Does a monopolist aim to maximize profit, or simply to hide from the antitrust laws?

AT&T's absolute monopoly in the switched long-distance telephone market ended in 1976 when MCI rolled out its Execunet service. Twenty years later economists still question whether AT&T can influence the market price of long-distance services.

Recent empirical studies are split on the question, sometimes finding AT&T has considerable market power, and sometimes finding it has none.

It appears that economists studying the long-distance industry may be misinterpreting the historical record. Rather than manipulating price for the sake of building profits, AT&T may have manipulated its market share for the sake of avoiding renewed regulatory and antitrust intervention.

The persistence of market power in deregulated markets is an important issue. Policymakers are engaged in transforming the local telephone monopoly into a competitive industry, a plan even more adventuresome and potentially more costly than opening the long-distance market. Couple that with efforts to restructure the electricity industry and the significance of the market-power issue becomes apparent.

Between 1990 and 1994, AT&T filed tariffs that other carriers quickly matched, avoiding in the process the sort of price wars that followed deregulation in airline and trucking industries. Market shares in each long-distance sector stabilized at roughly 60 percent (for AT&T), 20 percent (MCI), 10 percent (Sprint), and 10 percent (all other carriers). Industry concentration stabilized at a high level, with Herfindahl-Hirshman Index values ranging between 0.4 and 0.6 across long-distance services.

It is against this factual background that two recent empirical studies reach opposing conclusions on the question of AT&T's market power. This difference of opinion would not be surprising if the conclusions had been deduced from a priori oligopoly theory. Economics has something to say about oligopoly markets, but does a poor job of predicting the behavior of prices, quantities and market shares within them. Accordingly, economists regard oligopoly theory as indeterminate, %n1%n and so have sought other approaches to benchmarking economic performance in the long-distance market.

In one study, %n2%n Paul MacAvoy concludes that AT&T retains considerable market power. Moreover, he finds that the long-distance industry has performed as a tacitly collusive, three-firm oligopoly, with AT&T exercising market power to propel itself, MCI, and Sprint to super-competitive profits. MacAvoy shows that industry concentration and market shares stabilized around 1990, while prices and profit margins converged and then rose in lock step. Profit margins increased as industry concentration initially decreased (a divergence contrary to most a priori expectations), and continued rising as concentration stabilized. MacAvoy shows also that margins settled near 70 percent (i.e., estimated marginal cost equals 30 cents of the average revenue dollar), and compares these values with margins of 35 percent in other highly concentrated industries. He finds an unexpected absence of price shading among high-volume customers, who ordinarily would be expected to whipsaw competing carriers into making price concessions. The gravamen of MacAvoy's thesis is that long-distance markets never performed competitively under continued regulatory and antitrust constraints, they have moved systematically away from competitive performance. %n3%n

Another contemporary study reaches exactly opposite conclusions. %n4%n Rather than calculating profit margins directly, the authors (Kahai, Kaserman and Mayo) estimate a "dominant firm/competitive fringe" econometric model of the long-distance industry. They find that AT&T faces a demand price elasticity ranging between -3.73 and -7.81 in the least-competitive, long-distance market segment. They speculate demand is even more price elastic in the most competitive business service segments. These elasticity values correspond, respectively, to Lerner Index values of 0.27 and 0.13, %n5%n and so, by the "inverse elasticity" rule, to profit margins of 27 percent and 13 percent.

After comparing their findings with data for other industries, KK&M find that the long-distance industry might be the most competitive of all, a result that flatly contradicts MacAvoy's findings. The authors, therefore, conclude that AT&T lacks significant market power.

Why is the research split over this issue?

A Theory of Behavior

Empirical studies necessarily rely upon historical price and quantity data. They assume that the data reflects classical profit-maximizing conduct, as predicted by economic theory. Therein lies the potential for misinterpretation.

The studies described above both recognize that the long-distance industry does not operate unfettered, but rather under regulatory and antitrust rules. However, neither study seeks to control for the effect of these rules.

Moreover, neither study seeks to explain one fascinating aspect of long-distance market structure: Why did market shares stabilize where they did? In particular, why did AT&T's market share stabilize at roughly 60 percent across all service segments? Historical market shares may represent nothing more than an admixture of chance, cost and the subjective valuations of long-distance users. Or, they may help to explain the conflict among empirical studies. Here's why.

For roughly a decade beginning in 1976, AT&T assumed MCI's mantle of a law firm with antennas. During this period, AT&T diverted its financial, technical and managerial resources to a defense against antitrust actions. Some of these were little more than nuisance suits. Those brought by MCI, Sprint, and the U.S. Department of Justice were substantial by any standard.

