November 8 was a pretty good night for Democrats all over Connecticut. Ned Lamont won a second term behind the largest margin any Democrat has enjoyed since Bill O’Neill’s re-election in 1986, Richard Blumenthal easily swatted away the Trump-endorsed challenge of Leora Levy, Democrats once again swept all the constitutional offices, and held their big majorities in the legislature.
So why, in what turned out to be a better night than most Democrats had dared dream, was Jahana Hayes so nearly defeated by Republican George Logan in the 5th congressional district?
It’s a great question, and one I suspect strategists and analysis on both the Republican and Democratic side will be dissecting for the next two years. To try to answer it, let’s take a look at the map of the election.
On first glance, it seems like Hayes did everything right. She carried the Farmington Valley by a decent, if not stunning, margin. She won crucial towns like Newtown and Cheshire, and kept margins fairly low in towns like Southbury and Woodbury. So what went wrong?
The reason lies in the cities. Danbury, Waterbury, New Britain, and Meriden are the Democratic firewall in the 5th, lending their electoral heft to Democrats Chris Murphy, Elizabeth Esty, and now Jahana Hayes. It’s taken for granted that Democrats will win the cities by such a huge margin that no Republican will be able to get close without winning basically every suburban and rural town.
But what if that doesn’t happen? Two weeks ago in a piece speculating about how Logan could win, I wrote that Democrats usually carry the cities by somewhere between 7-10% of all votes cast. But:
So let’s say Logan keeps those margins down, either by getting voters to vote for him, or to dampen their enthusiasm for Hayes so much that they stay home. Let’s say Hayes wins the cities, but only by about 4-6% of the total vote. Logan can now make up that ground.
And that’s exactly what happened. Hayes won the four cities 38,159 to 29,065, which adds up to a 9,094 vote margin. That’s 3.6% of the total of 253,567 votes cast. The fact that the margin in the cities was so shockingly low is what allowed Logan to get so close while still losing the Farmington Valley and a few other big towns like Newtown and Cheshire.
What happened? Was it just low turnout? No. Turnout in the cities was definitely lower than in 2020 or 2018, but it wasn’t too far off from 2014 or 2010. This is well within the norms for the cities during a non-presidential year. Elizabeth Esty won 37,253 votes in the cities in 2014, similar to Hayes’s total this year, but her opponent, Mark Greenberg, only won 20,788 votes. Both 2010 and 2006 were very similar. No, in order to find an election in the cities that looks like this one, we have to go back two decades to 2002 and the face-off between incumbents Nancy Johnson and Jim Maloney.
2002 was the first year of the 5th as we know it. Before the redistricting process in 2001, Connecticut had six congressional districts. We lost a district thanks to anemic population growth, and so two incumbents would have to face one another. If you ever wondered why New Britain is in the 5th, it’s because that’s where Republican Rep. Nancy Johnson lived.
Maloney won the cities 34,320 to 28,122. He lost the race because he couldn’t win anywhere else; the only other town he won was tiny Cornwall.
George Logan managed to put up Nancy Johnson numbers. How did he do it? Relentless campaigning? Negative ads about Hayes? His adorable Spanish-speaking mother? Something else? Whatever combination of these things were the reason, he put a serious dent in the Democratic advantage in the cities.
Unfortunately for Logan, the rest of the district is very different now than it was in 2002. Running closer in the cities isn’t enough. If he had managed to win in places like Cheshire, Farmington, and Newtown, he would have beaten Hayes. But he didn’t.
Jahana Hayes survived, and Democrats can breathe a sigh of relief. But this should be a warning to them not to take for granted Connecticut’s cities and all their diverse constituencies.