Redistricting
By Nathaniel Rakich
Back in March, I started off an article with the sentence, “Congressional redistricting — the process of redrawing the nation’s 435 congressional districts to reflect the results of the 2020 census — is not quite finished, but it’s getting darn close.” Clearly, I jinxed it: Since then, the national redistricting landscape has changed substantially, thanks to a new Republican-drawn plan in Florida and a court-ordered remap of New York.
Moreover, it’s taken the 2021-22 redistricting cycle from a clear win for Democrats to something far more ambiguous — perhaps best described as the preservation of a Republican-leaning status quo. And a ruling earlier this month striking down Louisiana’s congressional map is yet another reminder that the 2021-22 redistricting cycle ain’t over till it’s over — and, in fact, very likely won’t be over until well after 2022. That said, the map below is probably the one that will get used in this year’s congressional elections (with Louisiana pending, of course).1
And as has been true for decades, this national congressional map is biased toward Republicans. Assuming Louisiana’s congressional map is reinstated upon appeal,2 the 2022 House map will feature 208 congressional districts with a FiveThirtyEight partisan lean3 of R+5 or redder, compared with 187 districts that have a FiveThirtyEight partisan lean of D+5 or bluer. Throw in the highly competitive seats, and 225 districts would be more Republican than the country as a whole, while 210 would be more Democratic. In other words, if the national House popular vote were perfectly tied, Republicans would theoretically win 225 seats and Democrats would win 210 (ignoring, for now, other factors like candidate quality).4
By this measure, however, Democrats are actually in a slightly better position than they were before, having added a handful of Democratic-leaning seats. Under the old congressional lines (those used in the 2020 election), there were 208 congressional districts with partisan leans of R+5 or redder and 181 with partisan leans of D+5 or bluer. Counting swing seats, 230 seats were redder than the nation as a whole, while 205 seats were bluer.
However, by other measures, the new map is better for Republicans. First, the “tipping-point” congressional seat — i.e., the majority-making 218th bluest and 218th reddest seat in the House — is slightly more Republican-leaning. Under the old lines, the tipping-point seat had a FiveThirtyEight partisan lean of R+2.3; under the new ones, the tipping-point seat will have a FiveThirtyEight partisan lean of R+2.5 (again, assuming Louisiana’s map is reinstated on appeal).5
In the short term (i.e., in this year’s midterms), Republicans are also more likely to pick up seats from redistricting. Remember that all the numbers above merely reflect each seat’s underlying partisanship; they don’t account for which party currently holds each seat. And when we do that, we see that more Democratic-held seats have been turned red this redistricting cycle than Republican-held seats have been turned blue. By my calculations, Republicans can expect a net gain of roughly three or four seats this November due to the effects of redistricting alone — not accounting for shifts in voter preference.6
Some of the House map’s GOP bias is due to geography (i.e., the Democratic tendency to cluster in cities, plus rural areas’ tendency to vote Republican). But a lot is also due to deliberate decisions by partisan mapmakers — namely, Republican lawmakers drawing congressional maps that advantage their own party. In 2014, a pair of academics created a metric called the efficiency gap, which attempts to quantify this phenomenon by measuring how efficient a map is at converting votes into seats for a given party. And using this measure, we find that seven of the 11 most biased congressional maps in the country were drawn by Republicans, while only one Democratic-drawn map (Illinois’s) provides Democrats with more than 1.2 undeserved seats.
How many extra congressional seats Republicans or Democrats theoretically gained from biased map-drawing in each state that has completed redistricting as of June 14, 2022, according to the efficiency-gap metric
Sources: U.S. Census Bureau, The Upshot, Voting Election and Science Team, Ryne Rohla/Decision Desk HQ
In fairness, this wasn’t because Democrats didn’t want to gerrymander. If given the opportunity, they may have tried to draw Democratic-friendly maps in states like Colorado or Washington where they have full control of state government. However, they didn’t have the chance; those states have vested the power of redistricting in independent or bipartisan commissions. (Though commissions are not a cure-all for gerrymandering — maps like California’s and New Jersey’s still have notable Democratic biases despite being drawn by commissions.) As it was, though, only five maps this cycle (worth 32 districts) ended up getting enacted unilaterally by Democratic politicians or institutions,7 while 18 (worth 171 districts) were enacted unilaterally by Republican ones.
That’s an even bigger disparity than we expected at the beginning of the cycle, when we observed that Democrats controlled the redistricting process in eight states worth 75 districts but Republicans controlled the process in 20 states worth 187 districts. That’s because liberal courts cracked down on gerrymandering this cycle and generally enforced fairer congressional maps — even when it hurt the Democratic Party. Most notably, Democrats initially drew maps with galling efficiency gaps in Maryland and New York, only to see them get struck down in court. However, courts in Republican-controlled states largely did not return the favor. For example, despite appearing to be slam-dunk illegal gerrymanders under established judicial precedent, Republican-drawn maps in Alabama, Florida and Ohio look like they will stand for at least the 2022 election.
