188体育

Giants of digital: separating the signal from the noise and the sound from the fury

The digital giants are blowing a Schumpeterian gale through our economies.

Amazon is shaking publishing to its core and opening new possibilities for authors and readers. Google is offering routes to market for huge numbers of new competitors in every corner of the economy, while serving up a monumental challenge to the old advertisers and the business models they sustained. Apple and Spotify have transformed access to music for millions of consumers. Skype has shaken a fist at telecoms oligopolies the world over. Facebook, Twitter and YouTube have brought entirely new forms of entertainment to billions.

And the digital giants are themselves not immune from being toppled over in this gale 鈥� remember MySpace and Bebo and look at Facebook鈥檚 somewhat defensive-looking acquisitions of Instagram and WhatsApp.

The storm shows no signs of abating. Sectors so far barely touched 鈥� like finance, energy, healthcare and logistics 鈥� are braced for impact.

Schumpeter championed competition which 鈥渟trikes not at the margins of the profits and the outputs of the existing firms but at their foundation and their very lives鈥� (1 鈥� see footnotes at the end). He would have admired the ambition of Larry Page鈥檚 鈥渕oonshots鈥� that strive for improvements of 10x rather than 10%.

The gain to consumers has been huge, and the importance of healthy digital sectors for Europe鈥檚 growth and productivity is clear.

So what should be the attitude of a competition regulator?

I think we see 4 types of problems, each of which requires different forms of action 鈥� or inaction 鈥� from us.

First, there is special pleading.

There is a risk of competition authorities鈥� and courts鈥� resources being diverted by special pleading. We need to be objective in our assessment of complaints, careful to distinguish cases of abuse from the working of effective competition. A competition legal challenge might come to be seen as a last chance saloon for a desperate business 鈥� at the extreme, the claim becomes that their disruptive competitor provides an 鈥渆ssential platform鈥� and that access to their secret sauce should be made available to all on fair and regulated terms. Anyone remember DR-DOS?

In these cases, the role of competition authorities is first and foremost to separate the signal from the noise. To ask: is this complaint a genuine concern that impacts the public good? Or is it the desperate plea of a special private interest? If the latter, then the competition regulator needs to be robust enough to let Schumpeterian competition run its course.

Second come big societal questions.

Disruptive innovation 鈥� particularly in consumer markets 鈥� can unsettle social norms and throw up questions which can be couched in competition terms but actually are much, much bigger.

Take the commercial use of personal information. This raises interesting economic and competition questions. Who owns our info-trails? What are individual incentives to provide useful information 鈥� are they too great, or too small? Are there behavioural biases by which firms can trick consumers into parting with something valuable? Will competition in today鈥檚 context lead to too much sharing? Or too little? Or the wrong sorts?

But these questions cannot be fully addressed until we have developed a new social contract around privacy. Does the public use of private data have knock-on costs and benefits in wider society? Does it have special value 鈥� like life and liberty 鈥� that will make us queasy about a fully economic analysis of the questions? Until our social and political processes have digested these questions more fully, competition authorities will have to play a more modest role on these wider questions 鈥� shining a light on competition trade-offs and consequences for the quality of the consumer experience.

Third, we must assess regulatory barriers to disruption.

Where business models are stable, laws and regulations have evolved around them. But digital disruption has overturned received wisdom in many sectors and might find itself impeded by extant rules and regulations. For example, the way you regulate an energy market in a world in which men drive around in vans once a year to read meters, and consumers cannot respond in real time to prices, may need to be very different from what should happen with smart meters and responsive appliances. Copyright rules could be another such case (2).

There will be many new but as yet unimagined business models to try in the new digital world. As competition authorities we need to recognise this and seek to make sure that they鈥檙e not impeded by rules designed for a different age.

Some of the most heartening examples of enabling innovation have come from removing informational barriers to participation in innovation 鈥� for example in the human genome project or, more modestly, Transport for London鈥檚 open sourcing of its public transport status information. Competition authorities can contribute to this information liberalisation effort in the way the UK Office of Fair Trading (OFT) did with its study

Finally, we have bread and butter competition issues.

Do the giants of digital tend towards dominance and if so, what is to be done?

Peter Thiel, PayPal co-founder, venture capitalist and libertarian evangelist has an answer to this question that鈥檚 a red rag to the competition authority bull:

鈥淐ompetition is for losers,鈥� he writes, 鈥淸鈥�] All happy companies are different: Each one earns a monopoly by solving a unique problem. All failed companies are the same: They failed to escape competition.鈥�

Matthias Machnig, Secretary of State at the German economics ministry, has a different answer: 鈥淲e need a review of our competition law because the classical competition law is no longer adequate for what is taking place in society through digitalisation.鈥� Axelle Lemaire, French minister for digital affairs, wonders whether 鈥渢he regulatory framework, especially on antitrust, allows us to respond to abuse of dominance鈥�. And the EU Parliament has expressed loudly and clearly the widely held belief that 鈥渟omething must be done鈥�.

