A long thread summarizing the fascinating launch event of our new @cepr_org #Competition_RPN on privacy & antitrust organized by @Caffar3Cristina. Economists wanting to impact first-order antitrust policy issues, focus on 40/onwards! Details/videos: bit.ly/35CRbEu.
2 panels. 1st w/ privacy experts sharing the data protection/privacy harms they see that aren’t being built into antitrust theories of harm for data-focused digital giants. 2nd w/ a fabulous panel of agency leaders describing (growing) progress on integration. 2/49 
@johnnyryan leads off [my clarifications/interpretations in brackets]: “We have 2 crises: a privacy crisis and a market [power] crisis. The two are related and in fact compound each other.” [Right!] 3/49
Emphasizes GDPR contains powerful tools to help [both], but they are not be enforced. Most important: Purpose Limitations == prevents any firm from re-using data collected for one specific purpose for any other purpose unless exacting requirements met. 4/49
Absent purpose limitations, [data giants] offensively leverage data from market to market by combining and cross-using it, both between subsidiaries and within subsidiaries across purposes. --> “Cascade their monopoly from market to market.” 5/49
“Q: What does this have to do with antitrust?” 2 answers: (1) When weak enforcement let [ex post anticompetitive] mergers through, it made it harder for data protection enforcers & (2) Data separation would bring more competition into the market. 6/49
@jason_kint next: Had high hopes for GDPR at the start, esp (1) users could choose not to be tracked outside a company’s existing apps [e.g. not across every page w/ a FB “like”] and (2) Purpose Limitations would limit dominant platforms’ ability to tie together verticals. 7/49
How does this harm consumers? (1) Publishers missed out on a chance to rebalance market power relative to platforms and (2) Forced the bar lower for *all* publishers. Told first-hand the story of Google’s broken promises w.r.t. Doubleclick data. 8/49
What can be done? Limit data collection/use across services, so dominant platforms can’t leverage market power in a single dominant product (e.g. Gmail or Facebook) to collect and use data across a portfolio of services. 9/49
Then @dinasrinivasan: Data rules should automatically be part of competition conversations. Many centralized electronic markets have access to data regulated b/c of conflicts of interest: financial exchanges, reservation/ticketing systems. [Relevant for adtech/ad exchanges] 10/
Again Google/Doubleclick: as Google gained market power in adtech, no regulatory requirement to act in best interest of clients, so switched data ownership from advertisers/publishers to themselves, consolidating and building market power into adjacent markets. 11/49
Emphasizes very poor experience in digital markets with behavioral commitments. Analogies again w/ financial markets: if there are conflicts of interest, divest assets or set v. strong firewalls/decouple compensation, etc. Andrea Coscelli in 2nd panel agrees. 12/49
Next Simeon Thornton @CMAgovUK on provisional commitments for Google Sandbox in the UK. [Google Sandbox == Google’s plans to remove third-party (3P) cookies from Chrome browser] 13/49
CMA worried about Google Sandbox changes distorting competition for third parties while keeping capabilities for itself & dressing it up as better privacy (“Privacy Washing”). 14/49
Emphasizes case jointly developed with UK’s DPA (@ICONews) and data protection principles are at the heart of the commitments. If accepted these would be legally binding, based on principles (to avoid loopholes), with CMA’s ongoing involvement. 15/49
Economic insights from Alessandro Acquisti (@ssnstudy): considerable research measuring consumers’ dislike for lack of consent and platforms’ exploitation of informational asymmetries and cognitive/behavioral biases. 16/49 papers.ssrn.com/sol3/papers.cf…
What about (lack of) privacy as a price? Could in principle be measured but not happening enough. Consumers pay prices with their data, but value proposition for consumers remains opaque: is it commensurate to the services and benefits they are getting back? 17/49
Conventional theory says more data is win-win, but little evidence for that. Platforms clearly winning, but behavioral targeting a zero-sum game for advertisers, own research shows very small benefit to publishers, & close to nothing known about consumer benefits. 18/49
Why so little known about consumer benefits and potential harms? Economists struggle to identify, measure, and aggregate all the many potential harms (inc. non-price discrimination, physical/emotional harms, [general lack of knowledge of how own data is being used], etc.) 19/49
The “Paradox of the Privacy Paradox”: lots of evidence that consumers demand privacy/engage in “privacy protective behaviors.” Emphasizes multidimensionality: privacy is a dynamic process of opening and closing, thus evidence of disclosure != lack of tastes for privacy! 20/49
On to the 2nd panel w/ many of the leading thinkers worldwide on integrating privacy/data protection issues into antitrust! 21/49
Andrea Coscelli, CEO of @CMAgovUK up first, making the case for historic underenforcement w.r.t. mergers, the failure of behavioral commitments to protect consumers, and sharing lessons from recent Joint Statement by @CMAgovUK and @ICONews. 22/49 gov.uk/government/pub…
These are: (1) Enforcing data protection rules will help level playing field, (2) Joint oversight in Google Sandbox novel and important, and (3) Ongoing collaboration important, e.g. w.r.t. data remedies and evaluating choice architecture. 23/49
Follow-up: how can we build data protection concerns into theories of harm? It’s important but appetite to take risks differs across agencies; this a process of learning by doing; Privacy Sandbox a good first example of what collaboration can look like and yield. 24/49
Elizabeth Denham, head of @ICONews, the UK DPA, next: existing statement and Google Sandbox work a long time in the making. Urgency to this work: “We have to get out of our silos. Policymakers expect us to be working together!” 25/49
