KEY POINTS FROM THIS ARTICLE
—Though former President Donald Trump easily won most GOP primaries this year, some areas stood out to us.
—By comparing how Trump did in each county to how he did overall in each state, we can get a better idea of how his coalition is shaping up.
—Geographically, Trump beat his statewide primary share in a majority of counties in most states.
—Meanwhile, and not surprisingly, Trump tended to struggle in areas that are Democratic, or blue-trending, in general elections, although there were some exceptions.
Looking back on Trump’s performance
With a bit of a break in the presidential primary calendar this month—the last major day of contests was more than two weeks ago, though Pennsylvania will hold its primary next week—we’ve been thinking about our bigger picture takeaways from this primary season. In terms of the actual results, it will probably go down as the least dramatic in recent memory. Still, that’s not to say we can’t find some interesting nuggets by digging a little deeper into the returns.
For basically the entire time that the Republican primary has been active, former President Donald Trump has been in the driver’s seat. But for this article, we wanted to look at where his primary performance was especially strong…or weak.
But wait—didn’t Trump win everything except for Vermont and D.C.? What could possibly be weak about his showing?
That’s a natural question to ask, but let’s take a look at Map 1. Going state by state, this map compares Trump’s share of the vote in each county to what he got statewide. Trump’s share in orange counties was greater than his statewide share, while the opposite was true in blue counties.
Map 1: Counties where Trump was over/under his total share in each primary state
Now, let’s get a few caveats out of the way. We only considered primary states for the purposes of this analysis—in caucus situations, raw votes aren’t always provided at the county level, which can make these types of comparisons difficult.
For the sake of an apples-to-apples comparison across the board, Map 1 considers Trump’s percentage share in a hypothetical two-way contest with former Ambassador to the United Nations Nikki Haley (so votes for other candidates or for various “uncommitted” options were not included). Think of this exercise as similar to the “two-party” vote we sometimes use for Crystal Ball articles, but instead of referring to Democratic and Republican votes, here the “two parties” are Trump and Haley. Though she dropped out early last month, Haley was his most prominent GOP rival for much of the time when the primary season was active. So our thinking is that if Republican voters were going to cast a “protest” vote against Trump, Haley would have been their most obvious choice.
For an example of how we arrived at Map 1, let’s consider New Hampshire, which was the first primary in the nation. Back in the state’s February contest, Trump took 55.7% of the two-candidate vote against Haley. In our actual timeline, Trump lost only one county, Grafton (which is perhaps most notable as the site of Dartmouth College). But, aside from Grafton, Trump’s share is lower than 55.7% in four other counties: Carroll (where he took 53.8%), Hillsborough (55.3%), Merrimack (53.6%), and Strafford (55.4%). So those four counties are blue on Map 1 while the other counties are orange. We repeated this for the other states.
One overarching if not entirely earth-shattering trend here is that Trump’s weakness—or, one could argue, Haley’s strength—was clearest in counties that either vote Democratic in general elections or have trended Democratic in the Trump era.
Let’s consider the author’s native state, Louisiana, which voted on March 23, a few weeks after Haley suspended her campaign. In the 2020 presidential election, Joe Biden improved slightly on Hillary Clinton’s statewide showing, although he still lost by 19 points. As part of that modest 1-point statewide improvement, he only gained ground in 19 of the state’s 64 parishes. All seven blue parishes where Trump was under his statewide share on Map 1 were among that group of 19 parishes that swung towards Biden in the 2020 general election.
Orleans Parish, which has long been the most Democratic parish in the state and is home to a growing white liberal bloc, cast a paltry 2,500 votes in the GOP primary (considering both its cobalt blue lean in general elections and the fact that the state has a closed presidential primary, it would stand to reason that the pool of GOP voters there is small). It was also, by far, Trump’s worst parish in the state—his 75% two-candidate share there was 18 percentage points under his statewide 93%.
Louisiana’s result also speaks to several other larger-scale trends that are apparent throughout the map. Jefferson and St. Tammany parishes, which border Orleans and are seeing the same types of pro-Democratic trends that are taking hold in other college-educated suburbs around the nation (albeit at a slower rate than most), also gave Trump a lower share than what he got statewide. Similarly, Trump’s familiar weakness in areas that house major college campuses shows up. East Baton Rouge Parish, home to Louisiana State University, and Lincoln Parish, which houses Louisiana Tech and is the state’s most college-educated parish, were parishes outside of the New Orleans metro area where Trump was relatively weak.
Another trend that Louisiana illustrates is Trump’s relative strength in rural areas that have large minority populations. For instance, St. Helena Parish, just northeast of Baton Rouge, is a small parish that is plurality-Black by composition and votes Democratic in general elections. But in last month’s primary, it was one of Trump’s best parishes in the state, giving him close to 97%. This is because nearly all the Republican voters who cast ballots in last month’s primary there were white. So, St. Helena, a blue parish in partisan races, looks very similar to Franklin Parish, which is deep red in partisan races, on this map of the Republican primary. We point this out because these types of counties, especially throughout the South’s “Black Belt,” stand out as some of Trump’s strongest performances in areas that have voted against him in general elections.
