Once all the votes are counted, it looks like Hillary Clinton will underperform Barack Obama by about 4 percentage points in the national vote. Was this an across-the-board loss, or was it concentrated among certain groups?
The quickest way to get a sense of what happened is to compare the exit polls from 2012 and 2016. What we’re looking for is demographic groups that differ from -4% by a significant margin. As it turns out, there aren’t very many. Clinton underperformed Obama across the board. She did somewhat better than -4% with seniors, college grads, married voters, and high-income voters. She did worse with low-income voters, union households, and unmarried voters.
This was not a “white revolt.” White men followed the national trend (-4% compared to 2012) and white women did better for Clinton (+1%). Black men and Latino women underperformed for Clinton by significant margins.
The big surprise here is that Clinton did so much worse with unmarried voters. She underperformed Obama among unmarried men by a whopping 10 points, and among unmarried women by 5 points. What’s up with that? I would sure like to see a crosstab of unmarried men by age, race, income, etc. Latino voters are also a surprise. Clinton only did slightly worse than Obama, but surely she should have done much better. What happened?
Here’s the full set of comparisons to Obama in 2012. For reference, 2012 exit polls are here. 2016 exit polls are here.
White born again