Andreas Mundt, head of @Kartellamt, next. The novelty of German Facebook case: (1) combining data into superprofiles (a violation of DP law) treated as an abuse of dominance (a violation of competition law) & (2) reduced consumer choice treated as a loss of competition. 26/49
Case now progressing through European legal system – looking forward to definitive answer to key questions referred to the ECJ, all about combining data protection and competition law. Shifts to new Section 19A of German Competition Law. 27/49
19A similar to proposals in UK’s DMU & EC’s DMA. Access to data critical to competition a core pillar, e.g. when dominance arises due to data collection/processing, @Kartellamt tp determine what is allowed/not (and can rely on GDPR for insights!). 28/49 bundeskartellamt.de/EN/Economicsec…
Justifications possible for companies; appeals go directly to German (Federal) Supreme Court. New proceedings already against Google, Amazon, and Facebook under 19A. Bottom line: Agencies have to be innovative! Persuading courts hard but important. 29/49
To the US, first with @RKSlaughterFTC on FTC’s tools to coordinate competition and consumer protection concerns. Dissented on 2019 Facebook settlement as was skeptical it could have disciplining effects on how Facebook treats data & privacy. 30/49
Goal to get at “root causes” of market power. Breakups possibly yes, but also guardrails on data collection/use/sharing. “User” control (e.g. notice and consent) will *not* solve the problem; need to shift burden to *firms*. Expect to need to fight in court! 31/49
Then to @chrismdangelo, contributing architect of State AGs Facebook complaint. “This is a policy debate we need to be having. There is nothing more important than this integration of privacy and competition [right now].” 32/49
Facebook complaints (FTC and State AGs) document how Facebook’s data practices worsened from a consumer perspective as they gained market power, both degrading consumer controls and increasing amount of data collected and how it was exploited. 33/49 ag.ny.gov/press-release/…
State AGs complaint establishes Facebook’s dominance, but also that its changes in data collection practices resulted in consumer harm comparable to “an increase in price in the traditional way you think about antitrust and competition law.” 34/49
Re: agency risk appetite, complaint designed to resonate with a US finder of fact [i.e. judge]. Harms from abuse of data and privacy have that effect. 35/49
Back to Europe with Wojciech Wiewiórowski, European Data Protection Supervisor (@EU_EDPS), who gives an important “reality check.” EPDS is supervisor of DP institutions and main advisor on legislation. Candid view: was more optimistic in 2014 than now (!). 36/49
Sees hesitancy from both DPAs and CAs. Cooperation not fast enough and no effective competition or protection of individual rights in digital economy until it happens. Each sees value of a little help/understanding, but real integration??? 37/49
Close with Henri Piffault, VP at @Adlc_. Cooperation with DPA (CNIL) very active in France, as witnessed by recent Apple case which considered whether providing users with Apps’ data collection practices w.r.t. advertising could disadvantage its rivals. 38/49
Rejected interim measures, deciding rules gave users more control, weren’t forbidden by regulation, and were proportionate (although case continues). Also active in @IntCompNetwork Steering Committee, with goals to share soft guidance w.r.t. key principles of collaboration. 39/
OK, what are the takeaways for the IO academics re: integrating privacy and data consideration protection into antitrust theories of harm? In my view, there are four. 40/49
FIRST: Greater market power --> Greater data collection. General argument by @johnnyryan on lack of merger enforcement worsening burden on DPAs supplemented by evidence from Andreas Mundt and @chrismdangelo re: Facebook and from @jason_kint and @DinaSrinivasan re: Google. 41/
SECOND: Greater data collection --> greater consumer harm, in two flavors. First, argument by @johnnyryan, @Caffar3Cristina, and me, that ``(lack of) privacy is a price,” supported by @ssnstudy and @chrismdangelo re: Facebook behavior. 42/49 voxeu.org/content/antitr…
Empirical economists: much more evidence is needed, both of the purported consumer benefits of big platforms’ data exploitation as well as estimates of consumer harms from their data collection, combination, and use. 43/49
Second flavor emphasized by @jason_kint: greater data collection and combination (in violation of purpose limitations) by data duopolists “lowers the bar,” forcing other publishers to do the same. Thus competition to *most* exploit consumers. 44/49
THIRD, the reverse of the first: Greater data collection --> Greater market power. A foreclosure story IO economists will be familiar with. Versions described by @johnnyryan, @jason_kint, @DinaSrinivasan, Simeon T, Andrea C, Andreas M, & @RKSlaughterFTC. 45/49
FOURTH, a conjecture. *Dominance* x Greater data collection --> exploitation (& thus greater harm). The third, foreclosure, story is about extending market power across markets. But greater data collection *in the presence of dominance* also allows … 46/49
… a monopolist to become a *discriminating* monopolist, increasing consumer harm. @ssnstudy described the early literature on privacy as analyzing the ability of data collection to permit price discrimination, but did it already allow for dominance? 47/49
Worst: per @johnnyryan, these are complementary strategies & this has brought the current crisis: Market power allows greater data collection, extending market power into new markets begetting ever more data collection, with consumer harms accumulating along the way. 48/49
Close with a “call to arms”! For competition authorities to progress integrating privacy/data considerations into antitrust, (esp. academic) IO economists must help to formalize & measure these concerns in ways that can be cited in casework! /END
@threadreaderapp unroll
• • •
Missing some Tweet in this thread? You can try to
force a refresh
