The Democrats’ Senate bill isn’t a real gerrymandering ban, either.
The Democrats’ Senate bill isn’t a real gerrymandering ban, either.
D emocrats see no problem in pursuing aggressive gerrymanders while also decrying the evils of gerrymandering. After all, they tell us, Republicans started it and Democrats want to ban it nationwide. But as I detailed in the first installment of this series, Republicans didn’t start it: Recent Republican advantages pale in comparison to prior and long-standing Democratic advantages. As I illustrated in the second column, the centerpiece of the House Democrats’ supposed gerrymandering ban — requiring each state to use a nonpartisan commission to draw maps — would be unconstitutional if imposed by the federal government, and has failed repeatedly when tried by state governments.
Democrats have substantially rewritten their redistricting proposals in the Senate. S. 2747, the “Freedom to Vote Act,” takes a dramatically different approach from the bill that passed the House. It removes all the requirements for commissions. But it is not even-handed. It is, in short, designed not to ban gerrymandering but rather to protect Democrats’ preferred tactics while allowing them to file lawsuits against Republican redistricting.
A better solution would be a more modest one: Ban racial gerrymanders and put outside limits on bizarrely shaped districts and districts that break up existing counties and cities unnecessarily. Such reform would eliminate the worst forms of mischief without spawning litigation, while leaving state legislatures with the traditional power to make fundamentally political decisions within reasonable bounds.
The Democrats Demand Racial Gerrymandering
Here is how the Senate Democrats’ bill works. First, the bill ranks the criteria that states must use for drawing districts. After requiring districts of substantially equal population (which is uncontroversial), there are three subparts setting forth the criteria for drawing districts. The language of the bill makes these mandatory and exclusive criteria: “A state may not use a congressional redistricting plan which is not in compliance with this section,” and a state’s redistricting plan “shall” use “the following criteria as set forth in the following order of priority.” Section 5003(e), following the list of prohibited criteria, underlines the exclusivity of its focus: “A State may not rely upon criteria not set forth in this section to justify noncompliance with the requirements of this section.”
What are those mandatory criteria? First, subpart (2) of Section 5003(b):
Districts shall comply with the Voting Rights Act . . . including by creating any districts where, if . . . 2 or more politically cohesive groups protected by such Act are able to elect representatives of choice in coalition with one another . . .
That’s right: The Democrats’ very first requirement for drawing districts, aside from their size, is racial gerrymandering. Note that the protected groups — some groups of American citizens, but not others — must be “politically cohesive” in order to receive gerrymandered districts. In other words: If a protected group divides its votes between Democrats and Republicans, it is no longer protected. But if it votes as a racial bloc for one party — which, in practical terms in today’s America, almost invariably means for the Democratic Party — then, the very first order of the Democrats’ bill is to guarantee that group the chance to elect a Democrat to Congress. The same goes if there are two distinct racial groups: In order to obtain legal protection, they have to vote together to elect Democrats.
Several of these terms and definitions are drawn from Voting Rights Act case law, notably Justice William Brennan’s opinion in the 1986 Supreme Court decision in Thornburg v. Gingles. That decision, however, originated these factors — most of which do not come from the actual language of the Voting Rights Act — in a very different setting: the use of multi-member districts as an alternative to single-member districts. And in any event, the Gingles factors derive from interpretations of federal statutes, rather than from the Constitution. A Congress that is sincere about doing away with race-conscious gerrymandering could simply toss them out and write its own criteria.
The need to do any of this in 2022, other than as a matter of protecting the interests of one political party, is dubious. The Voting Rights Act was written for a time and place in which race was meaningfully distinct from partisanship. Today, essentially every aspect of the voting, elections, and redistricting wars that is claimed to be racial is, in fact, partisan, and the combatants on either side of any local partisan battle can summon a national party to their side.
If Congress was able to limit or eliminate partisan gerrymandering on a race-neutral basis, that would act equally against racial gerrymandering. The fact that S. 2747 seeks to protect racial criteria in the drawing of district lines, and do so only to protect Democrats, while purporting to limit other forms of gerrymandering, gives away the bad faith of the whole game.
More of the same follows. Look at the next factor, in subpart (3) of Section 5003(b):
(A) Districts shall be drawn . . . to ensure [that] the practical ability of a group protected under the Voting Rights Act . . . to . . . nominate candidates and to elect representatives of choice is not diluted or diminished, regardless of whether or not such protected group constitutes a majority of a district’s citizen voting age population.
(B) For purposes of subparagraph (A), the assessment of whether a protected group has the practical ability to nominate candidates and to elect representatives of choice shall require the consideration of the following factors:
(i) Whether the group is politically cohesive.
