The AT&T consent decree: in praise of interconnection only.(The Enduring Lessons of the Breakup of AT&T: A Twenty-Five Year Retr

IV. CONDUCT AND STRUCTURAL REMEDIES FOR NETWORK INDUSTRIES

United Shoe Machinery Company was only an integrated company. At no point did it sell its products to downstream competitors. A telecommunications company, however, could be in the position of competing with its customers. Indeed, that situation could arise if the original and integrated AT&T were ordered to sell services to competitors in the long-distance markets. In that case, the pricing is no longer an internal accounting device within one giant AT&T firm. Rather, it represents a transfer of real dollars between unrelated entities. The firm that overcharges the outsider could easily cover its tracks by charging identical prices to its own downstream units. The upstream firm has its own costs, which, when added to the costs of its inputs, result in a higher cost than the single firm with the integrated product. (29) This form of a "price squeeze" is notoriously difficult to prove on the facts, even in unregulated industries, because the costs needed to produce complex goods and services are not easily calculated given the prevalence of joint costs and the inevitable difficulties in calculating marginal costs. But in the competitive context, there is a respectable theoretical concern that the "equally efficient competitor" will be unable to compete at the second stage of the business because of the overcharge it faces from the monopolist that supplies it with their initial input.

The 1982 Bell decree was fashioned in large measure to avoid these serious problems, for which neither Judge Greene nor Assistant Attorney General Baxter saw any conduct remedy. But the alternative substituted one serious problem for another. Under the decree scheme, many phone calls were routed through three separate companies: one LEC, one long-line company, and a second LEC (some calls in an extended local region--the so-called LATAs, or local access transfer area--were kept within a single company). Given the strong restraints on entry, no telephone company at the time could put together an alternative end-to-end package of its own to take on the established firms. Now the pricing issues are real. Assistant Attorney General Baxter's fatal miscalculation was to assume that the risk of cross-subsidy when AT&T remained an integrated firm with long-distance competitors was greater than the risk of mispricing access charges when all long-distance carriers were independent of the LECs. His choice of the structural remedy could not function like its ideal twin in a competitive market in which the regulator is able to cut the new firm loose and let it rip. (30)

Most critically, this structural remedy required someone to regulate all the various interconnections. Yet, as everyone immediately knew, the judicial application of the antitrust law could not mount sufficient expertise and firepower to deal with the pricing and access issues that Judge Greene's structural decree forced to the fore. The ratemaking and access chores had to be tackled by the FCC, which operated in tag-team style in conjunction with Judge Greene's nonstop judicial oversight of its output, backed by a team of experts from the Department of Justice. The complex multi-stage process led to clear delays on innovation. But the bad structure of the Bell consent decree put greater pressure on the conduct remedies. The conduct remedy in this context says that the regulated party should offer correct prices to its downstream rivals, and refrain from certain practices that shore up its monopoly position. It is often unclear, though, what that correct price should be, or which practices should be regarded as anticompetitive. The antitrust law has a shot at controlling some of these issues if there are certain contractual practices that it can sensibly ban, but it is hopeless on dealing with rate issues and technological innovation, as Assistant Attorney General Baxter and Judge Greene both understood.

At this point, the question is what other course of action could have been taken in these cases which did not require the conduct remedy. In this context, the first lesson that should have been learned from the previous history of litigation under section 2 of the Sherman Act is that there is something deeply incongruous in invoking the most potent remedy--structural change--in circumstances where it is often difficult to get a clear fix on the underlying liability. Stopping cartels from colluding gives the best of both worlds, high social gain at low administrative cost, but these cases are the low-hanging fruit.

The question is what should be done with more difficult cases, such as telecommunications, where no competitive solution is possible, no matter what the terms of the consent decree or settlement. The first point to recognize here is that once we leave the AT&T monopoly model, some form of regulation will prove necessary to deal with the question of interconnections between the parties. In this regard, some of the administrative turmoil associated with the 1982 breakup of the Bell System was unavoidable. But there is still a question of how this should best be done, to which there are two potential answers. The one chosen in 1982 created seven LECs and thus committed the system to a strong separation between local and long-distance carriers. The alternative approach would not have broken up AT&T, but would have required it to interconnect on just and nondiscriminatory terms with any and all telecommunications companies.

