Saturday, January 27, 2018

Border Control

In case you haven't read it yet, fivethirtyeight.com has a huge gerrymandering project up on their page, which has a lot of...interesting results. The map is linked above, but there's a lot of interesting things there.

(If case you've forgotten your eighth grade civics class, gerrymandering is the act of drawing Congressional district lines to benefit one party or the other. Since House districts are geography-based, they have to be manually set up every ten years after the census.)

They basically ran the math and came up with a bunch of different scenarios--the current state, emphasizing competitive districts, using computer algorithms, and so on. They also released a series of podcast episodes in conjunction with the project, which looked at blatant partisanship (Wisconsin), a creative "communities of interest" method (California); an impartial panel (Arizona); and dealing with minority-majority districts (North Carolina). If you are interested in that sort of thing, it's worth a listen.

Gerrymandering reform is always on the radar for a lot of reformers and activists, but this entire project seems to give mixed results--in fact, it's probably disappointing to a lot of people. It turns out there's two major factors, here:

  • Drawing districts is hard. No matter what, you can't draw it perfectly.
  • It probably also doesn't make that much of a difference.
For about two decades, reformers have blamed Republicans for gerrymandering their way to the House of Representatives. However, if you look at the maps, it turns out that the difference is minimal. The current state has an expected Republican seat count of 234. If we draw the districts with a computer algorithm--removing all biases--it's 232. In fact, nearly all of the cases they build have a difference of only about 10 seats total--and since there's 40-70 competitive seats in between, it just straight up doesn't make a difference.

It's also important to note that a lot of these scenarios ignore the Supreme Court's ruling for majority-minority districts--districts that must be drawn to make sure there is minority representation, which often has the practical effect of lumping Democratic voters together and pulling them from other theoretical districts, meaning that the GOP has an advantage in those suburbs. Or maybe; the effect of this is also pretty minimal, but it does exist and it does hamstring people who are trying to make effective borders.

Now, the data does show that there's a case for making more competitive seats. There are a few models that spike the GOP seats up by a lot, but it also spikes the Democratic seats as well--all the gains are at the expense of competitive seats. But even this isn't the fault of gerrymandering--the American population is already "self-sorting" enough that gerrymandering isn't necessary. Seats have become more partisan and 'safe' not because of redistricting, but because the people have started to think like their neighbors.

This has always been suspected by people who (like myself) were always skeptical of reform. A few papers had been done in the past estimating that the difference was 2 seats or less. Since there are several different goals, I won't say this confirms it, but it's pretty close.

What's the solution? Well, first we have to decide whether we need a solution. There's a case--not a strong one, but one nonetheless--that competitive seats aren't necessarily the best thing. What is better for democracy--a district where 50% of the people consistently don't like who is representing them, or a district that gets 90% of the vote for one party, meaning that nearly everyone in the district likes their representative? I'm not sold on that sentiment, but I'm not against it, either. Like most things in politics, it's probably somewhere in the middle.


An interesting solution I came up with is that all states have at-large representatives. Everyone still only gets to vote for one candidate, but you have as many candidates to choose from as slots in your state (times two, plus any third party). Then on election day the top Representatives win. Let's take the example of Virginia, which has 11 representatives. Probably around 25 people would run--11 from each major party and a few third parties. On election day, each voter votes for one, and the top 11 vote-getters win election. No districts (besides the state), no votes are "wasted" due to artificial districts, and you can feel represented even if you voted for someone across the state.

There are drawbacks. First is information--that's a lot of candidates to get to know. Chances are, then, we'd still see some form of localization. If you live in Fairfax, you're probably going to vote for someone around your area--but now no one has to worry if they live specifically on the right street or not. Parties would have to make concentrated efforts to make sure that votes don't get split--if all the minorities in a city split their vote too many ways, they may see no one elected at all. And finally the localization would be gone--there's no guarantee that 11 Representatives would come out of Northern Virginia and zero from south Virginia, something that is guaranteed now. That may not sit well with people. It's not likely, but with 50 states and an election every two years, it's a statistical likelihood that would eventually happen.There may also be a danger of strategic nominations--if the Republicans only nominate, say, 8 candidates, they are almost certain to win those 8 (since the Democrats will split their votes 11 ways). They won't sweep the state, but they can guarantee a majority--right up until the Democrats do the same.

And therein lies the problem. No matter what solution we present, it is at the expense of something else. If we want more competitive districts, the lines can get pretty wild and will split cities in two (or more). If we want minority representation, we have to accept that that means less minority representation elsewhere. If we want people to be represented by their interests, that may mean feeding into incumbency. And so on.

Turns out, this stuff is hard. 

No comments:

Post a Comment