Over at MoneyIllusion, I did a post discussing some odd election anomalies. Now that I’ve had a chance to look at the detailed election map more closely, certain consistent trends seem to show up.
Let’s start with the basic overview. In 2016, Clinton beat Trump by 2.9 million in the popular vote. In 2020, Biden won by over 7 million votes. So obviously the country shifted substantially toward the Democrats.
But many sub-groups went the other way:
1. Vietnamese-American areas in Orange County shifted dramatically toward Trump, by margins on the order of 40%
2. Working class Asian neighborhoods shifted strongly toward Trump.
3. Hispanic areas shifted strongly toward Trump. (But still Democratic.)
4. In Rockland County, New York, Orthodox Jewish areas shifted strongly toward Trump, with margins not usually seen outside Turkmenistan.
5. While data is sketchy, Amish areas in Pennsylvania seemed to shift substantially toward Trump.
6. African-American areas shifted mildly toward Trump (remaining strongly Democratic in absolute terms.)
I’ve found news articles discussing most of these shifts, and in almost every case the article mentioned resentment against Covid shutdowns.
So if all these groups shifted toward Trump, many quite strongly, how did Biden do so much better than Clinton?
It wasn’t rural white voters. That was a mixed bag, with no clear trend toward the Democrats. So where did the extra 4.1 million Democratic margin come from?
The detailed election map of vote shifts from 2016 is clear. Look at almost any urban area and you see the same pattern. Red for Hispanic areas (sharp shift to Trump.) Pink for African American areas (mild shift to Trump.) And blue for white suburban areas, (strong shift to Biden.)
All across America, white suburban areas shifted toward the Democrats.
These were often well-educated white-collar workers who worked from home during Covid. Whereas working class minorities resented Covid restrictions, these voters resented the fact that Trump didn’t seem to make much effort to control Covid, discouraging testing, pushing unproven remedies, ridiculing mask wearers, etc.
This might be the first election where the voters who did well under a president turned against him while those that suffered economically turned toward him.
In this post, I’m not taking a stand as to which voters were right, just trying to understand why some parts of the country swung one way while others moved in the opposite direction.
Here’s a map of the vote shift in Philadelphia area, which is typical. The bright red in the north side is Hispanic. The pink area just to the west is mostly black, and the further out areas are mostly white suburbs. Again, these are vote shifts. In absolute terms Trump still did fairly well among whites and poorly among minorities.
PS. You may wish to compare these vote patterns to the racial dot map of America, which is quite informative. I’m sure people will do regression analyses, but these vote shifts are so obvious you don’t even need to run regressions.
PPS. I certainly don’t mean to suggest Covid was the only issue. Various news articles suggest that communities such as Vietnamese-Americans and Orthodox Jews were also motivated by other issues.
READER COMMENTS
robc
Feb 4 2021 at 3:43pm
One other thing that you didnt cover.
Even though African-American voters shifted mildly toward Trump, did there overall vote numbers go up significantly relative to the population as a whole?
To pick a random example, if your vote percent goes from 95% to 90% but the number of voters goes from 1MM to 2MM, then you gained 850k votes, despite the loss in percentage.
I haven’t looked at the numbers in any detail, but that is my impression of part of what happened. It may have been an even bigger factor than the suburban shift. If so, I would think it is primarily due to mail-in balloting, and possibly ballot harvesting.
robc
Feb 4 2021 at 3:45pm
Two corrections:
their not there. I really do know the difference.
Also, while the vote increase in my example would by 850k, the net would go from +900k to +1.6k, so only a net gain of +700k, not quite as much as 850k.
Scott Sumner
Feb 4 2021 at 5:15pm
I believe that voting was up almost everywhere, but don’t have specific data on that.
TMC
Feb 4 2021 at 4:07pm
The effect is mostly from suburban white women. Bill Burr did a funny bit on how white women took over the BLM movement.
Floccina
Feb 4 2021 at 4:48pm
I also wonder how big a factor the anti-police brutality protest that some used as a cover for rioting and looting, where.
Did Trump have huge GOP tail winds and it was just him that made the loss. He did seem to under perform the party and his crazy refusal to accept the election results may have lost the 2 senate races in Georgia. Of course any of this is hard to know.
Scott Sumner
Feb 4 2021 at 5:16pm
Yes, I expect that was a factor, indeed it’s cited in some press reports.
Mark Bahner
Feb 7 2021 at 12:02am
I have to confess, I *still* don’t understand the Republican Party’s support for Donald Trump. Republicans used to consider themselves as conservatives, but there’s virtually nothing conservative about Donald Trump.
So the only reasonable argument for sticking with him that I can think of is, “He wins.” But even there, he failed. And not just a little bit…he hugely under-performed his party. And then he had a huge part in losing the Senate (at least as I see it).
Now, the Republicans in the Senate are almost certainly going to vote overwhelmingly not to convict in the impeachment trial. That will leave Trump free to run in 2024. If he gets nominated, it seems like he’d be an extreme long-shot to get elected. So why not just cut the rope?
