Chris Johns: Forecasts provide a buffer against inconvenient uncertainty

Perhaps with our new-found awareness of the fragility of forecasts, we can use them in more productive ways

Tradition demands we say things “are particularly uncertain” as we head into a new year. It is one of those phrases that is logically untrue – unless uncertainty is something that simply grows through time.

While it may feel that way, particularly if you are an economic forecaster, uncertainty about the future is something that just ebbs and flows. Benjamin Franklin famously said that the only things that we can be certain about are death and taxes (although Daniel Defoe is credited with the first use of those words). What is less well known is that Franklin was referring to the US constitution: a warning that nobody should take its permanence for granted.

Now that we have a US president-elect who clearly has little respect for constitutional matters, Franklin’s 230-year-old warning may come back to haunt us.

Political risk is on the mind of every analyst and commentator: some are discovering it for the first time, others are reminded of what it was like when Russia, America and China slugged it out with multiple proxy wars and a seemingly never-ending arms race.

Perhaps it is better to say that things have not been this uncertain for quite a while.

For all of the mendacity, half-truths and outrageous lies of the Brexit Leave campaigners, they did do us one small service. Michael Gove’s claim that the British people have had enough of experts was a serious wake-up call for a ruling class grown complacent if not hubristic about its ability to shape and forecast the future.

Truth-telling (by either side of the campaign) would have been a much more nuanced and subtle affair. Both sides should have begun their manifesto with the words, “we aren’t terribly sure about what happens next but here are the reasons why we will think it will be great/awful. We also have to acknowledge that it could be many years before we can definitively judge whether Brexit has been a success/failure. It might even be the case that we will never know: nothing much might happen. It is always terribly difficult to compare actual economic outcomes with what might have been: the former is always certain, the latter is pure conjecture.”

Sever a limb

Of course, if politicians ran campaigns using that kind of language most voters would, sensibly, prefer to sever a limb using a blunt penknife than engage with the debate. Perhaps

Jack Nicholson

was right, we can’t handle the truth, particularly when it is fuzzy, messy and highly contingent.

Gove has spawned a debate about how some people prefer beliefs to facts. And how those two things cannot be allowed to infect each other.

Economics has always known about the importance of beliefs, the ways in which they can shape the future via self-fulfilling mechanisms . If enough of us believe the future will be awful, we will stop spending and will generate the recession that we believe is coming. If enough of us believe house prices can go up forever we will borrow from German, British and French banks and drive property prices up until enough of us change our beliefs.

Once the importance of beliefs is acknowledged in this way, forecasting the future becomes well nigh impossible. The only thing that we can say about what happens next is that “it all depends”.

But we demand forecasts. Governments and companies need them for obvious reasons. Perhaps with our new-found awareness of the fragility of forecasts, their contingency, we can use them in better and more productive ways.


The media might be of some assistance here. It makes no sense whatsoever simply to report, say, that some agency is forecasting that GDP will go up 3 per cent next year.

If you told your doctor that you are a smoker and want to know when and if you will die of your habit, he will first ask you some questions. He will want to know how many cigarettes you smoke and for how long. He will ask you about some other of your lifestyle choices. Your family history will enter the conversation. He will tell you there is a very good chance that smoking will kill you. It could happen in a number of ways, and you are almost certainly shortening your life. There is a small chance that you will, a long time in the future, die peacefully in your bed of old age. That’s about as certain as he can be.

Economic forecasting is exactly like that. The future depends on many things. GDP might rise by 3 per cent in in 2017, but it probably won’t. Like the doctor’s advice to the smoker, economic forecasts, done properly, can still be hugely valuable.

Brexit, hard, soft or flabby, will be fiendishly difficult to pull off. If it is to be a negotiated settlement, two years is not long enough. Britain could flounce out the negotiations in 2017 (a 50 per cent chance according to some diplomats) or fall off the 2019 cliff. Either way, there is a very high probability that either event will be a large negative shock for the UK economy.


Without an interim deal that puts a lot of stuff on the long finger, there is going to be trouble. At the heart of the process is a fundamental contradiction: Britain as a bastion of global free trade can’t then disappear behind tariff barriers to protect what’s left of manufacturing. And so on: the more we think this through, the more we realise how contingent Brexit is. We also see more risks than benefits: both do exist but in very unequal quantities.

It is hard to prepare for a contingent future full of risks. Our fiscal strategy amounts to spending (today) next year’s tax buoyancy. I would suggest two innovations: budgets should be based on conservative assumptions (say 1 or 2 per cent growth), not forecasts. Forecasters could then do something else.

Budgets should also contain scenarios: a plan for what the government will do when those assumptions turn out to be wrong.