Separating telecoms? By Eli Noam Published: May 15 2009 23:42 Last updated: May 15 2009 23:42 A spectre is haunting European telecommunications companies: the prospect of being forcibly reorganized by governments through ‘functional separations’. Brussels and several countries want the national telecom incumbents to separate their infrastructure network from the service riding on it, which would be open to competitors on equal terms. BT in the UK was an early example, and the idea has caught on. The basic concept would be like requiring railroad companies to separate their tracks from their train service, to permit rival transportation companies to use these tracks at a low and non-discriminatory rate, and to lease to rivals selected parts of the track infrastructure. All this sounds sensible. It is promoted by competitors and internet companies. I advocated it myself for many years. But more recently I have come to have doubts, not from indulging in laissez-faire but from counting the transaction costs of such an approach, and from studying the historical record. The problem with discussing all this is that the debate has come to take on some aspects of a religious war, with true believers on each side issuing fatwahs on anyone who strays from their path to redemption. The truth, as often, is more complex. One needs to start, first, with the recognition that incumbents indeed often carry significant market power in some essential infrastructure and could capitalize on it through high prices, discriminatory practices, and gatekeeping behaviour. Recognizing the problem, however, does not mean that one particular remedy– functional separations –is the best way to deal with it. Second, one must recognize that the source for such market power is not nefarious behaviour or co-opted regulators. The basic issue that has always plagued telecommunications and its offshoots like cable TV and now the internet are the underlying economics of network provision, in particular the strong economies of scale and network effects which favour large providers. Thus, given the great importance to society of free flows of information, there is an eternal tension between the efficiency of large scale and the benefits of open competition. Given such fundamental tensions, separations policies have had a long history, and nowhere more than in the United States where private and part-competitive telecommunications go back for over a century. Brussels, seeking a ‘European solution’, should make the effort to learn from the American experience. Over the past decades, the US has tried almost any variant of separations possible – functional separation, lines of business separation, geographic separations, market structure separations, accounting separations, full operational separation. Yet none of these resolved the fundamental problem of market power. Eventually, the US was pushed to the most radical separations policy possible – the AT&T Divestiture 25 years ago, which split up the world's largest corporation into eight parts. But this hardly solved the problem. In the past, I strongly supported divestiture. But more recently I have studied the data on the actual performance of the US in comparison with Canada, which started out as a comparable system but did not institute a structural separation. Twenty-five years later, the US system had less of local competition, of R&D, sectoral growth, employment increases, and growth in market capitalization. Consumer prices in the US were comparable, and business and wireless prices were lower. The structural separation may have made sense in theory, but the numbers do not substantiate the benefits in practice. The problem was that the US divestiture was not content with a size reduction of a dominant company but tried to segment the company along the lines of competitive and non-competitive sub-markets. This kind of complexity led to ceaseless battles. In those 25 years, the US telecom industry was at war with each other, often less in the marketplace and more in the regulatory arena. Billions were spent in seeking approvals or seeking delays. Billions were spent in litigating and advocating the precise definitions of markets and permissible activities. And since telecommunications is in a highly dynamic phase of change, these battles never ended. What they did, however, was to slow down the introduction of innovation. Partly as a result, the US telecom sector which was by far the world's most advanced a quarter century ago is now just a decent also-ran. The Brussels approach will have a similar outcome. The definition of what can and cannot be done by separate parts of the same company will be a never-ending source of controversy. Nor will a separation resolve the problem of monopolistic prices that are non-discriminatory. Nor will it resolve gatekeeping power over content, as long as the same restrictive practices are exercised across the board. So these matters would still have to be regulated. If those prices are too high, smaller competitors will fail and complain bitterly. If the prices are low, on the other hand, there will be little incentive for the network providers to invest in upgrades, since they will in effect be asked to subsidize their competitors, while assuming the full risk. Furthermore, with low prices of usage of the incumbents' networks or elements thereof, there will be less incentives for rivals to build their own networks. Thus, regulators will be thrust forever into highly political balancing acts of conflicting goals. On the other hand, users and rival service providers will benefit from functional separation, at least in the short run. Lower network upgrades will affect them only later. So the question is whether one can achieve the positives with fewer negatives. The answer is yes. Separation is a tool, not a goal. There are other tools to achieve the same legitimate objectives, and they would be simpler and cheaper. The simpler way is to seek non-discrimination directly rather than through fiddling with a company's structure. This could be done through regulation and legislation, but a more effective way might be through contractual arrangements that are specific to a company. The time has come for incumbents, competitors, and regulators to fashion a grand bargain. In it, they will drop separations which are not a goal but a tool–in favour of specific incumbent commitments to further the goals that a separation tries to achieve: a. accelerated network investments, especially for rural areas. b. complete neutrality towards any kind of content from whatever source. c. terms of non-discriminatory practices towards rival companies' use of their networks. In addition to the carrot of non-separation, there is a stick. If these negotiated compacts would not be met in letter and spirit, then Brussels and European states should proceed speedily to an anti-monopoly challenge, seeking the full break-up of the offenders (and in ways simpler than in America a generation ago). Such a threat, credibly presented and administered, would achieve more than the awkward system in which separated parts of the same company operate quasi-independently yet report to the same top management, under a heavy regulatory regime. Such an arrangement of 'social compacts' will leave practically everybody unsatisfied. Competitors and consumers will point out that the market power of the incumbents would still be considerable. True. But its exercise would trigger the latter's dismemberment, which is a powerful incentive for incumbents to toe the line. And the commitments, if met, would benefit consumers and competitors. Incumbents, on their part, will prefer to be left alone entirely. Yet this is not likely to happen. The free flow of information has become too important to the information economy and society. Today is a good time for policy makers and companies to shape alternative futures, and the American history can provide both positive and discouraging lessons. It suggests that structural solutions, while intellectually appealing, create major transaction costs and retard network evolution. There are better ways to protect users and competitors. The writer is professor of finance and economics at Columbia University. Copyright The Financial Times Limited 2009 FT Online