in real policy situations.(7) Baumol's single-objective choice also excludes other meaningful objectives, including fairness, diversity, security, and competition itself as an economic and social objective.
Baumol claims that "static welfare maximization" as the exclusive standard is a "powerful analytical tool" and is almost universally accepted. But, again, that ignores major theoretical and empirical problems. It is acceptable mainly as a theorist's device, not as a practical guide.
B. Competition is the Standard for Obtaining Efficiency
Baumol then further narrows the topic by saying that "the proper task of economic regulation is to intervene where competitive forces are too weak to defend the public interest standard" (p. 27). Baumol notes that perfect competition yields static allocative efficiency, and so this pure and extreme version of competition becomes his standard.
Yet this is an odd and inappropriate forcing of competition as the mechanical standard. Regulation - economic or otherwise - may have many "proper tasks," such as the promotion of rapid innovation and the pursuit of fair outcomes. Baumol has offered only one, namely static efficiency. It is a position based only on his declaration. Conservative economists may also take that view, but it is a matter of preference, not of deductive validity.
Baumol also asserts that competition guarantees fairness in distribution, by preventing monopoly profits. Fairness may indeed result from competition. But competition may instead generate unfair outcomes if the prior situation is unfair. Competition tends to perpetuate the existing structure of distribution, rather than to guarantee fairness.
C. "Competition" Means Only the Pure Version of Perfect Competition
Baumol then narrows competition down exclusively to identify it with the old-fashioned 1930s-vintage model of atomistic perfect competition. To him, effective competition must involve "a large number of firms, each of which is too small to influence prices" (p. 31). Also, "by its nature, competitive equilibrium entails local constant returns to scale" (pp. 31-32).
That preoccupation with the pure and extreme case of competition is not really acceptable in a serious discussion of the nature and role of competition in real markets. Baumol's view reflects a narrow 1930s price-theory-textbook notion of competition. It ignores the rich literature in the field of industrial organization about the nature of effective competition.
D. Competition is Actually Unsuitable as a Standard, After All
Then, paradoxically, Baumol's exclusive focus on perfect competition leads him to reject competition as a standard, after all. That curious step becomes appropriate for Baumol only because - in his mistaken assumption - perfect competition is the only way to define competition.
I doubt that other observers or analysts would limit attention to this pure case as the sole basic guide. In fact, the theories and conditions of competition offer a thoroughly robust and appropriate standard for real markets. I have noted elsewhere (Shepherd 1990a) that effective competition requires only: (1) about five to eight comparable competitive firms, able to apply strong mutual pressure toward excellent performance, and (2) reasonably free entry of new competitors. Even if those conditions are not fully met, there will usually be enough competition to provide most of the efficiency and other benefits of full competition.
E. All Industries that are Regulated are Natural Monopolies
Baumol then claims that regulation applies exclusively to industries with large-scale economies. Such natural monopolies may indeed be regulated, but it is something of an archaic 1930s idea that regulation is limited in this way. In fact, the really interesting issues arise mainly in quite different situations, where the receding of scale economies has opened up possibilities for effective competition.
By restricting himself with the pure natural-monopoly assumption, Baumol creates a false apparent conflict between natural monopoly and perfect competition. By his odd approach, he comes to rule out all forms of competition as a standard, because - as he asserts - competition has to include a "multiplicity of miniscule firms" (p. 31). That is a simple error, as I have noted above. Other versions of competition generate effectively efficient and progressive outcomes without getting ensnared in the 1930s schoolbook assumption of atomistic competition.
In addition, Baumol claims "that perfect competition is unrealistic" (p. 31), not only because its firms are "miniscule" but because it sets price equals marginal cost as the criterion. The price-equals-marginal-cost goal cannot be relevant for regulation, he asserts, because - once again - he assumes that all regulated firms are natural monopolies, with scale economies that are vast. Because of this debatable assumption, Baumol asserts that any marginal-cost pricing would invariably force the firm into a financial deficit. So he summarily drops the fundamental pricing efficiency guideline.
Actually, Baumol's natural-monopoly view just accepts the old 1930s idea about declining-cost industries, without any independent verification. The unfortunate result is to place current policies in a straitjacket. Baumol may believe that those extreme ideas apply to current conditions, but neither here nor elsewhere has he provided any empirical basis for placing all "regulated" industries in the declining-cost box.
