Could Gerrymandering Backfire On Republicans?

I have written before that electoral college advantages are inherently unstable. The reason is that having an electoral college advantage — meaning that you win the electoral college while losing the popular vote — means that you managed to win a bunch of states by small margins, and lose others by big margins. Since small margins are likely to reverse in subsequent elections, an electoral college advantage can turn into a disadvantage pretty easily.

It’s harder to reverse a gerrymandered advantage, for two reasons. One is that a gerrymandered advantage usually engineers bigger margins of victory. An electoral college advantage means winning a state like Pennsylvania by less than a percentage point. A gerrymandered advantage means that you draw a bunch of districts your party can win by 10 points, and then a few your opponents can win by 20 points. Obviously, a 10-point advantage is more robust than a 1-point advantage. The second reason is that while state lines are pretty well set, districts get redrawn all the time, and you can use the redrawing process to regain your hold on a gerrymandering advantage that may have slipped away.

But I think it’s possible that Republicans could end up having gerrymandered themselves into a disadvantage in some states in 2018. To reiterate, I’m not saying this is likely. I’m saying it’s possible. And I think it’s likely that the Republican districting advantage will at least not be as strong as many people have assumed.

There are two forces that would push back against the Republican gerrymandering advantage. The first would be a wave election for Democrats. Imagine the following toy example. There are 300 people in the state, 150 of which vote Republican and 150 of which vote Democrat. There are three Congressional districts. Republicans draw districts so that district Blue has 65 Democrats and 35 Republicans, district Red has 60 Republicans and 40 Democrats, meaning that district Purple must have 55 Republicans and 45 Democrats. Note that in this case Republicans have engineered a districting advantage, since they win two districts out of three, even though the total votes are evenly split. They have done this by loading one district up with a big Democratic victory, and winning the other two more narrowly.

What if a big wave election happens and lots of people switch from Republican to Democrat (or turnout is higher among Democrats)? Suppose, for example, that 10% of Republican votes switch to Democrat. Then Democrats win the Purple district in my example: 10% of 55 is 5.5; round it up to 6 and take those votes over to the Democratic side, and Democrats win that district 51–49. Democrats win two out of three districts (67%) despite winning only 165/300 = 55% of the popular vote. Democrats end up with a districting advantage even though districts had been drawn by Republicans to favor Republicans.

Could this actually happen in 2018? Let’s look at North Carolina as an example. North Carolina has already been ordered by a court to redraw its districts for 2018, but let’s suppose the current districts were kept in place. The GOP won 53% of the House votes in 2016, and 10 of the 13 seats. This is because Democrats won by an average of 37 points, while Republicans won by an average of 21 points. How big would a Democratic wave need to be to wash away the Republican districting advantage?

Pretty damn big. Here’s a list of 2016 results, ordered from biggest Dem victory to biggest GOP victory:

Dems could win an extra 6.2% of the vote away from Republicans in 2018, winning the popular vote 53–47, and still — if those shifted votes are distributed evenly across these districts — win only one additional seat (Ted Budd’s seat in District 13 — if Dems win 6.2% more of the vote, and Rs win 6.2% less, Budd’s 12.2% margin of victory would be wiped out). If they won an additional 9.5%, then they would win the popular vote 56–44, and would win 9 of 13 seats (69%) — a districting advantage, in districts drawn by Republicans. Winning an additional 9% and 8 of 13 seats would also yield a districting advantage.

Winning an extra 9% of the popular vote is really hard. Especially for Democrats in a non-presidential year. Even in the 1974 midterms after Nixon was impeached, Democrats gained only 5.5%. It’s not impossible. Democrats started from a higher level in 1974, so maybe gains were more difficult. And surely in some states they gained 9%. But still, even with all the organizing energy surrounding the Democratic party now, and all the opposition to Trump and the Republican Party, gaining 9% is extremely difficult.

But remember I said there were two forces that could push against the Republican gerrymandering advantage. Increasing Democratic share of the popular vote is one. Another is that the Democratic share could stay the same, but the composition could change. Going back to my toy example, suppose that 6 votes in the Purple district switched from Republican to Democrat, but 6 votes in the Red district switched from Democrat to Republican. Then the overall popular vote would stay the same, but Democrats would win two of the three districts, and therefore have a districting advantage.

Of course, both things could happen at the same time, and reinforce each other — that is, Democrats could gain overall vote share, and the gains could be particularly concentrated in districts that are now Republican, but kind of close. Going back to the North Carolina example, suppose that Democrats gained 5% of the popular vote. This wouldn’t be enough to win any additional seats if that gain were evenly distributed across the state. But suppose they gained 10% in districts 13, 2, 9, 5, 8, and 6, and then lost a few percentage points in 10, 11, and 3, keeping the overall gain at only 5%. Then they would win 9 seats and have a big districting advantage.

Is it possible for that to happen? To that extreme of a degree, probably not. But some of that is happening already. A lot of those “close” districts that Republicans are winning include white suburbs with a large share of college-educated, fairly wealthy voters. That’s exactly the group that shifted away from Trump and towards Clinton in 2016.

Take, for example, District 2, represented by George Holding (full disclosure: George Holding is my uncle. I would be extremely happy if he lost to a Democrat in 2018). This district covers suburbs of Raleigh. Holding won by 13 points in 2016. Trump won this district, but by only 10 (Romney won it by 13). Of all North Carolina districts, District 2 had the second-biggest drop in Republican share of the presidential vote from Romney to Trump. Holding won the Republican primary against Renee Ellmers, who was the first member of Congress that Trump endorsed. This is Republican territory more than it is Trump territory. As the Republican Party gets more Trumpified and Republicans in Congress carry more water for the administration, some of these votes might go to Democrats. Or at least, of the votes that might switch, a lot of them could be found here. The same goes for Districts 8 and 9, which include Charlotte suburbs (although Trump won the same vote share as Romney in these districts).

The first big test case for this theory will be the special election for Georgia’s sixth district, the seat Tom Price left to become HHS Secretary. GA-06 is an Atlanta suburb that voted for Trump 49–48 after voting for Romney 61–38. Democrats are focusing a ton of resources there to support Jon Ossoff, who has already collected the endorsement of John Lewis, over $1 million in donations, and 3,500 volunteers. Special elections normally present the Democrats with turnout challenges, but the huge turnout and huge victory in Delaware’s State Senate election yesterday is encouraging.

Winning GA-06 is going to be very difficult. Even getting within 5 points should be considered an extremely encouraging sign for Democrats. Winning NC-02 would be even harder (it’s not even on the DCCC’s list of targets, although 8 and 9 are). But if Ossoff wins, or even comes close, Democrats will be able to recruit better candidates, and will devote more money and resources to these races in suburban districts. These are districts that form the backbone of Republican gerrymandering schemes. Flip a couple of those, even as Republicans run up their tallies in already-red rural districts that form Trump’s base, and the districting advantage starts to evaporate.

I think there is a small, but underestimated, possibility that GOP gerrymandering does end up backfiring, and Democrats end up with a truly huge victory in 2018 — like 40 seats — despite “only” gaining 5–6 percentage points in the popular vote. But Democrats do not need to turn the Republican districting advantage around entirely to win back the House. They could win the popular vote by less than the seven or eight points some argued they needed to win the House in 2016 and still win the House in 2018.

Courts are not the only limitation on gerrymandering. Gerrymanders are designed to maximize one party’s share of seats given the lay of the land when the districts are drawn. When the political ground shifts below the party’s feet, some seats can end up falling into the cracks.