Comments on the EDPS Big Data Conference

Comments on the EDPS Big Data Conference

Legal professional with strong interest in antitrust, strategy and new technology.

On Thursday 29th September I had the pleasure of attending a joint conference between the European data and consumer protection authorities, EDPS and BEUC, on big data and its interface with competition, privacy and consumer protection laws. The purpose of the conference was to facilitate discussion about large companies and their use of big data. The conference was very informative about the stance and outlook of various stakeholders, including consumer protection organisations, competition authorities and others. That said, I was disappointed by both the lack of appreciation among speakers that interfering with web platforms’ use of data can reduce the incentive for long-term innovation, and, relatedly, the overwhelming hostility towards price discrimination.

Web-platorm companies often collect data by having users sign broad and onerous privacy policies in exchange for using their services. For the most part, this personal data constitutes “payment” for the service which is “free” to the user (e.g. Facebook, Google search). Following this narrative, the firms then use the data to improve their services (e.g. Google showing you personalised results) and/or monetise their services (e.g. by showing you targeted advertising). This model is close to a socially optimal output, whereby the price of handing over data is close to zero, which is equal to the marginal cost of production. It is certainly creates a greater surplus to society than the alternative business model, whereby firms charge a positive price to consumers. In this latter model, there are welfare enhancing transactions that never take place and a deadweight loss to society results.

But enough Micro 101. This business model works because users get a great service at a very small cost (privacy harms aside). What the EDPS conference largely missed was that removing or meddling with web platforms’ ability to collect users’ data essentially removes or hinders their ability to make money, which, in turn, removes the long-term incentive for firms and individuals to innovate in the first place.

Take, for example, European Commissioner of Competition, Margrethe Vestager’s suggestion that data-collecting firms in particular industries could be forced to share data between them in order to improve all services for the customer. For instance, with self-driving cars, if all firms shared the data they collect, the operation of all cars right improve, resulting in fewer accidents etc. She noted that data-sharing could improve competition by creating a “level playing field” for all firms, big and small alike. But such a proposal would be a disaster for entrepreneurship and innovation. Under this regime, larger firms (which obviously collect more data), must immediately share such data with smaller firms. Since data is such a significant parameter of competition and revenue generator, there would be no incentive to innovate in the first place. In a similar vein, think of forcing Google to share its user-data with other search engines. Because Google relies almost entirely on its user data to generate revenue through advertising, handing over the fruits of its labour to Microsoft would entirely remove the incentive to innovate and create such a valuable service in the first place (and in fact might lead to express or tacit collusion between them with regards to the price set for targeted advertisements). This same simple intuition underlies the patent system and other legal protections for unique proprietary assets, like trade secrets, trademarks and copyrights. This was not mentioned by one of the many speakers, with the exception of Terrell McSweeny, Commissioner of the Federal Trade Commission, who mentioned it only in passing (but who nevertheless managed to earn the role of my new antitrust hero after speaking to me for 20 minutes during the break, despite me being the smallest fry there, and her the biggest).

Enforcers must also be careful when realising their determination to increase transparency for users who hand over their data. Everyone agrees that some companies’ devious tactics to get users’ “consent” to collect and use their personal data must be regulated to some degree. No one really thinks that true consent results from clicking “agree” when presented with a privacy policy that would take 32 straight hours to read. Due to a lack of transparency, many consumers do not know how, where or why data is collected or stored, and do not know whether the collection of data actually results in better service provision. But many of the ideas advanced at the conference to regulate this had a similar lack of regard for innovation incentives. For example, throughout the conference there was high scrutiny of the “take it or leave it” approach taken by companies to the acceptance of their terms and conditions and privacy policies, and many called for giving users the ability to choose how much data they hand over to web platforms in exchange for using their services, as opposed to just foregoing the service if users disagree with data collection policies. Again, we cannott lose sight of the fact that doing so would remove firms’ ability to monetise their services and therefore remove the incentive to innovate. Xavier Boutin, Vice President at Compass Lexecon, sensibly challenged the argument that consumers should be able to choose how much data they hand over. Consumers cannot self-select attributes like price or quality levels of regular products, after all.

