Wednesday, May 1, 2013
Posted by Heski Bar-Isaac
Separate bodies of literature and thought have addressed exclusion by a dominant firm (or through formal agreements by a dominant collection of firms), and tacit collusive behavior in pricing decisions; to date there has been relatively little that considers parallel behavior by firms to implement exclusion. Certainly, the issue has lacked a comprehensive treatment. In bringing together a wide range of examples of alleged behavior and tracing through a range of mechanisms, economic analysis, and doctrinal implications, Hemphill and Wu provide an original, clear, rounded, and provocative discussion that has implications both for anti-trust practice and academic research.
In reading the paper, there were several highlighted ideas that were new to me and contrasted with my understanding of (tacit) collusive pricing. I am certain that these will color my thinking going forward.
Stability of Parallel Exclusion
The authors make a strong case that parallel exclusion is likely to be much more stable than parallel pricing. First consider the similarities: Textbook models of collusion on pricing highlight that when several firms are maintaining elevated prices, there is an incentive for each of them to undercut the others to gain market share. Indeed, if firms weigh current revenues much more than future revenues (as may be the case for firms on the brink of bankruptcy), theory suggests that there will be no self-sustaining cooperation that assures elevated prices.
The authors highlight that exclusion schemes might also feature similar incentives: Exclusion may be costly for the firms colluding on exclusion; for example, it may involve transferring some rents to upstream or downstream agents who are the parties that assure the exclusion. Each of the excluding firms may be tempted to “cheat” and keep more rents. Alternatively, a vertically integrated firm that
is a colluding firm in an upstream segment, for example, may be tempted to accommodate an (efficient) entrant to gain an advantage in downstream competition.
However, Hemphill and Wu suggest that there are important differences between collusion on pricing and exclusion. In collusive pricing schemes, a challenge is to determine the prices (or market shares, or other outcomes) that the firms are coordinating on, and to observe compliance. In environments subject to variable demands, shocks to underlying costs, and where prices are privately negotiated, it may be difficult both to determine what to coordinate on and whether firms are complying. Instead, exclusivity is relatively easy to observe, and as a very coarse binary outcome (either the firm is operating exclusively or not) may also be a clear focal point for coordination.
Further, as Hemphill and Wu highlight, price cuts are reversible whereas accommodation tends not to be. This implies that the consequences of permitting entry may be more severe than the consequences of engaging in a temporary price war, and thus any individual firm would be more
reluctant to “cheat” from colluding on exclusion.
Interaction between vertical and horizontal structures
Collusion involves a horizontal agreement or understanding between rival firms—the market structure of these firms thus plays an important role in assessing the stability of a collusive agreement. Exclusion, instead involves vertical practices, such as restraints or agreements with suppliers or
customers—the market structure of this related segment can therefore by key. For example, a downstream monopolist may be able to internalize the gains from efficient entry making it more difficult to secure upstream exclusion; alternatively with very many downstream firms it may be difficult for an upstream incumbent monopolist to offer enough to each of them such that they would be willing to exclude a rival.
The latter observation (as mentioned in my recent paper Raising Retailers' Profits: On Vertical
Practices and the Exclusion of Rivals co-authored with John Asker forthcoming in the American Economic Review) highlights that an upstream monopolist or dominant group of firms may benefit
from limiting the number of downstream firms, and so there may be complementarities between upstream and downstream exclusion. Indeed one of the means that the upstream firm may use to effectively transfer rents to downstream firms, so as to encourage support of upstream exclusion, may be to support downstream exclusion.
Such considerations suggest that a thorough understanding of parallel exclusion may rely on the structure of both upstream and downstream segments and perhaps the structure of buyer-seller relationships. Bowman (The Prerequisites and Effects of Resale Price Maintenance, The University of Chicago Law Review, 22(4) 1955) discusses exclusive arrangements for pattern-makers and spark-plug manufacturers whereby a downstream firm dealt with only a single upstream firm; instead, Hemphill and Wu hint that parallel exclusion in the credit card industry may have been easier to achieve since most American banks issue both Mastercard and Visa (“duality”). Although, Whinston (Lectures
on Antitrust Economics, 2006, p.177) suggested that “further study of multiseller/multibuyer models should be a high priority,” the call thus far remains largely unanswered, and my intuition, at least, remains cloudy. However, a growing body of empirical academic work on buyer-seller network structure and contracting is emerging; for example the forthcoming work in the American Economic Review by my former colleague Robin Lee, Vertical Integration and Exclusivity in Platform and
There is a work on vertical integration and exclusion with dominant firms and some work on vertical integration and price collusion (both strands are reviewed in Mike Riordan’s chapter “Competitive Effects of Vertical Integration” in Buccirossi’s Handbook of Antitrust Economics). To my knowledge, however, there is no academic work on vertical integration and its effects on parallel exclusion.
