Historically, political party affiliation was a function of one’s income (i.e., the wealthier one was, the more likely that he/she was Republican). Even when Louisiana was a solidly Democratic state, the affluent voter populations in more urbanized areas of the state could be counted to vote Republican both at the top of the ticket and in “downballot” races.

This correlation between income and political party preference began to weaken in the 1990s, when social (as opposed to economic) issues became a more dominant part of the political discussion. More affluent areas in major metropolitan areas like Los Angeles, Chicago, Philadelphia, and New York moved towards the Democrats both in Presidential and “downballot” races, although “economic issue” elections like the 1994 and 2010 midterms tended to bring those voters back to the Republicans. This movement towards the Democrats was somewhat offset by movement towards the Republicans in historically Democratic rural areas. That “counter movement” arguably elected George W. Bush President in 2000: his upset win in West Virginia (which before then only voted Republican in landslide years) put him over the top. Without that victory, it would not have mattered how Floridians voted.

Last year’s candidacy of Donald Trump, however, introduced issues (protectionism and “building a wall”) that weren’t typically in a Republican platform, and this seemed to have further realigned the bases of support for both Democratic and Republican candidates. While Trump lost support (relative to what previous Republican nominees got) among white collar voters in the cities and suburbs, he also picked up critical support among rural and/or blue collar voters, and this additional support enabled him to break the “blue wall” in Michigan, Pennsylvania, and Wisconsin for the first time in a generation.

These realignments also occurred in the Deep South, and JMC Analytics wanted to test the extent to which this was the case by analyzing the 2012 and 2016 Presidential election results in Louisiana and Georgia – two Deep South states with a major metropolitan area (Atlanta and New Orleans) and a substantial black population. As part of this analysis (the extent that Trump lost white collar/urban support while picking up rural/blue collar support), JMC aggregated the data into “urban” (parishes/counties with a population of 250,000 or more) and “rural” parishes/counties.

ANALYSIS #1: GEORGIA

While Georgia voted twice for the Republican candidate, there was some slippage between 2012 and 2016 for the Republican nominee: the eight point margin by which Mitt Romney carried the state (53-45%) narrowed to a five point margin (50-45%) for Donald Trump.

When Georgia’s 159 counties are aggregated into “urban” and “rural” counties,  these contradictory political trends are apparent: the six “large” counties went from 60-39% for Obama to 63-33% for Clinton – not only did Trump perform six points worse than Romney, but his support dropped in every “large” county, with steepest declines (nearly ten percentage points less than Romney) occurring in once solidly Republican suburban counties. At the same time, Trump performed better than Romney did in the remaining 153 counties: Romney’s 62-37% victory widened to 63-35% for Trump.

ANALYSIS #2: LOUISIANA (SUMMARY)

Unlike Georgia, Louisiana voted solidly Republican both in 2012 and 2016: Mitt Romney carried the state 58-41%, while Donald Trump was favored 58-38%. However, the realignment apparent in Georgia was also apparent in Louisiana, even though the “top line” numbers suggested a static voter preference.

The five “urban” parishes (Caddo, East Baton Rouge, Jefferson, Orleans, and Saint Tammany), narrowly favored Barack Obama 51-48% but gave Hillary Clinton a somewhat larger 51-46% victory. And just like in Georgia, each “large” parish showed slippage for Trump relative to what Mitt Romney got in 2012, with the largest “slippage” occurring in East Baton Rouge. At the same time, the “rural” parishes went from 64-34% Romney to 66-31% Trump.

ANALYSIS #3: LOUISIANA (DETAIL)

Where the contradictory trends is really apparent is at the more detailed (JMC used state House districts, which typically have about 45,000 people) level. If this analysis is focused on those districts where the change in Democratic and/or Republican performance between 2012 and 2016 was five points or more, an interesting picture emerges:

Trump underperformance relative to 2012

Trump overperformance relative to 2012

The seats where Trump under performed are more affluent, white collar districts who have generally had Republican legislative representation for some time. On the other hand, the seats where Trump over performed were once the bedrock of Democratic voting strength, and as these seats have voted increasingly Republican at the top of the ticket, they have demonstrated a willingness to vote Republican down the ballot as well. It’s also worth mentioning that seven of the eight Democrats representing districts showing a noticeable movement towards Republicans in 2016 are term limited in 2019.

CONCLUSION

While political coalitions are fluid in nature, Donald Trump’s candidacy has somewhat scrambled the coalitions typically supporting Democratic or Republican candidates. And that shift (which was already occurring outside the Deep South since the 1990s) was noticeable in the Deep South in last year’s elections.