Introduction

We are pleased to present the inaugural issue of the Journalof Competition Law and Economics (JCLE), which the Oxford UniversityPress will henceforth publish quarterly. The aim of the JCLEis to publish cutting-edge academic     

A CONSUMER-WELFARE APPROACH TO NETWORK NEUTRALITY REGULATION OF THE INTERNET

"Network neutrality" is the shorthand for a proposed regimeof economic regulation for the Internet. Because of the trendto deliver traditional telecommunications services, as wellas new forms of content and applications, by Internet protocol(IP), a regime of network neutrality regulation would displaceor subordinate a substantial portion of existing telecommunicationsregulation. If the United States adopts network neutrality regulation,other industrialized nations probably will soon follow. As aresult of their investment to create next-generation broadbandnetworks, network operators have the ability to innovate insidethe network by offering both senders and receivers of informationgreater bandwidth and prioritization of delivery. Network neutralityregulation would, among other things, prevent providers of broadbandInternet access service (such as digital subscriber line (DSL)or cable modem service) from offering a guaranteed, expediteddelivery speed in return for the payment of a fee. The practicaleffect of banning such differential pricing (called "accesstiering" by its critics) would be to prevent the pricing ofaccess to content or applications providers according to priorityof delivery. To the extent that an advertiser of a good or servicewould be willing to contract with a network operator for advertisingspace on the network operator's affiliated content, anotherpractical effect of network neutrality regulation would be toerect a barrier to vertical integration of network operatorsinto advertising-based business models that could supplementor replace revenues earned from their existing usage-based businessmodels. Moreover, by making end-users pay for the full costof broadband access, network neutrality regulation would denybroadband access to the large number of consumers who wouldnot be able to afford, or who would not have a willingness topay for, what would otherwise be less expensive access. Forexample, Google is planning to offer broadband access to end-usersfor free in San Francisco by charging other content providersfor advertising. This product offering is evidently predicatedon the belief that many end-users demand discounted or freebroadband access that is paid for by parties other than themselves.Proponents of network neutrality regulation argue that suchrestrictions on the pricing policies of network operators arenecessary to preserve innovation on the edges of the network,as opposed to innovation within the network. However, recognizingthat network congestion and real-time applications demand somedifferential pricing according to bandwidth or priority, proponentsof network neutrality regulation would allow broadband Internetaccess providers to charge higher prices to end-users (but notcontent or applications providers) who consume more bandwidthor who seek priority delivery of certain traffic. Thus, thedebate over network neutrality is essentially a debate overhow best to finance the construction and maintenance of a broadbandnetwork in a two-sided market in which senders and receivershave additive demand for the delivery of a given piece of information—andhence additive willingness to pay. Well-established tools ofRamsey pricing from regulatory economics can shed light on whethernetwork congestion and recovery of sunk investment in infrastructureare best addressed by charging providers of content and applications,broadband users, or both for expedited delivery. Apart fromthis pricing problem, an analytically simpler component of proposednetwork neutrality regulation would prohibit a network operatorfrom denying its users access to certain websites and Internetapplications, such as voice over Internet protocol (VoIP). Althoughsome instances of blocking of VoIP have been reported, suchconduct is not a serious risk to competition. To address thisconcern, I analyze whether market forces (that is, competitionamong access providers) and existing regulatory structures aresufficient to protect broadband users. I conclude that economicwelfare would be maximized by allowing access providers to differentiateservices vis-à-vis providers of content and applicationsin value-enhancing ways and by relying on existing legal regimesto protect consumers against the exercise of market power, shouldit exist.     

WHEN DOES AN OPTIONAL TARIFF NOT LEAD TO A PARETO IMPROVEMENT? THE AMBIGUOUS EFFECTS OF SELF-SELECTING NONLINEAR PRICING WHEN DEMAND IS INTERDEPENDENT OR FIRMS DO NOT MAXIMIZE PROFIT

Optional or self-selecting tariffs allow customers to choosebetween an established tariff and an alternative outlay schedule.The possibility of making the vendor and at least one consumerbetter off, without making any other consumer worse off, makesoptional tariffs appealing to both economists and regulators.In economic terms, the introduction of optional tariffs makespossible a Pareto improvement in the allocation of resources.Unfortunately, the presumed desirability of such tariffs dependscrucially on assumptions that may not be fulfilled in the caseof a state-owned enterprise—in particular, profit-seekingbehavior on the part of the monopoly vendor and independenceof consumer demand functions. We analyze the economic implicationsand potential consequences, in general, of introducing negotiatedrate and service terms available to a sole user into a preexistingregulatory regime of uniform tariff rates and conditions ofservice. We identify the conditions under which it is economicallydesirable to introduce declining-block rates or other rate structuresthat discriminate among users of the affected services, withor without any basis in identifiable cost differences. We addressthe specific economic implications and potential consequencesof introducing negotiated rate and service terms available toa sole user where the affected service is provided under a monopolyestablished by federal statute, taking into account that suchnegotiated arrangements may include preferential pricing terms;that access to the negotiated terms may be limited to a smallnumber of users for administrative or other reasons; and thatcompetition may exist among users of the affected service orservices. Finally, we identify and describe regulatory measuresthat might be taken to accommodate potential concerns regardingthe impact of such negotiated rate and service arrangementson fairness in regulation and competition. We conclude thatit is not possible to derive sweeping propositions about theefficiency of optional tariff offerings. Instead, the welfareeffects of such pricing plans must be evaluated empiricallyon an individual basis. Our analysis has practical significancefor pricing policies in network industries, particularly thoseindustries served by state-owned enterprises that enjoy statutorymonopolies.     

