Our Blog
Forecasting Error:  How ‘Expert’ Forecasters Fail

Lady Gaga

In a refreshing admission this week the Bank of England confessed “our models are just not that good”.  The Bank’s economic models lack accuracy and sophistication that allows them to predict the future direction for the UK economy.  This is something we highlighted two years ago in a piece entitled “Lunch with The Old Lady”, in which we cited the Bank’s own internal report of 2012 stating “the current forecast process at the Bank seems under-developed as a vehicle for delivering analysis”.

Clearly, as the country’s central bank, they need to improve.

Uncertainty Principle

That said, we have some sympathy for their plight.  As Danish physicist Niels Bohr observed, “Prediction is very difficult, especially if it’s about the future”. 

At Equilibrium, we do not expend resources on predicting things like currencies or commodity prices where we believe there is very little chance of attaining any insight into their future direction.  Even central banks try to manipulate currencies and get it wrong, so what hope do we have?

Bad Calls

Consultants spend a lot of time forecasting trends and we have often used their reports to get some guidance as to future projections.  We use them as a guide rather than gospel and they are as prone to error as anybody.

Back in the early 1980s, the US telecoms company AT&T hired leading consultants McKinsey to estimate how many mobile phones there would be in the world by the Millennium.  Back then, mobile phones were large, unwieldy and unreliable.  After some number crunching, McKinsey concluded that by the year 2000 there would be 900,000 mobile phones in the world.  The growth of the device, they concluded, would be restricted by size, weight and low-life batteries.  As a result, AT&T pulled out of the mobile market. 

It turned out that McKinsey were a little off in their calculations. The consultant's prediction of 900,000 subscribers was less than 1% of the actual figure of 109 million phones that were in use by the year 2000, and the market was increasing by 900,000 devices every three days.

Last year, Apple alone sold over 900,000 iPhones every two days.

Fast Forward

Of course, many would argue that technology has increased the rate of change, making forecasting harder to get right.  To a degree this is true, after all there is considerable debate at the moment about the potential impact of robotics and artificial intelligence on the economy, with forecasts of anything from mass unemployment to massive productivity growth.

Although the focus is on the future, it is interesting to reflect that we are already in a technology age.

In an interesting article with economist Susan Houseman in The Washington Post last year* she highlighted that the computer sector accounts for only 13% of US manufacturing.  However, output has been weak or declining in the industries that make up the other 87% of manufacturing and if you exclude the computer industry from the numbers, manufacturing output is only about 8% higher now than twenty years ago. Indeed, non-computer manufacturing is about 5% lower than before the Credit Crisis.

The Future’s Bright-ish

Hopefully, this positive contribution from technology bodes well for the future. 

At Equilibrium we adopt a cautiously optimistic view of the future.  We are sceptical about many of the forecasts we see – after all, if consultants had such a great crystal ball, why don’t they simply invest in all the areas they are predicting and earn super-normal profits?

With an emphasis on data and empirical evidence, we focus on areas where we can add value and use scenarios to test our assumptions – we need to know that if we (like everyone else) get it wrong, that our clients will still be in a good position to enjoy a bright future.

 

*https://www.washingtonpost.com/posteverything/wp/2016/10/18/dont-blame-the-robots-an-interview-on-manufacturing-automation-and-globalization-with-susan-houseman/?utm_term=.6aaef8b180a1