In fact, the interesting questions arise precisely where costs are not strictly declining-cost but involve intermediate degrees of scale economies. But by his series of assumptions, Baumol blocks any such realistic evaluation. Instead, he puts himself in the odd posture of first advancing competition as the exclusive standard and then rejecting it.
This can be traced directly to his strictly theoretician's approach, which places complex situations into simple deductive boxes (and perhaps obsolete boxes at that): static efficiency only, perfect competition only, and total natural monopoly only. Nowhere does Baumol offer significant empirical evidence for his assertions nor does he discuss the interesting intermediate situations that others recognize as the heart of the problem. Those not impressed by his eminence and categorical confidence might wonder how he could advance these points so vigorously.
F. Ramsey Prices are a Valid and Superior Standard
Baumol then digresses in order to advance "Ramsey prices" as the complete answer to the supposed need for efficient pricing rules. The need exists because, by Baumol's assumption, the industries are natural monopolies and marginal-cost pricing is entirely dysfunctional. About Ramsey pricing, Baumol claims that the "validity of its arguments seems to command universal acceptance among economists" (p. 36). This is another instance of Baumol's occasional practice of asserting validity by claiming widespread authoritative support (perhaps the most extreme example being Baumol 1982).
Ramsey pricing is unfortunately not actually so helpful, because it offers little or nothing that is new; it is little more than a name for price discrimination, which has been well studied for many decades.(8) By merely labeling price discrimination after an obscure, long-deceased economist, Baumol essays to invest it with an aura of special meaning. Instead, price discrimination has long been recognized as a common activity, which may intensify monopoly exploitation and have strategic anticompetitive effects. And Baumol long advocated the "inverse elasticity rule" for natural monopoly pricing, before labeling it as "Ramsey pricing."
Baumol's explanation of Ramsey pricing seems to amount mainly to an apologia for monopoly price discrimination. He claims that "regulators have accepted the usefulness of Ramsey theory," a claim which is open to lively debate. No doubt some have accepted it, but Baumol provides no basis for establishing his assertion.
But then Baumol admits that, after all, Ramsey theory cannot be applied in practice! In a bit of euphemistic wording, Baumol says that Ramsey theory is accepted "as a source of general qualitative guidance rather than as a generator of precise and definitive prescriptions for pricing" (p. 39). A less sympathetic interpretation might be that the theory serves as a theoretical justification for price discrimination, even though it has limited practical use.
I have noted the limitations of Ramsey prices elsewhere (Shepherd 1992), but I note here that the theory is less valid and useful than Baumol claims. The idea of Ramsey pricing might have value only if there are absolute natural monopoly conditions, and only if the total profits of the firm can be rigidly constrained to competitive rates. But that attempt to constrain profits puts the task squarely back into the complex policy effort to regulate profits.
Those two conditions leave the theory with only a limited set of possible relevant industries. And, as Baumol notes, the theory gives little practical guidance to the selection of actual prices. But Baumol does not note that price discrimination can also be used by dominant firms in highly strategic ways so as to reduce competition (Shepherd 1990a; Scherer and Ross 1990).
G. Perfect Contestability is the Superior and Conclusive Guideline for Regulators
Baumol then arrives at his "second alternative" to the use of the perfect-competition model as a guide for regulation: recall his claim that perfect competition is unacceptable and Ramsey prices, though valid and ideal, are hard to measure.
He first disarms the reader by saying that "A perfectly contestable market is a fictional ideal, no more to be found in reality than a market that is perfectly competitive" (p. 43). But that comparison is deceptive, as I have noted, because perfect competition is not the only alternative for defining competition. Instead, there is an entirely satisfactory competitive standard: effective competition, rather than the pure theorist's textbook idea of perfect competition.
Continuing this flawed line of comparison, Baumol then claims that "Perfect contestability is a generalization of perfect competition, since both require the complete absence of barriers to entry" (p. 43). In fact, as I noted above (and 11 years ago), perfect contestability requires three extreme and inconsistent assumptions in order to theorize about super-free entry. Perfect competition has not required those assumptions, and the more practical concept of effective competition does not require them.
