The good news on redistricting isn't that good

The good news on redistricting isn't that good

By Matthew Yglesias

After months of worrying about a 2020 redistricting wipeout, Democrats now seem likely to be facing slightly friendlier maps rather than more hostile ones when all the boundaries are finalized.

This can be measured in different ways, but Joel Wertheimer at Data for Progress looked at each seat and asked whether it is to the left or to the right of Joe Biden’s 4.5% national margin in 2020. He finds, based on the maps that have been finalized so far, that 16 seats have switched from right of the margin to left of the margin, and it’s possible that a favorable court ruling in Ohio would result in there being slightly more seats to the left of Biden’s margins than to the right.

(Bill Clark/CQ Roll Call)

This will naturally turn down the heat on the idea of national gerrymandering reform, which is probably for the best since no such reform was passing anyway. There seems to be some prospect of a bipartisan agreement on the Electoral Count Act, and I hope they do that. But I also continue to think that gerrymandering is really bad even if 2020 redistricting turned out okay. And in particular, I think the situation whereby Dem-drawn gerrymanders balance out GOP-drawn ones to create a nationally balanced map is actually pretty bad.

The commission route that reformers have been pursuing is a bit of a dead-end, I think. If we reboot this discussion, we really ought to push for proportional representation along the lines suggested by FairVote.

The 2020 redistricting story

So what happened? Democrats had hoped for a huge national landslide in 2020 (as forecasted in the national polls) that would let them flip some state legislatures and redraw some GOP gerrymanders.

Those hopes were dashed, and Republicans were left with 193 seats subject to GOP partisan gerrymanders against just 94 for Democrats. That imbalance reflected not only performance in state elections but also the fact that California (and several other states with Democratic trifectas) adopted an independent redistricting commission system. So Democrats were prepping for a wallop, but it didn’t materialize for several reasons. One is that a lot of the GOP-drawn maps post-2020 were already heavily gerrymandered after 2010, so there was no change to them. Another is that Republicans got skittish about underlying political change in Texas and drew a defensive map shoring up their incumbents rather than an aggressive one going after Democrats.

And finally, several Democratic-run states (notably Illinois, Oregon, New York, and Nevada) drew more aggressive maps.

When talking about drawing new maps, it’s important to understand what we mean by gerrymandering. A lot of conservatives looked at the old maps in Illinois and Maryland as examples of egregious Democratic gerrymanders because they had lots of funny-looking districts.

But in both of these states, Democrats created elaborate snake-like districts to maintain the particular ethnic balances of different seats. So while the shape of something like Anthony Brown’s Fourth Congressional District in Maryland does indeed look absurd, this is not a map that is maximizing Democratic representation in the state. Far from it, in fact; drawing Brown such a lopsided D+29 seat is a gift to Republicans.

This map is funny-looking because it was drawn to ensure that Brown has a safe majority-Black seat and Steny Hoyer next door has a safe majority-white seat.

Maryland stuck with this plan, preserving a safe Republican seat on the Eastern Shore for Andy Harris. But Illinois and New York did not, instead moving toward plans that shift the balance toward partisan advantage. In addition, commissions in New Jersey and California gave Democrats favorable maps.

Long story short, the national map got fairer not because GOP gerrymanders were unwound, but because Democrats countered them with new gerrymanders of their own.

Mismatched voters deserve representation

Balanced gerrymandering is preferable to the alternative of unilateral disarmament.

But I still think it’s not that good. Among other things, the basic problem with gerrymandering is not unfairness to political parties but unfairness to voters. The GOP majority in the Utah state legislature has chosen to respond to population growth by dismantling the idea of a Salt Lake City-based congressional district that had become winnable for Democrats. That Democrats in Illinois have balanced that out by squelching representation for GOP-leaning voters in the Chicago exurbs doesn’t eliminate the harm to voters.

And in sociocultural terms, I think these practices undermine the national fabric by creating an exaggerated sense of national polarization.

Barack Obama became famous on the national stage with a speech about how “we coach Little League in the Blue States and yes, we’ve got some gay friends in the Red States.” The fact that there are about 500,000 Trump voters in Chicago and about 500,000 Biden voters in Utah is a relevant fact about American society that deserves to be made more salient so that people understand more clearly and vividly that opposite-party voters are members of their community and not people living somewhere else to be eliminated in Marjorie Taylor-Greene’s national divorce.

But the only way to really ensure representation for voters who are members of localized political minorities is with a proportional solution.

Representation matters more than aesthetics

Donald Trump got about a third of the vote in Massachusetts in 2020, but there turns out to be no “Trumpy” part of Massachusetts (I’m focusing on the presidential vote because, in recent elections, it explains 97% of the variance in House voting patterns). The western end of the state is, in political terms, an extension of Vermont — an eccentric slice of the rural United States that is very liberal in politics despite being full of white people. Trump lost every county in Massachusetts, and even in his closest county, he only got 40%.

Because this is New England, you can drill all the way down to the town level with this WBUR map and see a few red blobs, but they’re small.

