Composition of the House by Region, 1912-2022

[Antonio the Sixth]

So a while ago I went back and collected election results for House elections since 1912. I looked at how the seats broke down in each of the 4 basic regions as summarized below.

The chart below plots the % margin for Democrats for each region (in other words, if Dems won 70% of the seats in one region and Republicans won 25%, the margin will be 70-25=45%). The reason I did it this way is because, at several points, a significant numbers of independents won seats and I didn't want this to count for one party or another. When you assume no independents, though, the conversion into Democratic % is pretty easy: 0% means 50%, 10% means 55%, etc.

This is what it looks like, with data updated for 2018:


As you can see, Democrats made gains in every region of the country, but significantly more in the two regions they were already dominant in (Northeast and West), while making only modest gains in the Midwest and South (note: for the purpose of this analysis, I assumed NC-9 went Republican; I'll update the numbers if necessary, but the pattern doesn't change). This means that regional polarization is at a historical high: for the first time since 1946, different parties hold over 2/3rds of the seats in three of the regions. In the Northeast, Democrats are only slightly less dominant than they were in 2008 (which remains their best performance ever in this region). In the West, the Democrats control a far larger share of seats than they did in 2008 - in fact, it's the largest since the Great Depression era, when this region only had around 40 seats. In the South, meanwhile, while Republican dominance has subsided somewhat, Republicans are still stronger than in any election pre-2010. The Midwest is the most evenly divided of all regions, but even there, long-term trends look decently good for Republicans, since they are still well above their 2008 or even 2006 levels there. In fact, their performance this year, in the Midwest is on par with their 1994 blowout.

In terms of long-term historical trends, the obvious one is the steady realignment of the South, but everything has already been said about that. Other interesting patterns include the uneven but significant long-term growth of Democrats outside the South between 1952 and 1976. This period seems to have been crucial, as it allowed Democrats to maintain a majority until 1994 despite their eroding strength in the South. After the 1994 earthquake, these gains maintained themselves and even consolidated in the Northeast and West, but seemingly eroded in the Midwest, where the parties roughly are in a similar position as they were in the 1960s.

[Tartarus Sauce]

Good job on this, the data is quite stark when visualized. It's interesting that regional polarization is so high right now given that regionalism seems to be fading as a significant driver of partisanship compared to past trends. The urban-rural divide is all the rage now.

I wonder at what point Democrat seat representation will reach parity in the South and Midwest. As the Midwest continues to stagnate and it's urban areas decline while the urban areas in the South continue to grow and become more Democratic, it's likely bound to happen at some point.

[Antonio the Sixth]

I wonder at what point Democrat seat representation will reach parity in the South and Midwest. As the Midwest continues to stagnate and it's urban areas decline while the urban areas in the South continue to grow and become more Democratic, it's likely bound to happen at some point.

I figured this Southern breakthrough might happen this time around, given the amount of talk given to States like VA, NC and TX, but instead, Southern divergence from non-Southern regions is at its highest point since the Civil Rights Act (of course, back then it was in the opposite direction). And what's worse, there's really not much room for Democrats to grow, at least until they break some nasty gerrymanders. I don't think the South will be at parity again in our lifetimes, honestly.

Midwest could possibly happen in a massive wave (there are more seats with ancestral Democratic strength there) but even that is unlikely.

[Tartarus Sauce]

I wonder at what point Democrat seat representation will reach parity in the South and Midwest. As the Midwest continues to stagnate and it's urban areas decline while the urban areas in the South continue to grow and become more Democratic, it's likely bound to happen at some point.

I figured this Southern breakthrough might happen this time around, given the amount of talk given to States like VA, NC and TX, but instead, Southern divergence from non-Southern regions is at its highest point since the Civil Rights Act (of course, back then it was in the opposite direction). And what's worse, there's really not much room for Democrats to grow, at least until they break some nasty gerrymanders. I don't think the South will be at parity again in our lifetimes, honestly.

Midwest could possibly happen in a massive wave (there are more seats with ancestral Democratic strength there) but even that is unlikely.

Sorry, I wasn't clear, when I said parity, I didn't mean parity between Democrats and Republicans. I was pondering when Democratic representation in the Midwest and the South would be at parity in relation to each other. That's probably not too far away in the grand scheme of things, I could see it happening within the next decade potentially.

[Annatar]

Interesting look at the data, its interesting how the GOP really has become the party of the south and the Midwest. Something I noted this year is that that even in a D+9 climate, the GOP picked up 3 senate seats in the Midwest, lost only a net of 7 house seats in the Midwest, the fewest of any region in the country and split the governor elections in the Midwestern Obama Trump states 2-2. In the south as well, the GOP picked up a senate seat and the house losses were fairly small with some of the seats lost being real anomalies like OK 5 and SC 1. The west by contrast was a disaster, not only did the GOP lose 12 seats, but also lost 2 senate seats, the north-east as well where 11 seats flipped from an already fairly small GOP congressional delegation.