At the time of divestiture, AT&T was the dominant supplier of long-distance services, with market shares approaching 90 percent. Many industry pundits predicted that unfettered competition, coupled with uniform access charges, would quickly force AT&T's competitors out of business and push AT&T's market shares back toward 100 percent. How better for AT&T to protect itself against the cost of post-divestiture litigation and asymmetric regulation than to reduce its market shares below monopoly levels?

Economic theory does not pinpoint the level of market share needed to attain "monopoly" status. However, a practical answer can be found in federal court dicta. In United States v. Aluminum Company of America, %n6%n the court offered an offhand definition of monopoly while wrestling with the related issue of how to fix the size of the relevant product market. By one measure, the defendant held a 90-percent market share; by another it held only 64 percent. The court speculated that a 90-percent share "is enough to constitute a monopoly [though] it is doubtful whether sixty or sixty-four percent would be enough." This dicta remains the informal test for monopoly in the federal case law.

It may be only a coincidence that AT&T's market share has leveled off at 60 percent. Scholars, however, have focused on the fact that AT&T hemorrhaged market share until reaching 60 percent, at which point it acted to staunch its losses. Former Justice Department Consultant Peter Huber observed, for example:

In 1990, AT&T openly declared its intention to stabilize its share position ... AT&T can, and does, unilaterally decide what share it will hold or cede in the long distance market. %n7%n

If AT&T manipulated its market share to avoid being heckled by bureaucrats and competitors, then its "exercise" of market power over price may be merely an artifact of its preference for a safe harbor from renewed antitrust and regulatory action.

The economics of this market strategy are depicted in Figure 1. If AT&T were to exercise market power to maximize profits, then (by standard economic reasoning) it would choose price and quantity as determined by the intersection of its marginal cost and marginal revenue curves. By so doing, AT&T would fully exploit both its cost advantages and the subjective valuations of long-distance users.

By adopting a pure profit-maximizing strategy, AT&T would generate a market share that exceeded the presumed antitrust threshold. Instead, AT&T could choose to restrict output to Q*, a point at which its market share was equal to or comfortably near that threshold. How might that occur? Under one strategy, AT&T could reduce quantity by shifting demand (D) inward (on the graph), perhaps by tailoring a marketing program to this end. Or, the company might raise price from P to P*. Either approach would reduce AT&T's market share below the profit-maximizing level (where MC and MR intersect). This plan would not constitute a classic abuse of market power.

Regulatory and antitrust burdens are assumed to increase AT&T's operating cost sharply whenever output exceeds the level of Q*. Accordingly, a decision to price at P* exemplifies ex ante profit maximizing. Note, however, that industry data will not reflect this assumed increase in operating cost, so long as AT&T keeps its market share below the monopoly threshold.

If AT&T were to restrict quantity by raising price, then price P* would establish a market ceiling, which fringe carriers, including MCI and Sprint, would take as given. The market shares of these carriers would be determined in turn by output decisions reflecting each firm's unique cost and demand characteristics.

Pricing below P* by a fringe carrier could increase that carrier's market share (at least in the short run), and so might stimulate price competition. However, price cutting in long-distance markets might not be profit-enhancing. Industry demand is substantially price inelastic due to the absence of close substitutes for long-distance services at prevailing prices. %n8%n Accordingly, price cutting would not stimulate a firm's output enough to improve profits without a substantial, long-run shift in market share. The pursuit of market share through price cutting also could trigger a price war that could leave even the survivors worse off. Given these factors, fringe competitors might find it more sensible simply to set prices at or near P*.

Empirical work shows that the long-distance industry has behaved this way. Fringe carriers respond to AT&T's pricing decisions, but AT&T responds only slightly, if at all, to its rivals' choices of price and output.

At first blush, one might express surprise at prices set above the naive profit-maximizing level (P). However, a public policy that protects "competitors" at the expense of competition necessarily compels this result. Pricing at P* (the observed price level) cuts across the conventional presumption that regulation has held AT&T's prices below the profit-maximizing level; i.e., as KK&M note in their study, "To the extent that regulation may be binding, the profit-maximizing price of the dominant firm will be higher than observed prices." %n9%n

Finally, notice how the choice of method can dictate the conclusions of an empirical study. Any restriction in output brought about by raising price would increase AT&T's observed price elasticity of demand from E to E*. This violates the a priori assumption of a nexus between price elasticity and profit margin. A study that observes the high price elasticity (and, consequently, the low Lerner Index value) at E* may conclude that AT&T lacks market power. Conversely, a study that observes the high profit margin at price P* would conclude that market power and tacit collusion have boosted AT&T's profits, %n10%n but would miss the fact that P* actually lies above AT&T's unconstrained profit-maximizing price. Neither study necessarily would recognize that the effects being observed followed from a pricing strategy designed to avoid litigation and regulatory costs.

