[UPDATE: Several points in this analysis appear to be incorrect due to being based on preliminary vote tallies. In California and Washington especially, the counties can take several weeks to tally all of the mail in allots, and so several million ballots remain to be counted. 538 estimates that when all the ballots have been counted, the total number of votes in 2016 will be larger than in 2012, although the turnout percentage will be down. It is correct that Clinton got fewer votes than Obama, but in the final tally Trump will probably have received more votes than Romney.]
I've been wrong a lot during this election season. So have a lot of other people.
I was wrong when I thought that someone like Trump could never win the GOP nomination. I was wrong when I thought he could never win the general election. In this I had a fair amount of company. Nate Silver's 538 went into election night giving Trump just over a 30% chance of winning the election. That was a pretty fair estimation of the polling that was out there, and Silver actually took a lot of flack (and got into some profanity laced Twitter battles with other analysts) because venues like the Huffington Post were giving Clinton over a 90% chance of winning. The polls, however, were wrong, and Trump's victory arguably marks a reversal in polling-based expectations which is larger than the famous "Dewey Defeats Truman" screw up, because that was a time of relatively few and relatively primitive polls.
So here we are, with a real estate mogul and former reality TV star as our president elect, a man with no governing experience who managed to win despite massive opposition from many leaders and intellectuals within his own party. The GOP, which people were looking to write off as a fractured, disintegrating party, will control the White House, the House of Representatives and the Senate, as well as the majority of state governorships and legislatures. Trump will be able to replace Antonin Scalia with another conservative Supreme Court justice (assuming he bothers to nominate a conservative), thus leaving conservatives one Anthony Kennedy vote away from controlling the judiciary as well. The Democrats are left with a lot of thinking to do, though there remains a good deal of sorting out to do on the Republican side of the aisle as well. Will Trump's nationalist populism come to dominate the party, or will the fiscal and cultural conservatism of leaders like Paul Ryan keep the GOP on course?
Some thoughts follow under various sub-headers.
There's a lot being written about how Trump rode to victory on a wave of working class white anger at the state of the economy and the country. There's something to this. Trump achieved something impressive in winning key rust belt states such as Pennsylvania, Ohio, Wisconsin and probably Michigan (he's ahead by a small margin with all precincts reporting but the vote still has not been officially called.) I think there is a story out there to be told about how Trump connected with the white labor voting block in a way that the Democrats (despite their union ties) failed to. But it's also significant that in an election cycle dominated by vicious attacks on both sides, turnout was solidly down. Trump received fewer total votes than Mitt Romney or John McCain did. The reason that Hillary lost is that while GOP turnout dropped just over 1% in 2016 versus 2012, Democratic turnout dropped a whopping 8%.
The patterns vary a bit at the state level. Ohio, Wisconsin, and Michigan all had decreased overall turnout, but their GOP turnout was increased vs 2012. This means that either people who voted for the Democrats in 2012 voted for the GOP in 2016, or Democratic voters stayed home in large numbers while Republican leaning voters who hadn't voted in 2012 turned out. (I would assume it was in fact a mix of these, but deeper analysis would be required to see which dynamic predominated.) In Pennsylvania, overall turnout was up, with a huge increase in GOP votes and a slight decrease in Democratic votes.
2016 and 2012)
Just to fill in the corners, I also compared data on the most populous states: California, Texas, and New York. None of these are swing states, but the varying dynamics are kind of interesting. In California turnout was down a whopping 19% with GOP votes down a lot more than Democratic ones. In Texas turnout was up 9%, but the GOP votes were only up slightly while 17% more votes were cast for Clinton in 2016 than for Obama in 2012. In New York (from which both candidates hale) the number of votes cast was up 11% with Trump getting 19% more votes than Romney did and Clinton getting 7% more than Obama did.
I'm still not entirely sure what to make of this. Given the national changes in popular vote, it would be easy to say that this was a case of Clinton simply not being able to turn out the Obama coalition, and Trump actually not being any more enticing as a candidate than Romney. This would leave open the question of whether Clinton was a uniquely un-compelling candidate, or whether it took Obama's unique combination of personal eloquence, inspiration, and being a fairly young and attractive candidate for First Black President to get the massive youth and minority turnout which brought him to the White House twice.
Here's a longer term view of the popular vote totals.
As you can see, 2008 marked a historic high for the number of Democratic votes cast, and Clinton marks a continuing decline from that level. The highest number of GOP votes cast was in 2004, but the trend for the GOP has been relatively flat since then.
[Due to time constraints, this has become the first of several parts. Next part to come soon.]