INTRODUCTION

If territorial constituencies face the kinds of problems we outlined in the previous chapter, we need to ask, “what's the alternative?” In the next chapter, I will argue that randomized, national, permanent constituencies offer a promising one. Their promise rests upon the implications of some minimal aims of legitimate political representation, the subject of this chapter.

I begin the analysis by ignoring the costs of change and assume that we could redefine constituencies in any manner we wanted. If we were to build a large nation's electoral institutions from scratch, how should we think about constituency definition? Can there be a universal starting position, a baseline from which deviations must be justified? If so, what would that position be?

In this chapter, I argue that all things considered, electoral constituencies should be heterogeneous, stable, and involuntary, and they should create incentives for representatives to pursue the public good. These features form a default position derived from the limiting conditions of any plausible theory of legitimate political representation – conditions as noncontroversial as “representatives should be accountable to those they represent.” As a default position, these features of heterogeneity, stability, and involuntariness may be overridden by other more pressing values. But because they form a default position, the burden of justification will rest upon others who want homogeneous, unstable, or voluntary constituencies, including advocates of group representation or proportional systems.