The Realignment Happened While Obama Was Winning

In 2008 the country swung 9.86 points toward Democrats, from R+2.5 to D+7.4 in two-party terms. It was the largest one-cycle national move in our file. In that same election, 1,912 of the 3,111 counties with a complete series — 61.5% of them — moved toward Republicans relative to the nation. Not in raw terms — relative to a country walking hard the other way, which is the test that isolates the county.

Here is the argument. The rural realignment across Tennessee, Arkansas, Oklahoma and the coalfields is usually dated to 2016, and dating it there makes it a story about one candidate. Our file says half of it had already happened eight years earlier, in the cycle Democrats won by 7.4 points. That timing changes what kind of thing it is. A realignment that arrives while the other party is winning nationally is not a response to the winner. It is those places leaving on their own schedule.

The forty counties that moved furthest

Rank all 3,111 complete-series counties by total change in lean from 2000 to 2024 and take the forty that moved furthest toward Republicans. Their average shift is 82.92 points. All forty had a Democratic margin in 2000; twenty-eight were still carried by Democrats in 2004; six in 2008. That is the whole collapse, and this is when it arrived.

2008 is the largest single cycle in that column, and it is not close. It is also the only cycle other than 2016 in which every one of the forty moved right. The flip ledger says the same from another direction: of the forty, twelve first flipped Republican in 2004, twenty-two in 2008, five in 2012, one in 2016. By the time the 2016 campaign began, thirty-nine of these forty counties had already stopped voting for Democratic presidents.

Jackson County, Tennessee is the extreme case and it is legible in raw votes. In 2004, 2,996 people there voted for the Democratic ticket. In 2008, 2,224 did — a fall of 25.8% in the cycle the Democratic vote nationally rose 18.1%, as summed from the county file. Republican votes in Jackson rose from 2,024 to 2,185 over the same span. The county's raw margin went from D+19.2 to D+0.9, and its lean fell 28.34 points. Nothing about that is a 2016 event.

Knott County, Kentucky did it in one cycle. It sat at D+30.27 lean in 2004 and R+15.43 in 2008, a relative swing of 45.70 points, and it flipped Republican that year for the first time in the series. Mingo County, West Virginia flipped in 2008 on 35.10 points, Grundy County, Tennessee on 36.93, Buchanan County, Virginia on 23.27. Across the forty, Democratic votes fell 28.1% between 2004 and 2008 while Republican votes rose 9.4%.

The case against this piece

By almost every broad measure, 2016 was the larger cycle and the 2016 date is defensible. Across all 3,111 counties the mean relative swing in 2016 was 10.03 points toward Republicans against 3.67 in 2008; 543 counties moved more than 20 points against 229; and 85.6% moved right relative to the nation, the most of any cycle in the file, against 61.5% in 2008. If you want to know when the country moved, the answer is 2016 and we are not going to pretend otherwise.

What 2008 owns is the tail. Take the hundred largest single-cycle relative swings toward Republicans out of all 18,666 county-cycles in the file: 45 of them are 2008 and 42 are 2016. The leading edge moved first and moved hardest under Obama; the generalization came later. Both dates are real and they describe different events. Collapsing them into one is how the second gets credited with the first.

And it is not mainly coal country

Kentucky and West Virginia together hold 21 of the hundred largest shifts. Tennessee, Arkansas and Oklahoma hold 58. More than a third of every county in Tennessee — 37 of 95 — is in the national top hundred, roughly three times the rate in Kentucky or West Virginia. Tennessee's statewide mean shift is 50.78 points toward Republicans, the largest of any state with at least twenty counties, ahead of West Virginia at 41.03 and Arkansas at 40.05. We do not carry a regional boundary in this data, so we name states rather than a region. The states are enough: whatever moved these places was not confined to coal.

What would have to change

2024 is the first cycle in this series where the forty moved back. Ten of them swung toward Republicans; thirty swung toward Democrats; the group mean was 1.02 points toward Democrats. That is one cycle, it is small, and it happened in a year when 83.1% of all counties moved toward Democrats relative to the nation — so most of it is probably the national environment rather than these places. One reversal against a moving baseline does not support a prediction, and we do not publish them.

What would settle it is specific, and November is not it — this series is presidential, and the next observation lands in 2028. If the 2024 step is these counties genuinely coming back, their leans keep closing in a cycle where the national environment is not doing the work for them. If it is an artifact of the national swing, the leans resume falling the moment the country moves the other way. Until there is another cycle to look at, the useful correction is the smaller one: when someone dates this realignment to 2016, ask what happened in Knott County in 2008.