October 25, 2016

Lessons from Brexit

This is not just another presidential election season, since one of the candidates (Trump) has made the issues so black-and-white, so fundamental, and so urgent, that it is taking on the character of a national referendum on what direction we want the country to go in. The closest analog in time, theme, and cultural background, is the Brexit vote, and we would be foolish to ignore the outcome there when trying to predict what will happen in two weeks here.

First, the decision to Leave the EU was not geographically uniform -- there were entire areas like Scotland that went heavy toward Remain, as well as large urban areas throughout England. This will be akin to the red state vs. blue state pattern, as well as urban areas going blue in general.

That is to be expected from the past several elections, but it's worth emphasizing in order to make Trump voters realize that there will be no landslide akin to Nixon or Reagan. We (and the British) live in too polarized and partisan of a climate for there to be uniform support for any position throughout the country.

There also was not a landslide in the national vote -- 52 to 48, which is not unreasonable to assume for Trump beating Hillary. He will not win by 10-15 points.

The key region of support for Leave was the Midlands, which is whiter and more de-industrialized than other parts of England. Amazingly, a city here broke the pattern of urban areas voting to Remain, and in fact it is the second-largest metro area in all of England -- Birmingham voted to Leave. This is akin to the Rust Belt in America, which will be the key region for Trump. We may be in for a surprise where a major city will break the blue mold and vote Trump -- Pittsburgh or Cincinnati would be my guess.

Youth voting disappeared for the EU referendum, compared to the last two general elections: only 36% of 18-24 year-olds voted, and 58% of 25-34 year-olds.* It reached 83% for ages 65+. Younger voters were not only more toward Remain, but more extreme in the sense of farther away from the national average. Older voters were more toward Leave, but closer to the national average.

In other words, the climate turned off large chunks of normal people in the younger age groups, leaving only the hardcore extremists, while that same climate engaged an unusually high number of older people, including all normal people.

We're already seeing that among voters who are under 25 or 35 -- they could not be more disengaged, and viscerally repulsed by the election at this point. Polls have already found their desire to vote has fallen dramatically since 2012, but that was awhile ago. After the most combative debates in memory, most younger people are saying they hate both sides, there's nothing positive or uplifting on either side, they're feeling bullied into voting, and they can't wait for the whole election to just get over with, and they can get back to their regularly scheduled program of playing video games and watching Netflix. They have washed their hands of the election.

They were excited and talking non-stop way back when Bernie was still in the race, and they tried to fake enthusiasm for Gary Johnson after the Convention. But who's heard from him or seen him in the past month and a half, since "Aleppo"? His numbers are plummeting in the polls, from over 10% to just over 5%. Most of these will not bother voting on Election Day. Jill Stein will pick up the hardcore progressives, but they're not that numerous. And they all loathe Crooked Hillary Clinton, and have since the primaries, whose wounds have never closed (made fresh with each WikiLeaks revelation).

Young people have more or less gone back to fantasizing about "What if Bernie were still in the race?" or have checked out altogether emotionally. They're mostly not going to turn out, although those that do will be extremist SJWs supporting Hillary. Trump will get enough support among younger voters -- like Leave, it won't be a total blowout among young people -- but it won't end up making much of a difference.

Polls, polls, polls -- wrong, wrong, wrong. Here is a postmortem from YouGov about what aspects of polling made them better or worse. The main methodological finding is that phone polls did worse than online polls, and just about all of the public polling right now is done by phone (the USC / LA Times poll is a notable exception, and did much better than others in 2012, when it was known as the RAND poll).

Regarding turnout, apart from far fewer young people turning out than expected, the polls also missed how many less-than-highly educated voters would show up (even adjusting for the correlation of education and age). That is going to be a big surprise this time, too, particularly in certain white working-class areas in the Rust Belt where it will tip the balance in the state. As with Brexit, the non-college-educated are heavily in favor of Trump, where they would normally vote Democrat if at all (and their English counterparts would vote Labor).

We will undoubtedly be reading all kinds of spergery up through Election Day about what various poll aggregators are saying. How was their track record on Brexit? Here is quantitative tea-leaf-reader Sam Wang's summary on the very day of the EU referendum. He puts it at Remain +1, with a 95% confidence interval going only as far as Leave +2.6 on that side -- in reality, it was Leave +3.8. All that number crunching, and the outcome landed far outside his range of expected outcomes. Nobody has appreciated this failure, and everybody is going through the same motions with the polls on Trump vs. Clinton.

One under-appreciated source of error was what to do with people who said "don't know". No matter what pollsters did, their decision favored Remain and made them less accurate. I take that to mean that most people saying "don't know" at such a late date, on such a fundamentally important question, with polar opposite choices only, is simply unwilling to state their clearly held opinion. And that opinion will be whichever choice is less socially or otherwise acceptable to confess.