(ii) Whether there is racially polarized voting in the relevant geographic region.
(iii) If there is racially polarized voting in the relevant geographic region, whether the preferred candidates of the group nevertheless receive a sufficient amount of consistent crossover support from other voters such that the group is a functional majority with the ability to both nominate candidates and elect representatives of choice.
This is the same set of concerns, again drawing from the Gingles factors but going well beyond the text of the Voting Rights Act. Again, it is designed to favor only one party. Look at the factors that decide whether a group is protected if it is not the majority in an area large enough to be its own district. First, it asks whether the group is politically cohesive — read: Democrat-voting. Second, it probes whether there “racially polarized voting” — in other words, if the surrounding white voters are largely Democrats anyway, as in the bluest states, the legislature can ignore this factor; if they are Republicans, then the Democratic voters get their own district. (Although racially polarized voting often occurs in Democratic primaries, so this factor is not entirely partisan.) Third, it discerns whether there is “crossover support from other voters” — again, whether the protected group can elect a Democrat. If the non-white voters start voting Republican, they lose the protection of a separate district. If the surrounding white voters vote Republican, the non-white voters gain that privilege.
All of this is done in the name of the Voting Rights Act. But it has migrated far from the original concept or text of the VRA, which cautioned that “nothing in this section establishes a right to have members of a protected class elected in numbers equal to their proportion in the population.”
Finally, the third of the three factors is set forth in subpart (4) of Section 5003(b):
(A) Districts shall be drawn to represent communities of interest and neighborhoods to the extent practicable after compliance with the requirements of [the previous] paragraphs . . .
A community of interest is defined as an area [with] broadly shared interests and representational needs, including shared interests and . . . common ethnic, racial, economic, Indian, social, cultural, geographic, or historic identities, or arising from similar socioeconomic conditions.
The term communities of interest may, if the record warrants, include political subdivisions such as counties, municipalities, Indian lands, or school districts, but shall not include common relationships with political parties or political candidates.
This is the final factor in drawing district lines, and the first factor that is not explicitly tied to racial voting patterns. Still, it cites “common ethnic, racial, economic, Indian, social, cultural, geographic, or historic identities,” and only grudgingly mentions things like the lines of counties, cities, and towns. Legislatures and redistricting commissions are permitted (“if the record warrants”) to draw districts that adhere to the lines of existing counties, cities, and towns, but they are not required to.
That’s it. We’ll get to the prohibitions in a moment, but these are the only permitted factors that are listed. So long as the racial gerrymanders are protected, states need not even consider existing county, city, and town lines. There is no restriction whatsoever on the kinds of bizarrely shaped districts that gave “gerrymandering” its name back in 1812. No consideration need be given to creating competitive districts.
The Prohibitions
So, what in this bill bans gerrymandering? Democrats point to Section 5003(c): “A State may not use a redistricting plan to conduct an election that, when considered on a statewide basis, has been drawn with the intent or has the effect of materially favoring or disfavoring any political party.” Further, states are prohibited, in Section 5003(d), from considering “the residence of any Member of the House of Representatives or candidate” or “the political party affiliation or voting history of the population of a district.”
But even those prohibitions are suspended “as necessary to comply with the criteria described in” the sections that mandate racial gerrymanders. And “no redistricting plan shall be found to be in violation of” these prohibitions “because of the proper application of the criteria” for racial gerrymanders, “unless one or more alternative plans could have complied with such paragraphs without having the effect of materially favoring or disfavoring a political party.”
The result: Democrats are permitted to present and consider evidence of partisan voting behavior and incumbent protection in order to draw Democratic districts, so long as those districts are drawn to divide voters by race. That procedure, moreover, doesn’t count as partisan favoritism except as compared with other maps that would also create racially gerrymandered Democratic districts.
To call this a cynical sham is an insult to cynical shams. The stacked prohibitions against partisan gerrymanders coupled with mandatory Democrat-favoring racial gerrymanders bring to mind the remark attributed to Peruvian dictator Óscar Benavides: “For my friends, everything; for my enemies, the law.”
In determining whether a legislature has engaged in prohibited “intent or effect” of partisan favoritism, courts can consider modeling of the prior eight years’ worth of federal elections to determine how many seats each party would be projected to win, and find a presumptive violation where that analysis “results in partisan advantage or disadvantage in excess of 7 percent or one congressional district, whichever is greater, as determined by standard quantitative measures of partisan fairness that relate a party’s share of the statewide vote to that party’s share of seats.”