(WBUR map)

The upshot is that even if you gave Republicans carte blanche to gerrymander the state, there’s not much they can do. The current Democrat-drawn map creates nine safe Democratic seats. If you gerrymander it to maximize GOP partisan advantage, you get eight safe Democratic seats and one swing seat. By the numbers, Massachusetts should have a map that features five Democratic seats, two Republican seats, and two swing seats. But there’s no way to draw a map like that because Massachusetts’ Republicans are too well integrated into their communities.

By the same token, even though Biden won 34% in Arkansas, you can’t draw a 3-1 congressional map in that state. Arkansas happens to be a state whose largest city is small, but whose rural areas contain a fair number of African Americans. If every Democrat living in rural Arkansas moved to Little Rock, then you could create a Democratic seat. But you can’t. In Wisconsin, you get a similar result from an opposite process — too high a share of the state’s Democrats live in hyper-segregated Milwaukee precincts than make it possible to draw an efficient Dem-friendly map.

Some analysts, like The New York Times' Nate Cohn, argue that we should measure gerrymanders relative to these baseline facts about population distribution.

“What percentage of people live in Democratic precincts? What percentage live in Democratic counties?” he asks, “That's slightly different than vote share, but it's the one that really informs this problem.”

I think there’s a level on which that’s true. It is factually accurate to say that the 9-0 Massachusetts map, the 0-5 Oklahoma map, and the 0-4 Arkansas maps do not involve any particularly extraordinary line-drawing efforts. But I still think it’s bad to represent (or rather, not represent) citizens in this way, and the fact that in these cases the numbers even out only partially mitigates the unfairness.

Representation shouldn’t be based on segregation or where water is

Alabama’s congressional map is somewhat more proportional than Oklahoma’s because, in a state with very high levels of racial polarization in voting, there is also enough residential segregation that you can (and under the Voting Rights Act must) draw a “Black” district like this one that Terri Sewell represents.

But suppose Alabama becomes much less segregated in the future, even while people continue to disagree about partisan politics. Well, then Alabama’s Democrats get erased the way the Democrats in Arkansas and Oklahoma or the Republicans in Massachusetts are. Or suppose the voting becomes less racially polarized — same problem.

I don’t think it’s healthy for society to have huge levels of racial segregation or in general for people to be spatially sorted based on partisanship. But it’s doubly dysfunctional for our system of representation to rely on dysfunctional residential dynamics to generate a reasonable outcome.

The way a state like Massachusetts should work is the whole state is a single constituency that then elects representatives using a proportional method. There are a bunch of different proportional methods out there, and I don’t have strong feelings about which one is best — except that in my experience talking to German voters during the 2019 general election campaign, citizens cannot comprehend how mixed-member PR works. I think enacting any form of proportional representation on a state-by-state basis essentially achieves mixed-member’s goal, which is to incorporate some element of localism into the calculus. Massachusetts would usually send 5-7 Democrats to D.C. and 2-4 Republicans, but they would all have the specific obligation of representing Massachusetts.

A few states might be big enough that you’d want to chop them in half (one big district for Westchester + NYC + Long Island and a second one for the rest of the state) and California and Texas should maybe be cut in three. I also wouldn’t mind the idea of tripling or quadrupling the total number of representatives so you wouldn’t have so many states with only one or two House members and little room for proportionality.

If for some reason you’re addicted to small districts, one thing you could do is allocate congressional districts the exact same way we do now but have each one elect four members on a proportional basis. I think that’s unnecessary complexity, but I don’t hate it.

Multiple parties might be nice

In the public imagination, proportional representation is closely linked with multi-party politics.

I’m not sure this is correct. You see robust third parties in countries like Canada and the UK that use first-past-the-post voting in single-member districts. Indeed, in a lot of ways, that electoral system is advantageous to small parties with geographically concentrated bases of power. The separatist Bloc Québécois always gains way more seats in the federal parliament than their share of the national vote would warrant because single-member districts are good for regionalism.

I’m inclined to say the main reasons the United States doesn’t have multi-party politics are actually the fixation on the presidency and the openness of the primary process.

In other words, I think the reason Trump ran in the GOP primary rather than running as a third-party populist is that our parties don’t have institutional means to keep out undesired entrants. If we did have such institutions and the party establishment told him he wasn’t allowed to run, I don’t think fear of being a spoiler would have stopped him from going third-party or even necessarily stopped people from voting for him.

Long story short, I’m not actually sure how much of a game-changer proportional representation in House races would be for third parties. I think that Joe Manchin and Lisa Murkowski should leave their respective parties and form a small independent caucus (Democrats would be crazy to run spoilers against them), but whatever psychological or institutional dynamic stops them from doing that seems to be the key factor here, not a lack of proportional representation.

But if we did get some small additional parties in Congress, I think that would be a welcome change. If nothing else, making it really hard to ever have a one-party majority would help revive the idea of Congress as having powers of its own and not just partisan interests in supporting or thwarting the president. But the most important thing is that whether or not you got additional parties, proportional House voting would make everyone’s interests actually matter. Every state has marginal voters, and there are things politicians could try to do to appeal to those voters. But right now, the marginal voters in Massachusetts and Utah and Oklahoma don’t matter, so nobody cares what the GOP has to say to or about people in Massachusetts or what the Democratic Party has to say to or about people in Utah and Oklahoma. It’s a bad problem over and above the problems of partisan bias, and it’s crying out for a fix.