We’d also mention that Louisiana, and neighboring Mississippi, are states with only light shades of orange on Map 1. This is because Trump took 93% in the former and close to 95% in the latter—even in the counties/parishes where he won almost unanimously, it’s not really mathematically possible to do that much better than what he got statewide. Later on, we’ll get into some states with a little more color variation.
In what follows, we’ll move chronologically through the primary calendar and flag some things that stood out to us about Map 1. Though our observations won’t be exhaustive—for the sake of length, we won’t touch on every state—we’ll try to highlight what we think are the main takeaways.
Feb. 24 (South Carolina)
Trump carried his main rival’s home state with 60% of the two-candidate vote, and got a higher share than that in 36 of the state’s 46 counties. Some of his best areas were in rural but minority-heavy counties along the Pee Dee River Basin—as we mentioned with Louisiana, the GOP primary electorates here were likely whiter than those in general elections. In that general area, Horry County—where Myrtle Beach has become a magnet for retirees—cast the second-highest number of votes in the state, and gave Trump an especially strong share. Meanwhile, Trump was notably weaker in both Charleston County (Charleston proper) and Richland County (Columbia), as well as several suburban counties in their orbit—this group included Lexington County, which Haley represented during her time in the state legislature.
Feb. 27 (Michigan)
Just a few days after South Carolina’s late February primary, Trump took 72% of the two-candidate vote in Michigan. He ran best in northern Michigan and the Upper Peninsula, both areas that used to be competitive but are now solidly red. Though Trump beat his statewide share in Wayne County (Detroit), the state’s largest Democratic-voting county in general elections, he was several points under his statewide share in both suburban Oakland and collegiate Washtenaw (University of Michigan) counties.
In past editions, we have pointed to Democratic trends in western Michigan as a reason why Biden may retain an edge in the state. Trump was below his statewide share in several counties in the Grand Rapids area, as well as some near the state’s touristy northwestern “Cherry Coast.” Given what Map 1 implies, it seems plausible that Trump has more ground to lose in this traditionally Republican area.
March 5 (Super Tuesday)
On Super Tuesday, more than a dozen states that Map 1 includes held primaries—it was also the final primary date where Haley was an active candidate.
If the aforementioned Louisiana and Mississippi had some of the most muted colors on the map, we’d like to note two of the most vivid states: Colorado and Virginia. Both of these Super Tuesday states were textbook swing states a dozen years ago, but they have both since moved against a Trump-led GOP. Despite that, Trump is still quite clearly beloved in the most conservative parts of each state. In Colorado, several sparsely populated counties in the eastern High Plains region of the state gave Trump shares that were more than 20 percentage points higher than his statewide number. Meanwhile, Trump outright lost both Denver and Boulder counties to Haley, along with several other white liberal, ski-centric Western Slope counties.
Our home state, Virginia, saw the largest deviations of any state. Charlottesville, where the Center is located, gave Trump just 24% of the two-candidate share, or 40 percentage points worse than his statewide share, 64%. But if one travels just an hour or so south of the city, Trump was routinely beating his statewide share by close to 25 points throughout much of Southside.
But the extremes get even more pronounced in Virginia. Trump’s worst locality, relatively speaking, in the entire nation was the city of Falls Church, in Northern Virginia. He did 41 points worse than his statewide share there. If one limits things to just counties, adjacent Arlington County takes the cake. On the other side of the state, Trump’s best county, in relative terms, was southwestern Virginia’s Buchanan County—it was the only county (or equivalent) where Trump beat his statewide share by more than 30 points.
As a bit of an aside, for “official” purposes, the darkest blue county on Map 1 is Kent County, TX (it is in the middle of west Texas). However, considering how the demographically-similar counties around it voted, we suspect that its certified results, which show a two-to-one Haley lead, reflect some type of data entry error. The county cast only about 230 votes in the GOP presidential primary.
Texas, as it happens, was also a Super Tuesday state. It was not surprising that Trump ran furthest behind his statewide share (setting Kent aside) in Austin’s Travis County, as well as several nearby counties that follow Interstate 35. Trump was relatively weak in the core counties of Dallas’s Metroplex, although, in the Houston area, wealthy Montgomery County stands out as a populous county where Trump performed well. Montgomery was the sole county in the nation that twice gave Trump a raw vote margin of more than 100,000, and remains, basically, the reddest decently-sized county anywhere.
We have talked quite a bit in previous editions about Democratic erosion in South Texas—on Map 1, every county south of San Antonio’s Bexar County is orange. Considering that some of these counties are nearly entirely Hispanic by composition, this is a region where, unlike much of the Black Belt, Trump likely did earn many nonwhite votes.
