Population equality and the imposition of risk on partisan gerrymandering.

• Author: Buchler, Justin
• Position: Law Review Symposium 2011: Baker v. Carr After 50 Years: Appraising the Reapportionment Revolution

INTRODUCTION

The requirement for equal population across legislative districts constrains partisan ambition by imposing risk on partisan gerrymanders. This risk comes from the fact that a party attempting a partisan gerrymander must give itself relatively narrow and, therefore, potentially unstable majorities in a large number of districts. This Article examines the question of how much partisan advantage a party can take without running an uncomfortable risk of the plan backfiring. The Article estimates the size of the initial majority that a party must give itself in a district for that majority to be stable until the next round of redistricting and then calculates the number of safe districts the scheming party must cede to the disadvantaged party in order to guarantee the stability of its partisan advantage. This Article finds that actual redistricting plans frequently create less of a partisan advantage than parties could safely take under reasonable assumptions. Hence, partisan ambition may be constrained by some factor other than the risk imposed by the population equality requirement.

Prior to the equal population requirement for legislative districts, a partisan gerrymander was a relatively straightforward proposition, carrying little risk and limited only by the precision of one's data and one's own brazenness. Consider the position of someone charged with redrawing district lines without an equal population constraint. With partisan goals, perfect data, and no shame, the optimal strategy would be as follows: Group every voter affiliated with the disadvantaged party into a single, overpopulated district that maintains contiguity only by being so misshapen as to make Elbridge Gerry's salamander look like the state of Wyoming, and divide the rest of the state's population, consisting entirely of voters belonging to the advantaged party, into the remaining districts, which would be necessarily underpopulated. The resulting plan would give the disadvantaged party only one district, while giving the advantaged party perfectly stable majorities (by virtue of unanimity) in each of the remaining districts. Of course, such a plan would never be possible because one's data can never be so precise (particularly since data were less precise in the pre-Baker v. Carr era anyway), and shame, if not a conscience may prevent mischief-makers from attempting anything so crass. In the absence of an equal population requirement, however, there is no legal barrier to such a scheme without an objective standard by which partisan gerrymanders may be rejected. (1)

The purpose of an equal population standard for legislative districts is not to place a limit on partisan ambition, and there arc other philosophical reasons for such a standard. One of the interesting consequences of an equal population standard, however, is to limit partisan ambition by making the scheme described above illegal. Under an equal population standard, a party must be willing to run a risk in order to attempt a partisan gerrymander.

The strategy for a partisan gerrymander after the Reapportionment Revolution is the "pack-and-crack" approach, so named for the way that disadvantaged party voters are grouped. Disadvantaged party voters are "packed" into a small set of districts with inefficiently large supermajorities, guaranteeing them victory in these districts, but by larger margins than they need. In the remaining districts, disadvantaged party voters are "cracked" into relatively large minorities so that the advantaged party retains relatively small majorities in a large number of districts. Since advantaged party voters are spread more efficiently across districts than disadvantaged party voters, the advantaged party is likely to win more seats than its overall proportion of the state's population.

Unlike the previous scheme, however, this plan entails a risk. Consider an arbitrary group of thirty-three voters, consisting of twelve voters from Party A and twenty-one voters from Party B. If Party A must draw three districts of equal population, then, in principle, it can draw two districts consisting of six voters from Party A and five from Party B, and a third district consisting only of eleven voters from Party B. Such a plan gives Party A a majority in two out of three districts, despite the fact that it only holds approximately one third of the group's population. The problem with attempting to do so is that a slight shift in preferences can have devastating consequences. If just two voters from Party A switch allegiances, one from each of the two Party A-majority districts, Party A becomes the minority in all districts, and Party B wins all three districts.

Grofman and Brunell refer to such a backfired attempt at a partisan gerrymander as a "dummymander," and the potential for a dummymander means that under an equal population standard, a party must be willing to absorb a certain level of risk in order to take partisan advantage of drawing district lines. (2) A risk-acceptant party may attempt a pack-and-crack scheme, while a more risk-averse party would prefer a bipartisan gerrymander, in which voters of each party are packed inefficiently into their districts, thus guaranteeing each party a certain number of seats beyond which it can go neither above nor below. Interestingly, bipartisan gerrymanders have a number of positive representational consequences, which suggests that when risk-aversion is combined with an equal population standard, partisan ambition can be checked, with small-d democratic benefits. (3)

There is an important question, however, that has gone peculiarly unanswered. How much does the equal population standard limit partisan ambition? Put somewhat differently, how far can a scheming politician wade into the territory of a partisan gerrymander without incurring too much risk? This Article attempts to answer that question both theoretically and empirically. The results suggest that an ambitious politician could probably take more partisan advantage of controlling the process than most generally do while incurring relatively minimal risk.

1. HOW SAFE IS SAFE?

If the constraint that equal population places on partisan ambition is the imposition of risk, then our first task must be to measure that risk. Suppose that the party controlling the redistricting process has just over 25 percent of the population. In principle, that party can give itself a bare majority in a bare majority of districts (0.5 x 0.5 = 0.25), and win a majority of the seats despite having only one fourth of the state's population. A party with a bare majority in the state, in principle, can give itself a bare majority of the population in each district by making each district a microcosm of the state, possibly then winning every district with only just over half of the state's population.

The problem with each of these strategies is that a bare majority does not guarantee victory. So, we must begin with a simple empirical question. How big of a majority must a scheming party give itself in a House district when drawing the lines in order to count on holding that district until the next round of redistricting? Is 55 percent enough? It is a majority, but a party's 55-45 percent partisan advantage in a district does not guarantee victory in that district. The majority...