Will the Supreme Court Only Close the Door on Partisan Gerrymandering Halfway?

Even if the Supreme Court finally imposes limits on partisan gerrymandering this term, citizens could seem to win while democracy still loses. To fully close the door on partisan gerrymandering, the Court needs to make three specific decisions.

This fall SCOTUS will hear the most important partisan gerrymandering case in years: Whitford.

It is mindboggling, but unlike racial gerrymandering partisan gerrymandering is de facto legal today, because it is not specifically outlawed. North Carolina recently tried to defend their racially gerrymandering by arguing that districts were just optimized for partisan advantage!

In the past, justices have been unable to agree on whether partisan gerrymandering is something they can rule on (is ‘justiciable’) and how to detect it (a ‘manageable standard’). The Court has also clearly ruled that strict state-level proportional representation is not required. However, in the last major partisan gerrymandering case, Vieth, Justice Kennedy signaled that he thought partisan gerrymandering might be justiciable and telegraphed that ‘partisan symmetry’ might yield a manageable standard.

Partisan symmetry assesses the degree to which elections are fair, by looking at outcomes, votes cast, as opposed to intentions.[1] In our two-party system, partisan symmetry means both parties would win the same seat share for the same vote share. That does not require proportional results which is why it might succeed as a standard for partisan gerrymandering.

The Whitford case proposes a specific measure of partisan symmetry called the ‘efficiency gap’ designed specifically to appeal to Justice Kennedy. It is defined as the “relative wasted votes,” where the wasted votes for a party are the sum of all the votes they received in races they lost plus all votes over 50% plus 1 vote of the races they won. In other words, the efficiency gap measures the relative amount of “packing and cracking” in a set of congressional districts. Hence, the efficiency gap measures the amount of partisan gerrymandering in a simple, easy-to-understand way.

The popular narrative about the current case is that limiting partisan gerrymanders could “transform American elections” and “fundamentally change the way American elections are conducted.” Even in the best-case scenario though, this is not true. Moreover, the Court’s decision could leave substantial wiggle room for continued partisan gerrymandering.

This article explains three ways the Court’s decision could have the most positive impact.

1. What units should an efficiency gap standard be expressed in?

The creators of the efficiency gap proposed a 7% threshold for state legislative elections but a two-seat threshold for congressional elections, their logic being that what matters for control of Congress is the number of seats won or lost. There are three problems with this proposal:

First, while the number of seats won at the national level is, indeed, what ultimately matters for control of Congress, as Exhibit 1 shows a two-seat threshold conflates the degree of bias (the percentage gap on the x-axis) and the magnitude of the harm that bias causes (the number of unexpected seats on the y-axis).

Exhibit 1 — Efficiency Gap (%) vs. Unexpected Wins (#) (2012–2016)

Second, this difference is magnified, because the number of congressional districts per state varies widely, from 6–53 for states for which the efficiency gap applies. Hence, the disproportionality represented by two unexpected wins/losses in a state varies dramatically across states: a two-seat threshold would allow just ~4% bias in California but 8.5 times more (~34%) in Louisiana! Whatever degree of partisan asymmetry the Court decides is (un)acceptable should apply equally to all states.

Finally, an efficiency gap standard of two seats for congressional elections would have missed nearly 60% of states with egregious partisan gerrymanders in the 2012–2016 elections.[2] As Exhibit 2 shows, a two-seat threshold would have caught just 14 of 33 individual election results where the efficiency gaps were 10% or more. Besides the conceptual issues, a seat-based standard is not stringent enough.

Simply put, the efficiency gap measures the degree of partisan gerrymandering in a state and should be expressed as a percentage.

Exhibit 2 — Congressional Elections by State (2012–2016)

2. What should the specific threshold be for unacceptable asymmetry?

In addition to adopting the efficiency gap standard expressed in percentage terms, the Court must decide what the specific threshold should be, i.e., what separates unacceptable from acceptable asymmetry. There are two reasons for the Court to set a low, stringent threshold:

First and foremost, efficiency gaps are functionally equivalent to malapportioned districts. Specifically, every set of similar-sized districts with variable winner surpluses is functionally equivalent to a set of variable-sized districts with similar winner surpluses. While what it means for a state to have a significant efficiency gap can be hard to grasp mathematically, it is exactly the same as that state having variable-sized congressional districts which is against the law (PL94–171). This idea is elaborated more fully in Efficiency Gaps as Variable-sized Districts.

At most, 5% deviations are sometimes allowed in some kinds of legislative districts to achieve “some legitimate state objective.” Requirements for congressional districts are more exacting though. Within a state they have to be “roughly equal,” which in practice has meant almost exactly equal. North Carolina’s 13 congressional districts for its 10 million residents vary in size by single digits, as do the districts in many other states. (See Exhibit 1 in A Quick Way to Enable More Natural Congressional Districts.)

