I wanted to look at trends from 2012 to 2016 in a different way. Generally I see maps that show a change in percentage. I wanted to see the change in raw votes. I wanted to see where Clinton received more votes than Obama, less, and the same for Romney and Trump, and whether Clinton’s loss or gain was more than Trump’s. If both sides saw a decline, which was worse? If both sides gained, which gained more? If one side lost votes, does it look like most of them shifted to the other party, or did most of them vote third party or not vote?
Of course some of the change isn’t from voters changing sides. Some of it is from voters dying between elections, and other people voting who hadn’t voted in 2012 because they were too young, had been indifferent but were activated, or other reasons. Some of the losses and gains can also be explained by migration. Democratic voters are moving from rural areas to cities and suburbs, in particular. So to an extent, regardless of how they voted in 2016, voters who were there in 2012 may not be in the same place, or anywhere at all, in 2020, and there will be new voters in 2020 who weren’t anywhere in 2012.
But anyway, let’s get to the colors so you know what you’re looking at.
1. Dark red: Trump’s increase over Romney > Clinton’s decline from Obama
2. Medium red: Trump’s increase over Romney > Clinton’s increase over Obama
3. Light Pink: Trump’s increase over Romney < Clinton’s decline from Obama
4. Orange: Trump’s decline from Romney < Clinton’s decline from Obama
5. Green: Trump’s decline from Romney > Clinton’s decline from Obama
6. Light Blue: Trump’s decline from Romney > Clinton’s increase over Obama
7. Medium Blue: Trump’s increase over Romney < Clinton’s increase over Obama
8. Dark Blue: Trump’s decline from Romney < Clinton’s increase over Obama
And let’s see look closer at some areas with smaller counties, that are harder to see.
The Midwest is really what started this. I was interested in seeing what happened in Michigan and Wisconsin in particular, below the statewide level. As you can see in Wisconsin there is not a single dark red county, and in Michigan outside of Macomb County and the base of the thumb there are few. I was particularly surprised there was only one in the Upper Peninsula. On the other hand, there are no medium red, and medium or dark blue counties in Wisconsin, either, and few in Michigan. The Wisconsin result is entirely a situation of lower turnout mainly for Democrats, with some rural/outstate Democrats switching to Trump but mainly Obama voters opting out of voting for either Trump or Clinton. Minnesota and Iowa are mainly the same, with only a few dark red counties and few areas of real improvement.
Michigan had poor Republican turnout in the Grand Rapids area, and poor Democratic turnout almost everywhere outside of the Ann Arbor area. Ohio is similar, with poor Democratic turnout in most of the state except the Columbus area more so than a swing to Trump, outside of more southern rural parts of the state where there are more dark red counties.
Illinois looks like the states it borders. Aside from the Chicago area, which is an aberration for the region, with a strong shift to Clinton. That’s why Illinois is the only state in the region that didn’t swing toward Trump pretty strongly.
What happened in bluer parts of the Midwest is not what happened in Pennsylvania and New York, or not exactly. There was a general shift in rural areas to Trump and urban areas to Clinton, so that’s similar. In the rural areas, though, in Pennsylvania, it was not mainly a decline in Democratic votes from 2012 (though that was very consistent), but mainly an increase in votes for Trump. In New York there was a enormous shift to Clinton in most of the city (hard to believe given how already supremely Democratic it is), but upstate most areas saw a bigger surge in Republican votes than a decline in Democratic ones. It’s not a uniform situation, though, with especially the counties in the far north nearer to Vermont, like most of Vermont and New Hampshire, having lower Trump gains than Democratic decline.
In the South there is also the urban/rural split, but again that’s not all. In the black belt from Virginia to the Texas/Louisiana border there was depressed Democratic turnout, but in many counties, especially in Mississippi, both sides saw a decline in turnout. Not surprising after Democratic highs of 2008/2012 with a black Democrat leading the party. In rural but heavily white areas Trump tended to ascend to new heights. Bigger urban areas like New Orleans, Atlanta, the Triangle and Charlotte in NC, and Richmond and the DC burbs in Virginia, saw big swings to Clinton and declines in Republican support. Other cities like Little Rock, Baton Rouge, Birmingham, Montgomery, Tallahassee, Huntsville, and Nashville saw declines in Republican support, but not real increases in Democratic support in 2016.
If you just add El Paso, Tarrant, Travis, Bexar, Harris and Dallas counties, Trump’s vote total declined by nearly 92,000 from Romney’s, and Clinton’s increased by nearly 377,000 votes over Obama’s in 2012. Not enough to change the state as a whole, but definitely a big change.
So there’s a lot left to interpretation here. I take hope in the pink counties, because I think there are a lot where we can come back from the 2016 declines, and some of the defectors will lose faith in Trump. There is another side, too, where you can say that the Democrats in those areas may’ve died or moved to urban areas that shifted in our direction. My choice of colors is my own, and follows my interpretations. Oh well. I like it.