H. Barriers are Merely a Matter of Sunk Costs
Moreover, Baumol advances an oversimple notion for defining barriers: "any costs that must be borne by an entrant but not by an incumbent" (p. 43). That may apply to some situations, but it is not an adequate description of many forms of entry barriers. It would displace the whole rich literature on the causes of entry barriers with a theorist's simple definition of costs. Once again, Baumol offers a simple and absolute term, in place of a satisfactory discussion of complex phenomena.
I. Contestability is a Robust Theory
Because competition is an exceedingly robust theory, contestability would have to be equally or more robust in order to displace it. In fact, the literature has established (Schwartz 1986) that contestability is not a robust theory. The Baumol group has offered no effective rebuttal, nor could they.
The required assumptions are exceedingly strict, and any departure from them would set aside the efficient outcomes. By comparison, competition is a very robust standard. Even when there are major departures from the assumptions of strict competition, the remaining competition (e.g., among five, six, or more firms) usually generates most of the benefits of fully effective competition.
Neither Baumol nor his coauthors have discussed competition in these terms, and so he can provide no basis for offering comparative assertions about the robustness of perfect contestability. Yet he still asserts that it is robust. Indeed, he occasionally seems unaware of the realities of his theory and of the issues that he is dealing with: "It is sometimes said that the theory of contestable markets is a controversial subject. But that is not really so" (p. 45).
That claim cannot be sustained. Baumol
1. seeks to place a flawed, untested, and largely unverifiable theory at the center of complex, high-stakes economic issues;
2. makes strong claims about the theory's practical lessons, while also admitting that the theory is a "fictional ideal"; and then
3. says that the theory is not controversial.
The reader may be pardoned for feeling baffled by this mixture of ideas.
J. Contestability Offers Definitive Lessons
As just noted, Baumol declares that contestability yields a series of powerful lessons, but they are out of proportion to the notion's defects. The lessons assume away all real-market conditions; and yet Baumol claims that they are entirely superior to competitive criteria. "Thus the concept escapes the impediments to applicability to regulatory issues that are the shortcomings of perfect competition in this arena. . . . a perfectly contestable market can serve as a model for regulation, because it offers all the guarantees of socially beneficial performance that perfect competition brings" (p. 43).
Here, particularly, a confusion between abstraction and reality is evident. He correctly admits that contestability is a "fictional ideal," and yet he declares that this theory is supreme as a guide for actual cases.
This degree of illusion recurs in the Baumol group's use of their ideas about contestability, pricing, and the related regulatory and antitrust implications. They may admit engagingly at professional meetings that the theories are unrealistic, intended only for "insights." But then they often appear as sponsored witnesses at regulatory and antitrust hearings, speaking emphatically about the theory's powerful lessons. Contestability and the related theories, they say, conclusively prove that monopoly firms actually have no market power, and that price discrimination is never anticompetitive but only "efficiency promoting." And they frequently say that all reputable economists agree with them, and that those who may disagree are merely obsolete or unable to understand the correct theories.
So the contestability notion persists among exaggerated claims and militant forensic attacks.
IV. POLICY USES: A RAILROAD MERGER AND LONG-DISTANCE TELEPHONE SERVICE
Contestability theory may divide the profession and lack content; yet it has become extremely valuable to certain companies and their witnesses as a weapon to use against regulation and antitrust.
Although some regulators have been unmoved, others have not. State-level regulatory officials have been perhaps the most vulnerable. In the 1980s they rapidly removed regulation from many important utility markets across the United States, in the hope that contestability would nullify all monopoly power held by entrenched dominant firms. And in case after case, witnesses still advance claims about the conclusive power of entry, as "shown by" contestability theory.
I will briefly note two practical situations that have reflected the theory's uses and misuses.
A. The Proposed Merger between the Santa Fe and Southern Pacific Railroads
In the early 1980s, the Santa Fe railroad sought to merge with the Southern Pacific railroad (see Shepherd 1988b). They were the only railroads in the "southern corridor" between Texas and southern California, and so the merger would create a railroad monopoly. Of course, a true monopoly would not result in the corridor, because trucking already provided important competition for many categories of freight. But for about half of the dollar volume of freight carried on the two railroads, trucking had distinctly higher costs than railroads. That freight included low-value, high-bulk, uniform freight (e.g., coal, cement, minerals) that railroads could carry at low costs. Therefore, trucking competition would not provide direct competition to protect the shippers of that freight.
