The U.S. Census reports that 224 million people were U.S. citizens aged 18 and over eligible to vote in the 2016 presidential election. Perhaps contrary to popular belief, a solid 70% of those eligible to vote were reported as registered. The polls saw 60% participation among all Americans who were eligible to vote (both registered and non-registered). In total, approximately 137.5 million people cast a ballot in the 2016.
Despite losing the election, Clinton won the national popular vote by a difference of approximately 2.9 million votes. This margin of victory in the popular vote represents the greatest of any losing candidate in U.S. history. The difference was equal to just greater than 2% of the total votes cast nationally. Yet, despite her unequivocal and decisive win of the NPV, Clinton lost to Trump even more decisively and unequivocally after the tallies of electors from the states were in. It would seem reasonable to expect at least some correlation between the numbers from the popular vote and the winner of the race to 270--so why the disconnect in 2016? And what could might this mean for future elections?
In 2016, most votes did not matter. Why? Votes for the loser in any state do not count toward a national total. They are essentially thrown away. All the votes for a winner in a state do not count either: whether a candidate wins in a state by a landslide or by very few votes, the number of electoral votes he or she receives is the same. So why would candidates bother campaigning in states they are sure to win? They wouldn't, and they don't because votes in those states do not matter. Our elections come down to the results of ten or fewer competitive states, a figure that has not changed much over several election cycles. Obama campaigned in only ten states in 2008 and in 2012. Republican challengers mirrored this strategy. The result? Candidates are getting good at winning with the fewest number of votes. And how does that benefit voters?
So, in future elections, we might very well see even greater disconnects between margins of victory in the popular vote and electoral margins of victory after all state electors are tallied, at least if the current winner-take-all model of electors by state remains constant.