I wouldn’t press the point so hard if the alternative business model, that of pay-walling content on the internet, was a viable option. Hakon Wium Lie, CTO of Opera (the web-browser), reckons that people are willing to pay for services so that they don’t hand over their personal data. He intends to create a “safe space” on the internet that will be free of data collection and targeted advertising. I asked how this would be monetised, and he replied that he believes users will be willing to pay to access services safe in the knowledge that their privacy is protected. I challenged that users would not realistically pay for such services when free alternatives are available at the small cost of handing over user data. Sure, if users were educated about the privacy harms taking place (or not taking place), and had perfect information such that they knew which firms are good and which are bad when it comes to the collection and use of data, their data may become more valuable to them and their willingness to pay for data-free services might increase. But, in this world of perfect information, firms would compete on privacy anyway (outside of such a safe zone). Plus, hark back to Micro 101: the pay-wall model prices certain users out of the market and creates a deadweight loss, thereby resulting in a smaller total welfare in the long-run.

My other big criticism was the conference’s hostility towards price discrimination. For those who don’t know, price discrimination is the practice of charging users exactly how much they are willing to pay for a service, thereby maximising profits of the firm. This applies to Amazon with regards to consumers, and to Google, Facebook etc. with regards to advertisers. The more data the firms collect, the better they can gauge users’ willingness to pay, and the more money they can extract. The criticism is clear: consumers pay more for products than they otherwise would. And as data collection and analysis becomes more and more sophisticated, prices will rise in tandem. But high prices have never (or, I should say, should never have been) a concern of competition law. A firm that innovates by itself to lawfully achieve market power (i.e. a degree of inelasticity of demand that allows it to charge above marginal cost) should not be punished once it gets there. Why not? Why not ban a pharmaceuticals company from charging astronomical prices for life saving drugs (where market power is derived from the state in the form of patent protection)? Because doing so removes the long-term incentive to innovate. No industry I can think of will produce new products, services or technology absent a promise that their ability to make money won’t be tampered with (cf. the open-source movement online). The U.S. Supreme Court put it best in United States v. Grinnell Corp., 384 US 563, 571 (1966):

The mere possession of monopoly power, and the concomitant charging of monopoly prices, it not only lawful; it is an important element of the free-market system. The opportunity to charge monopoly prices — at least for a short period — is what attracts “business acumen” in the first place; it induces risk taking that produces innovation and economic growth.

(Though European Law has a provision for preventing dominant firms charging high prices, it is rarely employed). So stopping firms from charging high prices endangers the necessary thrust of innovation at its source. And, because the internet has opened the world of business to global competition and forced prices down for consumers, it is ever more necessary to look at price discrimination as a legitimate method of making profit and preserving the incentive to innovate. Now, price discrimination is a controversial topic and when there’s such a direct trade off between high prices (in this case theoretically as high as possible) and long-term dynamic efficiency, there is a balance to be struck. But I was surprised that nearly every speaker at the conference who mentioned price discrimination presupposed that it was a “problem” and did not consider the effects that banning it could have on long-term incentives to innovate.

We all value our privacy, and the interface between data privacy and competition law is increasingly taking centre stage. But the purpose of competition law is to serve consumers and consumer welfare in the long-run, which includes, as part of its mandate, getting the balance right between maximising consumer surplus in the short-run and ensuring that more and more new products and services reach the market in the long-run. A necessary part of achieving this is allowing firms to charge above-competitive prices, such that they have a pecuniary interest in innovating. But this notion of long-run efficiency, along with other Chicago School concepts, were openly shunned at the EDPS Big Data conference, making me — and, I’m sure, the US antitrust practitioners and scholars in the room — hot under the neck.

Antitrust Chicago School Data European Commission Facebook Google Platform Technologies Privacy