SHOULD ANTITRUST CONSENT DECREES REGULATE POST-MERGER PRICING?

Competitors proposing to merge sometimes propose price regulationin a consent decree as a condition of receiving merger approval.Antitrust enforcement agencies in the United States have beenreluctant to use such price-regulating decrees, as they sufferfrom practical problems in implementation. It is less recognized,however, that the use of consent decrees to regulate post-mergerprices may be unlawful. Such decrees exceed the scope of antitrustlaw and blur the distinction between the legislative power toregulate prices and the executive power to enforce the antitrustlaws. Despite the willingness of merging parties to accept priceregulation in consent decrees, economic and constitutional considerationscounsel against antitrust enforcement agencies adopting thispractice.     

PATENT HOLDUP AND OLIGOPSONISTIC COLLUSION IN STANDARD-SETTING ORGANIZATIONS

Current controversies over patent policy place standard-settingorganizations (SSOs) on a collision course with antitrust law.Recent theoretical research conjectures that, in an SSO, patentowners can "hold up" patent users in the sense of demandinghigh royalties for a patented input after the SSO has adoptedthe patented technology as an industry standard and manufacturerswithin the SSO have incurred sunk costs to design end productsthat incorporate that standard. Consistent with this conjecture,actual SSOs have recently sought no-action letters from theAntitrust Division for a variety of amendments to SSO rulesthat would require or request, at the time a standard is underconsideration, the ex ante disclosure by the patent owner ofthe maximum royalty that the patent owner would charge underthe regime of fair, reasonable, and nondiscriminatory licensing.This price information—which is characterized as the "cost"of the patented input—would, under at least one recentSSO rule modification, be a permissible topic for potentialusers of the patent to discuss when deciding whether to selectit in lieu of some alternative standard. This exchange of informationamong horizontal competitors would occur ostensibly becausethe cost of the patented technology had been characterized assimply one more technical attribute of the standard to be set,albeit an important technical attribute. The Antitrust Divisionand the Federal Trade Commission have jointly stated that suchdiscussion, by prospective buyers who are competitors in thedownstream market, of the price of a patented invention thatmight become part of an industry standard should be subjectto antitrust scrutiny under the rule of reason rather than therule of per se illegality. The rationale that the antitrustagencies offer for applying the rule of reason to such conductis that such horizontal collaboration might avert patent holdup.The Antitrust Modernization Commission (AMC) similarly endorsedthe view that rule-of-reason analysis is appropriate for exante discussion of royalty terms by competing buyers of patentedtechnology. This rule-of-reason approach, however, is problematicbecause it conflicts with both the body of economic researchon bidder collusion and with the antitrust jurisprudence oninformation exchange and facilitation of collusion. Put differently,because of their concern over the possibility of patent holdup,the U.S. antitrust agencies and the AMC in effect have indicatedthat they may be willing in at least some circumstances to forgoenforcement actions against practices that facilitate oligopsonisticcollusion by encouraging the ex ante exchange of informationamong competitors concerning the price to be paid for a patentedinput as an implicit condition of those competitors' endorsementof that particular patented technology for adoption in the industrystandard. However, neither the proponents of these SSO policiesnor the antitrust agencies and the AMC have offered any theoreticalor empirical foundation for their implicit assumption that theexpected social cost of patent holdup exceeds the expected socialcost of oligopsonistic collusion. This conclusion does not changeeven if one conjectures that such collusion will benefit consumersby enabling licensees to pass through royalty reductions intheir pricing of the downstream product incorporating the patentedtechnology. Proper economic evaluation of the plausibility ofthe pass-through conjecture will require information about thecalculation of royalty payments; the demand and supply elasticitiesfacing the licensees; and the structure of any industries furtherdownstream between the manufacturer and the final consumer.Consequently, the magnitude of this effect will likely be amatter of empirical dispute in every case. Moreover, such ajustification for tolerating horizontal price fixing finds nosupport in antitrust jurisprudence. Given the analytical andfactual uncertainty over whether patent holdup is a seriousproblem, it is foreseeable that antitrust questions of firstimpression will arise and affect a wide range of high-technologyindustries that rely on SSOs. However, there is no indicationthat scholars and policy makers have seriously considered whetheroligopsonistic collusion in SSOs is a larger problem than patentholdup.