[Antonio the Sixth]

I wonder at what point Democrat seat representation will reach parity in the South and Midwest. As the Midwest continues to stagnate and it's urban areas decline while the urban areas in the South continue to grow and become more Democratic, it's likely bound to happen at some point.

I figured this Southern breakthrough might happen this time around, given the amount of talk given to States like VA, NC and TX, but instead, Southern divergence from non-Southern regions is at its highest point since the Civil Rights Act (of course, back then it was in the opposite direction). And what's worse, there's really not much room for Democrats to grow, at least until they break some nasty gerrymanders. I don't think the South will be at parity again in our lifetimes, honestly.

Midwest could possibly happen in a massive wave (there are more seats with ancestral Democratic strength there) but even that is unlikely.

Sorry, I wasn't clear, when I said parity, I didn't mean parity between Democrats and Republicans. I was pondering when Democratic representation in the Midwest and the South would be at parity in relation to each other. That's probably not too far away in the grand scheme of things, I could see it happening within the next decade potentially.

Ooh, I see. In that case, yes, I could see that happening at some point in the near future. That would probably require some of the R gerrymanders to be broken, but they seem on their way to be in VA and NC and might be at some point in GA and TX too.

[Orser67]

Interesting look at the data, its interesting how the GOP really has become the party of the south and the Midwest. Something I noted this year is that that even in a D+9 climate, the GOP picked up 3 senate seats in the Midwest, lost only a net of 7 house seats in the Midwest, the fewest of any region in the country and split the governor elections in the Midwestern Obama Trump states 2-2. In the south as well, the GOP picked up a senate seat and the house losses were fairly small with some of the seats lost being real anomalies like OK 5 and SC 1. The west by contrast was a disaster, not only did the GOP lose 12 seats, but also lost 2 senate seats, the north-east as well where 11 seats flipped from an already fairly small GOP congressional delegation.

Yeah, the regional variation this year was striking. Democrats did pretty much as well as they possibly could have in the Northeast and the West (except against a few popular incumbent Republicans and in AK), while the results in the South and the Midwest were underwhelming (or even downright bad in a couple states) for a midterm.

Also, great graph OP.

[Antonio the Sixth]

Interesting look at the data, its interesting how the GOP really has become the party of the south and the Midwest. Something I noted this year is that that even in a D+9 climate, the GOP picked up 3 senate seats in the Midwest, lost only a net of 7 house seats in the Midwest, the fewest of any region in the country and split the governor elections in the Midwestern Obama Trump states 2-2. In the south as well, the GOP picked up a senate seat and the house losses were fairly small with some of the seats lost being real anomalies like OK 5 and SC 1. The west by contrast was a disaster, not only did the GOP lose 12 seats, but also lost 2 senate seats, the north-east as well where 11 seats flipped from an already fairly small GOP congressional delegation.

Yeah, the regional variation this year was striking. Democrats did pretty much as well as they possibly could have in the Northeast and the West (except against a few popular incumbent Republicans and in AK), while the results in the South and the Midwest were underwhelming (or even downright bad in a couple states) for a midterm.

Also, great graph OP.

Thanks!

It's worth noting that, while Democrats gained more seats in the South in absolute (10 or 11) terms than in the Midwest (7), in relative terms their gains were lower in the South, because it has by far the most total seats of any region (152, vs only 94 in the Midwest).

[Comrade Funk]

Thank you for this.

[Oryxslayer]

If you count things by seats, then Dems gain 12 in the West, 10(+1?) in the South, 11 in the Northeast, and 7 net from the Midwest (9 gains two losses). Just makes the Midwest stick out more.

[Nyvin]

Any chance of creating a cumulative flow diagram with the data?   I think that'd look awesome!

http://toolsforagile.com/blog/archives/177/interpreting-the-cumulative-flow-diagram

I guess you'd need eight colors (Democrat South, Republican South, etc) and it might be a bit crammed, but worth trying.

[Antonio the Sixth]

Any chance of creating a cumulative flow diagram with the data?   I think that'd look awesome!

http://toolsforagile.com/blog/archives/177/interpreting-the-cumulative-flow-diagram

I guess you'd need eight colors (Democrat South, Republican South, etc) and it might be a bit crammed, but worth trying.

Sure, it looks easily doable in Excel.

Would you prefer the regions to be stacked together (ie Dem NE, then Rep NE, then Dem South, then Rep South, etc.) or parties stacked together (ie Dem NE, then Dem South, then Dem MW, etc.)?

