What Are the Swing States? Part 3

A post-midterm update.

The 2018 midterm elections have come and gone. Aside from one House race in North Carolina (NC-09 remains vacant due to election fraud), we now have a lot of interesting data to pick through.

One big takeaway from the midterms was that the vote for House was highly correlated with 2016 presidential vote, even more so than it was in 2016. You can infer that will probably be true in 2020 as well, so I wanted to use that information to predict how swing states will behave.

I pulled the total vote for all of a state’s elections to the House of Representatives. Then I sorted all of the potentially competitive states by the share of the vote won by Democrats. It stands to reason that a higher share of the vote for Democratic House candidates means a higher proportion of voters willing to vote for a Democratic president.

There are not too many surprises here. States with a pronounced Democratic lean (New Mexico, Minnesota, Virginia) are at the top of the list. States with a pronounced Republican lean are close to the bottom (Georgia, Texas, Ohio). Wisconsin is right in line with the national environment, reflecting its swing-state status perfectly. I’m a little surprised that Colorado is right there with it, but the Republican share of the vote is almost 3 points lower than Wisconsin’s. Florida is regrettably not informative here. The state does not report vote totals for uncontested elections, so Democrats are missing all of their House votes for FOUR districts. This artificially lowers Florida’s placement on the list. If you want to think about this graph in a different way, I have also sorted it by Margin instead of vote share.

This table more accurately ranks the states, in my opinion. Colorado was too low on the first one, and Pennsylvania was too high. However, the placement of Florida still bothers me. It’s fundamentally inaccurate since there are plenty of Democratic voters who showed up for the election but just don’t have a vote reported in this particular type of contest.

I’m not skilled enough to reverse engineer how many Democrats voted in those FL districts (which is technically possible, using votes for Governor and Senate as a proxy). Instead, I decided to estimate. I added 100,000 votes for Democrats and Republicans in three districts, with 50,000 and 10,000, respectively, in the last one. I then added half the 2016 raw vote margin for the first three districts onto the first 100,000. This ends up producing an additional 478,000 votes for Democrats, and 310,000 for Republicans. The Republican vote is also included because I’m trying to estimate votes for 2020, where Republicans will be voting for President even if they don’t have a candidate for House. I realize this is a very crude and messy way to estimate, but for this type of analysis it won’t matter too much. I just want to attempt to get a more accurate idea of where Florida falls in the ranking of swing states.

This final chart is my idea of what a ranking by House vote would look like.

A few notes from this analysis:

This is all I have for now. Soon I hope to summarize each of my analyses into a unified ranking that incorporates all the methods I’ve looked at in parts one, two and three. Maybe by then we’ll have some polling data to look at too.

Related posts:

My immediate thought was, “Isn’t that the point?” What am I supposed to do? Hope that it happens? Anything that I want, I work hard to make it happen — isn’t that the way it goes? Like, really…

This year was amazing. My life started falling in place in every aspect. This was the year where I would wake up happy to go to work, This was the year where I started living with a purpose, This was…

is it because love is messier that we’re readier for order? is that why we remain stranger though no one knows us better than we do about each other? i know you don’t like waiting table but still…