Many thanks to Rick Hasen for the opportunity to blog about my forthcoming article, Spatial Diversity, 125 Harv. L. Rev. (2012). I’ll be writing a series of posts this week that track the article’s main claims and findings. Today I’ll introduce the concept of spatial diversity and explain its relevance to electoral districts. Tomorrow I’ll discuss the role the concept has played in the Supreme Court’s case law on gerrymandering and the Voting Rights Act. And then on Wednesday and Thursday I’ll summarize the results of my empirical analysis.

So first things first. By the term, “spatial diversity,” I mean the variation of a given factor (e.g., income, race, age, etc.) over geographic space. If the factor takes on different values in different areas within a larger entity, then the entity is spatially diverse (or heterogeneous). But if the factor stays relatively constant throughout the entity’s territory, then the entity is spatially non-diverse (or homogeneous).

Crucially, spatial diversity is not identical to the usual notion of diversity. Consider a district that is fifty percent rich and fifty percent poor. This district typically would be deemed highly diverse, in terms of wealth, since it contains large shares of both rich and poor voters. But this same district could be very spatially diverse or very spatially non-diverse depending on its geographic composition. The district would be highly spatially heterogeneous if most rich voters lived in one area and most poor voters lived in another. But the district would be highly spatially homogeneous if both rich and poor voters were dispersed evenly throughout its territory.

So why does spatial diversity matter? The answer, at least in the districting context, is that it is linked to a variety of democratic injuries. Both in theory and empirically, voters are less likely to participate in the political process, and elected officials are less likely to represent their constituents effectively, in districts that are highly spatially diverse along key dimensions. Such districts also do not appear to generate any offsetting increases in electoral competition.

Starting with participation, voters tend to become confused and disengaged when they are placed in districts that merge dissimilar geographic groups. They do not identify naturally with districts that disregard their underlying residential patterns, and their capacity for political mobilization decreases when they lack both shared interests and geographic proximity. Consistent with this reasoning, the empirical evidence shows that voters are less informed in districts that fuse different political subdivisions, and that turnout is higher in states that respect geographic communities when they redraw district lines.

With regard to representation, elected officials cannot easily identify or advance their constituents’ interests when those interests vary widely by location. The officials often receive different political signals from different geographic groups, and their actions typically please some groups but antagonize others. Again, the empirical evidence is quite supportive. Key district characteristics and the position of the median voter are worse predictors of politicians’ voting records in heterogeneous districts than in homogeneous districts. It is partisan affiliation, not voters’ attributes or opinions, that best explains the behavior of representatives from heterogeneous districts.

Finally, it turns out (somewhat counterintuitively) that spatially diverse districts are not any more competitive than spatially non-diverse districts, and that if anything the opposite might be the case. According to a number of studies, the better that districts correspond to political subdivisions and media markets, the better challengers do in them. In these more spatially homogeneous districts, challengers find it easier both to communicate with voters and to craft campaign messages that appeal to a broad swath of the electorate.

Spatial diversity therefore has adverse implications for a number of important democratic values. What is more, scholars are not the only ones who have noticed. Tomorrow I’ll explain how the Supreme Court has also seemed to intuit, in many election law cases over many years, that spatial diversity is an undesirable district attribute.