Used for decades to predict the future of markets, economies and even fashion trends, has forecasting become nothing more than an exercise in futility?
For Andy Haldane, Chief Economist at the Bank of England, the fact that economists didn’t predict the 2008 financial crisis was a ‘Michael Fish moment’. In conjuring the failure of the TV weather presenter to anticipate the 1987 hurricane, Haldane admitted that the profession of economic forecasting was in trouble.
Fast forward to 2016, and that sentiment wasn’t helped by the Bank’s own prediction that a vote for the UK to leave the European Union in last year’s referendum would lead to a technical recession, which then duly failed to materialise.
Last year will go down in history for its unexpected events – the Brexit vote and the election of Donald Trump being just the most notable. So it’s hardly surprising that the credibility of forecasting has been put under the spotlight and – to some, at least – is in question as never before.
The World Economic Forum in Davos, usually the place where bold forecasts are made, this year held a session on ‘Forecasting failure’.
Forecasting is a multibillion-dollar industry. Businesses, governments and organisations of every kind are prepared to pay huge sums of money for information that helps them peer into an otherwise uncertain future, and to plan accordingly.
It’s hardly a new concept, of course. Back in the time of Ancient Greece, the Delphic Oracle was consulted about the future, while Romans examined the entrails of sacrificed animals.
In modern times, the Wall Street Crash of 1929 was correctly predicted by one financial pundit, Roger Babson, while another, the celebrated economist Irving Fisher, got it completely wrong. He infamously pronounced that “stock prices have reached what looks like a permanently high plateau”.
Economic number-crunching is arguably the most high-profile form of forecasting. Fortunes can be won and lost, and the prosperity of millions can hinge on knowing what’s going to happen to GDP, interest rates, tax receipts and a slew of other data. But it’s far from being the only type of crystal ball gazing that businesses are interested in.
Trends forecasters look at what’s going to be hot in fashion this season, and how consumer behaviours are likely to change. Market analysts forecast demand for products and services by sector and help companies to make investment decisions.
In the housing market, for example, forecasting focuses mainly on predicting movements in property prices. And scenario planners help companies play out the ‘what if’ consequences of events and understand the impact on their organisation.
All of the above share a desire to look forward, yet in terms of approach they’re quite different. And science, or logic, plays a part in varying degrees. At the heart of economic forecasting methodology is regression analysis – developing a model that attempts to describe the relationship between different variables.
The trouble is, according to Mark Cliffe, Chief Economist of ING Group, economists delude themselves that the resulting forecasts are precise and certain, when they’re anything but. Cliffe says: “Having constructed models to predict the future, forecasters tend to forget that they’re only approximations.”
Overconfidence gives an illusion of control, says Cliffe, while forecasts are typically flawed by other factors, such as biases and preconceptions, a reliance on what’s happened in the past, or a simplistic focus on a single factor.
And there’s a natural tendency to be influenced by what others are saying. “Most forecasters prefer to stick close to the consensus of their peers,” Cliffe points out. “They seek safety in numbers, figuring that it’s better to get it wrong in a crowd.”
Trend forecasting is a somewhat different cup of tea. It’s more a question of watching closely what’s happening and then following your intuition.
In a blog interview, fashion forecaster Geraldine Wharry explained: “Identifying a trend is a continuous effort of compiling observations. I call it ‘hunting and gathering’. Once I see there’s a flurry of images with one particular thread running through, it’s very easy to see this as a trend, although sometimes a single image can be so powerful that it triggers an instant conviction.”
Blue sky thinkers, sometimes known as futurologists, are another sub-species of the forecaster. Their focus tends to be on ideas – inventions that might happen – rather than hard and fast predictions about events.
Ian Pearson is a self-styled futurist/futurologist with a background in engineering, who used to work for BT and now specialises in the impact of technology through his consultancy Futurizon. Among his many inventions, he came up with the idea of text messaging in 1991, but he says BT wouldn’t back it.
One area forecasters can help with, he says, is identifying which product innovations will take off and which won’t – companies themselves often fail in this. “One recent example of a company getting it wrong was Google Glass,” says Pearson. “And I see the same thing happening with the Apple Watch.”
The human element
Reading the runes in this type of forecasting is seldom a case of developing mathematical models, but more a question of logical reasoning, Pearson says. “You need to be able to see the big picture, to understand the human factors involved, which can’t usually be reduced to an equation,” he says. “For example, people just don’t want to wear a watch they have to recharge every day.”
As for Brexit, he says economists got it wrong because they weren’t taking account of the societal factors, because they couldn’t write equations for those.
He says neural networks help computers to model complex issues better. “You end up with a large number of forces in play, and if a couple of those forces are slightly inaccurate, it doesn’t throw the whole thing awry – that’s how your brain works,” says Pearson.
The UK referendum vote in favour of leaving the European Union is a classic example of an outcome that few forecasters predicted, but which has serious implications for all businesses – especially in the financial sector.
EY carried out research among Channel Islands companies in the wake of the vote to find out how many had considered the impact of Brexit. Some 73 per cent said they’d carried out a review to assess how it had affected their business – the surprise was the number that had not.
“This may be symptomatic of the difficulties that businesses encounter when attempting to model the implications of the numerous possible scenarios that flow from the future UK departure from the EU,” says Stuart Gardner, EY’s Channel Islands Restructuring Director.
EY’s survey set out to encourage companies to undertake scenario planning. In financial forecasting, for example, individual business units are assigned revenue targets. “When you’re forecasting, you have to consider the sources of uncertainty, and which factors will impact those uncertainties,” says Gardner.
“Scenario planning comes in when you consider outliers across the distribution curve of expected outcomes, and what impacts they might have.”
What if one of the drivers for your revenue forecasts is taken away? “You consider how you can influence your drivers, which helps you reach your objectives,” he says.
This kind of forecasting has got much harder in today’s economic climate, says Gardner, with negative interest rates and the pound at its lowest level for years.
So are businesses over-reliant on forecasting in its many guises? Tidjane Thiam, Global CEO of Credit Suisse, told US TV channel CNBC: “I don’t like forecasting. I so deeply believe that the world is random, that management is about getting ready for a broad range of outcomes, rather than trying to predict the outcome. Sometimes it’s useful to predict the outcome, but the primary rule is to make your organisation resilient, so that it can cope with a number of outcomes.”
In any case, the writing may be on the wall for the individual star forecaster who seems to have a hot line to the future. The Good Judgment Project, initiated in 2011 by the University of Pennsylvania’s Philip Tetlock in response to a US government forecasting tournament, applies aggregation methods to turn individual forecasts into ‘the wisdom of the crowd’.
Participants are weighted according to their competency, and even trained in forecasting methods. The Project emerged as the clear victor in the tournament, building the case for a future in forecasting that relies on the aggregation of expertise to give a higher degree of confidence in its predictions.
This is just one response to the crisis forecasters face in an age of uncertainty. Artificial intelligence and ever-more sophisticated computer algorithms may also help to provide more accurate results in future – but it seems that for businesses, the best advice is to plan for everything.