Taking Virtual Representation Seriously

Virtual representation-the representation of people who cannot vote-has come to occupy a distinctly disfavored position in modern conceptions of democracy. For most of our history, this was not the case. Before women's suffrage, most U.S. citizens were represented only virtually. Today, virtual representation remains a substantial part of our democratic order, but it is a part that we tend to ignore. For instance, about a quarter of all U.S. citizens cannot vote because they are children. They, and others who cannot vote, are virtually represented. This Article argues, first, that virtual representation is an inevitable part of any democratic system; second, that it has value, even if it is generally second-best to actual representation; and third, that as long as we are going to do this, we ought to try to do it as well as we can. The Article traces the origins of the American aversion to virtual representation, then begins to build an account of how some forms of virtual representation are better or worse than others. Surprisingly, under conditions of sufficient geographic segregation along various politically salient dimensions, it turns out that our system of electing nearly all representatives from single-member districts may actually ensure non-voters tolerably good virtual representation. That is, as long as we count all the people, not just the voters, for purposes of apportionment-and similarly, as long as we count prisoners at their last address rather than at the location of their prison (this is the problem of so-called "prison gerrymandering"). Finally, the Article argues that the United States Constitution itself constructs a mixed system of actual and virtual representation. The question of the reach of this federal constitutional model-and in what contexts, if any, it mandates the current practice of drawing districts with equal numbers of people, rather than equal numbers of voters-is likely to come before the Supreme Court sometime after the 2020 Census.