An interesting feature that arises in one of the cases (conduit for electric wiring), perhaps the most blatant discussed by Hemphill and Wu, is that rather than seeking to achieve exclusion through contracts with upstream or downstream firms, the incumbent manufacturers sought to achieve exclusion of plastic conduit through the standards process.
Pro-competitive effects and harm beyond elevated pricing
As Hemphill and Wu, parallel exclusion can have pro-competitive effects—for example, standard setting often aims to ensure safety or assure quality for consumers with poor information. Though,
oftentimes, and as particularly illustrated in the conduit example, standards are not without controversy. As another example, while many argue that professional licensure ensures quality, there are others who suggest it may have anti-competitive in raising high entry barriers—indeed this was a central topic of Milton Friedman’s PhD thesis (Income from Independent Professional Practice with Simon Kuznets, 1945) and remains a point of controversy (a good recent overview is Morris Kleiner’s book Licensing Occupations: Ensuring Quality or Restricting Competition?)
More broadly, an influential body of work has highlighted pro-competitive effects of vertical restraints and influenced legal practice; for example, there is a wide-ranging discussion of RPM in the Leegin decision (Leegin Creative Leather Prods., Inc. v. PSKS, Inc., 551 U.S. 877, 897, 2007).
Hemphill and Wu suggest that harm from parallel exclusion might be more severe than harm from parallel pricing. Both support elevated prices—if exclusion had no effect whatsoever on prices, then it is hard to imagine why firms should take even minimally costly actions to achieve it.
However, parallel exclusion has a further effect in blocking or slowing down innovation of either
lower-cost substitutes or substitutes that are more desirable to at least some consumers (this may involve higher-quality that may be more desirable to all consumers or simply a differentiated offering of more appeal to a segment of consumers). Note that collusion on pricing might also be expected to have some effect on product innovation: incumbent firms might have muted incentives to innovate to the extent that innovation might prove disruptive to the collusive arrangement. Such harm, in practice, is likely to be very difficult to estimate. It is intuitive that the harm should be more severe in the case of parallel exclusion, to the extent that a rival product or technology exists; moreover, with an existing potential entrant, there is much greater scope to measure the effect.
By noting that parallel exclusion may be stable (through a simple point of coordination and relatively cheap implementation), and may offer severe harm (through stifling efficient products), and highlighting several cases, Hemphill and Wu make a convincing case for actively investigating behavior
of this type. Particularly as the interconnectedness of firms in emerging technology markets increases, whether through standard setting organizations or other interactions, it seems possible that parallel exclusion may become an increasingly important form of anticompetitive abuse. Hence, whether through government enforcement or the actions of the private plaintiff's bar, increased attention to parallel exclusion in practice seems likely and important.
On the academic side, unsurprisingly, I would argue that there is more research that could usefully be undertaken. For economists, in addition to the theoretical issues surrounding the network of upstream and downstream relationships, empirical work examining this form of conduct can help us to better understand the range of competitive harm that results, and develop methods for the assessment of damages. For legal scholars, further development of the underlying legal theories of liability seems useful would continue the substantial progress made in this fundamental contribution by Hemphill and Wu.
 In particular, the example that Hemphill and Wu provide in the telecommunications market (p. 1221-2) of wireless carriers can be seen as an example in this spirit where the carriers in effect sell both bandwidth and network access to consumers, but also services that employ these.
 Indeed, previous work on joint dominance (such as Iaccobucci and Winter Abuse of Joint Dominance in Canadian Competition Policy, 60 U. TORONTO L.J. 219, 2010) has highlighted that “contracts or strategies that are adopted by cartel members as facilitating devices to aid in sustaining the cartel price can serve simultaneously to deter entry.”