One other Super Tuesday state that we’d single out is Alabama. Towards the end of last month, we profiled a special legislative election where the Democrats overwhelmingly flipped a Trump-won seat in the Huntsville area. A few weeks earlier, on Super Tuesday, Huntsville’s Madison County was also Trump’s worst county in the state. While we’re not sure Madison County is ready to vote Democratic at the presidential level just yet—it was Trump by 8 points in 2020—it is an area worth watching in this red state.
March 12
We already touched on Mississippi a bit, but elsewhere in the Deep South, Georgia voted on March 12, as did Washington state.
In Georgia, Trump was strongest outside the Atlanta metro area— Savannah’s Chatham County was one of the few counties south of Atlanta where he fell under his statewide 86.5%. Within the Atlanta area itself, there was something of an economic component to the breakdown. Trump was over his statewide share in several Biden-won counties south of the city (such as diverse Henry County) but under in a few red, but Democratic-trending, counties north of the city (Cherokee, Forsyth, and Hall). The latter group generally has higher incomes than the former.
Trump underperforming in places like Cherokee and Hall counties while holding serve in demographically similar Montgomery County, TX may be a small sign that a blue Texas may be further away than some Democrats would hope.
In Washington state, Trump, as expected, fared worst in Seattle’s King County but fell under his statewide share in just one county east of the Cascades (Washington State University’s Whitman).
March 19
States that voted on March 19 included Arizona, Florida, Illinois, Kansas, and Ohio. By this point, as we should note, with Haley out of the race for multiple weeks, Trump won each contest with more than three-quarters of the vote. But some interesting patterns were still taking hold.
In Arizona, Trump performed two percentage points worse in Phoenix’s Maricopa County than he did statewide—this was exactly what happened in the 2020 general election, Biden carried the state’s mega county 50%-48% while running just better than even with Trump statewide.
In another Sun Belt state, Gov. Ron DeSantis’s Florida, the opposite was the case. Miami-Dade, the state’s most populous county, was the sole county in South Florida where Trump beat his statewide 85.4%. This probably speaks to Trump’s continued strength with Cubans, a bloc that swung heavily towards him in 2020.
Though much of the rest of the state broke down as one would expect, Jacksonville’s Duval County made for a surprising source of Trump strength. Of the three Trump-to-Biden counties in Florida, Duval was the only county where Trump beat his statewide share (the other two flips were Pinellas and Seminole, which are both blue on Map 1).
In Illinois, Trump ran just under his overall 84.8% in Chicago’s Cook County. Moving out into its suburbs, he underperformed more noticeably in upscale DuPage and Lake counties but did better in Will County, which is just south of Cook and has a more working class flavor.
Trump beat his statewide share in every Illinois county south of the state capitol, Springfield. The area was the scene of a down-ballot GOP primary skirmish between Rep. Mike Bost (R, IL-12) and his challenger, former state Sen. Darren Bailey. Trump’s endorsement was probably instrumental in Bost’s 51%-49% win over the more anti-establishment Bailey. But Trump himself actually ran better in the eastern part of the district, which was Bailey’s homebase—perhaps his endorsement allowed Bost to pick off just enough votes there to hold on.
April 2
The most recent major primary day, April 2, saw contests in Wisconsin as well as a sampling of northeastern states.
Wisconsin, looking to the general election, is one of—if not the—most critical states. In a development that should shock absolutely no one who is even remotely familiar with the state’s political geography, Trump’s worst county was Madison’s Dane, where his 73% was 13 percentage points lower than his statewide share. Aside from Dane, there was not too much deviation from the statewide result among the state’s 71 other counties, with only light shades of blue and orange in Wisconsin on Map 1. Interstate 94, which connects the counties in Madison’s neighborhood to the Milwaukee metro area, appears as a marker of Trump’s relative weakness. However, Trump fell under his statewide share in just four counties north of that stretch: Door, Eau Claire, and La Crosse, which all went to Biden in 2020, as well as Winnebago, which is the most Democratic of the “BOW” Counties (although it still voted for Trump twice).
Finally, a quick word on New York. There was a bit of continuity between this primary and 2016, when Trump easily carried what was then his home state against then-Gov. John Kasich (R-OH) and Sen. Ted Cruz (R-TX). As with the 2016 primary, Trump’s best county was Richmond (Staten Island), an area where he could, at least temperamentally, hardly be a better fit. In both years, Manhattan was easily Trump’s worst borough within New York City. In 2016, Trump won all Upstate counties, but did so with only pluralities in many cases. In this month’s contest, his showing there was markedly worse than what he posted in New York City and Long Island.
Conclusion
While Trump dominated in nearly every state during the 2024 Republican primary season, his strength was not uniform. Even in states where he beat Haley by a better than nine-to-one margin, there were clearly areas where he was especially strong…and ones where he lagged. While we would expect the Biden campaign to try to capitalize on Trump’s pockets of relative weakness, don’t bank on Biden carrying every county that had a pro-Haley lean during the primary season. In fact, in a future issue, we may do something similar looking at some of Biden’s own (relative) strengths and weaknesses from the primary season.