As a counterweight to the Silicon Valley Utopians, let鈥檚 consider the opposite, a dystopia that everyone can agree would be a failure. The useful question will then be how far from that world is the one we live in today.

Please follow me in this short flight of fantasy to a possible (unwished-for) future, say in 2050:

Like everyone else, when I want something, I 鈥楤ynglezon鈥� for it.

Some years back, I鈥檇 sometimes use other services for shopping, like eBay, Bing, Google or Amazon, but they either merged to create Bynglezon or just disappeared from my screens.

That鈥檚 a pity, because I quite liked the way some of those other services did it. And they probably kept Bynglezon honest, too.

When I Bynglezon something, I get presented with a wonderful array of choices, including many things that I didn鈥檛 know I wanted. But, once informed, I often find them pretty irresistible.

I鈥檝e happily given Bynglezon access to my bank account, my smart electricity meter, my thermostat, my tax returns, my mobile phone contacts, my car black-box, my health monitor. They handle all the messy, annoying bits of paper work that used to clutter up my life. I like the simplicity of it, and whatever criticism one might level at Bynglezon, they do make life easier.

I am told (by an economist friend) that they make their money from collecting data about consumers and about goods and services; they then match goods to desires 鈥� and they鈥檙e better at that than anyone has ever been. They then sell sales leads to companies. Apparently they charge each listing a carefully determined fee that is exactly equal to the maximum willingness to pay for the positional listing given what everyone else is paying. No supplier can afford to ignore this essential channel to market. Of course, the cost of the Bynglezon leads are passed through to final product prices.

Whenever I get a result from them, it certainly seems as if everyone鈥檚 competing hard for my custom, although prices don鈥檛 seem to have come down.

Bynglezon are said to be excellent employers. They make huge profits but they do seem to spend their money on worthwhile things.

OK. End of fantasy. If we ever get to that place, then the competition authorities will have failed.

So what are the essential features of this dystopia that demarcate it from the current reality of the giants of digital?

There is no competition between platforms in the dystopia, whereas there is competition between platforms today, at least on certain parameters. Indeed the platforms argue that there is an intense battle between Amazon, Google, Facebook, Apple, Twitter and any number of other more specialised platforms for the attention of customers; and that this battle for eyeballs is driving competition between platforms and forcing them to offer customers on both sides of their markets ever more and better service. They argue 鈥� with at least some justification 鈥� that platform competition is creating value and driving further innovation.

On the other hand, there do seem to be forces driving the new business models towards natural monopoly or natural oligopoly. The more we use particular online platforms and associated products, the harder it may be for other firms to come up with a competitive offering, because quality improves with use.

And this happens on both sides of a platform. For example, search results are improved for the searcher the more information it has; and click-through rates are improved for the advertiser the better a link is targeted. The giants of digital may supplement these informational and network economies of scale with other arrangements, like a physical logistics network that might have similar scaling properties.

We need to remember that these are still early days. Innovation continues to move very fast in these markets, and we may yet find the giants toppled by new disruptors or by other giants. Equally, we need to keep in mind that the costs of 鈥榝alse positives鈥� 鈥� premature, unmeritorious interventions 鈥� are likely to be quite high, given the basic back story of welfare-enhancing innovations.

But let鈥檚 work with this hypothesis for a moment and assume that there is a tendency to natural monopoly or oligopoly and that it may be persistent. What then should we do about it?

Let me start with an example of a recent decision we made at the CMA related to price comparison websites (PCWs) for private motor insurance (PMI) policies. PCWs are a feature of the web in the UK more than in any other country 鈥� they are specialist search platforms helping private consumers find complex products 鈥� like insurance 鈥� and helping upstream suppliers find consumers. They have had a phenomenal impact on competition between insurers. There are 4 large PCWs in the UK motor insurance market, all of which have become household names through extensive advertising. On the surface, this seems like a straight win for competition and consumers. But our investigation showed that some of the insurers had at least one price parity clause in their contracts with PCWs, restricting competition on price between platforms. Our analysis showed that although competition was fierce between insurers, a good part of the benefits of that competition might never get to consumers.

The same logic might come to apply to a digital giant. It might stir up competition on one side of the market 鈥� as in the way that Amazon puts publishers in competition not just with each other, but also with self-published authors 鈥� but fail to pass on its benefits to the other side, the end consumers.

What if 鈥� hypothetically 鈥� one of the digital giants were to come to exercise market power for its own account? We can imagine a scenario where consumers come to them and there is the appearance of a competitive marketplace, with a multitude of offers for goods and their close substitutes. But in reality, the platform is extracting all the rent from that competition through an unchallenged control over listing fees and sales commissions.

What should competition authorities do in the face of such a possibility?

Because competition between platforms restrains platforms from exercising dominance, behaviours that threaten to restrict vigorous inter-platform competition should be challenged. This is exactly the approach that the CMA鈥檚 predecessor organisation, the OFT, took in its investigations into price parity clauses in Amazon鈥檚 鈥楳arketplace鈥� third party seller platform and, before that, into the agency arrangements between Apple and the publishers relating to ebooks. In both cases, the OFT intervened promptly, which is of course highly desirable in such fast-changing markets. In the first case Amazon decided to withdraw the particular clauses across the (3). In the second, given concurrent European level work offering strong prospects of a comprehensive resolution of the issues, the OFT closed its case, passing the matter on to who then brought to successful conclusion.

