A year of shocking forecasting inaccuracy

clouds0315 026 50

There have been huge forecasting errors in 2020.  While this year may be an extreme case, we are experiencing very poor accuracy about one year in ten.  Even in other years, forecasting accuracy can be poor.  How then should business and government plan?

Forecasts for the number of deaths attributable to COVID-19 were very pessimistic.

Australia’s Deputy Chief Medical Officer Paul Kelly said that the number of infections would be in the range 20% to 60% of the population.  Deaths would range from 50,000 to 150,000 (The Age 17 March 2020).

Australian economist Warwick McKibbin  estimated that almost 100,000 Australians could die from COVID-19 (the range is 21,000 to 96,000).  This is based on modelling seven different scenarios, building on the experience of the SARS outbreak in 2003 and Spanish Flu in 1918 (Australian Financial Review, 3 March 2020).

As at 16 April, the number of deaths attributable to COVID-19 is 63 and the daily number of new deaths is declining.  The number of new confirmed cases is declining, despite increased testing.  On current trends, and assuming continued adherence to restrictions such as social distancing, there will be no active cases by the end of May.

The lowest of the predicted number of deaths, 21,000, is likely to be at least 210 times too high.  The highest of the predicted number of deaths, 150,000, is likely to be 1,500 times too high

These are dreadfully inaccurate forecasts.  They may not have allowed for how Australia would respond to the pandemic – if so, they should have stated this assumption (or the reporters should have asked about the assumptions).

Economic growth forecasts were not much better.

On February 6 2020, the Reserve Bank of Australia’s forecasts were for GDP growth of 2.75% and an unemployment rate of 5.0%.  The 90% confidence interval on their GDP growth forecast included +5%, but not 0% or lower!

A survey of economic forecasters by The Age, published on 1 February 2020, produced an average predicted GDP growth of 2.18% (range 1.0 to 3.3) and an average predicted unemployment rate of  5.23% (range  4.8 to 6.4).

On March 18, the governor of the Reserve Bank said “While it was not possible to provide an updated set of forecasts for the economy given the fluidity of the situation, it was likely that Australia would experience a very material contraction in economic activity, which would spread across the March and June quarters and potentially longer. The size of the fall in economic activity would depend on the extent of the social distancing requirements, and potential lockdowns, put in place to contain the virus”.

Australia will almost certainly be in recession and the unemployment rate is likely to exceed 10%.

We only have to think back to 2008 and 2009 for another example of very poor economic forecasting.  In 2008, the Reserve Bank was lifting interest rates until March, seemingly oblivious to the imminent global financial crisis.  By September 2008 they were cutting rates as fast as they could.  In May 2009, the federal budget predicted a recession.  So too did the Reserve Bank and 99% of economic forecasters.  The recession never arrived!

Economic forecasting seems to experience very bad economic forecasting errors at least once every decade.

With frequent large forecasting errors and an increasingly turbulent world, how can planners prepare for the future?

They could look at risk assessments, such as that produced by the World Economic Forum (WEF) at the start of each year – it is an assessment of long term risks.  At the start of 2020, the WEF Global Risks Report, prepared in partnership with Marsh & McLennan and Zurich Insurance Group, rated a set of 30 risks on the basis of impact and likelihood.  Infectious disease was rated 27th out of 30 on the likelihood scale and 9th on impact.  No doubt many planners would not include infectious disease at the centre of their plans.  The WEF most likely risk was extreme weather followed by climate action failure.  The highest rated risk in terms of impact was climate action failure.

Within weeks of the release of this report, an infectious disease has had devastating impacts on lives and economies with long term repercussions.  Events which are considered unlikely do happen!  Accordingly, there is a need for a well-rehearsed plan ready for implementation to minimise the impacts of such events.

Scenarios are a valuable input to developing robust plans in the face of uncertainty.  Scenarios are plausible futures, based on a range of assumptions.  While scenarios can map out an envelope of future possibilities, they also provide a basis for developing plans which are resilient across all plausible futures.

A frequent refrain from some business leaders is a call for certainty, especially from governments.  There is no such thing as certainty, as shown by the global financial crisis and the COVID-19 pandemic!  Even government policy must change in the face of uncertainty

Tracking surveys of Australian consumer expectations about the year ahead, conducted by foreseechange, showed that in November 2019 the average estimated likelihood of a major disease outbreak in the year ahead was 42% (the Wisdom of the Masses).  This is a perceived risk that people would expect governments to manage.  The Wisdom of the Masses also indicated economic downturn and a rise in unemployment.

The general public has information that is important in the development of scenarios.  The Wisdom of the Masses should complement information from experts in the development of scenarios.

Future posts will provide more information on scenarios and the Wisdom of the Masses.

Charlie Nelson