[Nyvin]

Any chance of creating a cumulative flow diagram with the data?   I think that'd look awesome!

http://toolsforagile.com/blog/archives/177/interpreting-the-cumulative-flow-diagram

I guess you'd need eight colors (Democrat South, Republican South, etc) and it might be a bit crammed, but worth trying.

Sure, it looks easily doable in Excel.

Would you prefer the regions to be stacked together (ie Dem NE, then Rep NE, then Dem South, then Rep South, etc.) or parties stacked together (ie Dem NE, then Dem South, then Dem MW, etc.)?

I'm not sure,  I would guess parties stacked together would be better, but would have to see both, lol.

[Antonio the Sixth]

All right, why not both! First stacked by region, then stacked by party.

Democrats are reddish shades, Republicans blueish, Independents greenish. I varied the luminosity to indicate regions (Midwest is the darkest, South the lightest, NE and W in between).





Personally I actually think the regions-stacked chart looks better, since the regional breakdown is relatively stable over time while the party breakdown fluctuates a lot.

[Nyvin]

That's great!  I agree the region breakdown is better,  you can see the NE and MW shrink gradually over time while the South and West grow.    Great job!

[Antonio the Sixth]

That's great!  I agree the region breakdown is better,  you can see the NE and MW shrink gradually over time while the South and West grow.    Great job!

Thanks!

[Antonio the Sixth]

An update now that we pretty much know how the House is going to end up:


Redistricting cycles always create some fun patterns in terms of seat distribution. This year, they helped Democrats hold on surprisingly well in the West and Midwest. In both regions, Democrats wound up holding the same number of seats as they did after the 2020 elections. Since the West gained seats and the Midwest lost some, the net result is a slight improvement for Republicans in the former but for Democrats in the latter. Democrats' performance in the Midwest is remarkable - it's actually their best result in the region in a year when they lost the House since 2000. Of course, this has a lot to do with favorable redistricting - the egregious IL gerrymander combined with a genuinely fair map in MI and a fair-ish one in OH meant that the playing field was a lot more level than in the past. Still, Democrats did overperform expectations in that region, as they did in much of the West (even CA, while disappointing, still effectively flipped only 1.5 seats to the GOP if the current numbers hold).

Meanwhile, the Northeast was brutal, with Democrats losing another 5 seats there (Republicans gained just 3 since the region as a whole lost seats) on top of the 3 they lost last time. Of course, these losses come down almost entirely to New York, and there's still a lot to be learned about how Democrats f**ked up so badly there. Sadly, this obscured their solid performance in PA and their continued dominance of New England. The South also saw some significant losses, partly due of brazen Republican gerrymanders and partly to their underperformance among Hispanics in Texas and Florida. Democrats had actually held up well in the South in 2020, but this year showed that the bottom hasn't yet entirely fallen out (and things could get worse if NC Republicans revert to a gerrymander).

[Skill and Chance]

An update now that we pretty much know how the House is going to end up:


Redistricting cycles always create some fun patterns in terms of seat distribution. This year, they helped Democrats hold on surprisingly well in the West and Midwest. In both regions, Democrats wound up holding the same number of seats as they did after the 2020 elections. Since the West gained seats and the Midwest lost some, the net result is a slight improvement for Republicans in the former but for Democrats in the latter. Democrats' performance in the Midwest is remarkable - it's actually their best result in the region in a year when they lost the House since 2000. Of course, this has a lot to do with favorable redistricting - the egregious IL gerrymander combined with a genuinely fair map in MI and a fair-ish one in OH meant that the playing field was a lot more level than in the past. Still, Democrats did overperform expectations in that region, as they did in much of the West (even CA, while disappointing, still effectively flipped only 1.5 seats to the GOP if the current numbers hold).

Meanwhile, the Northeast was brutal, with Democrats losing another 5 seats there (Republicans gained just 3 since the region as a whole lost seats) on top of the 3 they lost last time. Of course, these losses come down almost entirely to New York, and there's still a lot to be learned about how Democrats f**ked up so badly there. Sadly, this obscured their solid performance in PA and their continued dominance of New England. The South also saw some significant losses, partly due of brazen Republican gerrymanders and partly to their underperformance among Hispanics in Texas and Florida. Democrats had actually held up well in the South in 2020, but this year showed that the bottom hasn't yet entirely fallen out (and things could get worse if NC Republicans revert to a gerrymander).

I know mapmaking rules complicate this, but this chart makes it look like 2016/18 trends have reversed.

[Antonio the Sixth]

I know mapmaking rules complicate this, but this chart makes it look like 2016/18 trends have reversed.

To some extent, yes, although it shouldn't be overstated. The House seems a little less polarized now than it looked in 2018, although there are plenty of districts going the opposite way (see Republicans clearing up in Iowa and outer Minnesota). We'll probably need a couple more cycles before we get a clear sense of where some of these Congressional districts are headed politically.