The Popular Vote Is Likely Settled. The Electoral College Not So Much.

 I don't live and die by the election prognostications of wunderkind Nate Silver, but he's been closer to right than just about anybody else in gaming out the last two Presidential elections.  This past week, he crunched the numbers on an issue I've long been speculating on relative to this election cycle.....the growing disconnect between the popular vote and the need for the 270 electoral votes that ultimately decide who becomes President.  The bottom line is that with the current distribution of votes nationally, Donald Trump has a historically unique advantage in the Electoral College.

Recall that in 2000, Al Gore beat George Bush in the popular vote by approximately 0.5%.  Even though the Electoral College outcome was different, Bush's victory over Gore was in the same general ballpark, an advantage of 0.6%.  Fast forward 20 years later and the divergence has decidedly grown, according to Silver's calculations.  If Biden managed a 1% popular vote win over Trump this year, Silver gives Biden only 6% odds of becoming President.  Six percent!  A Biden popular vote of 1-2 points would give him only 22% odds of victory.  Even a 2-3 point Biden popular vote victory would give him only 46% odds of victory.  Let that sink in.  Biden could beat Trump by 3 points in the popular vote, even more than the 2 points Hillary won by in 2016, but Trump would still be more likely than Biden to be re-elected.  If Biden won by 3-4 points, his odds to win the Electoral College climb to 74%.  But that's still an amazing number.  Biden could win the exact same comfortable and decisive popular vote margin that Obama won in 2012, but still would have a 26% chance of losing to Trump in the electoral vote count that picks Presidential winners and losers.  Even with a 5% popular vote victory by Biden, Trump would still have an 11% chance of being re-elected.

How is this possible, any sane person would ask?  At the onset of a historical realignment of party affiliation, Republicans have an electorally ideal distribution of votes for Presidential elections.  It's the same reason Trump won a decisive 306 electoral votes in 2016 despite losing the popular vote by three million votes.  Democratic votes are concentrated in heavily populated coastal states while the swing states are dispersed throughout Middle America, and if they break narrowly in the GOP direction as they did in 2016, it can massively skew the national outcome.  Since 2016, there's some indication this distribution has grown even more favorable for Republicans, at least for the time being, because of the aforementioned realignment.  Several large red states are becoming more competitive while long-standing blue states are trending Republican.  

The most terrifying scenario for the Biden campaign is continued rightward trends in some combination of Wisconsin, Michigan, Pennsylvania, and Minnesota, leading to another scenario where Trump ekes out narrow majorities in most of them, while ekeing out equally narrow victories in some large red states moving away from him like Texas, Georgia, Arizona, and North Carolina, all while holding the perennial swing state of Florida.  In a worst-case scenario, Biden could win the popular vote by a full 6 points and still not become President if the vote distribution in six of those eight states fall in Trump's direction.  You can call that a warped perversion of a democratic election and I wouldn't disagree, but it's the system we got and it will be respected if it plays out that way.

This GOP Electoral College advantage is something of a historical anomaly and probably won't last for long, but there's no reason to doubt it will work to the continued advantage of the luckiest man in American political history, Donald Trump.  This is a guy who avoided a war with Iran only because the retaliatory missile Iran intended to use against America for a military attack hit its own passenger plane instead, as just one example of an unthinkable streak of luck the incumbent President has maintained.  Betting against him is a fool's errand.

So how likely is this scenario of a wide chasm between popular vote and electoral vote?  It's still probably odds-against, but the chances seem a bit larger every day.  Even Trump's people quietly acknowledge they have no chance at winning the popular vote barring a massive game changer, but there's some indication that polls are tightening in key battleground states, and that the remaining undecided voters look pretty Trump-friendly from a demographic standpoint.  Under normal circumstances, undecided votes tend to break heavily against an unpopular incumbent President, but the country hasn't been as tribalized as it is today in nearly 150 years, so when most undecided voters are noncollege whites who voted for Trump in 2016, it's a decent bet that more go Trump than not.  If you couple that with the likelihood that state polling is undercounting support for Republicans, as has been the case in the last three cycles, and you can start to cobble together a scenario where Trump can still win, possibly with relative ease if everything breaks his way as it did last time.

I concur with Nate Silver's arithmetic that the odds of this are about 25%.  Biden still has a lead in most battleground states, enough to get him over 300 electoral votes if they hold.  But so did Hillary.  Biden tends to be at or above the 50% threshold in many of these polls in a way that Hillary wasn't, but that's not to say he can't lose a couple points of that support....or that polls are overstating it by a couple of points.  Both parties tend to be pretty confident going into election night how things are gonna go, but it doesn't go entirely according to script very often.  The exception was 2008 where state and national results were pretty much perfectly in line with pre-election polling, but polling has gotten harder since then with the disappearance of land lines, a sharply declining response rate from callers, and the rapid shift of partisan allegiance between college-educated voters and noncollege voters.  I, for one, am going into this election night expecting pretty much anything, with very little confidence that I have the American electorate figured out, and I doubt anything that happens in the next seven weeks will change my thinking on that.