Differential pricing: Use an axe, not a scalpel to stop these anti-competitive practices

By Mishi Chowdhary The Internet creates new problems in competition policy, but they are not unprecedented. Traditional patterns of anti-competitive conduct recur, but the complexity and unfamiliarity of the technological details affords rich opportunities for obfuscation and confusion. In the current inquiry, two different problems, neither particularly intricate, have been conflated by Telecommunications and platform companies, seeking to avoid limitations on anti-competitive conduct who believe they benefit from the resulting confusion.  The two problems, simple in themselves, are: 1. Should it be prohibited for a regulated telecommunications network operator to sell the traffic it carries for its low-income subscribers, in bulk, to a data-mining company for surveillance and analysis, which thereby achieves a competitive advantage in the sale of Internet-based advertising?  If this practice should be forbidden as anti-competitive in general, is it legitimized by applying the purchase price of the packets to the subsidization of some fraction of the packet transmission involved, offering “partly-free” data service, at the data-miner’s option, to the network operator’s low-income subscribers?  In simple words: FreeBasics by Facebook. 2. Should telecommunications network operators be common carriers, or should they be allowed to offer discriminatory pricing structures to favored operators of data services, including themselves or their business partners? Conflating these inquiries, and applying the gloriously indefinite, non-specific label of “network neutrality” to the combination, results in an uncertainty merely apparent, not real.  Vast expenditures on public relations, including the best scholarly commentary money can buy, have created the confusion intended, but it is one to which no well-informed regulator need succumb.  The greatest asset of the expert is firm grasp on the obvious, which is all we need confidently and correctly to resolve the issues posed. “Network neutrality,” for present regulatory purposes, means neither more nor less than “rules prohibiting anti-competitive routing practices.” The phrase denotes a regulatory conclusion, not an analytic category or a standard of review. Trai is presently asking for consultation on whether the behaviors described above involve anti-competitive practices in network routing that should be prohibited for their socially-undesirable effects on the market, what the US antitrust law calls “unreasonable restraint of trade.”  For different reasons, connected with their different anti-competitive effects in different markets, these two practices (which we can call for convenience “free basics” and “ differential pricing ”) should be prohibited. Practice #1, the sale of packets for surveillance and analysis by third party “advertising platform companies” involves a routing step (the transmission of all traffic through the routing proxy of the data-miner) unnecessary to the technical completion of the interconnection service the customer buys.  Any step in routing or retransmitting customer data unnecessary to the actual interconnection requested by the customer (to the particular endpoint address and port to which connection is sought) should be subjected to more stringent analysis in any situation. In this instance, the routing step of no functional value to the interconnection of endpoints enables the network operator and its data-mining business partner to capture the full value for advertising purposes of surveilling the customer.  The network operator would be prohibited from spying on the customer in this fashion for its own profit: no one has argued or believes that this is an acceptable practice for a regulated operator.  Competing operators of advertising placement businesses are excluded from the market, or face substantially higher barriers to entry, because of the collusion between the surveillance agent and the network operator. This anti-competitive effect is not negated by rebating a portion of the value to the network subscriber that makes some of the subscriber’s packet traffic “free.”  This is a fixed price for the subscriber’s valuable personal information, set by the collusion of the operator and the data-miner.  We can expect the price to be substantially below the competitive price that would be paid in a free market for the right to spy on the customer. Naturally, this practice should be prohibited for additional reasons beyond its directly anti-competitive effect in leveraging the network operator’s “natural” oligopoly in packet carriage into unrelated markets for personal data and advertising placement.  The practice’s noxious effects on personal privacy and network security, its particular implementation when nationally-regulated operators are selling citizens’ data abroad to foreign corporate surveillance platforms, are independent reasons why prohibition should be imposed. But the nature of the anti-competitive effect in this situation is familiar from competition policy analysis in more traditional industries.  In the US, for example, federal and state regulators overwhelmingly prohibited the telephone operating companies from owning the “Yellow Pages” business directories through which telephone subscribers located businesses by category, and where businesses bought display advertising.  Even those that did not prohibit ownership of Yellow Pages by telephone operating companies would never have approved exclusive-dealing arrangements by which the phone company could pick and choose among potential publishers of Yellow Pages.  The fact that the directories were provided at no cost to households, which was invariably the case, had no effect on that analysis. The second question, whether to permit discriminatory pricing of packets in the network for favored endpoints, is even simpler, despite all the lavish dancing to distract our eye.  Packet-transmission is a form of transport business in which the number of units moved is astronomical and the marginal cost of moving each individual packet approaches zero. Any operator, therefore, can–if allowed to discriminate freely in the price charged per packet among potential transmitters and receivers–price any particular fraction of the immense volume it moves at zero.  By eliminating cost of communication between businesses and customers in any market for digital goods or services, the network operator gains the power to pick winners and losers, for substantial periods, as it chooses.  Obviously, whether the rules permit the operator to discriminate in favor of its own businesses is irrelevant: unless the network operators are common carriers, harms to competition are eventually inevitable. No party, no matter how audacious its taste in obfuscation, has ever asserted that vertical integration and price discrimination will *never* result in serious competitive harm.  Instead, the expensive talent and fancy titles are employed to argue that in specific situations a wise regulator might discover occasional reasons why the dangers of such arrangements are absent or can be behaviorally confined. The cost of such a system, however, falls on the society and its regulators: the regulators must be committed to an indefinite series of complex reviews of individual deals, which the regulated parties evolve to make the task of the regulators as difficult and time-consuming as possible.^[Any acquaintance with the processes of merger approval or rule-making at the Federal Communications Commission in the United States, and the years of additional litigation that usually follow, shows the depths into which such regulators can be dragged.] Any failures of regulatory scrutiny are visited on society through the costs in diverse markets of collusion between some market participants and the transportation infrastructure providers. The alternative is to prohibit all such arrangements.  This deprives society of some value, which accrues from the small number of arrangements facially suspect, but which a perfectly vigilant regulator with infinite resources for analysis and deliberation would have discovered to be justified exceptions to the general rule. Despite all the shaded rectangles in all the graphs in all the articles by the best and highest-priced “independent” scholars, there is little reason to believe that game is worth the candle.  Blanket prohibition on vertical integration of telecommunications network infrastructure with “layer seven” services, and similarly broad prohibition on discriminatory or differential pricing–that is, the treatment of telecomms operators as common carriers–is a better and simpler solution. Some speculative welfare loss from over-regulation can be identified, more easily in theory than in reality.  Vast harms otherwise difficult to reverse and productive of significant dangers to democracy, are cheaply and reliably avoided. Mishi Choudhary is Technology lawyer and Executive Director at SFLC.IN .