How Trump Won: The Midwest
Part 4 in a series
We now turn our attention to the biggest prize of the 2016 election: the Midwest. This region has largely been where our national elections have been won or lost since the Civil War. Republicans knew that they could not win the presidency without this area; hence all but three of their nominees from 1868 through 1920 hailed from the region (Democrats tended to counter with presidential candidates from electorally rich New York, but all but three of their vice presidential candidates during this time period were from the Midwest).
Donald Trump picked up a whopping 50 more electoral votes from this region than Mitt Romney did, and very nearly added 10 more. Moreover, these pickups largely surprised analysts. It is therefore useful to spend some time looking at how it happened.
We’ll start this time with maps, then charts, then some more maps. What we’ll see is that the contagion that RealClearPolitics identified in a 2009 series on the Obama coalition spread throughout the region, just as it did the other regions of the country: Democratic strength in rural areas collapsed. But unlike the West and Northeast, it mattered here.
So we start with Michael Dukakis’ map. What I’ve done here – as with most of the articles in this series – is subtracted out the national vote. This allows us to compare counties where Democrats won by large margins (as in 1996) with elections where they didn’t fare as well. So what this map shows are counties that Dukakis won, as well as counties that he lost, but by less than his national margin.
As you can see, the Massachusetts governor performed well in the region, especially in the western division. This is in part because of the farm recession, but as we’ll see, it wasn’t limited to this. We note Democratic strength in eastern Ohio, which is part of Appalachia, along the Lake Erie coast (reflecting the strength in old industrial cities), in the Upper Peninsula of Michigan, and in the Balkans (a rural, yet ethnically diverse section of southeastern Kansas). But in general, it is difficult to identify any particular home for the Democratic Party here. Democrats perform well in all sorts of places. This helped Dukakis carry Iowa, Minnesota, and Wisconsin, while coming close in South Dakota, Missouri and Illinois.
This is true if we back out and look at gradations of red and blue:
The 1996 election featured the famous “MOM” strategy, where the Bob Dole campaign focused on Michigan, Ohio, and Missouri. It failed, rather spectacularly, but the region did begin to shift toward the Republicans. You can see the broad ocean of blue beginning to retreat to islands:
This is also apparent in the overall map of the region:
In 2008, Barack Obama won a substantial victory in the region, comparable to Bill Clinton’s win in 1996. Yet his coalition was much narrower, and dependent upon strong showings in the cities.
By now you really can pick out individual cities and regions in many of these places. By 2016, the Democratic coalition has really become islands of blue in a sea of red:
If we filter out just the counties that were more Democratic than the country as a whole, we can almost count the counties:
We’ll look at Minnesota, Ohio, Michigan, and Wisconsin in some detail, but outside of those states, we can see that the Democratic coalition has been reduced to: two counties with Indian reservations in North Dakota; four such counties in South Dakota, plus Vermillion County (home to University of South Dakota); one in Nebraska (Douglas County, which includes Omaha); two in Kansas (Douglas County, which included the University of Kansas and Wyandotte); three in Missouri (Boone, including the University of Missouri; St. Louis City, and Jackson, which includes Kansas City); and three in Indiana (Lake, including Gary; Marion, including Indianapolis; and Monroe, which includes Indiana University). In Illinois, the counties are more numerous, but follow the same basic pattern; the only difference is Chicago – a massive metropolitan area – along with a few blue-collar towns where Democrats maintain residual strength. Overall, the shift in this region since 1996 has been dramatic:
Note that South Dakota and North Dakota were both single-digit states in 1996 (South Dakota went for Dole by three points), but no longer. Note too the pattern in Missouri; this is consistent with the state’s settlement pattern, where the band through the center of the state was settled by Republican-leaning Germans, while the northeast and southeast were settled by people from Mississippi; the electoral patterns there follow, to some degree, the trends in the Magnolia State. These voters kept the Democratic Party of Missouri afloat for decades, but no longer.
Of course, counties don’t vote; people do. We don’t mean to suggest that this coalition is useless, as many of those places that cast a lot of Democratic ballots also have a lot of voters. We simply use these maps to illustrate how wide-ranging some of these changes are. This becomes more apparent when we look at the distribution of votes here.
By now, these trends should look familiar. Hillary Clinton improved over even Obama’s vote share in the “Mega City” areas. But she fell back to 2000 and 2004 levels in the “Small City” and “Large City” areas, and then fell behind Dukakis in the towns. In fact, in large towns she had the worst Democratic showing since 1988, while in rural areas and small towns she had the worst showing since 2004. In the rural areas, she ran over 15 percentage points behind Dukakis.
