Seems to me that the counties with big swings towards the D nominee (in terms of percentage margin) between 2012 and 2016 are (i) those around Dallas, (ii) those around Houston, (iii) those around Austin, and (iv) El Paso. That sounds like underperformance among high-income suburbanites and Mexican-Americans. The counties with smaller swings probably can be explained by Gary Johnson and Egg McMuffin getting far more votes than did right-of-center third-party candidates four years ago.
Trump’s vote percentage in GA-06 (the original subject of this thread) was 48%, down from 61% for Romney. The GA-06 are the high-income suburbs north of Atlanta, the area represented by Johnny Isakson and Newt Gingrich prior to Tom Price. Less populist Republicans than Trump did much better in high-income suburbs in 2016, although obviously they did worse than Trump in blue-collar areas (compare the victory maps for Ron Johnson in WI and Pat Toomey in PA to those of Trump). We need to combine Trump’s performance in blue-collar areas with traditional GOP performance in high-income suburbs, or our control of Congress, the presidency, key state governorships and key state legislatures will be brief.
I just was alarmed with noting that Trump failed to increase his % performance over Willard in 33 out of 35 TX CDs. I found it to be an astonishing fact, given the opposition and contrasting that with other GOP states. Hopefully this will just be a one-off and 2020 will see Trump break through 60% there.