Here are two arguments in favor of the second option. First, this very approach was slowly gaining traction in the FCC at the time of the breakup. Indeed, Judge Greene acknowledged that FCC policy was tending in that welcome direction by allowing both interconnection and resale, and sharing of AT&T services. (31) The question, therefore, was why not continue with the efforts to open up this network by taking one of these routes, preferably through interconnection. Second, this approach was ultimately embedded in the 1996 Telecommunications Act, in section 251, which creates a general duty of telecommunications carriers "(1) to interconnect directly or indirectly with the facilities and equipment of other telecommunications carriers; and (2) not to install network features, functions, or capabilities that do not comply with the [applicable FCC guidelines and standards]." (32) In order to make this system go, it is also necessary to impose a duty to negotiate in good faith over the terms of the interconnections, which the Act did, (33) subject to a set of provisions that dealt with the negotiation, arbitration, and approval of agreements, here done by state communications commissions, subject to rules and regulations promulgated by the FCC. (34) These provisions have provided the least controversy under the 1996 Act. Indeed, virtually all the confusion under that Act came from the same mode of aggressive intervention that undermined the effectiveness of the 1982 consent decree. The chief villain of the piece was section 251(c)(3), which created an "additional" obligation that swamped the stated obligation by forcing each of the carriers to sell off in bits and pieces the unbundled network elements, which created endless disputes over pricing. (35) There is no point to restate the pathology here. Suffice it to say that the statute gave a free option to the new entrant that a set of regulations were allowed to be exercised in such a manner that if the entire network were sold off piecemeal, the incumbent carrier could not recover the cost of its investment. (36) The ostensible benefit of this provision was that it prevented needless duplication of network elements. Its far greater vice was its utter inability to control the key pricing decisions. Sale at a forced valuation is dangerous even in the simplest eminent domain context; it is far more dangerous with tiny components of key switches.

The subsequent history shows the superiority of interconnection. Had Judge Greene decided to stress that theme, the entire path of the telecommunications industry would have been altered, and for much the better. The key decisions on how to expand the network would be made by market players with guaranteed entry and not through an administrative process filled with major uncertainties and lengthy delays. One key consequence of this new pattern would have been the more rapid rise of cell phone use in the United States, which market actors could have developed in connection with the existing network. This free entry approach would, in principle, have allowed entry at both the local and the long-distance level, or a combination of the two, at the discretion of the applicant, not the judge. Unfortunately, the proper treatment of mobile phones was nowhere mentioned in the backward-looking 1982 decree. (37) It therefore fell to the fine art of legal interpretation to see whether, in the end, they would be allocated to the LECs or treated as new competitive elements in the long-distance side of the equation. At the time the decision was made, the range of mobile telephony was limited so that all this business was allocated to the RBOCs, just as if it had all the characteristics of the last mile monopoly of the land phones.

Naturally, the high stakes led to extensive litigation. Judge Greene held in 1986, four years into the decree, that the LATA boundaries were applicable to cell phones and pager services, even though their technology bore no relationship to the then-dominant land lines: there is no last mile control for cell phones. (38) That decision was, not surprisingly, overturned in the court of appeals, which found no evidence in the decree or the circumstances of its negotiation of an intention to so limit cellular services. (39) But even with that appellate assist, the entire administrative path was filled with potholes, for Judge Greene still remained at the helm, a veritable one-man regulatory monopoly. Even though he routinely approved requests for waivers from the LATA provisions, all this took time (eight months for pager waivers and nineteen months for cellular waivers). (40) These numbers are very large in a world in which the useful life of some technical innovations is measured in months, not years.

Print Get the Mag Weekly Updates