High-speed broadband networks are key to the growth of digital markets as well as most modern forms of communication, and have been subject to far-reaching regulation in many countries. In this piece, we’ll review the rationale behind a cornerstone of the prevailing regulatory paradigm: forced access to incumbent operators’ network infrastructure by alternative operators on regulated terms, so-called “unbundling”.
Underlying the regulation of broadband access networks is a presumption of “natural monopoly”, which means a market on which the “efficient” firm size is so large relative to demand that the “efficient” market configuration is monopoly, owing to large fixed costs which are mostly sunk. Consequently, in an unregulated world, there would be either “ruinous” competition, or – in case one firm has a significant head start, perhaps by virtue of being an erstwhile state monopoly – no competition at all. The European telecommunications landscape, for instance, was characterized by recently privatized, typically vertically integrated state monopolies in the 1990s, and the preconditions for sustainable competition were, according to the natural monopoly paradigm, found wanting, resulting in a perceived need for regulation to prevent overpricing.
In theory, assuming sufficient information about costs and demand (a bold assumption, see below), a retail price cap could have been used to address concerns of monopoly pricing. But regulators wanted more. Their idea was to create retail competition between internet service providers (ISP) by forcing the incumbent infrastructure owner to grant access to its network upon request, typically at a regulated price, thus allowing competitors to the incumbent’s downstream branch to offer their own internet services to end customers. Hence the term “unbundling” – a production chain is split into a wholesale and a retail component, where the wholesale segment is presumed to be a natural monopoly, at least initially, and the retail branch potentially competitive. Such forcible downstream competition to self-supply was considered desirable and conducive to innovation and beneficial product differentiation.
A compounding justification for unbundling derives from the so-called “ladder of investment” hypothesis, known in the US as the “stepping stone” theory. The idea is that unbundling allows entrant ISPs to build up a customer base, thus reducing demand uncertainty, as a stepping stone to building their own competing infrastructure. Regulators are supposed to push the entrant ISP in that direction by making access to incumbent infrastructure less and less attractive (read more and more expensive) over time. The assumption here is that the presumed natural monopoly problem can be engineered away step by step with the help of a rational regulator possessing of all relevant information and the right objective function. At the end of this process, regulation would then become unnecessary, allowing the regulatory agency to abolish itself.
How has the unbundling approach fared so far? Certainly, alternative operators have exhibited a robust demand for the regulated wholesale access product. In a purely numerical sense, the number of alternatives for consumers on a given incumbent network has increased. One can also argue that quality of service has improved since the beginning of the unbundling era, although it is difficult to attribute this causally to unbundling over other factors (technological progress, competition from mobile networks, etc.), and quite conceivable that incumbents would have improved or differentiated their services even absent forced retail competition. Meanwhile, the “ladder of investment” approach has been criticized theoretically and, more importantly, not been vindicated empirically, as incumbent regulation seems to deter rather than stimulate entrant investments (link 2, link 3, link 4). ISP relying on wholesale access have settled nicely into their customer role and displayed little initiative to build their own competing networks, even citing the availability of regulated wholesale access as a reason for not doing so. Ironically, therefore, unbundling may in some cases have perpetuated incumbents’ infrastructure monopolies, thus achieving the exact opposite of its stated objective.
One cannot immediately conclude that this invalidates the “ladder” approach as such since national regulatory authorities in many countries have in fact been reluctant to follow its prescriptions by making access conditions less comfortable over time. Nonetheless, to the extent that such reluctance is structurally rooted in the political economy of regulatory practice, one can at least say that either the postulated theoretical mechanism is weak or non-existent, or a crucial assumption underlying it is unrealistic.
More generally, it is unclear whether even a “correctly” implemented stepping stone approach – or, for that matter, retail price regulation – would improve long-run prospects for innovation and competition compared to laissez-faire. The very concept of natural monopoly has been subject to criticism for theoretical shortcomings and implausible assumptions (link 1, link 2, link 3, link 4). Empirical research has pointed out that monopolies that are not protected by government are temporary and tend to erode rather quickly. Industries that are often presumed natural monopolies tend to be surprisingly competitive when left to their own devices. Standard procedures defined in legal frameworks to establish “dominance” of particular firms run a serious risk of defining relevant markets too narrowly (link 1, link 2), underestimating the relevant demand elasticities and ignoring relevant substitutes.
Firms tend to organize in response to transaction costs, and splitting up the production chain to force them into a particular business model almost certainly has a direct efficiency cost. Further, a regulatory commitment to expropriating hypothesized monopoly rents and guaranteeing competitors regulated access to existing infrastructure discourages investment into new and more efficient networks and business models. In fact, and in contrast with received “perfect competition” wisdom, it stands to reason that temporary private monopolies benefit society. Among other things, they (a) maximize incentives to innovate and produce a product or service in the first place, (b) internalize demand externalities between close supply substitutes as well as network effects, and (c) lead to lower, not higher prices in the long run as markets with winner-takes-all characteristics put a huge premium on cost reduction through efficiency gains (link 1, link 2, link 3, link 4).
It should be noted that even conceding the possible existence of natural monopolies would not automatically imply that regulated access would be advisable. Regulators face a formidable knowledge problem vis-à-vis the regulated operators which gives rise to problems of asymmetric information in rate setting and moral hazard in choosing the level of productive efficiency (link 1, link 2), as well as the well-known risk of regulatory capture. Moreover, from a public choice perspective, regulators may have career and other incentives which conflict with the public interest (assuming the latter can be coherently defined). Ignoring these issues means running into the so-called “nirvana fallacy” of comparing market outcomes to an unachievable first-best rather than a feasible alternative.
In sum, while network unbundling may have brought about a wider variety of choices on a given network, it is far from clear that the approach is in consumers’ long-run interest. Notably, the United States, which started unbundling with the Telecommunications Act of 1996, largely abandoned the practice for broadband networks in 2005. In Canada, Australia, New Zealand and many European countries, on the other hand, the policy remains firmly in place.