Separation, Pooling, and Predictive Privacy Harms From Big Data: Confusing Benefits for Costs

ABSTRACT:

Privacy is about being “let alone,” which boils down to being able to conceal certain details about oneself from the world. In economic jargon, it means to “pool” with others. Economists, however, tend to prefer separating to pooling equilibria, as the former supply the market with more information and lead to concomitantly more efficient decisions. This tension between privacy and market efficiency—between pooling and separating equilibria—is on full display in the burgeoning privacy law scholarship surrounding big data. The privacy concerns raised in the big data context in large part have shifted away from the more traditional domains to so-called “predictive privacy harms,” which arise as big data allows firms to make granular distinctions based on predictive algorithms and to tailor offers to customers, employees, and borrowers accordingly. The increasing use of algorithmic predictions based on big data has led some to call for limits on their use or collection of data in the first instance. Privacy is valuable intrinsically, but is also crucial that policy discourages dissipative privacy—strategic concealment of facts relevant to a transaction in hopes of getting a better deal. This paper brings to bear insights from the economics of contracts and torts to develop a positive framework for thinking about big data regulation; one that helps identify dissipative and productive privacy and revelation. It also attempt to allay some of the concerns that big data is likely to have a disproportionate impact on the economically disadvantaged by bringing economic theory and empirical evidence into a debate.