The internet is not a utility, and should not be regulated like one (Part 2: unbundling and platform regulation)
Unbundling is a term borrowed from the regulation of, most saliently, electricity and telecommunications companies. As part of the liberalisation of these markets, governments across the world have introduced provisions requiring erstwhile (almost universally state-owned and -operated) monopoly providers to offer new challengers access to certain parts of their network at set prices.
The thinking is that these markets exhibit the features of a natural monopoly, where the investment and cost structures would make it most efficient to have a single provider. A typical example is the electricity grid, where fixed investment costs (building the grid) are high but marginal costs are relatively low. In that context and for a given (small enough) market, it is least costly in the long run to have one single grid transmitting electricity to customers, rather than several competing ones.
But, the reasoning continues, natural monopolies have wide-ranging, and potentially negative, implications for the markets in which they operate. As an illustration: BT owns the telephone network, which other providers need to access in order to offer communications services to their customers. At the same time, BT is competing with these very same providers for telephone and broadband services. Absent government regulation, what is to prevent BT from charging its competitors prohibitively high rates for access to its network, and using the proceeds to subsidise its phone and broadband operations, thereby driving its competitors out of the market? These are the anti-competitive practices feared by proponents of unbundling, because BT, having bankrupted its competitors, could then go on to increase prices, harming consumers. Unbundling, it is argued, is needed to prevent BT from engaging in such practices.
Sceptics of regulation have convincingly pointed out that in the absence of legal barriers to entry – from the state’s sanctioning of the firm’s monopoly to entry requirements which raise costs for newcomers – the nightmare scenario of ‘price dumping’, followed by sharp hikes and a permanent monopoly, need not necessarily take place. But let us take the case for unbundling as given, and apply it to internet platforms, specifically to Google, which far surpasses its competitors in its market share for search engines. According to proponents of regulatory intervention, Google is currently able to use its dominance in search to direct users to its own services: Google Maps, Google+ (its social network), Google Books, Google Flight Search and Hotel Finder, Google News, and many more. This would come at the expense of competing providers of the respective services, such as Expedia, TripAdvisor, online news sources, Facebook, etc.
Thus in Google’s case, the equivalent to BT’s telephone network would be its search engine, which customers use to find information on airline fares, the weather, news reports, and so on. But these services are also offered by Google itself, which (like its competitors) derives advertising income from traffic to its sites. By using its search dominance to drive traffic to its own products, Google would drive users away from its competitors, making them economically unviable and establishing a monopoly not just on search, but on the myriad other services that rely on web browsing.
It might seem, then, that Google is able to engage in the kinds of anti-competitive practices that only unbundling can solve. Google’s search engine needs to be unbundled from its other services to ensure a level playing field, proponents argue.
But not quite. This is framing the unbundling issue in a fundamentally flawed way. The question is not whether Google can use its search engine to drive demand for its other services. That is just a product of economies of scale and convenience, in the same way that BA can offer full travel packages to customers who were initially just looking for a flight, and that John Lewis can sell towels and espresso makers to customers who were originally just looking for a new pair of shoes. This makes economic transactions less, not more costly, and more, not less efficient.
The real question is whether Google’s ability to offer multiple services via its search engine has adverse consequences for competition and consumer welfare. This does not appear to be the case.
First of all, the dynamics of a natural monopoly that provide the rationale for unbundling regulation are simply not there. Yes, Google has established itself as the dominant search engine, and it may even have used this dominance (as well as the prestige and name recognition coming from users’ interacting with it dozens of times a day) to successfully compete in multiple other markets. But this decidedly does not preclude competitors from entering the search market. The high fixed costs and low marginal costs characteristic of a natural monopoly do not apply to Google’s business model, where the adequate maintenance of its search algorithm to offer accurate and timely results is a key ongoing investment. And indeed new competitors have emerged in recent years, seeking to outperform Google in many different ways, catering specifically to e.g. tablet owners, or those particularly concerned about privacy.
Secondly, the relationship between Google and its competitors for internet services is more complex and interactive than that between, say, electricity providers and the owner of the grid. Like the latter, Expedia, TripAdvisor etc rely on Google’s search engine for traffic to their websites. Yet Google also relies on their success to keep and expand its market share – and, critically, for the advertising revenue that is such a fundamental part of its business model. The main reason Google is so popular with web users is because its algorithm is seen as exceptionally good, getting users the information they are looking for quickly and effectively. But if Google started offering results purely on the basis of whether or not they were Google services, rather than on the basis of pertinence, users would quickly lose trust in the search engine and turn to alternatives. This means that Google has a stake in the continued success of many of its competitors, decisively weakening the incentives to promote its own products exclusively.
(Of course, it could well be that some Google services, now or in the future, are better than even its specialised competitors, in which case increased traffic to Google services and a better ranking of them by its algorithm should be expected to occur. But this would not be due to anti-competitive practices, and it would only remain the case so long as those Google services outperformed the competition.)
Finally, one of the internet’s revolutionary features is that it makes available to users all around the world not just the general information that one used to be able to glean from textbooks, travel guides, newspapers, etc. but also highly specific, localised and individualised insights that no single physical source could feasibly convey. In the same way that Wikipedia is continuously built through the dispersed knowledge of its thousands of contributors, following the insights of Hayek’s seminal paper ‘The Use of Knowledge in Society’, the internet caters to the needs of billions of users in a remarkably personalised way. Its ability to offer users the information they are looking for (and nothing more) makes it an enormously efficient vehicle for economic and social interaction. The upshot is that there is an ongoing demand for ever more specific, niche information – the kind of information that a large collection of small specialised sources can usually best provide. A big conglomerate like Google may be able to use economies of scale to cut costs, but there is little reason to believe that it could substitute for the gazillions of sources currently populating the web. Ten years ago, Google might have been likened to a big fish in a lake – now it’s more like a whale in the ocean. And as the Internet carries on growing organically, Google’s relative size can be expected to shrink, even while it becomes bigger in absolute terms.
Efforts to regulate the web seem to be motivated by a mixture of good intentions, a lack of understanding of the processes of innovation and market competition, and a misguided desire to shape the future of the internet in the same way in which industrial policy in the 1960s sought to shape the future of industry. Yet, just as in the 1960s, attempts to design a complex system that is built on trillions of daily human interactions are guaranteed to fail – with the addendum that the economic justifications used for the regulation of utilities in the last century do not apply to the digital sector.
Policymakers would be well-advised to leave the internet well alone. Over the past two decades, it has revolutionised economic life in a fundamentally benign way, and on a scale unseen since the great industrial breakthroughs of the 18th and 19th centuries. Well-meaning efforts to address perceived unfairness and alleged market failure, from net neutrality to unbundling, are highly likely to result in the opposite of what proponents intend. They are also built on mistaken notions of what drives the creative process and what constitutes market abuse. When it comes to internet technology, it is more urgent than ever for politicians and bureaucrats to follow that old but oft-ignored dictum: “If it ain’t broke, don’t fix it.”