Recently, Scott Peppet, Dan Solove, and Paul Ohm appeared in a great Al Jazeera comic on big data and privacy, called “Terms of Service.” The comic covered the growth of data-driven companies from scrappy startups to the behemoths we know and fear today. It’s also a good introduction to the problem of discrimination by data and algorithm. For those who want to continue the conversation, Nathan Newman‘s article is an excellent guide to the issues.
Newman has already made several important interventions into the scholarly debate over the effects of big data. Marketing industry leaders have argued that data-driven marketing increases the accuracy of ad targeting. Critics have contended that the opacity and complexity of data flows makes it impossible for the average citizen to understand how they are being rated, ranked and judged. The White House Big Data Report from 2014 was a major validation for critics, compiling numerous problems in the big data economy and taking seriously threats on the horizon.
Newman opens How Big Data Enables Economic Harm with the observation that the “increasing loss of control of private data by individuals seems to be leaving them vulnerable to economic exploitation by a range of corporate actors.” He then explains how classic economic theory may exaggerate the positive effects of price discrimination, while glossing over its negative consequences.
Our common mental model of price discrimination is first class seats for airlines costing more than, say, economy: the presumption is that wealthier passengers pay more for the more comfortable ride, and in some way cross-subsidize the flight as a whole. Price discrimination is assumed to be a way reflecting ability to pay in prices. In demotic models, we imagine high prices matched to a small group (for example, those with business accounts, or those who care deeply about having more legroom on a flight).
But what if price discriminating firms take a different tack, structuring options in ways that are less easy to parse? Pushed into areas where smaller purchases are made, price discrimination can exacerbate (rather than ameliorate) inequality. For example, Newman notes that a report found that, for a certain retailer, those living in “higher-income locations were offered better deals than low-income communities, because those poorer areas had fewer local retail outlets competing with the online stores.”
According to Newman, it is not just the poor who should worry. While internet retailers may advance price transparency in some contexts, in others “price obfuscation strategies are designed to frustrate consumers and keep prices up.” Newman does a close reading of Google economist Hal Varian’s article in the industry-based academic journal Marketing Science, showing how each corporate advantage attributed to extensive data collection can amount to a consumer disadvantage, as the firm determines the “pain point” just below which it can maximally charge for its product. Microtargeting also applies to groups: as Newman notes, big data-driven “search advertising is especially attractive to companies looking for micro markets of vulnerable targets for scams,” because it allows “targeted access” to likely victims. “Vulnerability-based marketing” is also a hot new strategy. Want to find lists of the impotent, the depressed, rape victims? Data brokers have sold those and more, at 7 to 15 cents a name.
Newman’s counterintuitive take on price discrimination joins important pieces by James Boyle and Julie Cohen—both of which undermined classical economic analysis in the intellectual property field. As Boyle observed, pricing can be a manifestation of power, particularly where there is monopoly provision of a service. Newman brings these important insights to internet law. And as he has shown in other work, the market power of big data platforms can be substantial, giving consumers few ways of escaping major firms’ power.
In How Big Data Enables Economic Harm to Consumers, as well as others on the economics of search advertising, Newman is developing a very important counternarrative to the usual stories were here about Silicon Valley efficiencies and marketing magic. To be sure, there will be many who will continue to frame big data as a “tool for fighting discrimination and empowering groups.” But they will have to grapple with Newman’s work, and concede many critical points he makes, before rehabilitating big data.
