Since 1994, the GOP is getting a lot more bang (House seats) for its buck (votes)

As is the case when I see an interesting graph, I occasionally want to re-create it. So I re-created the one above, pulling data from the House website and using Cook Political Report’s 2020 House vote tracker. Instead of looking at Democratic vote share, I decided to look at Republican share, a shift that in practical terms doesn’t matter, given that the vast majority of House seats are held by one of the two parties.

I also decided to go back a bit further, from 1962 forward. In other words, to compare each decade’s vote-to-seat ratio. Then I broke out each decade (from the first election after a redistricting to the year of the following Census) and looked to see what happened.

There’s a lot happening there, so let’s break it down. In the first three decades, from 1962 to 1990, the dots (which are scaled to total votes cast) are below the diagonal line, meaning that the GOP was getting a higher percentage of the vote than it was House seats. Then, after a good election for Democrats in 1992, the dots jump above the line.

Here’s a static version, showing the elections before 1992 (light orange) and from 1992 on (dark orange).

1994 is a pretty clear demarcation. That year was a Republican wave that brought the party back to power in the House for the first time in four decades. Since that election, only once, in 2008, has the GOP had a lower percentage of the vote than seats. In every other election, it’s secured a greater percentage of seats than it has a percentage of votes.

Interestingly, that ratio hasn’t been as beneficial for the GOP in the last two elections. They’ve been two of the four least unbalanced elections since 1994.

Why is this happening? One factor may simply be that the GOP is earning a higher percentage of the national House vote than it used to. From 1960 to 1992, it earned an average of 45 percent of that vote. Since, it’s earned an average of 48 percent. Because the percentage of votes cast is necessarily more finely represented than the percentage of House seats (you can’t win precisely 50 percent of House seats, for example) that necessarily means that there will be some disconnect in most years.

You can see on Morris’s graph that the percentage of the vote won by Democrats has been under 50 percent for most of the recent elections, which correlates to a lower density of seats won.

Neither my graphs nor his answer the question of why completely and unequivocally. But both point in the same direction: The…