Fraser Nelson

Don’t ask the experts

Pronouncements from wonks are no substitute for proper political argument

Don’t ask the experts
Text settings

Michael Gove never intended to make his most famous remark. In an interview during the EU referen-dum campaign, the then justice secretary was told that the leaders of the IFS, CBI, NHS and TUC all disagreed with him about Brexit. He had tried to reply that people have ‘had enough of experts from organisations with acronyms saying that they know what is best and getting it consistently wrong’. But he was picked up mid-sentence by his appalled interviewer. ‘Had enough of experts? Had enough of experts?’ Gove’s partial quote was held up to ridicule, as if it embodied Trump-style populist rage; the battle of emotion against reason.

As it turned out, the experts were all wrong — or, at least, everyone who predicted a instant and immediate recession after the referendum ended up feasting on humble pie. The Bank of England cut interest rates and printed money in response to an economic slowdown that now turns out to have been imaginary. Indeed, economic growth seems to have accelerated after the vote. Andy Haldane, chief economist of the Bank of England, recently spoke about the economics profession having had its ‘Michael Fish moment’ — a reputation-shredding failure of prediction. Economists, he said, are now ‘to some degree in crisis’.

They’re in good company. Look at political scientists: on the eve of the US election, Princeton University academics calculated that Hillary Clinton had a 99 per cent chance of victory. This finding mattered, they said, because activists should know that they would be wasting their time campaigning for her and should help a Democrat congressman instead. The bookmakers gave Donald Trump a 20 per cent chance, and a Brexit victory only 11 per cent. Their error was based on opinion pollsters, who were confounded. A cartoon in The Spectator summed it up. It depicted three homeless men, each with a placard. They read: ‘ex-bookie’, ‘ex-pollster’, and ‘ex-pert’.

Economists are not clairvoyants, which is why none of them saw the 2007 crash coming, nor the extraordinary British employment boom, nor the immigration rise, nor the spending splurge that followed the Brexit vote. But a recent Financial Times poll of 120 economists showed they do not feel they were proved wrong about Brexit, even though two thirds of them were wrong about the aftermath of the vote. By revering experts as Romans once did augurs, certain politicians had come to see the actual voting as a formality. On election night, Hillary Clinton didn’t even switch on her television until eight states had declared their results. She signed a celebratory issue of Time magazine, printed to commemorate her victory. (That copy would be worth rather a lot of money now.) She should have known better: Mark Penn, her pollster, famously subdivided America into dozens of categories of people (soccer moms, etc) and worked out who to target in order to win her the 2008 nomination. But then along came Barack Obama with an inspirational message that appealed to everyone.


The fault doesn’t lie with the data scientists, but with the politicians who wrongly treat them as prophets. David Cameron, for example, notoriously hired the pollster Andrew Cooper as his chief strategist, an appointment that seemed to be like putting a weather-vane where a compass needle should be. But Cameron had faith in ‘the data’ and would scrap policies if the polls looked inauspicious. After a while Cooper returned to Populus, his polling company, and came up with the ‘predictor’, which calculated the likelihoods for the 2015 election. It gave a 0.5 per cent chance of a Tory victory. The ever-forgiving Cameron then appointed Cooper as the Remain campaign’s pollster: on polling day, Populus predicted a ten-point victory for Remain.

It’s not surprising, or new, for opinion pollsters to get things wrong. What’s new is how quickly they’re forgiven, then trusted again by politicians. The explosion of data has made the temptation worse, because there have never been more ways to divide, analyse and model society. Faith in data, computer models and prediction lay behind the 2007 crash: speculators believed the new algorithms could predict the future and made massive wrong bets. The same hubris emerged with the rise of the political class; young men who went from Oxbridge to Westminster and knew they were out of touch with working-class voters. So they sought out polls and focus groups to tell them what the C2D voters were thinking — underestimating what a blunt tool this has always been.

The best economists do stress uncertainty, and emphasise that a guess ending in a decimal point is still a guess. Robert Chote, who runs the Office for Budget Responsibility, is a case in point. He was once invited on to BBC Newsnight and told that his forecasts for tax and spending could be wrong. Of course they’re wrong, he replied: economic forecasts usually are. They’re a bunch of assumptions plugged into a computer. Get one wrong, and the whole thing is rendered useless.

This is the point politicians had forgotten — or wilfully ignored. In recent years, a narrative has emerged of elections being contests between the educated and the knuckle-draggers, the ‘post-fact’ loons. This has led to the musings of experts being presented as fact, and to the rise of the argument-without-argument: X is correct because experts say so. And if you disbelieve X, you’re a post-truth Flat Earther. The Remain campaign tried this method: the fake dossiers, celebrity endorsements and multi-signatured letters to newspapers. It sought to say: don’t try to work Brexit out for yourselves, dear voters. The intellectuals have excogitated the problem, and their conclusion is unanimous.

Fraser Nelson and James Forsyth on the end of experts

The strategy was stunningly ineffective. You don’t need a PhD to notice the gap where argument and a postive case for membership ought to have been. And to wonder why, if economists all failed to predict the 2007 crash, they should be right about a post-referendum crash. Being proved utterly wrong tends not to deter public intellectuals. As the philosopher Eric Hoffer once put it, ‘one of the surprising privileges of intellectuals is that they are free to be scandalously asinine without harming their reputation’. Professor David Blanchflower, a labour market expert, predicted that a million jobs would be lost as a result of Tory austerity. When two million jobs were created instead, he lost none of his hauteur or enthusiasm for offering advice. In 2015, after David Cameron’s austerity programme, we were told by no less an authority than Joe Stiglitz, a Nobel laureate, that inequality had ‘gotten much worse’ in Britain. Instead, as it emerged this week, income inequality has been falling and now stands at a 30-year low. An incredible, historic achievement — but who predicted it?

Professors Blanchflower and Stiglitz are men of fearsome intelligence and ability. But there is what Thomas Carlyle called ‘unwise intellect’: smart people can believe in daft notions. One of the groups of Americans identified by Mr Penn for Hillary Clinton was the ‘impressionable elite’ — people who consider themselves empiricists yet who are driven by emotion and personal bias. How many economists, trade experts or diplomats might fall into that category? Of those, how many would admit it?

The people of Britain have not really had enough of experts: scientists and number-crunchers are still respected, valued and needed — in their fields. But their ability to foresee the future has been thrown into question, and their predictions sought a bit less as a result. When Theresa May decided to govern without the services of a pollster, she was settling down to life in a post-expert political world. She hasn’t yet won an election. But as she will have noticed, the people who do so tend to be those who realise that the algorithm for human nature has not yet been invented — and that there’s no substitute for good old-fashioned politics.

inline-ad-experts3 Fraser Nelson and James Forsyth discuss the end of experts.