article main image
My advice for Rishi Sunak: superforecasters won’t make your Budget betterBy Evening Standard

Rishi Sunak, the Chancellor of the Exchequer, is presenting his first Budget on March 11. To do so, he’ll need some vaguely credible economic forecasts. Without them, the fiscal arithmetic is no more than guesswork. In the bad old days, chancellors simply looked at themselves in the mirror and came up with numbers that best suited their political purposes. Gordon Brown kept changing his forecasts to prove there could be “no more boom and bust”, singularly failing to spot the looming global financial crisis. In the late Eighties, Nigel Lawson persuaded himself that Britain was about to embark on a prolonged period of faster growth and lower inflation, conveniently ignoring the housing boom that led to the early-Nineties recession.

George Osborne changed the rules of the game by introducing the Office for Budget Responsibility, an independent team of forecasters whose job was to come up with projections for the UK economy free from the burden of political influence. All future chancellors would receive the OBR’s forecasts like tablets of stone. There would be no more room for wishful political thinking.

Yet even this process is now in doubt. Dominic Cummings, Boris Johnson’s right-hand man, suggested just the other day that journalists should stop wasting their time reading the thoughts of “political pundits who don’t know what they’re talking about” and instead turn to a book by Philip Tetlock (and Dan Gardner) called “Superforecasting: The Art & Science of Prediction”.

Being in the forecasting game myself, it just so happens I’ve read Tetlock’s book. As such, I need to offer a health warning. I work for a bank, I occasionally turn up on TV and (as you can see) I write columns for the Standard. I might therefore be regarded as a bit of a pundit. Tetlock rightly points out that, while pundits often tell a good story, few of them express their views with measurable precision. They don’t learn from their mistakes because they’re simply not aware of them. And, too often, they’re like stopped clocks: right twice a day but not in any meaningful way.

Tetlock runs forecasting “competitions” designed to identify people who are free from the various psychological biases that lead most of us astray. These “superforecasters” don’t provide precise estimates — unlike Albert Einstein, they know that God does indeed play with dice — but they have an uncanny knack of beating so-called “experts”, including, famously, the US intelligence community, at their own game.

You can see why Cummings might be a bit of a fan. After all, the 2016 Brexit campaign was partly based on the idea that — as Michael Gove once put it — “people in this country have had enough of experts”. Coming from all walks of life, superforecasters are, by definition, not experts: they may be more numerate and intelligent than the average citizen, but they don’t have any kind of political or economic axe to grind.

Yet I have my doubts. Tetlock regards John Maynard Keynes, the great economist behind the General Theory, as someone who behaved a bit like a superforecaster: “One of Keynes’s many remarkable accomplishments was his success as an investor.” He notes that Keynes was financially almost wiped out in the early Twenties and again after the 1929 Wall Street Crash “but he bounced back and did even better than before….try, fail, analyze, adjust, try again. Keynes cycled through these steps ceaselessly”. If, however, Keynes had died not in 1946 but in, say, 1931, we might have reached a rather different conclusion. Dying in 1931 would have left his investing reputation in tatters. His apparent investing “success” is determined in part by the random year in which he met his maker. Had he lived a lot longer, he might well have screwed things up again. We’ll never know.

Another problem. Imagine — as Nassim Nicholas Taleb does in The Black Swan — that a superforecasting airline safety officer ruled that, from September 10, 2001 onwards, all airline cockpit doors would be locked from the inside to prevent the kind of terror attack that — in reality — took place on 9/11. Some would doubtless have been mightily hacked off: the need to install secure doors would have cost airlines money and dampened returns to shareholders. And, by preventing the 9/11 attacks, the superforecaster’s — justified — actions would have looked to most of us like no more than vacuous scaremongering.

Some things just aren’t forecast for the simple reason that it’s easier to pretend they won’t happen

Now back to next month’s Budget. Pretend Rishi Sunak replaces the OBR with a team of superforecasters who, for reasons best known to themselves, conclude that the chances of a UK recession are, say, 80 per cent.

Sunak would now be faced with one of three awkward choices. He could accept we now are living in straitened times and cut his cloth accordingly, making him a deeply unpopular rookie Chancellor. He could decide to launch a massive fiscal stimulus in a bid to prevent recession (in which case, the superforecasters would prove to be “wrong” in the same way Taleb’s airline safety officer was “wrong”). Or he could simply replace the superforecasters with a process that is more politically “acceptable”.

And that’s ultimately the problem with economic forecasting, particularly in the public realm. Some things just aren’t forecast for the simple reason that, until they actually happen, it’s easier to pretend otherwise. In May 2008, midway between the failures of Northern Rock and Lehman Brothers, the Bank of England apparently regarded the risk of recession as very low — in hindsight a seemingly ridiculous conclusion.

It would be easy enough to believe that the Bank’s team of forecasters just wasn’t sufficiently super. It’s more likely, however, that the Bank’s senior players didn’t want to be dragged in front of the Treasury select committee in order to explain why they were in the business of talking the economy into recession. In economics and finance — unlike the weather — forecasts have consequences. Which is why some forecasts are never made.

Stephen King (@kingeconomist) is HSBC’s Senior Economic Adviser and author of Grave New World