That 7 percent standard, at long last, sounds like something concrete. But it must be compared with some baseline of “standard quantitative measures of partisan fairness.” What are those? There are multiple models proposed by progressive law professors and activists. As Chief Justice John Roberts noted in abandoning the task of judicially imposing a partisan fairness test in Rucho v. Common Cause (2019), “the initial difficulty in settling on a clear, manageable and politically neutral test for fairness is that it is not even clear what fairness looks like in this context.” Some tests will look to maximize competitive seats to give each party as many potential seats as possible, while other tests will seek — as the racial gerrymanders do — to lock in a baseline minimum of safe seats for the minority party. These are diametrically opposed goals. Without clearer legislative guidance, these standards will, as Roberts concluded, lack “a limited and precise standard that is judicially discernible and manageable” in litigation.
The models used in the baseline are not the only problem; there is also the data. Models are to be derived from the results of prior statewide federal elections for president and Senate. Districts are to be compared with the last two presidential elections (which will often involve at least one party running the same candidate twice) and the last two Senate elections (which may include an incumbent running effectively unopposed). There will, every cycle, be states where that baseline itself introduces a significant bias compared with the state’s actual political climate.
Of course, rather than establish a bright-line rule, courts are instructed to consider “the totality of circumstances” in determining whether a gerrymander is partisan. That standard is great news for Marc Elias, the Democrats’ chief litigator in these cases, because it offers a nearly limitless field for expensive, fact-intensive litigation. It is great news for activist district judges who want to impose their own rules — and for challengers who can pick the right judge to hear their case by strategic forum-shopping. It is not good news for the consistent application of a clear rule of law that applies evenly across the nation to all.
Democrats also propose a ban on mid-decade redistricting, so that whatever maps are locked in at the start of a decade have to stay that way — a way of protecting maps approved by Democratic governors elected in the 2018 wave from the prospect of a red wave in 2022. But again, it exempts the sort of mid-decade redistricting by litigation that Democrats pursued in the past decade in Pennsylvania and Virginia. This is reminiscent of the Illinois Democrats’ approach, which claims to be even-handed but actually tilts the scale on things like voter registration toward the needs and convenience of Democratic voters.
A Better Way Forward
There is, in fact, a better way for Congress to address partisan gerrymandering, if it is serious about taking a party-neutral, race-neutral stance that adheres to the traditional goals of redistricting, and leaves the job in the hands of state legislatures while placing some outer limits on the worst abuses. Indeed, in Rucho, Roberts noted with approval state and federal efforts to mandate “traditional districting criteria” such as “standards of compactness, contiguity, and respect for political subdivisions in redistricting.” Congress even mandated districts of “contiguous and compact territory” in the 1901 redistricting, albeit without providing any sort of judicially enforceable guidance.
There are three simple principles that Congress could mandate as outer limits, and flesh them out with specific, bright-line rules that are easy for courts to apply:
- Race-blindness: Ban consideration of race in drawing districts, doing away with the entire industry of race-minded districts and litigation to create them. This would accord with Justice Clarence Thomas’s view, to which the Supreme Court has nodded without adopting it, that the equal-protection clause bans racial gerrymanders even when states adopt them in an effort to comply with the Voting Rights Act.
- Respect for Existing Boundaries: Make it illegal to split a county or city into more than two districts unless the county or city has more people than will fit in a single district. Require that any city or county that is large enough to form its own district must contain as many districts as can be drawn entirely within the county or city, leaving only the remainder to extend outside the county or city limits.
- Compactness: Establish a minimum standard both for the average shape of a state’s districts and the shape of a state’s least-compact district. This is not unworkable, although exceptions would be needed for districts that follow the boundaries of geography (you can’t turn Cape Cod or Manhattan into a square district) or existing county lines.
Consider New York for an example of how the boundaries rule would work. In New York City, because the city itself covers five counties, the rule would require three House seats entirely within Brooklyn, three entirely within Queens, two entirely within Manhattan, two entirely within the Bronx — all ten of them safe Democratic seats, and all representing constituents who have a lot in common. Because the city has enough people for twelve districts, that leaves two districts’ worth of people that the state’s Democrats could allocate as they like to shore up suburban districts — but they could not divide Staten Island into more than two districts, and they would have to draw at least one district entirely within Nassau County and another entirely within Westchester. In other states, this rule alone would limit both Republican efforts to crack smaller cities into multiple districts to swamp them with suburbanites, and Democratic efforts to use megacities such as Chicago to extend tendrils far outside of city limits.
These should be outer limits; let states do what states have always done, but simply require them to abide by some clear rules that prevent the worst abuses. It is a reasonable and workable approach.
It is also the opposite of what Democrats are trying to do.