Mark Z
Feb 4 2021 at 6:49pm
I’m hesitant to try to infer much from changes in percentages. I doubt many Vietnamese or black Democrats converted to Republicans; more likely Republican-leaning black or Vietnamese voters were much more enthused than last election. And if that’s what caused the shift, it’s less likely to persist in future elections because relative voter ‘morale’ seems much more fickle than latent political affiliation. Maybe black non-voters tend to lean more conservative than black voters, and an inordinate number of black non-voters actually voted in 2020? If so this could easily disappear in future elections. Converting people from one team to another is a much more long-lasting accomplishment, but is much harder and probably less of a factor than variation in ‘enthusiasm,’ imo.
Scott Sumner
Feb 4 2021 at 9:04pm
Lots of Vietnamese who voted against Trump in 2016 voted for Trump in 2020. A very large number.
Mark Z
Feb 4 2021 at 9:51pm
The Vox article only gave changes in overall percentages of Vietnamese voting/supporting (it was written before the election) Trump, I don’t think it had any stats what % voted for Clinton in 2016 and Trump in 2020. Asian voter turnout is generally low (49% in 2016) and apparently increased dramatically (it’s hard to find national statistics for 2020 but in Georgia the Asian vote grew by 63% relative to 2016), so it seems ambiguous to what extent the shift is due to people changing their votes vs. changes in who voted and who didn’t.
Scott Sumner
Feb 5 2021 at 12:23pm
In Orange County’s Little Saigon the vote shifted from roughly 60-40 Clinton in 2016 to 60-40 Trump in 2020. That’s a huge shift.
Rajat
Feb 5 2021 at 1:20am
Isn’t this broadly consistent with what you said a couple of months ago?
Scott Sumner
Feb 5 2021 at 12:24pm
I’m old, so I repeat myself. 🙂
Thomas Hutcheson
Feb 5 2021 at 7:58am
This is pretty depressing if shutdowns — exclusively state and local policies — were used as issue in elections to national office. Granted that Federal institutions like CDC failed to provide states and local governments with data and recommendation on how best to regulate social distancing, mask wearing, and restriction on pubic commercial and social activity and failed to advocate for investing in Tabarrok-Romer massive testing of the asymptomatic, this is hardly the fault of Biden.
KevinDC
Feb 5 2021 at 9:41am
Depressing, perhaps, but not remotely surprising. The typical American voter tends to be very badly informed (or outright misinformed) and is very bad at figuring out who or what institution is responsible for what policy or outcome. Just think about how common it is to hear “X happened while Z was President” treated as if it was equivalent to “X happened because Z was President.” Statements like “during Obama’s presidency, 7.5 million new jobs were created” ends up becoming “Obama created 7.5 million new jobs while he was President.”
Federal government officials have spent decades assuring Americans that they can solve all the nation’s ills. Claiming that the Federal government can ameliorate X necessarily entails that X is something the Federal government has the knowledge, ability, and authority to control in predictable ways – otherwise, how could they promise to bring about improvements? Everything negative which occurs in the country is claimed to have been entirely created by the Other Tribe’s Failed Policies, and everything good which happens is the result of Our Tribe’s Wise Policies – even if whatever occurs has no plausible connection whatsoever to the Federal government. Federal politicians still find it in their interest to claim themselves to be in control of these things, to drive up the stakes and convince ever more voters that it’s critically important for Our Tribe To Have All The Power.
Given how Americans have spent decades being fed the message that anything good or bad that happens ultimately comes down to who is in Washington, and how loudly those in Washington take credit or assign blame among themselves for anything that happens, and given the overall level of political ignorance among Americans – it would be a miracle if the thing you find so depressing didn’t happen.
sd0000
Feb 7 2021 at 1:43am
Scott – you specifically point to COVID, but there’s a broader theory that’s gained momentum recently that the Republicans are becoming a multiracial working class party and the Democrats a white collar professional party.
Best article that I saw on this was an election postmortem with Democratic strategist David Shor on Politico:
https://www.politico.com/news/magazine/2020/11/12/2020-election-analysis-democrats-future-david-shor-interview-436334
“The joke is that the GOP is really assembling the multiracial working-class coalition that the left has always dreamed of,”
I think part of it is also just that the GOP had the more populist candidate this round. Working class voters are attracted to populism. If the election was Kasich vs. Sanders, I think you’d see the exact opposite outcomes. That being said, the fact that Republicans nominate Trump while Democrats nominate Clinton and Biden says a lot in itself.
Hazel Meade
Feb 7 2021 at 2:09pm
There is something to this, that Trump shifted the Republican party towards the working class. I’m not so sure about the multiracial part. Seems like Trump was aiming pretty directly at the *white* working class, but that might have had some spill over into other racial groups, because today’s working class is much more multi-racial than it once was (which is one of the things that angers the white working class). Policies designed to boost the income of working class white guys may unintentionally also boost the incomes of working class blacks, hispanics, and asians.
Mark Bahner
Feb 7 2021 at 2:02pm
Republicans and “multiracial working class.” They certainly have a long way to go!
Let’s look at 2020 presidential results, with Trump’s then Biden’s percentage:
NY Times 2020 exit poll results
Ethnicity:
Black: 12/87
Hispanic/Latino: 32/65
Asian: 34/61
Age:
18-29: 36/60
30-44: 46/52
65+: 52/47
Education:
College: 43/55
No college: 50/48
It seems to me that this indicates the Republican Party is presently the party of elderly white people.
Comments are closed.