Behaviour of the largest Trump 2016 breakthrough counties in 2020

In 2016, a number of counties voted Republican for the first time since at least 1988. Of these counties, 27 cast over 40,000 votes in 2020 (a somewhat arbitrary cutoff to make for a decently-populated, but manageable, list--although it also appears to have formed a rough cutoff for the counties on-air analysts saw as important on election night [e.g., Chuck Todd singled out Erie County, Ohio, the smallest of these and casting only barely above the threshold (40,416 votes), at one point]). 

Trump increased his vote share in 20 of them (bolded; the colour is of the party the county voted for in 2020):

Pueblo, CO (+1.8%)
Windham, CT (+0.4%)
Monroe, FL (+2.4%)
Dubuque, IA (+3.3%)
LaPorte, IN (+2.9%)
Androscoggin, ME (-0.9%)
Kennebec, ME (+0.5%)
Saginaw, MI (+1.1%)
Bay, MI (+1.8%)
Gloucester, NJ (+0.3%)
Niagara, NY (-2.3%)
Broome, NY (-0.5%)
Rensselaer, NY (-1.1%)
St Lawrence, NY (+3.9%)
Robeson, NC (+8.1%)
Montgomery, OH (+0.2%)
Trumbull, OH (+3.8%)
Portage, OH (+3.3%)
Ashtabula, OH (+4.2%)
Erie, OH (+2.9%)
Erie, PA (+0.8%)
Luzerne, PA (-1.3%)
Northampton, PA (-0.7%)
Kent, RI (-1.6%)
Jefferson, TX (+1.3%)
Cowlitz, WA (+5.8%)
Kenosha, WI (+3.5%)

The seven counties where his vote share went down were all in one of New York, Pennsylvania, Maine, and Rhode Island.

In 2020, Trump got the highest vote share of any Republican this century in all but one of the 20 bolded counties, despite losing a number of them back to the Democrats. (The exception is Montgomery County, Ohio, where he got 47.7% in 2016 and 47.9% in 2020; Bush got 49.0% in 2004.)

Of the above 27 counties, Trump won with a plurality in 2016 in 13 of them. This was the two-party vote in 2016 in those counties:

Pueblo, CO: 91.7%
Kennebec, ME: 92.1%
Rensselaer, NY: 92.8%
Kent, RI: 92.8%
LaPorte, IN: 93.1%
Dubuque, IA: 93.2%
Broome, NY: 93.2%
Kenosha, WI: 94.1%
Montgomery, OH: 94.6%
Saginaw, MI: 94.8%
Gloucester, NJ: 95.1%
Northampton, PA: 95.4%
Jefferson, TX: 97.3%

Trump generally got fairly ample credit for a number of these counties. In May 2017, David Jarman wrote, 

[I]f you’d told me that the Republican candidate was going to win places like Kenosha County, Wisconsin, Trumbull County, Ohio, or Monroe County, Michigan (bastions of organized labor that kept the Democratic faith over the decades, even during their 1980s low-water mark), that would have sounded like a catastrophic Dem wipeout, probably wore than Walter Mondale’s benchmark of futility.

(Monroe County, MI is an odd choice; it voted Republican in all three elections of the 1980s and had last voted for a nominee of either party that was losing the national popular vote by more than 4.5% in 1944, for Dewey. Kenosha and Trumbull, on the other hand, had last voted Republican in 1972. A more fitting third county--if it was to be taken from a Rust Belt swing state--might have been Erie, PA, or Saginaw, MI--both had voted for Reagan in 1984 [although, of course, so did Monroe, MI]--or Dubuque, IA, which had last voted Republican in 1956. A third Trumb breakthrough county that voted Democratic throughout the 1980s and cast closer to 100,000 votes than Dubuque would be Pueblo, CO.)

Of course, the dramatism was helped by the juxtaposition of counties that surprisingly went the other way; Jarman also wrote,

If you’d told me before the 2016 election that the Democratic candidate was going to finally win Orange County, California (the most legendary Republican stronghold of all), as well as Fort Bend County, Texas, and Cobb County, Georgia (the suburbs that gave us Tom DeLay and Newt Gingrich, respectively), it would have been safe to assume that would be part of a crushing Democratic victory, something on the order of LBJ vs. Goldwater.

But if you’d told me that both things would happen in the same election, if I hadn’t slowly and cautiously backed out of the room, I might have assumed you were talking about a distant future election, one that exaggerated trends we were only starting to see.

Orange, Cobb, and Fort Bend were all Romney-HRC counties (and all casting over 100,000 votes). There was 'only' one Dole-Gore county that cast over 100,000 votes--Orange, FL--but it, like the three Romney-HRC counties Jarman names, had a long, loyal stretch of Republican voting behind it when Gore flipped it, having last voted Democratic in 1944. And, like Cobb and its name-mate in California, it had provided Republicans with their largest raw-vote margin in its state (albeit only once, in 1988--which, however, was also the only time Florida gave the Republican nominee his largest raw-vote margin of any state in the country). And it was sufficiently 'legendary' as of 2000 that Lawrence Levy focussed on it in an October 2000 article about how contemporary conservatism was faring in Southern suburban areas. In 2004, Ken Mehlman told Wolf Blitzer that 'one of the reasons you called [Florida] was because a county that was called a bellwether county was a place called Orange County. It's Orlando. Four years ago we lost it...'.

However, there were a number of large counties that had voted to re-elect George H. W. Bush in 1992 and then switched to Clinton in 1996, that went on to vote for Gore in 2000. Some of these counties had last voted Democratic no more recently than 1964 before 1996. In such cases, I think it's fair to say that Gore carrying them was fairly dramatic (certainly more so than Trump carrying Monroe County, Michigan). The largest such counties were Oakland, MI, Suffolk, NY (which, ironically, Trump won back for the GOP, despite that it had, like most of these counties, 'benefitted immensely from the global order that Trump promised to upend'), Franklin, OH, Bergen, NJ, and Fairfield, CT. Given their criticality to certain large swing states, Oakland, MI and Franklin, OH would have made a nice triplet combined with Orange, FL, akin to Orange (CA), Cobb, and Fort Bend in 2016. (Dole had also managed to keep the margins closer in Oakland and Franklin than in the three large Northeastern Republican suburban counties he dropped; Bill Clinton won outright majorities in 1996 in Suffolk and Bergen.)

Overnight on election night 2016, John King asked Wolf Blitzer to look at Erie County, Pennsylvania, saying 'if you're a Democrat, that's supposed to be blue.'  Bill Hemmer pointed out that Trump was slightly leading in Trumbull County, and was able to contrast the preliminary Trumbull County results with what the county had done in the previous three elections. In December 2016, Jud Lounsbury pointed out a plethora of counties Trump had taken in the key states of Iowa, Wisconsin, Michigan, Ohio, and Pennsylvania that had last voted Republican no more recently than 1988 (in most cases, no more recently than 1984).