A number of features of partisan competition force the use of different methodologies for defining and measuring partisan fairness than we do for other violations of equal protection, such as the protection of racial groups or determinations of one person, one vote violations. Thus, it will be possible for courts to adopt appropriate tests of partisan bias and to incorporate them into standards for unconstitutional partisan gerrymandering without any necessary need to rethink how the Supreme Court has dealt with equal protection in other domains of voting rights in the light of the tests to be used in this area. Three key distinctions are necessary here.
First, just as we previously showed that the concept of partisan symmetry is only appropriate for competitive situations where there is a potential for a change in partisan outcomes (majority control, in particular) as a result of shifting electoral tides, a concept of “racial symmetry” would not be the appropriate way to evaluate equal protection claims involving racial or ethnic groups with fixed ascriptive characteristics. To evaluate “ethnic group symmetry” we would need to ask what would happen to the seat division between the ethnic groups if a state with, say, 10% African Americans suddenly became 90% African American. But that really doesn’t make sense. While people regularly change their support for political parties on the basis of candidate qualities, issues, and the performance of incumbents, they do not (or cannot) similarly switch ethnic group memberships. And although immigration, emigration, birth, and death do produce changes in the ethnic composition of a state, these changes tend to be very gradual over time and do not really make applying a symmetry standard to racial categories something that makes conceptual sense. 115
Second, and relatedly, because, as we noted above, there is no direct analogue for measuring the impact of districting on racial groups to the role that partisan bias plays as a direct measure of partisan gerrymandering effects, we would anticipate a more central role of subsidiary criteria such as geographic compactness in the initial phase of voting rights claims that involve racial impact than is found in the legal review of claims related to partisan gerrymandering. In the racial context, courts have examined the shape and appearance of district lines in order to resolve whether a State was engaging in a racial gerrymander.116 Similarly, the shapes of districts as they are related to the geographic concentration of minority group populations and to existing jurisdictional boundaries has been found relevant in cases stemming from Shaw v. Reno,117 where the issue is inferring whether race has been a preponderant motive in line drawing.118 But the role of these subsidiary criteria appear to be much less direct in cases involving partisan gerrymandering since there we have, in partisan bias, a direct measure of the consequences we are most concerned about vis-à-vis fairness and equal treatment.
Third, whether for partisan or for racial groups, we would emphasize that the concept of symmetry is not an appropriate basis for addressing equal protection claims in situations involving pure at-large election systems, since absent voter fraud, symmetry will essentially always be satisfied for such at-large systems. In pure at-large systems (i.e., one where the entire legislature is chosen from two competing partisan slates, with each voter in the entire polity having as many votes to cast as there are seats in the legislature), if voters cast all their multiple votes along partisan lines, thus electing either an entire slate of Democrats or an entire slate of Republicans, changes in partisan vote share can affect partisan balance, but they do not affect either partisan bias or electoral responsiveness. That is because in such pure at-large systems there can be no partisan bias, since the two parties are treated symmetrically in that each will win 100% of the seats with a plurality of the votes, and the level of electoral responsiveness in pure at-large systems is effectively infinite in that an infinitesimal change in party vote share from 50%- to 50%+ will change all the seats and thus change party control.
In contrast, in districted systems that use plurality voting (including both ones where all seats elect a single representatives and ones involving multiple districts from some or all of which different numbers of representatives are elected), changes in partisan vote share can affect both partisan bias and electoral responsiveness.119 Thus, the possibility of long-term bias in a districted plan is completely different from what we find in an at-large plan.120 In a pure at-large election, if there are only two slates, then, as shown above, bias is normally zero, but electoral responsiveness is huge, because there is either feast or famine – either a party (the one with the majority of the votes polity-wide, however slight an edge they might have) wins all the seats, or it wins none.121
One way to think about the basic ideas in this paper about how social science can be useful to the courts in making determinations about partisan gerrymandering is in terms of four different contributions, involving (1) conceptualization, (2) measurement, (3) legal relevance, and (4) legal specificity, respectively.
Conceptualization We have shown that a widespread consensus exists in the social science literature supporting a specific definition of partisan fairness in legislative redistricting in plurality-winner district-based elections based on the concept known as partisan symmetry.
Measurement The social science literature offers statistically grounded, well accepted, and highly informative methods for measuring the deviation of a redistricting plan from partisan symmetry. The degree of deviation from symmetry of treatment is known as partisan bias, and is easily quantified, and made specific as to direction. Expert witness testimony can readily be obtained about the level of partisan bias, and areas of disagreement among competent experts about the magnitude or durability of partisan bias will be limited to fact-specific issues of a sort that courts regularly deal with in the numerous domains where expert witness testimony is commonly accepted as relevant.
Legal Relevance The social science concept of partisan bias is relevant to courts seeking to craft legal standards for unconstitutional partisan gerrymandering because it is built on the most basic and common notions of fairness in elections, and because it offers a straightforward and direct way to understand “equal protection” in the context of redistricting and political party competition. Thus, courts can draw on social science evidence in making fact-specific and case-specific legal decisions. Of course, however, the ultimate decisions need to be made by the Supreme Court about exactly what criteria, thresholds, and fundamental legal rules should guide lower courts.122
Legal Specificity The menu of choices available to the Court for defining unconstitutional gerrymandering, although based on ideas of equal protection and fairness, are specific to the domain of partisan gerrymandering and do not necessarily carry over into other domains of election law or legislative redistricting. In particular, measures of symmetry are less relevant in domains where we have racial groups with fixed ascriptive characteristics and for which well-established legal concepts such as minority vote dilution are applicable.123