Second-Order Diversity

Abstract

Much scholarship on democratic design is preoccupied with a single problem: how to treat electoral minorities in a majoritarian system. A term often deployed in those debates, particularly those focused on demographic difference, is "diversity." When scholars use the term, they usually mean that something - a class, an institution, a decisionmaking body - should roughly mirror the composition of the population. The problem with this debate is that its participants often unthinkingly extend theories about diversity derived from unitary institutions to disaggregated ones - institutions in which the governance system is divided into a number of equal subparts (juries, electoral districts, appellate panels, schools committees, and the like). Thus, despite the prevalence of such institutions, scholars have not systematically considered how to tailor our normative commitment to diversity to their unique features. This Article is a first step toward providing such a conceptual framework. It argues that we can seek at least two kinds of diversity in disaggregated institutions - first-order and second-order. First-order diversity mirrors the conventional intuition; it is the normative vision associated with statistical integration. The notion of second-order diversity, proposed here, posits that democracy sometimes benefits from having decisionmaking bodies that look nothing like the population from which they are drawn but instead reflect a wide range of compositions. The Article then deploys these two notions to examine a recurring set of trade-offs we face when designing disaggregated institutions.