Flawed assumptions, erroneous conclusions
Britain after Brexit: UK forecasters have got it wrong
by Brian Reading in London
Tue 6 Sep 2016
In early June the average forecast for UK GDP growth in 2017, according to a survey conducted by the UK Treasury, was 2.1%. The July post-referendum forecast cut this to 0.5%. In August the forecast bounced back a smidgeon to 0.7%. With half the year almost over before the 23 June referendum decision to leave the European Union, the revision to 2016 growth was modest, to 1.6% from 1.8%.
When year-on-year changes are converted to a quarterly path, in early August the crème de la crème of UK forecasters were still expecting a post-referendum recession. So far the hard evidence suggests they have got it wrong.
I am looking forward to seeing whether forecasters change their minds in September.
Since January 2012 the UK Treasury has published a monthly survey of around 37 UK independent forecasters. Details of all individual forecasts are provided – and not just for the UK. It is an economists’ and market analysts’ goldmine of who expected what and when. Yet very few people actually mine it!
I have looked at the impact of the unexpected referendum Leave vote on GDP growth forecasts.
The immediate conclusion is that this was an unprecedented downwards spike. From month to month, the average forecast for next year’s GDP rarely changes by more than 0.1 or 0.2 percentage points. Yet the June-July 2016 surveys showed the biggest recorded change. The May 2015 Conservative election victory had no observable impact, despite being equally unexpected. But between the June pre-referendum and July survey reports, the forecasts for 2017 were written down on average by no less than 1.6 percentage points.
There must be something terribly wrong with conventional forecast models. The new hard data inputs in early August could hardly justify such a dramatic change of view. Forecasters normally tend towards optimism. They generally lack the ability to predict cyclical turning points and crises. But, when taken by surprise, they swing violently to doom and gloom.
Forecasting models argue with impeccable logic from flawed assumptions to erroneous conclusions. They are based on regression equations that explain the past but cannot foretell the future. Apart from accounting identities to ensure consistency, every equation contains an error between the latest data and the regression equation fit. The forecaster must then decide whether the error will be larger, smaller or remain unchanged.
The forecasts have no more validity than the assumptions that forecasters feed into them. Every number that comes out is determined by a number they put in. When faced with an unprecedented event, likely to distort historical relationships, forecasts are simply guesses about an unknown future. Moreover, forecasters tend to move as sheep. One crosses the road and the rest follow.
Even worse, the impact of single events can never be known. Even when the future is known, it is still impossible to know what it would otherwise have been.
To claim in advance to know what Brexit causes is ridiculous. Putting precise numbers on it presupposes popular gullibility, especially when so-called experts agree. Project Fear – the effort by Remain campaigners to convince voters that quitting the EU would bring disaster – failed as a result of justified popular scepticism about whether these people could be trusted.
Economies have momentum. Financial markets go from bull to bear overnight. Economies go from boom to bust over months. But they are influenced by forecast-driven markets. It is early days yet to assess the post-Brexit future. We should watch how consensus forecasts evolve. My bet is that they will be revised upwards again pretty soon.
So, with apologies to TS Eliot’s Macavity: The Mystery Cat, I am guided by the following thoughts:
Brexit is the mystery cat, he’s called the guilty paw,
He’s the master criminal, much worse than all before.
The bafflement of HMT, Project Fear’s despair,
For when the future’s known, Brexit is not there.
Brian Reading was an Economic Adviser to Prime Minister Edward Heath and is a Member of the OMFIF Advisory Board.
Tell a friend