The resulting national congressional rat’s nest doesn’t just hurt Democrats; it also hurts the average voter who just wants their vote to matter. As Republicans drew maps to help protect their vulnerable incumbents, they decreased the number of competitive House seats around the country. The number of swing seats has been declining for years due to both gerrymandering and simple polarization, but this year, we’re on pace for the smallest number of non-safe seats (defined as having partisan leans between D+15 and R+15) in decades.
By our reckoning, the new maps8 have six fewer highly competitive seats (partisan leans between D+5 and R+5) than the old ones. The number of competitive Republican districts (partisan leans between R+5 and R+15) has decreased by even more — 13 seats! However, the number of competitive Democratic districts (partisan leans between D+5 and D+15) has increased by 12.
The change in the number of “solid” and “competitive” seats from the old to the new House maps in states where Democrats enacted the new map, Republicans enacted the new map and both or neither party enacted the new map as of June 14, 2022
“Solid Democratic” seats have FiveThirtyEight partisan leans of D+15 or bluer; “competitive Democratic” seats have partisan leans between D+5 and D+15; “highly competitive” seats have partisan leans between D+5 and R+5; “competitive Republican” seats have partisan leans between R+5 and R+15; “solid Republican” seats have partisan leans of R+15 or redder.
Partisan lean is the average margin difference between how a state or district votes and how the country votes overall. This version of partisan lean, meant to be used for congressional and gubernatorial elections, is calculated as 50 percent the state or district’s lean relative to the nation in the most recent presidential election, 25 percent its relative lean in the second-most-recent presidential election and 25 percent a custom state-legislative lean.
SOURCES: U.S. CENSUS BUREAU, THE UPSHOT, VOTING ELECTION AND SCIENCE TEAM, RYNE ROHLA/DECISION DESK HQ
This reflects the different goals the two parties had in redistricting this year, as can be seen plainly when we break down the seat change by which party enacted the new maps. Having won the 2011-12 redistricting cycle, Republicans had little left on their redistricting to-do list other than to reinforce Republican-held seats that weren’t as red as they could be. As a result, they drew a whopping 18 fewer swing and light-red seats in exchange for 16 more dark-red ones.
On the other hand, the number of solidly blue seats decreased by three in states where Democrats unilaterally redistricted. But the number of light-blue seats grew by six. That’s because Democrats’ goal in redistricting wasn’t to draw safer seats for themselves; it was to draw more seats that leaned Democratic, whether by a lot or a little. The way they did this in states like Nevada and New Mexico was to dismantle solidly blue districts and spread their wealth of Democratic voters out among more districts.
To be sure, it’s a risky strategy. Democrats could hold several seats they might otherwise have lost — but, in a particularly good election year for Republicans, they could also lose seats that wouldn’t have been in danger under the old maps. But for a party trying to dig its way out from under a Republican-biased House map, it may be worth risking more losses in a worst-case scenario in order to make it possible for the party to win a majority in (more common) neutral scenarios.
Up to this point, I’ve been focusing on the partisan impact of the 2021-22 redistricting cycle. But I’d be remiss not to mention the racial impact too. As part of their efforts to draw the best possible maps for their side, both parties — but mostly Republicans — neglected to provide full representation to people of color.
Sometimes, this took the form of denying seats to racial minorities even when their numbers could support them. For example, Texas’s nonwhite population increased by almost 4 million people between the 2010 and 2020 censuses, almost single-handedly earning the state two new congressional districts. But the state did not add any new districts where people of color were the largest racial or ethnic group. Similarly, it is readily possible to draw two predominantly Black districts in Alabama and Louisiana, but the maps passed by those states’ Republican legislatures contained just one predominantly Black district each. (This is why Louisiana’s map has been found illegal, for now.) The new maps in Arkansas, Georgia and South Carolina are also currently subject to lawsuits over racial gerrymandering.
Other times, this took the form of actively decreasing the clout of nonwhite voters. This is difficult, if not impossible, to quantify since a racial group does not have to constitute a majority of a district in order for that district to elect that racial group’s preferred candidate. But districts in Florida, Georgia, Maryland, Michigan, Nevada and North Carolina all got whiter to the extent that their ability to consistently elect Black or Hispanic voters’ candidate of choice is now in question. Most egregiously, the predominantly Black congressional district between Tallahassee and Jacksonville — which was created explicitly to elect Black voters’ candidates of choice and both Democrats and Republicans alike agreed was constitutionally protected — no longer exists under Florida’s new map.
Voting-rights advocates are suing over that maneuver, too, and the case is currently pending before a mid-level Florida appeals court. A final decision isn’t expected before the midterms (especially since the losing side would almost certainly appeal to the Florida Supreme Court anyway), but that doesn’t mean the lawsuit will disappear. In fact, it is just one of many lawsuits that could continue to change the face of the national congressional map for years after the “2021-22” redistricting cycle is supposed to be complete.