Yet Baumol argued (1984) that contestability theory proved that the merger would cause no increase whatever in market power. He portrayed trucking as providing complete new entry against any effort by the merged railroads to raise price even slightly above competitive costs. His conclusions were stated in categorical terms, and he did not concede or even discuss the inability of trucking to match costs on a large share of the railroads' traffic.
The Interstate Commerce Commission rejected this reasoning, and the merger, in 1985.
B. Long-distance Telephone Service
AT&T was the sole provider of long-distance service in the U.S. before the 1970s. By the time of the divestiture in 1984, MCI and Sprint had entered, but they accounted for less than 5 percent of traffic. Their shares have slowly risen, to approximately 18 and 10 percent, respectively, in 1994. Only one other significant firm (LDDC) has entered and remained in the national market (with about 5 percent of the market).
AT&T's market share naturally declined substantially in the later 1980s, by about 4 points per year. That is well above the common rate of about 1 point per year, by which dominant firms often lose market share in industrial markets (Shepherd 1990a). AT&T immediately pressed for complete deregulation, since its market share was slipping and competition was, it said, already fully effective. The Federal Communications Commission largely complied in 1989, removing virtually all restrictions and relying on loose "price caps." Many state commissions also withdrew all or most of their regulation of intrastate long-distance service, starting in 1986.
But AT&T largely stabilized its market share in the 60 to 65 percent range after 1989. MCI created a major innovation (the "Friends and Family" program), which AT&T was not quickly able to match.
For this paper, the lesson arises from widespread use of the contestability idea to press for deregulation. Willig and other AT&T witnesses have repeatedly claimed that the national and state-wide markets are contestable, so that there is no prospect whatever of monopoly pricing.
Yet the facts belie the assertion. MCI and Sprint initially set their prices some 30 percent below AT&T's, which should have enabled them instantly to oust AT&T entirely from the market, according to contestability theory. Yet the result has been starkly different. It has taken MCI and Sprint over 10 years to edge painfully and slowly toward substantial shares of the market. Despite MCI's brilliance after 1990 in devising the "Friends and Family" program, AT&T has largely stabilized its market share at over 60 percent or more of the market, both nationally and in many states.
Therefore, much of the deregulation of AT&T can be seen to have been premature and risky. Few other significant firms have entered, there has been significant exit, and the market may have settled into a stable lopsided three-firm oligopoly. That would be far from the conditions of genuinely effective competition.
Undeterred by the clear evidence of high barriers to entry, AT&T's economic witnesses still cite the theory of contestability in asserting that entry is wholly free and that AT&T has no market power.
1 Leading criticisms include Schwartz and Reynolds 1983; Weitzman 1983; Brock 1983; Shepherd 1984; and Schwartz 1986.
2 For example, William J. Baumol's Biographical Statement and Testimony List for the years 1982 through 1993 include at least 611 individual antitrust and regulatory cases in the U.S. in which he apparently testified about contestability, Ramsey prices, and other related concepts that are discussed here. Robert D. Willig has also maintained a comparably active schedule of testimony in the U.S. Their work in other countries is in addition to this.
3 Recent prominent criticisms of antitrust include Baumol and Ordover 1986, 1992.
4 Baumol is of course a distinguished and prolific economist, with significant writings on economic theory, operations research, the history of economic thought, the economics of the arts, and other topics. My strictures here apply solely to his contestability ideas and policy lessons. The same disclaimer applies also to the other work of Willig and others in the Baumol group.
5 For example, General Motors and IBM escaped any real corrective actions under Section 2 of the Sherman Act, so that the opportunities to cure the internal inefficiency and sluggish innovation were missed. See Shepherd 1994.
6 Hereafter cited as Baumol.
7 The theory of consumers and producers surplus is controversial, since it cannot be proven that the surpluses have any finite volume. Moreover, the inclusion of producers surplus as a social maximand is also rejected by many economists.
8 See J. M. Clark 1922; Machlup 1955; Shepherd 1990a, ch. 11; Scherer and Ross 1990, ch. 13. For a critique of the Ramsey pricing approach, see Shepherd 1993.
