Hillary Clinton is on pace to secure about a 1 or 2 percentage point margin over Donald Trump in the popular vote totals in the 2016 presidential elections. As of this moment, Mrs. Clinton has about 1.6 million more votes than Mr. Trump in that tally. (UPDATE: this post was last updated Nov. 20.)
Of course, this margin is meaningless. Except, I suppose, in pithy tweets designed to prove a point that is... well, meaningless.
First, campaigns would behave differently if they won elections based on the popular vote rather than the Electoral College. Jonathan Adler ably makes this point. Campaigns are designed to eek out, at any margin, electoral votes, not popular votes. And if the popular vote mattered, then campaigns would be designed differently. The most common analogy is to look at the 1960 World Series. The Yankees outscored the Pirates 55 runs to 27 runs. But the Pirates won the Series, 4 games to 3. That's because it doesn't matter if the Yankees won a game 16-3 or 12-0; the only thing that matters is winning 4 games. The rules define the contest.
Second, voters would behave differently, too. Would New Mexicans have cast over 73,000 ballots for former Republican Governor-turned-Libertarnian nominee Gary Johnson? Would Utahns have cast over 175,000 ballots for Evan McMullin? Would the Great Plains and upper Rockies have voted in such high numbers for the libertarian nominee (over 5% of the vote in Montana, Wyoming, North Dakota, and South Dakota)? You see, if their votes "mattered" in a national popular vote total, they may well have voted differently. Instead, because their results were restricted to their home states--often fairly reliably Republican or Democratic--voters may have behaved differently.
Third, our laws would have to be different to have a true popular vote tally. Consider, for instance, that Mr. McMullin was only on the ballot in a handful of states, including Utah; or that Green Party candidate Jill Stein was not on the ballot in all fifty states. Or, consider that some states have strict forms of voter identification, and others have none at all; some allow incarcerated felons to vote, and other prohibit them from ever voting if they have been convicted of a felony. We run fifty-one elections in the presidential election; dumping them into a single basket of the "national popular vote" simply doesn't tell us anything meaningful. (For more on that, consider my article in the Arizona State Law Journal on the topic.)
Fourth, while Mrs. Clinton may have the most popular votes, she will be far from a majority of the popular vote. She is likely to secure something around 48% of the popular vote total--meaning 52% of Americans voted for someone else. We tend to prefer majority winners, even though each state in the Electoral College can be carried by a plurality, and many other elections also occur by plurality winners. Nevertheless, note how the Electoral College requires an outright majority to win. And fairly narrow margins can quickly turn into apparent Electoral College landslides--consider 2012, in which Barack Obama defeated Mitt Romney by a popular vote margin of 51.1% to 47.2%, but won the Electoral College soundly, 332 to 206. Until a system with a runoff is in place, we might prefer a system that offers a fairly clear majority winner, whatever the rules may be.
Fifth, this result is exactly what the Electoral College was designed to do! One reason for the Electoral College was to protect the smaller states by guaranteeing them a meaningful say in the outcome of the presidential election, as each state receives three electoral votes, and the smallest states pack a greater punch in the Electoral College than their populations would otherwise suggest. But the smallest states are quite diverse in their partisan allegiances in recent years, and the Electoral College is doing something else.
But as another way of protecting smaller states, the Electoral College ensures that a candidate must have broad geographic support. That is, she cannot "run up the score" in a small number of jurisdictions. Indeed, at the Founding, some worried that New York or Pennsylvania would simply dominate the elections. That's the flip side of guaranteeing some say to smaller states--it's to ensure a broader base of support across the country. Trump looks to carry the plurality in 30 states. That's very broad support in a country of 51 jurisdictions! Granted, some of the support was somewhat narrow, of course--it's the reason he'll lose the popular vote total.
Indeed, a couple of charts displaying the sheer disparity of performance in two states--California and New York--effectively overwhelm the entire rest of the country.
Yes, California and New York two of the largest states. But their margin for Mrs. Clinton will likely exceed 4 million votes; the rest of the country combined will offer something like 3.8 million votes in favor of Mr. Trump. It is a deep geographic imbalance, reflecting a strong intensity of preference in two states for one candidate.
Undoubtedly, there is an appeal to the cry, "The person who gets the most votes wins." But given many complications in our federalism-driven election regime, the answers are far more complicated. In particular, there remain good (at least, good to me) normative reasons for the Electoral College--a requirement of broad geographic support for the presidency rather than pockets of intense support in a couple of places being one of them. For more, reach some of my articles on the subject.... (For more on this election and the Electoral College, see John McGinnis's perspective here.)
(I should note that I do not believe we'll see a major push to abolish the Electoral College. The National Popular Vote effort has moved forward in a handful of jurisdictions, mostly Democratic-controlled, and it's pretty much run out of new places to go.)
(UPDATE: these charts were updated with results as of Nov. 20.)