Second and equally important, eliminating egregious partisan gerrymandering will not, by itself, solve the problem of misrepresentation in Congress. The problem of misrepresentation in Congress is much broader and deeper than just gerrymandering.

With our polarized two-party system, Congress now has national winner-take-all dynamics like the Presidency with one party in control. Moreover, because all unexpected wins/losses count the same, i.e., whether they’re from a gerrymandered state or not, the unexpected wins in states have become interdependent, raising interesting constitutional issues.

As Exhibit 3 shows, only 2/3’s of the average net misrepresentation of almost 20 seats comes from egregiously gerrymandered states, i.e., average efficiency gaps for 2012–2016 of 10% or more. An average of seven net seats per election come from other states, with either average efficiency gaps less than 10% or fewer than six congressional districts.[3]

Exhibit 3 — Average Net Unexpected Wins (#) (2012–2016)

But, as Exhibit 4 shows, there’s very little slack in the system to absorb such unexpected wins/losses: an average of just 4 seats per election since the redistricting in the wake of the 2010 Census. In other words, it takes very little bias in the overall results to tilt control of Congress in favor of the party that did not garner the most votes, and the average net misrepresentation not due to egregious partisan gerrymandering is greater than the average slack in the system.

Exhibit 4 — Congressional Election Results (2000–2016)

Even if partisan gerrymandering could somehow be eliminated entirely and delegations could somehow be constrained to not deviate more than one seat from proportional, 43 states have more than one congressional district and can produce non-proportional results. As Exhibit 2 shows, over 80% of those states that can have non-proportional results typically do, on average 36 of 43 states in the 2012–2016 elections. Even if these states only break 2/3’s, 1/3, that’s ~14 seats which is, again, much bigger than the average slack in the system.

The bottom line is that outlawing partisan gerrymandering is necessary but not sufficient.

Consequently, to have a meaningful impact on the problem of misrepresentation that is loosely consistent with existing redistricting law, the Court should establish a low tolerance for partisan asymmetry of no more than 5–10%.

3. What should the scope of the decision be?

Even if the Supreme Court upholds the efficiency gap, expresses the standard in percentage terms, and sets a stringent threshold, there’s one more way the Court could unwittingly leave the door open to continued partisan gerrymandering.

Since Bandemer, gerrymandering claims can only succeed if they show both discriminatory intent and effect. The efficiency gap only measures discriminatory effect though. So, an efficiency gap standard is, by itself, necessary but not sufficient to curb partisan gerrymandering. In other words, even if Whitford is affirmed, gerrymandering can continue when challenges fail to separately demonstrate discriminatory intent. In the absence of a broader ruling, you can be sure that states will simply work harder to obscure their intent.

The Court can close this loophole though, by also recognizing that excessive discriminatory effect prima facie demonstrates discriminatory intent. The logic is simple.

If you believe:

  • Every state has a set of districts that are sufficiently politically symmetric; and
  • States can reasonably consider some of those plans when redistricting; and
  • States can evaluate the degree of partisan symmetry/asymmetry of different plans;

Then you must believe that excessive partisan asymmetry in resulting plans also demonstrates intentional discrimination for partisan advantage.

Simply put, while the efficiency gap measures discriminatory effect, it also shows discriminatory intent. The Court can kill two birds with one stone, by collapsing the standard for discriminatory intent into the standard for discriminatory effect


With the Whitford case this fall, the Supreme Court can finally impose limits on partisan gerrymandering. In addition to adopting the efficiency gap standard for measuring partisan gerrymandering, the Court should:

  • Express the standard in percentage terms for congressional elections;
  • Establish a low tolerance of no more than 5–10% inefficiency; and
  • Also rule that excessive discriminatory effect demonstrates discriminatory intent.

Even this best case ruling wouldn’t fully solve the problem of misrepresentation in Congress, but it would be a good start.

This is the first in a series of articles on redistricting and partisan gerrymandering.

[1] This discussion of partisan symmetry paraphrases Andrew Gelman in “Measuring partisan bias in redistricting: our amicus brief and some other thoughts” [http://bit.ly/2v1sNrC].

[2] The next section discusses what the threshold for an “egregious partisan gerrymander” should be and proposes a maximum of 10%.

[3] The creators of the efficiency gap did not compute it for states with less than eight districts. The Brennan Center extended the analysis to states with six or more districts. So, practically speaking, the efficiency gap is not defined for states with fewer than eight (or six) districts. While it covers 79% (86%) of congressional districts, it only covers 49% (61%) of states.