Compatibility and interconnection pricing in the airline industry: a proposal for reform.

• Author: Weidenhammer, Bradley H.

CONTENTS INTRODUCTION I. NETWORK ECONOMICS A. Direct and Indirect Network Effects B. Tipping C. Compatibility 1. The Economics of Compatibility Choices 2. Interconnection Pricing II. NETWORK EFFECTS, COMPATIBILITY, AND ANTITRUST ANALYSIS III. COMPETITION AND COMPATIBILITY AMONG AIRLINE NETWORKS A. The History of Airline Competition B. Network-Compatibility Theory and the Airline Industry 1. Modeling Compatibility and Interconnection Pricing Among Airline Networks 2. Observing Compatibility and Interconnection Pricing in the Airline Market IV. ASSESSING EFFICIENCY, PROPOSING REFORM A. Compatibility Through Regulation B. Welfare Analysis CONCLUSION INTRODUCTION

Since the deregulation of the airline industry in the late 1970s, the behavior of firms in the industry has often baffled analysts and stymied prognosticators. In the years preceding and immediately following deregulation, scholars theorized that the deregulated airline industry would approximate perfect competition. (1) Scholars reasoned that because the airline industry did not appear to manifest significant economies of scale, firms would be able to enter the market at a relatively small scale and still attain competitive costs and discipline large firms' pricing. (2) Moreover, because existing firms had bloated costs and suboptimal route structures, new entrants were expected to achieve drastically lower costs and to unseat incumbents quickly. (3)

Yet none of these predictions has come to pass. Firms existing before regulation (so-called "legacy firms") did not disappear. Instead, they survived and grew to dominate the industry without bringing their costs down near the level of entrant firms. The anticipated contestability of the airline market--the ability of firms to engage in hit-and-run market entry wherever supracompetitive rates are being charged--never materialized. In addition, the incumbents soon reorganized their route plans into hub-and-spoke networks. Airlines turned to hubbing ostensibly as a cost-saving measure, though scholars have since shown that hubbing is also an effective entry-deterrence strategy. (4)

This Note suggests that network-compatibility theory can explain some of the perplexing conduct we observe in the airline market. In particular, I argue that competition in the airline industry functions at a suboptimal level, in part because systemic incompatibilities between airlines permit competition only at the network or full-itinerary level, foreclosing competition at the component-flight level. Logistical difficulties involved in connecting from one airline to another have thwarted compatibility among airline networks. Furthermore, the dominant airline networks have used pricing policy to anticompetitive effect, setting high interconnection fees that reduce consumer welfare.

Scholars have paid little attention to the effects of incompatibility on competition in the airline industry, despite the well-established correlation between network incompatibility and the creation of market power by a dominant network. (5) Moreover, where incompatibility exists among competing networks, free entry may not erode a dominant network's market power. (6) Hence, scholars taking the orthodox view that free entry assures optimal market performance fail to account adequately for the network aspect of the airline industry.

This Note attempts to remedy this gap in the literature by using network-compatibility and interconnection-pricing theory to analyze firm behavior in the airline industry and to propose a corrective regulatory regime. Part I introduces relevant elements of network economic theory. Part II examines important antitrust cases to construct a framework for analyzing the welfare effects of compatibility decisions. Part III then discusses network effects in the airline industry, surveys the history of airline competition since deregulation, and demonstrates the ability of network-compatibility and interconnection-pricing theory to describe and account for airline behavior. Part IV employs network theory to formulate a regulatory solution to the market failures identified in Part III and discusses the costs and benefits of the proposed regulation.