Bailey, Elizabeth E. and William J. Baumol. 1984. "Deregulation and the Theory of Contestable Markets." Yale Journal on Regulation 1:111-37.
Bain, Joe S. 1956. Barriers to New Competition. Cambridge: Harvard University Press.
Baumol, William J. 1982. "Contestable Markets: An Uprising in the Theory of Industry Structure." American Economic Review 72 (Mar.):1-15.
-----. 1984. Verified Statement in the Santa Fe-Southern Pacific Railroad Merger, Interstate Commerce Commission Finance Docket 30300, April 2, 1984.
-----. 1994. "Contestable Markets." In The McGraw-Hill Encyclopedia of Economics, 2d ed., ed. D. Greenwald, 215-18. New York: McGraw-Hill.
Baumol, William J., and Janusz A. Ordover. 1986. "Use of Antitrust to Subvert Competition." Journal of Law and Economics 28 (May):649-69.
-----. 1992. "Antitrust: Source of Dynamic and Static Inefficiencies?" In Antitrust, Innovation and Competitiveness, eds. T. M. Jorde and D. J. Teece. New York: Oxford University Press.
Baumol, William J., and J. Gregory Sidak. 1994. Toward Competition in Local Telephony. Cambridge: MIT Press.
Baumol, William J., and Robert D. Willig. 1986. "Contestability: Developments Since the Book." Oxford Economic Papers, Special Issue, 38 (Nov):9-36.
Baumol, William J., John Panzar, and Robert D. Willig. 1982 and 1988 (rev. ed.). Contestable Markets and the Theory of Industry Structure. San Diego: Harcourt, Brace Jovanovich.
Brock, William A. 1983. "Contestable Markets and the Theory of Industry Structure: A Review Article." Journal of Political Economy 91 (Dec.):1055-66.
Clark, John Bates. 1887. "The Limits of Competition." Political Science Quarterly 2:45-61.
Clark, John Maurice. 1922. Studies in the Economics of Overhead Costs. New York: Macmillan.
Machlup, Fritz. 1955. "Characteristics and Types of Price Discrimination." In Business Concentration and Price Policy, ed. G. J. Stigler. Princeton: Princeton University Press.
Scherer, F. M., and D. Ross. 1990. Industrial Market Structure and Economic Performance. 3d ed. Boston: Houghton Mifflin.
Schwartz, Marius. 1986. "The Nature and Scope of Contestability Theory." Oxford Economic Papers 38 Supplement (Nov.):37-57.
Schwartz, Marius, and Robert J. Reynolds. 1983. "Contestable Markets: An Uprising in the Theory of Industry Structure: Comment." American Economic Review 73 (June):488-90.
Shepherd, William G. 1984. "'Contestability' vs. Competition." American Economic Review 74 (Sept.):572 87.
-----. 1988a. "Three 'Efficiency School' Hypotheses about Market Power." Antitrust Bulletin 33 (Summer):395 415.
-----. 1988b. "Contestability and Railroad Mergers." International Journal of Transport Economics (June).
-----. 1990a. The Economics of Industrial Organization. 3d ed. Englewood Cliffs, NJ: Prentice Hall.
-----. 1990b. "Potential Competition versus Actual Competition." Administrative Law Review 42 (Winter):7-34.
-----. 1993. "Ramsey Pricing: Its Uses and Limits." Utilities Policy 1 (Oct.):295-8.
-----. 1994. "Antitrust Repelled, Inefficiency Endured: Lessons of IBM and General Motors for Future Antitrust Policies." Antitrust Bulletin 39 (Spring):203 34.
Tirole, Jean. 1991. The Theory of Industrial Organization. Rev. ed. Cambridge: MIT Press.
Weitzman, Martin L. 1983. "Contestable Markets: An Uprising in the Theory of Industry Structure: Comment." American Economic Review 73 (June):486-87.
William G. Shepherd is professor of economics, University of Massachusetts, Amherst. I am grateful to Johannes M. Bauer, Kenneth D. Boyer, Robert D. Cairns, and Harry M. Trebing for discussions on this topic.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Special Issue: Public Utilities Regulation|
|Author:||Shepherd, William G.|
|Date:||Aug 1, 1995|
|Previous Article:||The values jury to aid natural resource decisions.|
|Next Article:||Electric utility holding companies: the new regulatory challenges.|