Notice how much the overall trends here resemble what we saw in the South. Geographically, the Midwest voted a lot like the South:
The problem here for Democrats is twofold. First, when we talk about “Mega Cities,” we are talking about Chicago. Most of the states don’t have “Mega Cities,” which means that the Democratic coalition necessarily recedes in 11 of the 12 states. Second, while rural areas and towns aren’t as powerful here as they are in the South, they still cast a near-majority of the votes:
This is what we mean by the contagion in the Democratic coalition spreading. Over the years, Democratic strength in Midwestern rural areas has been on the decline, just as in the South. This tendency became particularly acute during the Obama years. We just didn’t really notice, because the rural areas don’t outvote the cities, as they do in the South. But the polarization reached a point in 2016 where it mattered.
Let’s take a look at some of the states in the Midwest that were particularly problematic for the Democrats. We’ll start with Iowa. Iowa really is dominated by rural areas and towns; cities make up less than a quarter of the vote here. As you can see, the Democrats have stayed alive by managing their losses in the cities and towns; the decline there has been very gradual. Until 2016, that is, where things dropped off in the rural areas and small towns, which combined for almost a majority of the vote:
The result is a creeping “redness” across the map, as the Democratic coalition is pushed gradually eastward, before retreating almost entirely to a few small cities and college towns. These maps aren’t necessarily the easiest to read, but in the interest of efficiency we’ve paneled them in this manner. They all show the Democrats normalized vote share in 1996, 2004, 2012, and 2016 (reading left to right, by row):
In Minnesota, we actually see a similar pattern:
The difference is that Minnesota sees a lot of votes cast in its large city, which approaches “Mega City” status; Minneapolis casts about 60 percent of the votes there:
In this respect, Minnesota is not unlike a Western state. With that said, if Democrats suffer any losses there, this state will slip through their fingertips as well. As you can see, their coalition has gradually been reduced to the Iron Range and Minneapolis-St. Paul. This is sufficient for statewide elections, but will wreak havoc on their congressional/statehouse delegations:
Wisconsin has seen one of the more dramatic transformations over this time period. While the Republican coalition was initially restricted to the Milwaukee suburbs and some rural areas, it has since spread into the northern portion of the state. After 1996, the Democratic coalition is basically three old industrial counties on Lake Superior, the City of Milwaukee, and Madison and its environs:
You can see the Republicans’ strength in rural and small-town areas; note too that Milwaukee has been relatively flat (probably due to crucial Waukesha County):
The difference between Wisconsin and Minnesota is that its large city is not anywhere near as dominant as Minneapolis (the “Mega City” is a single county that now falls in greater Chicago):
In Michigan, we note similar trends. Remember, the Upper Peninsula was once a Democratic redoubt, as was the northeastern quadrant of the “Mitten.” But in 2016, the Democratic coalition basically consisted of Battle Creek, Flint, Oakland County, Ann Arbor, Lansing, and Detroit. Again, this is a lot of votes, but not enough to carry the state.
Again, the Democratic decline in rural areas is apparent, although Clinton performed worse than any Democrat since 1988 in almost every area (and worse even than Dukakis in rural and small-town areas):
And again, while Detroit casts a majority of the state’s ballots, it isn’t quite as dominant as Minneapolis is in Minnesota:
Finally, Ohio. I will admit, I was absolutely shocked to see Clinton lose there by eight points. This is a state that has always been near the center of American elections. But when you look at the maps, it isn’t hard to see what happened:
As late as 1996, Bill Clinton ran strong in Appalachia, and dominated in the industrial Northeast and along Lake Erie. The state had strong east/west and north/south splits that gradually disappeared over the course of the decades. In 2016, Democrats ran strong in Lucas (Toledo), Cuyahoga (Cleveland), Summit (Akron), Athens (Ohio University), Franklin (Columbus) and Hamilton (Cincinnati) counties. Again, these are significant prizes. But the drop-off in rural areas and the towns looks much like it does in the southern region, with Clinton running almost 15 points behind Dukakis in the rural areas, and about 10 points behind him in the towns:
And, again, while large cities cast a lot of votes in Ohio, unlike in Minnesota, they cannot carry a candidate to victory. If Democratic weakness in rural areas and towns persists, they will have a hard time bouncing back here. It is a real problem for them.
This completes our granular study of the 2016 election. Because this has run quite a bit longer than we anticipated, tomorrow we will add some closing thoughts in a bonus fifth part of the series.