For our latest “Voices & Choices” podcast, we invited FairVote Law and Policy Director Drew Penrose to talk about FairVote’s congressional race projections for 2020, as part of our ongoing Monopoly Politics project. The following is an excerpt from that interview, slightly edited for clarity.
Rich Robinson: Before we get into the projections for 2020 tell us a little bit about how FairVote started doing these.
Drew Penrose: FairVote has been doing this since 1997. It started with the insight – which was really very new at the time – that you could do a lot of the work of projecting election results, well in advance of the election, just by looking at how people had voted in each district or the presidential candidates. So, if people voted for the Democrat at a higher rate than the nation as a whole voted for the Democrat, then we'd call that a Democratic district and vice versa for the Republicans. You can measure that out and if a district is overwhelmingly Democratic or overwhelmingly Republican, it gives you a good sense that that's a safe district.
So that was a way of measuring at a time when people really thought that [winning] all came down to campaign tactics: how much money you raised and spent, how smart your campaign was and that sort of thing. Our insight showed that that can move the dial a little bit, but at the end of the day the underlying partisanship of the district is everything when it comes to a winner-take-all-system. That’s sort of common knowledge now, which is why we have this epidemic of Gerrymandering. Gerrymandering uses that as a premise, essentially drawing the lines to shape what the results will be. But our insight showed that once lines are laid down, you've got the election results for a lot of the districts already without anybody casting a vote.
Robinson: Were there any surprises when the first projections were made and they came back and we realized how closely that we could forecast these races?
Penrose: Well honestly there really wasn’t, because we only forecast the very, very safest of races. I mentioned we’ve been doing the general project since 1997. It was early in 2012 that it took on the form that it has now, with a really rigorous methodology and the release of the projections two years in advance. Monopoly Politics 2012 was the first time we did it in a much more rigorous way. And that year it was a little surprising. I think that we came out with 100 percent accuracy in our high confidence projections. So, we projected a smaller number of seats that year, 333 seats for 2012. But every single one of them was correct.
Robinson: Wow. And that level of accuracy has maintained since then?
Penrose: For the most part. In 2014, there was one projection that was wrong, that was a big surprise. But basically all of these elections are all statistical things, so there's always some chance that something will be wrong. The more abnormal the year, the higher number of things we expect to get wrong. We got one wrong in 2014 which was a pretty heavily Republican year. In 2016, we again got everything right which as much as people think of 2016 as this change year with the election of Donald Trump, when you look at the congressional results, it actually was about as boring a year as you can get. It was nearly evenly divided 50/50. Nearly all of the incumbents were reelected. 2018 on the other hand, was the most unusual congressional year I've ever seen. And we got a larger number of them wrong. We made projections on 379 seats going into 2018. We were correct on 367 of them, so only a 97 percent accuracy rate for our highest confidence projections, which shows how even in the most wild of years, it still remains the case that the district lines are fundamentally determining our election outcomes rather than all of the energy that goes into the voting and campaigning and all that.
Robinson: I don’t know, Drew. It sounds like we’ve gone from an A+ to an A.
Penrose: Well, maybe (chuckles). Over four elections cycles we’re at 99 percent accuracy. I’m not shaken as to 2020. I think we’re still well within the bounds of reason to say we can call these races two years in advance.
Robinson: I’m glad you got there. Let’s talk about 2020…