Surging Democratic turnout in swing-state suburbs carried Joe Biden to the US presidency, according to a Financial Times analysis of detailed election results from thousands of voting precincts.
In a contest defined by record turnout for both candidates, the outcome of the race hinged on who was able to attract the most new votes — and where.
A blue wave in the majority-white suburbs of Pennsylvania and Georgia was pivotal to Mr Biden’s win in the electoral college. In Arizona, both majority-white and majority-Latino areas in Phoenix and Tucson contributed tens of thousands of net votes to Mr Biden, overturning a swing towards Mr Trump in majority-white rural areas.
This year has also underscored the diminishing utility of umbrella terms such as “Black”, “Latino” and “Asian” for voting groups, as race has played an increasingly nuanced role.
“You can’t overgeneralise our community,” said Lorella Praeli, president of Community Change Action, during a recent panel on Latino voter trends. “You need to understand that we are different in New Mexico, and we are different in Nevada, and different in the state of Florida.”
Detailed data on turnout by demographic group will not be publicly available until next year, so the FT has analysed turnout by looking at shifts in voting patterns at the local level. Our analysis compared tallies in majority-minority neighbourhoods to those in majority-white areas between 2016 and 2020.
While large cities remained Democratic strongholds, it was the suburbs that accounted for Mr Biden’s biggest swings in many states, including key battlegrounds in the rust-belt and the south.
In 2016, Mr Trump saw a surge in both turnout and support among white voters, while turnout and support rates for Hillary Clinton among black voters, particularly in battleground states, declined.
Mr Biden underperformed Mrs Clinton in Philadelphia’s majority-black neighbourhoods while making gains in its majority-white areas. Further south in Atlanta, both majority-black and majority-white precincts added more new votes to the Democrats than the Republicans, but the blue shifts were much larger in majority-white areas.
Suburban swing in crucial states
Across the country, Mr Biden made his biggest gains in city centres as urban America turned even deeper blue.
But in key swing states Pennsylvania, Michigan, Wisconsin and Georgia, he gained a net 355,210 votes in suburban areas, more than double his gains in large metro areas. This shift allowed him to win the suburbs in these states outright after Mrs Clinton lost them by more than 200,000 votes in 2016.
By contrast in the biggest cities in all of these states, Mr Biden either lost ground or remained stagnant compared with 2016.
Mr Biden flipped Wisconsin, the “tipping point” state that pushed Mr Trump over the 270 electoral vote threshold in 2016. This year, record turnout topped 3.2m statewide, but in the city of Milwaukee the number of votes cast stayed flat. Instead, it was Milwaukee’s inner suburbs that delivered the “blue wave” for Mr Biden.
It was a similar story in Pennsylvania and Georgia, where the suburbs drove Mr Biden to victory.
The suburban swing was opposed by a red rural wave. The Democrats lost ground to the Republicans in the rural portions of 37 of the 40 states for which sufficient data exist, continuing the country’s steady geographical polarisation.
In Arizona, the dense cities of Phoenix and Tucson prevailed against a red rural surge, propelling the Democrats to victory for the first time since Bill Clinton’s 1996 win.
A suburban swing for Mr Biden is consistent with a previous FT analysis showing the president-elect gained ground with college-educated white voters. At the same time, America’s suburbs are becoming increasingly racially diverse, and the higher numbers of non-white voters in these areas appear to be contributing to the shift towards Democrats.
Non-white areas were critical for Biden in some places, but key for Republicans in others
Race and place are tightly interlinked in the US, meaning different demographics were key to the outcomes in different regions.
In Georgia, Stacey Abrams and other Democratic party organisers have been credited with a drive to register 800,000 new voters across the state, more than half of whom are non-white. The data show that these additional votes — which skewed heavily Democrat — played an important role in Mr Biden’s victory there, though perhaps not the pivotal one.
Majority-black precincts in the state capital Atlanta actually swung very narrowly towards Mr Trump compared with 2016, but still backed Mr Biden by 93 per cent to 6 per cent. Combined with a 7 per cent rise in turnout, these neighbourhoods added 14,490 more votes to the blue column than the red compared with 2016, and without them the Republicans would have won the state.
Even more critical in Atlanta, however, were the majority-white suburbs, where a 20 per cent jump in turnout was matched by an 11 point swing towards Mr Biden.
In Wisconsin and Pennsylvania — states that helped bring Mr Trump victory in 2016 — many saw black turnout as the key to the election outcome. In the biggest cities in these two states, Milwaukee and Philadelphia, majority-black precincts were where Mr Biden had the highest victory margins.
However, Mr Biden underperformed Mrs Clinton in these areas, which, despite being overwhelmingly Democratic, turned out less and shifted slightly Republican since 2016. 
Many factors could be at play, including the closure of colleges and universities due to the pandemic, which could have depressed turnout. Several majority-black areas have also seen population declines or changes over the past four years, such as more white voters moving in.
In Arizona, majority Latino precincts swung 2.6 percentage points towards the Republicans while majority-white ones swung 4.3 points towards the Democrats. But as in Atlanta, existing partisan skews meant that both added more votes to Mr Biden’s column than Mr Trump’s.
Although Mr Trump’s support in Latino precincts increased on 2016, 72 per cent still backed Mr Biden, meaning that their increased turnout added 26,725 more votes to the Democrats than the Republicans in the state’s five most populous counties for which the FT obtained data.
Away from the battleground states, precinct-level data in California suggest portions of the Asian community have shifted slightly towards Donald Trump since 2016.
Orange County, where Republicans won back two House seats, as a whole swung narrowly towards Mr Biden, but in the diverse cities of Westminster and Garden Grove there was a sharp swing towards Mr Trump. While majority-white precincts swung for Mr Biden, the large Vietnamese American community increased its vote for the Republican by 35 percentage points, suggesting a similar dynamic to Miami, where Cuban Americans voted Republican as an anti-socialist move.
Methodology and sources
Demographic data in this article came from the US Census Bureau’s 2014-18 American Community Survey. We aggregated this data from the block group level to precincts by calculating the fraction of the population in a block group that fell into each precinct, then using these fractions to apportion the census counts. We used the National Center for Health Statistics’s urban-rural classification for counties.
Election results by precinct were obtained from state and local election offices where possible. For some states, we used geographical boundaries and historical results data compiled by the MGGG Redistricting Lab and the Open Elections Project.
Many places have changed their precinct boundaries over time. For Milwaukee, we used a data set compiled by John D. Johnson of Marquette Law School derived from the Milwaukee County Election Commission and the Wisconsin Legislative Technology Services Bureau. For Philadelphia and other cities that have changed their boundaries since 2016, we reapportioned historical data into 2020 boundaries using a similar method to the one described on the Philadelphia elections blog Sixty-Six Wards.
Using group data — such as precinct- or county-level results — to infer individual behaviour is a rough way to estimate voting patterns; a clearer picture will emerge in the months ahead, as vote tallies are finalised, states update their voter lists and the US Census Bureau publishes its survey estimates.