The dominant narrative since the election has been about the rural (read white) vote.  I think that is right and it is important to understand why we lost voters who voted for Obama previously.

Less discussed, though equally important is what happened with African American turnout.  I first noticed something was amiss when I looked at the African American vote in Philadelphia.

Since that first investigation,and at the suggestion of someone at Dkos, I have updated the analysis I did on Philadelphia:
photo AAPHIL_zpsmcucituu.gif

.

One can say with confidence lower African American turnout cost Clinton Michigan (her margin in Detroit was more than 40K lower than Obama’s), Wisconsin (turnout in Milwaukee was down), maybe Pennsylvania (see above) and Florida.  
Initially most of the discussion about Pennsylvania assumed African American turnout had held up since Clinton’s margin out of Philadelphia was equal to Obama’s.  But the precinct numbers told a different story.

Exit polls are not granular enough to really fully understand what is happening with the African American vote.  Even there, though, you see disturbing trends.  Clinton’s margin in Florida among African Americans in the exit poll was +75, a 15 point decline from Obama who won the African American vote by 90. If the exit polls are right — and they are notoriously unreliable in certain subgroups — the shift from `12 was enough to cost Clinton Florida.

The best analysis of minority voting comes from the data files.  State data files contain the race of the voter and whether they voted.  

This chart is a comparison of the turnout by race in the battleground states versus what we would have based on prior elections and population vote. This comes from Patrick Ruffiani — a Republican to be sure — though his work here is respected by liberals as well.  Having said that it would be REALLY GOOD to have a group here analyze these files (I only have access to Florida — they aren’t cheap — but I do see the same problem).

photo racetrend_zps1lz0s3eu.gif

I want to draw your attention to the turnout among young African Americans, which was substantially below expectations.  The problem was most acute among young black men

There is good news with respect to the Hispanic vote and the young — which was over expectations.

But we need to really understand the reasons for the decline in this turnout.  With Senate races looming in most of the battleground states, reversing this turnout decline is vital.

The temptation is to blame voter suppression for part of this.  This can’t be the case in Florida, where the restrictions in place where in place in 2012. North Carolina and Wisconsin are a different story.

But that alone cannot, I think, explain all of this under performance. Why would it be so concentrated among the young?  Why the gap between male and female performance among young African Americans?  Similar gaps between genders do not exist to nearly the extent it does between young African Americans.

There are about 5 readers of these diaries and honestly I post it as much to keep track of the data that explains why 2016 happened.

If I were to list the diaries they would show:

  1.  There is evidence in rural counties that very white counties that saw an influx of immigrants saw bigger swings than those that did not.  This is consistent with UK research post Brexit. See also discussion about immigration in the focus groups that Greenberg did on Macomb County in Michigan, and the generalized racial animosity he found.
  2.  There is evidence that the economic recovery was much weaker in rural counties.
  3.  There is significant evidence that counties with stagnating populations in the Midwest where the mostly likely to shift from Obama to Trump in a significant way.
  4.  There were significant swings in opposite directions among income groups, particularly in states in the Midwest.
  5.  There was significant erosion in the battleground states among millennials.
  6.  In Florida, there were enormous swings in the exurbs around Tampa.  These shifts were as large as 10 to 15 percentage points, and account for Trump’s margin in Florida.
  7.  In my polling diaries I showed deterioration in Clinton’s numbers after the third debate.  Some of this deterioration was the result of GOP voters coming home to Trump, some of it occurred after the Comey letter. There was clear slippage in the last 72 hours in key states.  There is evidence that much of this late slippage was the result of shifts away from third parties.  Other polling has suggested that voters moved away from Clinton on the grounds that her economic message had not gotten through. Others have suggested the second Comey letter angered GOP voters and knocked Clinton off stride.

It is hard to substantiate a lot of this movement in the Florida early vote, which was reasonably stable in partisan makeup. Trump did perform well on election day in Florida, but some of this may the result of the large Democratic Early vote.

  1.  It is clear that Clinton’s organization was not as good as advertised, and in fact very thin in the Blue Wall.  In Florida Steve Schale who ran Obama’s ’08 Florida campaign has suggested that the lack of organization outside of deep blue counties may have had a role in Clinton’s loss of Florida.

  2. Clinton’s advertising was very focused on personality and was the least focused on policy in recent history.  One can argue that Donald Trump was her opponent and this makes sense, or one can argue it represents a failure to address the economic anxiety of voters.