The 2016 Presidential Election witnessed a sizable breakdown in the highly polarized political patterns seen post-2000. Traditionally Republican suburbs with high levels of college degrees broke away from their partisan mold for the first time in decades, while the Democrats fell through the floor in white rural (and some urban) regions with low levels of college completion, some of which had been Democratic stalwarts for nearly a century. While demographic trends have long pointed toward such a realignment, the uniquely repulsive attributes of Clinton and Trump spurred on its rapid realization long before anticipated in quasi-Maoist stage-skipping fashion. But how much of this breakdown was solely a one-off phenomenon contained to 2016 versus a longer lasting change? While we can’t know for certain until the 2018 Midterm Elections at the earliest, the early special elections of 2017 provide a few clues to an answer.
In Georgia’s Sixth Congressional District, Rorschach-like Democratic wunderkind Jon Ossoff built a campaign based off the strong anti-Trump/pro-Clinton trend in Atlanta’s suburbs, believing his trolling motto “Make Trump Furious” could lead to an upset in the traditionally Republican district. If Ossoff could persuade anti-Trump Republicans to support him as they did Clinton while maintaining a strong turnout advantage, a path to victory becomes apparent, but this relies on quite weighty assumptions.
How much did the patterns of 2016 carry over in the GA-06 race? A fair amount, actually. The precinct-level correlation between Ossoff’s vote share and Clinton’s vote share was 95.4%, slightly higher than the 94.7% correlation with Tom Price’s Democratic challenger in 2014. Ossoff clearly retained a number of anti-Trump Republicans, but we also saw some reversion to the mean: the correlation between the 2016-2017 swing and the 2014-2016 swing was -41.7%, implying Ossoff retained approximately three-fifths of the trend toward Clinton in the district. However, in only one of the three counties were Ossoff’s patterns more similar to Clinton’s than Montigel’s: Fulton. In Cobb and DeKalb Counties, Ossoff’s patterns had more in common with 2014 than 2016, suggesting partisan Republicans in Cobb may have reverted to form whereas partisan Democrats in DeKalb may have turned out as a higher percentage than in 2016. The reversion was not symmetric, however; the reversion in Cobb was substantially stronger than the reversion in DeKalb. Ossoff only carried over 20.6% of Cobb County’s 2014-2016 change in patterns while he carried over 74.1% of DeKalb’s swing.
In Kansas’s Fourth Congressional District (or at least in Sedgwick, Harvey, and Butler Counties where precinct results were easily available), Democrat James Thompson sought to employ the tremendous unpopularity of Republican Governor Sam Brownback to his benefit, as Democrats almost successfully did in 2014, only to be undone by a national (or at least east-of-the-Rockies) Republican wave. As in GA-06, the Democratic challenger’s patterns in KS-04 more closely resembled those of Clinton’s performance than of previous races; the correlation between 2017 and 2016 was 91.9% whereas the correlation with Paul Davis’s performance in the 2014 Gubernatorial Election was 87.9% and the correlation with Perry Schuckman’s House-level performance was only 82.0%. This time, all three counties had a stronger relationship with 2016 than 2014; however, unlike in Georgia, this may be more beneficial to the Republicans than the Democrats given the larger rural share of the population and the fact that rural areas swung harder for Trump and against Clinton than suburban areas swung against Trump. Specifically, rural Butler County looked more like 2016 than Wichita-based Sedgwick County, although both were more like 2016 than Harvey County. There was also very little carryover one way or the other between the 2014-2016 swing and 2016-2017 swing, suggesting little longer term trend behavior.
Together, these two race paint a similar story: 2016’s trends are carrying over into 2017 races, but in a more muted way than last year. There is little evidence that Democrats have been able to (at least so far) extend their gains of 2016, but appear to be retaining slightly more than half of the gains (or losses) they made previously. This suggests that suburban districts with narrow Trump victories might be just out of reach at the moment while those that flipped to Clinton altogether may be somewhat more fertile ground, but even such districts won’t flip automatically without a very strong turnout game. Similarly, rural districts along the Mississippi in the upper Midwest might also be just out of reach for Republicans, but any open seats in rural Minnesota could be strong pick-up possibilities.
The game hasn’t completely changed, but it’s certainly progressing forward.