The entire financial and political Establishment, all of the media, educated elites, and even powerful foreign leaders like the President of the United States, were all bearing down on anyone thinking of voting Leave. While most of them could not have cared less what the elites thought, some small chunk was intimidated into public silence -- but privately voted enthusiastically for Leave.

I think that's what's going on with Americans who are answering "don't know" -- it's the "Shy Trump" voter. By now, you know pretty well who you're for, there is no middle ground for fence-sitting with such polar opposite choices, and the matters are of fundamental importance (immigration, trade, war, corruption, and so on).

The entire world is hammering you over the head to vote Clinton or be cast out of respectable society forever, so if that were your choice, why not proudly admit that you're going with the socially sanctioned choice? Some of these "don't know" voters could decide to vote third party (also not very acceptable in what you perceive to be a do-or-die election), but most are going to vote Trump. And of course some may wimp out of voting at the last minute, despite passing through the "likely voter" screen on the poll. At any rate, Clinton will get very few of these, and Trump will get the plurality.

That's all pretty good news for the Trump movement.

There are, however, a few things working against us that were not at play in Britain. They have to do with rigging the election. All ballots were paper and counted by hand in the EU referendum. We are using mostly electronic machines, and in many deep blue states the company making them is owned by George Soros. We will need good exit polling to see if they've been mechanically rigged -- if there's more than a 2-point difference between the exit polls and the reported result.

We also have more endemic corruption in our cities, specifically relating to elections, than they do in England. These cities with one-party rule and large non-white populations are the biggest threat to an honest and fair election, and unfortunately they are located in several states that we are looking to win -- Philadelphia and Detroit, for example.

I'd say that means a closer race than Brexit, as far as the initial reports are concerned. There could be voter fraud and election theft involved if we lose Pennsylvania or Michigan (definitely if we lose Ohio). That would take awhile to resolve. But with all the preparation Trump and others have been doing to deal with those contingencies, there's not much else we can do or worry about.

As far as last-minute shocks, the GOP Establishment launched a failed coup with the hot mic tape and coordinated desertion. That's why Trump's numbers have been holding more or less even over the past week or so, when they should be entering the final upward phase of the wishy-washy cycle. I don't know whether that's better or worse than some wacko on the Leave side killing a Remain MP a week before the vote. And yet the Leave side survived that last-minute crisis.

What else can we say but buckle up and enjoy the end of this very close race.

* A later poll done weeks after the vote found youth turnout that was nearly double the estimate of the day-of poll. I dismiss that as the respondents lying about having voted, in order to not appear that they had shirked their civic duty to keep Britain in the EU. Young people were widely blamed when their minimal turnout was reported, giving them quite an incentive to lie about it weeks later.


  1. I second your opinion that Pittsburgh could flip in this election. In the greater Pittsburgh area, Trump signs outnumber Hillary signs by anywhere from 3:1 to 9:1 roughly. That's true even in staunchly Democratic Allegheny County. In surrounding counties, such as Beaver, Butler and Westmoreland, there are almost no Hillary signs. I don't know to which extent these signs indicate overall enthusiasm, but I've noticed a rough correlation between these signs and which way an area votes.

    I have also noticed many lifelong Democrats switching their parties around the primaries so that they could vote Trump.

    So Trump might win PA by "the fuzz on a tick's ear" and that same fuzz might be enough for a win overall.

  2. Even the Iron Range in northern MN looks poised to turn to Trump. Thus area is typically D+20 above the national average and has been forever. Trump winning here would be the final dagger through the working class slice of the dem coalition.

    Check out my analysis:

  3. I'd say if exit polls favor Trump by ANY amount more than the vote results, we should be suspicious, because if anything voters will under-report their votes for Trump.

  4. John Killmaster10/25/16, 3:13 PM

    What a sweet thing it would be if Trump were to pull it off. Just the look on the talking heads faces would be like the sipping the finest of wines. However, I can't help but think that even a Trump victory would only be a last hurrah. The coming generation has been turned against their own bloodline and will hand the country over to the barbarians. Who in the alt-right-o-sphere is talking about that impending disaster? We are heading towards a holocaust or a civil war methinks, and in the last one 600,000 men had to die.

    Our only hope is for peaceful secession, which seems slightly more possible than in 1864, as I don't think either side would fight to keep the other in the union. We need to be thinking about the new borders and migration to red states, even if Trump pulls out the victory.

  5. Birmingham voting leave was a big shock too as it is the 2nd most "diverse" major city in England.

    Full of blacks and Pakistanis


You MUST enter a nickname with the "Name/URL" option if you're not signed in. We can't follow who is saying what if everyone is "Anonymous."