- REUTERS/Lucas Jackson
Back in 1998, the world of finance learned a very painful lesson: Models break and markets aren’t efficient.
And with the rise of Donald Trump from sideshow to presumptive Republican nominee, politics has learned the same lesson.
Long-Term Capital Management was a hedge fund staffed by several Nobel Prize winners that possessed a supposedly unmatched grasp on how markets work. The firm had the most sophisticated methods for exploiting any and all inefficiencies, millions and millions of times over. And it blew up.
Chronicled at length in Roger Lowenstein’s brilliant book “When Genius Failed,” the short version of LTCM’s blowup is that a series of misplaced bets that certain interest rates would converge over time – because they always had in the past – went against the firm until they were out billions of dollars.
LTCM’s core conceit was that it believed markets were efficient and any inefficiencies would be corrected in due course.
They were wrong.
Trump’s efficient market
And so Trump was not only written off as a marginal and unserious candidate because he himself is so unserious, but also because no outsider candidate with no establishment support and using only his money to fund his campaign could, the theory goes, win a nomination.
In politics there’s an axiom that says “the party decides.” This idea that vague, entrenched powerful interests – not the voters – determine an election’s outcome is the political field’s efficient-market hypothesis.
And it was essentially this idea that underwrote months of Trump commentary that effectively followed the simple line that this can’t happen because the party won’t allow it. The voters and the media can have their fun (look at Bernie Sanders), but eventually reality sets in and the practical candidates that Very Serious People believe are best for the job will be put to the voters.
And yet here we are, not a year later, with Trump as the presumptive Republican presidential nominee.
- REUTERS/Jonathan Ernst
Buoying themselves against this market ideology of deciding parties were not only the traditional pundit class but also the newly crowned top dogs of the political commentariat: the data hounds.
In 2008, Nate Silver rose to fame by correctly calling the election in favor of outsider Barack Obama, a junior senator who overcame the establishment in defeating Hillary Clinton but also topped the face of the GOP, revered Sen. John McCain.
In 2012, Silver again nailed the election – which was never really that close – that so many believed to be a toss-up between Obama and GOP establishment choice Mitt Romney.
But this time around Silver failed.
On Wednesday, Silver wrote: “To me, the most surprising part of Trump’s nomination – which is to say, the part I think I got wrongest – is that Trump won the nomination despite having all types of deviations from conservative orthodoxy.”
Nate Cohn at The New York Times also drew similar conclusions in a reflection on what data-based predictions about the Republican contest got wrong.
Which are both ways of saying that it seems the party itself failed – the market failed to self-correct its inefficiency, in other words – and thus the arguments undergirded by a belief in a coherent party structure inevitably fell apart.
There was simply no there there.
The statistician George Box once wrote, “Essentially, all models are wrong, but some are useful.”
This is true.
The failure to predict of Trump’s rise to the nomination, then, is not the fault of the work done by folks like Silver, but a manifestation of the hubris involved in trusting the party over what was happening on the ground. Trump dominated Republican polls for months, but his place in the race as a self-funded outsider who was clearly not the choice of The Party seemed entirely untenable.
The incoming data was doubted all the way. The model broke.
In a great tweetstorm Wednesday, former Wall Street trader Chris Arnade – who was among the slick, model-wielding upstarts to hit finance in the 1990s – broke down the problem with models, with beliefs, and why Trump’s imminent nomination is, really, a pie in the face for everybody.
The success of Silver in 2008 and 2012 at the time appeared to be the triumph of math over feeling or inspiration. The classic political pundit could – still can! – anecdotally outline their case for or against a certain candidate. Silver instead brought the data to back up his view. And he was very right.
But where a Silver-style model eventually broke down this cycle was in doing what all models do: using the past to predict the future.
And this is ultimately why Box’s quote endures. All models – even those that are useful and correct for long stretches – will eventually reach a point at which the current inputs no longer yield results that look anything like the past. The model’s guiding light goes dark. The model breaks.
Arnade argued Wednesday that this affirms the need for on-the-ground reporting, meeting voters in real life, getting a feel for just how serious the Trump thing is by talking to people who take it seriously.
Maybe this is the answer. Maybe not.
But Arnade’s point is that using the model as a backstop to affirm your priors – that Trump can’t win because he’s not the party’s choice, that he’s too unserious, too racist, too inconsistent, too everything – is exactly the point at which the model begins to fail.
Long-Term Capital Management thought all arbitrage opportunities would eventually revert to some efficient equilibria. Then they incurred a revision of belief, and then they were out of business.
Markets, in general, are pretty efficient.
But they are not absolutely so. There’s an old John Maynard Keynes quote that says markets can remain irrational longer than you remain solvent. Trump rendered the Republican Party insolvent.
Business Insider’s Henry Blodget wrote Wednesday that Democrats who are so confident that our next president will be Clinton should be a bit more humble.
And what I think his post really drives at is that to believe Trump can’t beat Clinton would be to once again trust the model, trust the market. This position would require you to believe, just like the GOP did for the last year or so, that the party would decide, that the sensible decision would be made, and the theory against which you balanced your world would not fall apart into nothing.
And then, finally, Trump lost Iowa. Peak Trump! The model was going to be right!
- REUTERS/Charles Mostoller
But this was and remains Donald Trump. His campaign of schoolyard-style insults exploited what my colleague Allan Smith has called Trump’s biggest strength: finding someone’s weakness quickly and hammering on it.
So Iowa winner Ted Cruz became “Lyin’ Ted.” A multi-decade Bush political dynasty was destroyed with two unforgettable words: “low energy.” Marco Rubio, the preferred establishment choice for the GOP nomination, became “Little Marco.” John Kasich was “1 for 38 Kasich,” which isn’t even that catchy: It’s merely true.
Cruz, as Josh Barro noted on Tuesday, ostensibly admitted that he was finally done deceiving voters after he knew he wouldn’t win their support.
Jeb Bush? He’s actually pretty low energy. Which is fine, but it’s also true.
Marco Rubio, at 5-foot-10, isn’t short at all, just shorter than Trump, who’s 6-2.
John Kasich did, in fact, win only one state.
But with these schoolyard insults – which, again, were mostly statements of slightly inconvenient facts – Trump galvanized his base against his opponents, against the party he hoped to represent, and against the truth we all took to be self-evident about modern presidential politics.
He upended the party and the “truths” that come along with a political establishment using its heft to nudge voters toward the candidate that has been vetted, supported, put in position to succeed.
Trump broke the model, and now he’s one vote from the White House.