On that note, we already know that at least one state will have to go through the redistricting process again during 2023-24: North Carolina. The North Carolina Supreme Court struck down the legislature’s first two attempts at drawing a congressional map and eventually imposed its own, but the map is only valid for one election cycle. Several other states could join North Carolina. There’s Florida, of course, where past rulings by the Florida Supreme Court (albeit a less conservative one) suggest that the new map should be struck down as both a racial and partisan gerrymander. This would necessitate a complete redraw ahead of the elections in 2024 — or, if the case really drags on, 2026.
The Ohio Supreme Court is also currently considering the legality of Ohio Republicans’ second attempt at a congressional map, which is not all that different from the one that was struck down as a partisan gerrymander in January. If they strike it down again, Ohio will have to adopt a new map for the 2024 elections. And regardless, Ohio will have to adopt a new map for 2026 anyway, since this one was not passed with bipartisan support, which is required in Ohio for a map to last the entire decade.
There are also Alabama and Louisiana, whose maps seem destined to be decided by a single U.S. Supreme Court decision next year, as they both hinge on the question of whether the Voting Rights Act requires them each to draw a second Black district. Because this is a federal case, the decision here could also inspire — or force — other states to redraw their maps to include fewer or more minority-opportunity seats, depending on which way the ruling goes.
But the case with the biggest potential repercussions is a federal lawsuit brought by Republicans in North Carolina that makes a radical argument: that only state legislatures, not state courts, have the power to draw new congressional districts. The Supreme Court will decide on Thursday whether to take the case, and if they decide to embrace its argument to the fullest, every congressional map not enacted by a legislature could be invalidated. This would mean not only throwing out court-ordered maps in Connecticut, Minnesota, New Hampshire, New York, Pennsylvania, Virginia and Wisconsin, but potentially also commission-drawn maps in Arizona, California, Colorado, Hawaii, Idaho, Michigan, Montana, New Jersey and Washington. In other words, the 2020s could see a historic amount of mid-decade redistricting.
And this underscores the folly of ever declaring redistricting complete; maps are never set in stone. The national congressional map might be “darn close” to final for 2022, but it will continue to evolve in the years after. While it’s tempting to wrap up our redistricting coverage with a neat little bow, I’m not going to do that. Redistricting doesn’t work like that. Maps could still change; partisan and racial biases could still get better or worse. This isn’t goodbye; it’s see you later.
Granted, that’s also what I said before Louisiana’s map was struck down. However, there’s something special about Louisiana: It doesn’t hold a traditional primary, so unlike every other state, it doesn’t need to use a new congressional map before the November election.
A likely outcome given the conservative nature of the 5th U.S. Circuit Court of Appeals and the U.S. Supreme Court, the two courts above the one that struck down Louisiana’s map.
Partisan lean is the average margin difference between how a state or district votes and how the country votes overall. This version of partisan lean, meant to be used for congressional and gubernatorial elections, is calculated as 50 percent the state or district’s lean relative to the nation in the most recent presidential election, 25 percent its relative lean in the second-most-recent presidential election and 25 percent a custom state-legislative lean.
If Louisiana’s old map is not reinstated and the legislature draws a new map that has two predominantly Black seats — as ordered by the lower court — you can add one seat to Democrats’ totals in this paragraph and subtract one seat from Republicans’.
However, if Louisiana’s map is not reinstated and the legislature draws a new one with two predominantly Black seats, the tipping-point seat would go back to having a partisan lean of R+2.3.
Here’s the gory math behind that calculation. Between nine and 12 Democratic-held districts have been redrawn to favor Republicans (two in Arizona, three in Florida, one in Georgia, one in Michigan, one in Ohio, one in Tennessee and potentially one each in Kansas, Maryland and New Jersey), while between four and seven Republican-held districts have been redrawn to favor Democrats (two in California, one in Illinois, one in Michigan and potentially one each in Louisiana, New Mexico and Ohio). That’s a net gain of five seats for Republicans, give or take. However, Republicans also lost one or two seats on net from reapportionment: Population trends added one red seat in Florida, one in Montana, one in Texas and potentially one in Colorado, but they also subtracted one from Illinois, one from Michigan, one from New York, one from Pennsylvania and one from West Virginia.
Although it’s worth noting that Democrats could have enacted Maryland’s and Massachusetts’s unilaterally had they wanted to, since they have veto-proof majorities in the legislature in those two states. As it happened, though, Republican Govs. Larry Hogan and Charlie Baker, respectively, agreed to sign their Democratic-drawn maps.
Other than Louisiana’s, for now.
Nathaniel Rakich is a senior elections analyst at FiveThirtyEight. @baseballot
Filed under
2022 Election (321 posts) Redistricting (83)
2022 Election
2022 Election
Supreme Court
Supreme Court
2022 Election
2022 Election
See all newsletters
© 2022 ABC News Internet Ventures. All rights reserved.