1. NETWORK ECONOMICS

   Economists have long recognized that the consumption value of some products depends on the number of other consumers who use those products. For example, a telephone has value only as a paperweight unless there are other telephone users with whom one can communicate. Similarly, a car is more valuable if repair shops are plentiful, and the number of repair shops depends on the existence of other car users within a reasonable proximity. As the number of people who speak English increases, English language skills become more valuable. (7) Economists use the term "network effect" to describe the benefit consumers experience as the result of others' consumption of the same product. (8)

   1. Direct and Indirect Network Effects

      A network consists of two or more interconnected nodes. Connections between the nodes may be physical, as in the case of telecom or highway networks, or based upon common standards, such as language (for example, Portuguese or the software language C++) or design (for example, electrical outlet configurations). Analytically, however, the means of connectivity is less important than the type of consumption benefits the network creates. Since Michael Katz and Carl Shapiro first suggested the distinction, economists have generally distinguished between direct and indirect network benefits. (9) Though the distinction is not always easily made, (10) direct network benefits generally exist where the good being consumed is connectivity itself. (11) A paradigmatic example of a direct network benefit is the benefit realized when a communication network is expanded: Each member realizes a direct benefit in the form of expanded communication opportunities. (12)

      Indirect network benefits are realized by consumers of goods that are not instruments of connectivity but that increase in value as consumption of the product increases. (13) For example, a Ford Taurus is valuable regardless of whether anyone else drives a Taurus, but economies of scale dictate that complementary goods (such as repair shops and replacement parts) will be more widely available if the pool of Taurus drivers is larger. Thus, a Taurus owner experiences positive consumption effects indirectly--through an increasing variety and availability of complementary goods.

      Consumers joining a network create an externality insofar as they create benefits to other users of the network that they themselves do not internalize. (14) As a result, networks tend to be smaller than optimal because marginal consumers do not sufficiently value their participation in the network and may fail to join even when total (internal and external) benefits outweigh total cost. (15) Where a network is proprietary, however, the owner is able to internalize all network benefits by subsidizing marginal consumers and extracting the additional network value from existing members. (16)

      Though a debate is ongoing as to what constitutes a network benefit versus a network externality, (17) it is sufficient for our purposes to use the terms "network benefit" to denote the portion of a good's value that is attributable to network size and "network externality" to denote the difference between the social marginal benefits and the private marginal benefits of an additional consumer joining a network.

   2. Tipping

      In an industry that exhibits network effects, the convergence of consumers to a single standard or network will maximize network effects. (18) Thus, in industries in which multiple incompatible networks compete (assuming uniform quality across competing networks), even a small size advantage will make one network more desirable than the others. (19) The emergence of a size leader produces a demand-side feedback loop where each new adopter increases the relative value of the network, thus increasing demand for the network. This, in turn, further increases the relative value of the network, and so on, producing a dominant firm. (20) This phenomenon is referred to as "tipping."

      Once a firm attains a dominant size, the network benefits realized by its consumers vastly outstrip the network benefits other firms in the market may offer. (21) While this does not foreclose entry completely, it provides the dominant firm with a competitive advantage that shields it somewhat from direct competition over productive efficiency or product quality. (22) For an entrant network to draw customers away from a dominant network, the entrant must offer substantive benefits that exceed the value of the dominant network's substantive benefits plus its network benefits. (23) Thus, where rival, incompatible networks compete, and where one network has a first-mover advantage, the dominant network may preserve its market share over time, in spite of an entrant's superior productive efficiencies.

   3. Compatibility

      Compatibility is perhaps the most important concept in understanding network industries. Total compatibility exists when two components are combinable and function together without extra costs. Partial compatibility is also possible. For example, American electrical plugs are compatible with outlets in the United Kingdom, but an adapter is required. The compatibility of any two components falls along a continuum stretching from total compatibility, when two components function together without additional cost, to total incompatibility, when the cost of achieving compatibility outweighs the benefit.

      When two systems exhibiting network effects are compatible, an aggregate network is formed, composed of the total membership of all compatible systems. (24) For instance, land-line telephones and citizen-band (CB) radios are incompatible communication networks; therefore, the value of the telephone network versus that of the CB-radio network is a...