Sign up for How To Read This Chart, a weekly data newsletter from Philip Bump
That is not the point of this article, though. Instead, the point is this tweet from the New York Post.
When I first read it, I skipped over the word “big” and jumped right to the “tight in suburbs, upstate” part. Knowing a bit about politics and a bit about New York, my immediate thought was: If Hochul’s tied upstate, she must be crushing him. And only then did I see the word “big” and learned of the 14-point gap upon clicking through.
But that framing stuck with me. The Siena poll breaks out the regions highlighted by the Post and, sure enough, shows that Zeldin has margin-of-error leads both upstate and in the counties immediately surrounding New York City. In the city, though? Hochul’s up by nearly 50 points. Since the city is nearly half of the population of the state, that’s all you need.
Consider just presidential voting. Since 2000, New York has consistently voted for Democratic presidents by about a 20-point margin. The gap in support between the city and those upstate counties has consistently been about 50 points. In 2016, with native son Donald Trump on the ballot, the upstate counties voted Republican, leading to the biggest upstate-city gap in the past six elections. On average, 4 in 10 votes cast in New York in those elections came from the five counties of New York City.
So why did the Post highlight the “upstate” and “suburbs” part? Who cares? It’s like being behind 50 at half time but trumpeting that you scored just as many points in the second half. Uh, cool?
Here, though, the effect is probably something else. By separating out the city from the rest of the state, the Post is drawing a culturally familiar line for New Yorkers. Those people in the city want one thing but the rest of the state — real New Yorkers — want something else. What’s more, it’s highlighting a differentiation between city and non-city, between urban and rural, that has come up a lot in recent years.
Remember when Trump won that race in 2016 and started showing everyone that red-hued map of his victory? He was taking advantage of the fact that large, rural areas with low populations are more visible than compact, heavily Democratic cities to suggest that his victory was broader than it actually was. (After all, he lost on the metric that map depicted: county-level popular vote.) In another sense, though, he was offering a judgment: That so much of the country was red was a reflection of how important those areas were. It was right to submerge the results in cities to those in rural areas since those rural areas voted correctly. Voted right.
This impulse isn’t simply metaphorical. In 2018, Wisconsin Assembly Speaker Robin Vos (R) argued that “if you took Madison and Milwaukee out of the state election formula, [Republicans] would have a clear majority.” Excise those counties from the state and you can finally measure its real predilection. When officials in Wayne County, Mich., considered rejecting certification of the 2020 election results, one was direct: Maybe they could reject just the votes from heavily Democratic Detroit and accept the rest. Before a House special election in Ohio in 2018, the Republican candidate argued that voters shouldn’t “want somebody from Franklin County representing us” — that is, from the city of Columbus. He won.
Part of this stems from the fact that rural and urban areas of the country have been moving in different directions in recent years. The Agriculture Department has a nine-category rating system for counties to evaluate how urban or rural they are. Comparing the presidential vote in the most-urban and most-rural counties nationally and in key states since 2000 shows that divergence — though not as much in New York.
Another categorization, from the University of Wisconsin’s Population Health Institute (PHI), clusters counties into fewer groupings. The effect, though, is largely the same: The margins in rural counties often shift to the right, as the margins in large urban counties shift to the left.
See that weird Arizona result? Maricopa is the state’s most-populous county, by far. In 2020, more than 60 percent of the votes cast in the state were cast there. It has long been the main driver of presidential results — and has generally voted Republican. It’s also the only “large urban” county in the PHI categorization.
Importantly, this shift hasn’t been uniform. Both nationally and in states like Michigan, more-rural counties have shifted right faster than more-urban ones have shifted left.
Overall, on either metric of county urbanization, the trend has been consistent: a growing gap in presidential vote margins between urban and rural counties. (The only exception is Arizona, where Maricopa County in 2020 voted more like other urban counties.)
Interestingly, the gap between urban and rural in New York has actually not grown as rapidly as in other states — in part because it was already wide back in 2000. Other parts of the country are now becoming as polarized on the urban-rural measure as New York has long been.
We have to note one element of subtext here. Cities vote more heavily Democratic and are therefore seen by some on the right as unwelcome participants in the democratic experiment. But the residents of those urban counties also often don’t look like rural residents, quite literally. Trump’s 2016 election was confounded heavily by racial views; the extent to which muffling or excising cities also disempowers non-White voters is hard to overstate.
In Wisconsin, for example, removing Milwaukee and Dane (home to Madison) counties from the mix dumps a quarter of the state’s population. It also removes a fifth of Wisconsin’s White population — and half of its non-White population. New York state without New York City is 43 percent smaller and loses 63 percent of its non-White population.
There is an added way in which the New York Post’s summary of the gubernatorial contest in that state was odd: The Post is itself a product of the city. It was playing off a resentment of the city despite being an emblem of it. But then it also often targets the city’s substantial number of moderate or conservative residents. And there are a lot of them: More residents of New York City voted for Trump in 2020 than did residents of Alaska, Idaho, Montana, Nebraska, North Dakota, South Dakota, West Virginia or Wyoming.
Cities are not as simple as might be assumed.