Price parity clauses and restrictive agency arrangements are just two potential sources of unnatural barriers to entry. There could be others. Bundling, foreclosure, defensive acquisition, and predation could all be used to turn innovative oligopolies into sclerotic natural monopolies. They need case-by-case consideration. But it must be remembered that an economy in the throes of creative destruction needs to be kept on a knife-edge (a precarious place to be, especially during a gale). Markets needs to contain enough promise of profit to spur innovation, while being competitive enough to keep strong incentives to continue to innovate and serve customers.

This creates both an opportunity and a danger for competition regulators.

Consider the role of antitrust in the last two mega-storm cycles that IT has produced, with the IBM and Microsoft cases (4). It is easy to think in both cases that antitrust was too late, or even that it got the wrong remedies which ran the risk of entrenching the status quo. And that is clearly a danger for competition authorities: the storm moves too fast, and before we know what鈥檚 happened, it鈥檚 already after-tomorrow and we have a useless 鈥� or even harmful 鈥� solution to yesterday鈥檚 problem.

But that鈥檚 too one-sided a view of those cases. It was an antitrust-weary IBM that ceded control of the personal computer. IBM executives learnt from experience who it was they did not want to spend any more time with 鈥� people like us. Microsoft executives may have felt similarly battered in the mid to late 1990s when small-scale product design decisions crucial to the maintenance of an open web were made.

I am sure that today鈥檚 giants of digital, with platforms that have the potential to become natural monopolies, are fully aware that actions which risk serious harm to inter-platform competition will be scrutinised carefully by the competition authorities.

So back to Peter Thiel. Is competition 鈥� and by inference, competition policy 鈥� really for losers? No. I believe that competition authorities are essential to keeping our economies on the knife-edge that鈥檚 so important for innovation, dynamism, jobs, consumers 鈥� and to being ready to solve humanity鈥檚 great challenges of tomorrow.

What exactly is this knife-edge?

When markets don鈥檛 reward innovation and investment, you get too little of it. But when once innovative firms are entrenched in their positions and do not face credible threats to their comfortable positions, they will also tend to pass up great opportunities for innovation. So you have this 鈥榠nverted U-shaped鈥� graph for competition and innovation (5).

As I am trying to separate the signal from the noise, let me end with a cautionary tale from the pioneers of digital signal processing: AT&T.

Founded in 1874 on the basis of Alexander Graham Bell鈥檚 patents on the phone system 鈥� Web 0.1, as it were 鈥� AT&T grew through network economies of scale and acquisition into the vertically integrated behemoth that was broken up after 100 years of dominance in 1984. By the time of the break-up, AT&T鈥檚 Bell Labs had notched up 4 Nobel prizes and had to its credit some of the great pieces of technology that have powered the ICT revolution: the transistor, the laser, the UNIX operating system, C, C++ and many digital signal processing breakthroughs.

But extracting the value from all that invention and research mostly fell to other companies. At the time of break-up, AT&T was generating piles of cash, producing fabulous research, but failing to deliver innovation to consumers.

In that one story, we have the essence of the knife-edge. AT&T had been innovative and a great R&D investor. But it had so little fear of any disruption to its basic business that improving the lot of consumers got hardly a side glance in its strategy.

Did antitrust contribute to welfare and innovative dynamism through the1984 break-up of AT&T? Almost without a doubt, as even Peter Thiel would agree.

Should it have been done any earlier? Probably yes.

That鈥檚 why we need vigilant and agile competition authorities, that while careful not to intervene too soon, don鈥檛 leave it too late either (our own knife-edge).

And competition authorities with the ability to sort real market distress signals from all the noise of creative destruction; and to tell the sound of actual market failure, from the fury of the out-competed.

Footnotes

I would like to express my warm appreciation for the help of Tony Curzon-Price and other CMA colleagues in preparing these remarks.

1. Joseph A Schumpeter, 鈥楥apitalism, Socialism and Democracy鈥�, 1942.
2. Ian Hargreaves, 鈥楧igital Opportunity, A Review of Intellectual Property and Growth鈥�, 2011.
3. The Bundeskartellamt had a parallel case in Germany.
4. The latter case, it must be said, generated a debate among economists that is very similar to the one we鈥檙e having today: if innovation comes in waves, maybe the debate about policy does also. A classic 鈥� Microsoft-inspired 鈥� recommendation of an antitrust 鈥榣ight touch鈥� in the early years of a wave of innovation is provided in David Evans & Richard Schmalensee鈥檚 , 鈥楽ome economic aspects of anti-trust analysis in dynamically competitive industries鈥�.
5. Philippe Aghion, Nicholas Bloom, Richard Blundell, Rachel Griffith and Peter Howitt, .