Disqualify only those plans that (can be expected to) translate a minority of the votes into a majority of the seats
In general, it is not possible to directly infer the existence of partisan bias from the fact that a given party receives a lower seat share than its vote percentage. That is because, as we have pointed out earlier, seats-votes relationships are a combination of electoral responsiveness and partisan bias, the two components of which must be jointly estimated by appropriate statistical methodology. Only statistical analyses that make uses of election outcomes in particular units of geography aggregated to the level of the individual districts can give us the appropriate measurements of these key parameters.
But there is one situation in which it is possible simply by comparing vote share and seat share to reliably infer partisan bias in an election that has already occurred, namely a two party contest in which a party with a majority of the votes gets less than a majority of the seats. As long as this difference cannot be ruled out on the basis of random variability that is not likely to persist in subsequent elections (a problem easily avoided with the methods outlined above), we can think of such a situation as violating a fundamental tenet of majority rule which we may call the majoritarian criterion.91
In such a situation there must be partisan bias against the party winning the majority of the votes. This situation may be thought of as distinct from other examples of partisan bias because majoritarian ideas are so deeply engrained in our political system. It somehow seems more heinous to prevent a majority from exercising its mandate, then merely to exaggerate the size of a majority. Moreover, exaggerating the size of the majority is virtually inevitable under plurality-based legislative elections because of electoral responsiveness values above one (the bonus effect). As we have previously remarked, Justice Kennedy, in his LULAC opinion (at p. 12) views these two types of outcomes (vote majorities into seat minorities; vote majorities into even larger seat majorities) as conceptually distinct. He also clearly regards the former case as the potentially more serious violation of equal protection in terms of the impact of partisan bias in generating the outcome.92
If we take this distinction seriously, and regard only the former type of outcome as (potentially) constitutionally violative it suggests a distinct (prima facie) test for when partisan gerrymandering rises to the level of an unconstitutional violation, namely that the level of partisan bias must be such as to (reliably) convert a minority party in terms of votes into a majority party in terms of seats. Because this condition will only rarely be found in situations involving partisan gerrymandering —after all, normally the party which controls the districting is a majority party and it will merely draw plans that will extend the magnitude of its dominance – under this test, even a prima facie finding of unconstitutional partisan gerrymandering might be very rare. Nonetheless, had this standard been in place earlier, however, the evidence reviewed in Justice Kennedy’s opinion suggests that the 1991 pro-Democratic gerrymander might have been struck down in mid-decade or later in looking at the results of elections that took place under that plan.93 This districting was one whose circumstances could give rise to a violation of the majoritarian criterion because it was a partisan plan drawn by a party in power able to control districting whose leadership was, arguably, aware of the electoral tides running against the party and who sought to preserve that party’s dominance through creative gerrymandering. Thus, even under this remarkably strict standard suggested by some of Justice Kennedy’s observations about partisan bias, there is no reason to think that partisan gerrymandering would remain a legal “dead letter.”94
Disqualify only those plans whose partisan bias is both severe and greater than that in the plan being replaced
A central issue with respect to any standard for equal protection is the specification of a baseline against which equal treatment is to be judged. As long as the U.S. makes uses of plurality-based single seat districts for its most important elections, the issue of determining a baseline will be with us. For two party competition, a good way to specify such a baseline is to look at deviations from symmetry in translating votes into seats at the point where each party receives exactly 50 percent of the vote. Equal treatment is this context implies that each party would get an (approximately) equal share of the seats. Another approach, suggested by ideas in Justice Kennedy’s opinion, is to compare the magnitude of the deviations from partisan symmetry found in the challenged plan with that found in its immediate predecessor (or predecessors). A possible test based on this idea would give us a prima facie case for unconstitutional partisan gerrymandering only if the plan’s deviations from symmetric treatment were both severe and more egregious than those in previous plans.95
Just as most of the prima facie tests for gerrymandering that we identified above are inspired by other areas of voting rights case law (e.g., standards used for one person, one vote), so this latter test could be thought of as inspired by how courts have often dealt with the issue of compactness When it arises in cases involving state constitutional requirements. While there are many measures of compactness, for none of them is there a clear metric to determine what is egregiously ill-compact.96 In coping with this problem, courts frequently compare the compactness of a given plan (or given districts) to that of competing plans (or districts within the same or other plans) or to compactness measurements in a predecessor plan. Here, analogously, what we would be doing is comparing the (absolute) magnitude of deviation from partisan symmetry in a challenged plan from what was found in earlier plans – usually plans that had been held to be constitutional.97
We take no position on which of these five approaches is constitutionally most appropriate or which best fits the notion of “egregious” unconstitutionality first enunciated in Bandemer.98 As noted earlier we regard that determination as a legal issue that is outside the purview of this essay, and a choice the Supreme Court can and should make. Whichever of these legal thresholds is adopted, it will be straightforward to determine if a constitutional violation has occurred (at least prima facie), and States and those drawing the lines in each state will know in advance whether any plan they propose is likely to pass court review.
For purposes of empirical calibration, we would note, however, that studies of past partisan gerrymanders have shown that most gerrymanders have a partisan bias of 1-3 percentage points in favor of the party controlling the redistricting. This effect is typically persistent over the decade following the redistricting, and accounts for measurable differences in the representation of the state’s population in the state legislature or Congress. Occasionally, the difference is greater than 5 percentage points, and only in rare cases does a gerrymander results in a difference of over 10 percentage points.99