The places that had the biggest swings toward and against Trump

With most of the slow-to-report votes tallied, we finally have a clearer picture of last month’s presidential results. Despite the high polarization in the country that carried over to the reaction to the results — with 70% to 80% of Republicans still saying they disbelieve that Joe Biden won — in some respects the vote itself was less polarized than in 2016.

Compared with 2016, in 2020 there was less difference by race or ethnicity, and urban areas and suburban areas voted more alike. But the economic and education partisan divides widened. Biden gained in well-educated suburbs and exurbs, often in places that have tended to vote Republican in recent decades, like the Atlanta, Dallas and Phoenix areas.

In addition to these broad patterns, some specific shifts reflected local political loyalties.

As of Sunday, reporting was complete enough to look closely at nearly all individual counties and metropolitan areas (which consist of one or more counties). Though some important patterns may show up only in more fine-grained precinct or individual survey data, counties make it easy to compare the entire country, over time, though many lenses.

These near-final county totals confirm a few patterns in the 2020 vote.

Stability

For all the ways that this election broke with precedent, the county-level vote pattern in 2020 was overwhelmingly similar to that in 2016. The correlation between counties’ vote margins for President Donald Trump in 2016 and 2020 was 0.99. (A correlation of 1 represents a perfect relationship, and 0 represents no relationship.)

Strikingly, the vote was slightly less polarized than in 2016, breaking a pattern of increasing geographic polarization in the 2000s. The standard deviation measures how much an individual county’s vote typically deviated from the national average; this measure was a bit lower in 2020 after rising steadily from 2000 to 2016.

Furthermore, this year the vote in many counties was a small shift toward the center. The bigger that Trump’s victory was in 2016 in a given county, the more ground Biden picked up in 2020 relative to Hillary Clinton in 2016. In counties that voted for Trump in 2016, the 2020 margin shifted 3.2 percentage points toward Biden, versus a 1.9-point shift toward Biden in counties that voted for Clinton in 2016.

Education

Among metropolitan areas with a population over 250,000, three Colorado metros — fast-growing Colorado Springs, Denver and Fort Collins — topped the list for swings against Trump in two-party vote share. Colorado has one of the nation’s highest rates of people with a four-year college degree, and this year it solidified its status as a blue state. Colorado’s shift against Trump was 8.6 points, second only to Vermont’s 9 points.

A part of conservative-leaning Kentucky posted one of the bigger anti-Trump shifts. The Lexington-Fayette, Kentucky, metropolitan area, which flipped from red to blue and is home to the University of Kentucky, has the highest percentage of college graduates in the region and is in the top fifth of metros for college attainment.

And what about the big swing in the Huntsville area in deep-red Alabama? Fast-growing places with brighter economic prospects — correlated with a higher number of people with college degrees and more jobs in professional, tech and creative fields — moved toward Biden. A highly educated workforce in Huntsville, first put to use in industries like aerospace, has become attractive to other businesses in recent years.

Suburban areas in Nebraska swung significantly toward Biden. Nebraska’s 2nd Congressional District, which includes well-educated Omaha, contributed one electoral vote to Biden’s tally. It “swung against Trump more than any swing state,” according to Dave Wasserman of the Cook Political Report. The swing was 8.8 points, as the district flipped from red to blue.

The Hispanic Swing

Most metro areas swung Democratic, but the most extreme swings were toward Trump, and the biggest of those were in heavily Hispanic metros, like Miami and areas along the Texas border.

Initially it appeared that Trump’s strength in 2020 relative to 2016 in heavily Hispanic counties might have been specific to South Florida and Texas border areas, despite the very different national origins and identities in those two regions. But data from later-reporting counties points to a national trend.

The correlation between a county’s percent Hispanic population and its swing toward Trump was 0.41. Even excluding Miami-Dade County and all of Texas, the correlation is still 0.30. The Bronx, New York; Los Angeles; Osceola County near Orlando, Florida; and Hudson and Passaic counties in northern New Jersey are all at least 40% Hispanic and swung toward Trump by at least 5 percentage points.

Urban versus Suburban

Biden’s gains were largest in suburban areas, as was immediately apparent after the election. But because many urban counties have been among the slowest to report their votes, the picture in these counties has become clear only recently.

Trump did better in 2020 than 2016 not only in Miami-Dade, Los Angeles and the Bronx, but also in New York's Brooklyn, Queens, Manhattan, and the counties encompassing the cities of Chicago, Philadelphia, San Jose and San Francisco.

Though urban counties remain overwhelmingly Democratic, Trump’s margin improved by 1.1 percentage points in these counties — an even bigger pickup than Trump achieved in non-metropolitan counties, which are largely rural. Thus, the urban-suburban divide narrowed in 2020 — and the urban-rural divide did as well.

Regional Factors

Some regional and local patterns offer other possible explanations for 2020 swings.

— Home-state loyalties: Arkansas is one of only a half-dozen states that swung toward Trump. It’s possible that Clinton benefited in 2016 from some residual good feeling in the state based on her decadelong term as first lady of Arkansas — something that Biden couldn’t compensate for. (The Fort Smith, Arkansas, area was among the top 25 metro areas that swung toward Trump.)

On the other hand, Biden appeared to receive a home-state bonus in Delaware, which had the third-biggest swing against Trump of any state.

New Castle County, with Wilmington as its county seat, moved almost 8 points toward Biden. Why did the Salisbury, Maryland, metro area have such a big swing against Trump? One possible reason is that a little more than half of the population of the metro area is across the state border, in Sussex County in Delaware.

— Religion: Trump enjoyed big gains in the Utah metros of Provo-Orem and Ogden-Clearfield, but much of it can be explained by the drop in third-party voting. Evan McMullin won 21.5% of the vote in his home state in 2016, but he didn’t run this year. Utah, the home of a large Mormon population, had the largest decline in third-party vote share in the nation.

There were sizable swings toward Trump in areas with high percentages of Orthodox Jews, like Rockland County, New York, and parts of Brooklyn — perhaps because many Orthodox Jews were angered by coronavirus-related restrictions.

— Military presence: Overall, Biden performed well relative to Clinton’s 2016 showing in the Northeast, particularly in smaller metro areas, like in upstate New York and in places like Portland, Maine, and Norwich-New London in Connecticut. Connecticut and Maine each moved about 6 points toward the Democratic ticket.

But there’s another possible reason for the swing in Norwich-New London. New London County; the Bremerton, Washington, area; the Virginia Beach metro; and Colorado Springs (home of the Air Force Academy) have military bases and a fair share of military veterans. Counties classified as military posts by the American Communities Project swung by around 6 points against Trump, still voting for him in 2020 overall but by a smaller margin than in 2016.

Some of the correlations are hard to untangle. What we do know more clearly is that there has been about a 2.4-point swing nationwide toward Biden relative to Clinton’s 2016 showing and a similarly modest swing in many of the critical battleground states. But even modest shifts in a roughly evenly divided nation can be decisive.