Prediction and the Moral Order: Contesting Fairness in Consumer Data Capitalism

Abstract

Corporations increasingly gather massive amounts of personal data to mathematically predict how individuals will behave and then treat people differently based on these predictions. This dissertation unpacks the moral foundations of such market arrangements by analyzing the case of U.S. car insurance pricing and regulation. I focus on a series of public policy debates about which sorts of data are fair for car insurers to use in deciding who to sell insurance to and what prices to charge, starting with the introduction of credit scores in the early 1990s and going through debates about “big data” today. To understand how people justify and contest the use of personal, predictive data, I draw on more than 13,000 pages of documents and 40 hours of recordings from regulatory and legislative hearings and investigations; 50 interviews with policymakers, members of industry, and other powerful actors; observations from more than a dozen regulatory events; and an original, nationally representative survey of 1,095 Americans. I advance the literature in economic sociology by showing that the idea of actuarial fairness—that data which predict are, by definition, fair to use—functions as a moral argument by obscuring why prediction works and what leads people to show up in the data the way they do. Policymakers and others who mobilize a competing conception of market fairness, one rooted in notions of moral deservingness, seek to recover such distinctions, which wind up shaping the law and regulation that follow.