Implications

Four implications follow from this interpretation of conduct and performance.

First, empirical studies that use historical price and quantity data drawn from tariffs and other industry sources may yield results that cannot be interpreted, at least in AT&T's case, in the simple economic sense. Embedded in these data are the effects of either an avoided regulatory cost or a significant X-inefficiency %n11%n (em i.e., insurance against renewed regulatory and antitrust action. These effects must be controlled in empirical studies lest they vitiate conclusions.

Second, the contention that the long-distance industry performs as a tacitly collusive oligopoly is less sinister than at first appears. A market price at P* clearly sits above the competitive price, which is expected to lie somewhere between price and marginal cost on Figure 1. But price P* results, not from a tacit conspiracy to build profits, but from a decision by AT&T to constrain its market share. Conjectural variations by fringe competitors still come into play, %n12%n but these too are fueled by AT&T's pursuit of safe-harbor objectives. Tariff filing requirements do facilitate the coordination of price and output levels among fringe carriers, but the consequences are less ominous than might be imagined.

Third, public policies that expose large firms to tactical rent-seeking by predatory rivals and ambitious bureaucrats may provoke super-competitive prices as a perverse and unintended consequence. %n13%n

Fourth, past antitrust policy to exclude local Bell companies from the long-distance market may be the most significant cause of super-competitive pricing. A market survey commissioned by one Bell company, Pacific Telesis, shows the firm would supply upwards of 35 percent of all intraLATA service originating in its operating area if (when) permitted to enter this business. If so, then nationwide entry into long distance by the Bell companies could have reduced AT&T's market share below the antitrust threshold, without any artificial restriction in output or increase in market price.

Another Bell company, SBC (which recently merged with Pacific Telesis), ironically is poised to enter the long-distance indsutry through a peculiar, though narrowly rational, merger with AT&T. Public policies that foster such market conduct are unlikely to serve the public interest. t

James A. Montanye is a consulting telecommunications economist. His work last appeared in 1996, under the title, "Give No Credence to Lemons! Some Lessons in Market Research."

1For discussion of oligopoly theory, see for example: Dennis Carlton and Jeffrey Perloff, Modern Industrial Organization, Harper Collins (1990); F.M. Scherer, Industrial Market Structure and Economic Performance, 2nd ed., Rand McNally (1980); Jean Tirole,The Theory of Industrial Organization, MIT Press (1988).

2Paul MacAvoy, The Failure of Antitrust and Regulation to Establish Competition in Long-Distance Telephone Services, MIT Press and AEI Press (1996).

3In MacAvoy's view, the country has been mislead into believing the long-distance industry is competitive, due to the following: a large, but inconsequential, competitive fringe; ubiquitous marketing without meaningful price competition; periodic price reductions attributable to regulatory fiat rather than to competitive pressure; frequent price structure changes without price level reductions; and subscriber shifting among rival carriers in pursuit of elusive savings.

4"Is the 'Dominant Firm' Dominant? An Empirical Analysis of AT&T's Market Power," by Simran Kahai, David Kaserman and John Mayo, Journal of Law and Economics 39 (October 1996): 499-517.

5The Lerner index, which is simply the (absolute value) inverse of the demand price elasticity, provides a relative measure of market power. Market power increases as the index value increases through its theoretical range of 0 to 1.

6148 F.2d 416 (2d Cir. 1945).

7Peter Huber, Michael Kellogg, and John Thorne, The Geodesic Network II, The Geodesic Co. (1992).

8MacAvoy, note 2 above, p. 155; Kahai and others, note 4 above, p. 509; Lester Taylor, Telecommunications Demand in Theory and Practice, Klewer Academic Publishers (1994), p. 17.

9Kahai, Kaserman, and Mayo, note 4 above, p. 514.

10The problem of estimating long-distance marginal costs notwithstanding.

11An "X-inefficiency" is any deviation from the maximization of money profits. See Harvey Leibenstein, "Allocative Efficiency vs. 'X-Efficiency,'" American Economic Review 56 (June 1966).

12"Conjectural variations" are oligopolists' strategic responses to each other's price and output decisions. See MacAvoy, note 2 above, pp. 99-103, 155-157.

13Rent-seeking is systematic pursuit of windfall profits through political means.

Articles found on this page are available to Internet subscribers only. For more information about obtaining a username and password, please call our Customer Service Department at 1-800-368-5001.