Sixth hearing on Competition and Consumer Protection in the 21st Century features disagreement over FTC’s enforcement priorities for consumer data.
The Federal Trade Commission (FTC) recently held the sixth hearing in its series of nine planned hearings on “Competition and Consumer Protection in the 21st Century.” Speakers, moderators, and panelists convened from November 6 to 8 to focus on several topics involving the intersection of competition with privacy and big data. Discussion (and disagreement) at the hearing covered the collection and use of consumer data, the role big data can play in supporting innovation and commercial growth, and attendant privacy and competition concerns. Panelists also touched on several related topics, including how antitrust laws should analyze and treat data, and the appropriate scope of US privacy regulatio
Hearing #6’s Big Idea: Is Consumer Data an Antitrust Problem?
There was very little disagreement during the sixth FTC hearing that consumer data is now, and will only continue to become more, central to businesses in the tech industry and beyond. But the agreement ended when participants considered the extent to which the FTC should apply its antitrust authority to regulating the use of consumer data. Several panelists endorsed the application of antitrust to address concerns about access to big data (or lack thereof) impeding competition. Others recommended that big data be assessed as its own relevant product market, while others argued that such treatment should be limited to instances when data or access to data itself is commoditized. However, there were also a number of participants who urged skepticism and caution about expanding antitrust enforcement to address privacy concerns about consumer data.
Hearing #6 addressed topics ranging from the relationship between consumer data and price discrimination to competition concerns involving the accumulation and cordoning off of data. Notable remarks from regulators and industry leaders that reflect the topics discussed during this hearing include the following:
• “Competition law offers at best a convoluted solution to address privacy concerns with respect to big data.” Maureen K. Ohlhausen, former FTC Commissioner and Acting Chairman
The FTC’s Bureau of Competition currently directs most questions about data privacy to the agency’s Bureau of Consumer Protection. In practice, this means that questions or concerns about data privacy, including questions about the use (and misuse) of consumer data, have largely fallen outside the scope of FTC merger review, conduct investigations arising under Sections 1 and 2 of the Sherman Act, and the prohibition in Section 5 of the FTC Act against unfair methods of competition.
Speaking on the same panel, Bill Baer, former head of the DOJ Antitrust Division and former Director of the FTC Bureau of Competition, agreed with Ohlhausen about the need to distinguish consumer protection from competition problems. He said that “it is important analytically ... [to] separat[e] what is an antitrust problem and what is a consumer protection problem.” However, he also noted that “the competitive market can create these externalities where competition isn’t taking into account certain costs to society and to consumers from a lack of competition on privacy and data security.”
• “Is there any sense in which data can be a barrier to entry?” Jeremy Sandford, Economist in the FTC Bureau of Economics
In the panel “The Economics of Big Data and Personal Information,” moderator Jeremy Sanford of the Federal Trade Commission, Bureau of Economics, led a panel that discussed, among other topics, whether and to what extent data access should be considered a barrier to entry. This question has important implications for data-rich companies that potentially could face more antitrust scrutiny relative to their data-poor competitors. Florian Zettelmeyer, Professor of Marketing at Northwestern University, Kellogg School of Management, argued that saying firms can “just go out and buy data” in order to compete with firms that already have a trove of big data is an inadequate defense. Further on this point, Zettelmeyer noted that that the amount of data required to train algorithms effectively is “astronomical,” reflecting his skepticism that home-grown data resources are competitively distinguishable from data available on the open market.
However, Ginger Zhe Jin, Professor at the University of Maryland, Department of Economics, disagreed and urged caution about how regulators ought to view the competitive significance of big data access. She said “we have seen entrants overtake incumbents” even without accessing data, and that accordingly, “even where data does give an advantage, agencies still need to think hard about how that translates into regulatory action.”
• “That is where we are with respect to advanced analytics, individuals cannot understand how these technologies work, and so cannot use traditional privacy protections — notice and choice — to protect themselves.” Dennis Hirsch, Professor, The Ohio State University Moritz College of Law
In his talk “Corporate Data Ethics: Risk Management for the Big Data Economy,” Professor Dennis Hirsch of The Ohio State University Moritz College of Law suggested extending the use of the FTC’s Section 5 unfairness authority to reach instances in which companies rely on machine learning and algorithms to analyze consumer data. He suggested these technologies allow companies to do more with consumer data than consumers may expect. Hirsch argued that big data analytics renders compliance with current privacy laws insufficient to address data ethics.
According to Hirsch, “Privacy law [is] premised on the idea that given accurate notice individuals can make choices about what companies can do with their data, and by making such choices individuals can protect themselves. But big data analytics changes this, it allows [companies] to take surface data and to infer latent information from it. ... Given this ability ... people cannot know what they are really revealing when they decide to hand over the surface information, as a result they cannot use notice and choice to protect themselves ... from big data analytics, machine learning, and AI.” As a result, Hirsch suggested that the FTC should encourage strong “data ethics” (using Section 5 guidelines, best practices, or other measures) and consider ways to address “issues of bias, procedural fairness, and manipulation” that derive from use of big data analytics, like advanced inference making
. Whether the FTC will consider using its regulatory authority to further regulate the use of algorithms and other machine learning tools will be a key topic at the seventh hearing in the series, which is taking place this week.