UPDATE: So it turns out, "100%" reported on the New York Times election result for California does not refer to 100 percent of votes counted but to 100 percent of precincts reporting incomplete counts. There remain millions of ballots to be counted and the currently estimated Clinton popular vote majority is MORE THAN TWO MILLION VOTES. So it turns out my speculation regarding turnout is mistaken but the resulting estimates of popular vote margin are nevertheless in the right ballpark.
According to the New York Times, Hillary Clinton currently leads Donald Trump in the popular vote by around 280,000 votes. Donald Trump leads in electoral votes, 290 to 228 with Michigan and New Hampshire yet to be declared. Trump leads in Michigan and Clinton in New Hampshire. Adding in those two states, Trump's marginal goes to 306 to 232.
If electoral votes were allocated proportionally, by state, rather than winner-take-all, Clinton would garner 271 electoral votes compared to 267 for Trump. That would be rather closer to the popular vote outcome than the current result.
But here is another issue that I haven't seen discussed at all. Florida, a swing state with 29 electoral votes had a total of 9,386,750 votes cast, while Californians, in a safe Democratic state with 55 electoral votes cast a total of only 8,9320,68 ballots. See how that works? Candidates campaign, spend money and run get-out-the-vote operations in swing states but neglect safe states, skewing voter turnout and thus popular vote totals in favor of the supposed swing states.
Assuming voter turnout proportional to population in every state and assuming that the split between parties remains constant, Hillary Clinton's margin of popular vote victory would increase to over a million votes.The electoral college not only contradicts the popular vote outcome, it also skews the popular vote outcome by discouraging -- or not encouraging -- participation in supposedly safe states. If all states had the same level of turnout as Florida but retained their proportionate share between parties, Clinton's popular margin would increase to 1.4 million.
The conventional assumption is that low turnout hurts Democratic candidates most because Democrats receive a larger share of lower income voters and lower income voters have a relatively lower participation rate. Based on this conventional assumption, then the estimate of a 1.4 million vote margin for Clinton may be conservative.
Remember, though, this hypothetical popular vote margin of 1.4 million would not change electoral vote counts, which would still be 306 to 232 in Trumps favor on the current winner-take-all basis or 271 to 267 for Clinton if awarded proportionately by state.