The credibility of economic predictive analysis

Algemeen Dagblad
Erasmus School of Economics

Economic forecasts have been subject to change multiple times in the course of 2020. Even though the accuracy of these models is questionable at times, Professor of Public Economics Bas Jacobs argues that these models are undoubtedly of added value.

Economic decline seems to be one of the most important themes of 2020: many economists and institutes have continuously been analysing the economic climate to grasp the effects of the global pandemic. For instance, Rabobank estimated the shrinkage of GDP in March to be 0.2 percent over this year. In October, the estimates are as big as five percent. However, ABN Amro predicts decline to be 3.5 percent and the CPB Netherlands Bureau for Economic Policy Analysis gives a confidence interval ranging between 1.2 and 7.7 percent. One thing can be concluded: the range of predictions is wide.

Limited explanatory power

One of the causes of limited explanatory power of many economic models, is because some phenomena are not internalised. Elwin de Groot, euro area economist at Rabobank, gives an explanation: the roots of the global pandemic lie in China. When the virus caused a part of China to go into lockdown, factories had to close their doors. Then a chain reaction was set in motion: the virus spread across the globe. This caused a supply shock, which impacted our economy. However, the model of Rabobank is mainly based upon changes in demand. This resulted in an inaccurate prediction.

People might ask themselves: why do institutes keep updating their models when the outcome of their different scenarios is uncertain? Jacobs establishes that estimations of economic growth are always uncertain; however, it is still beneficiary to use these estimates to determine which reactions are adequate. During crises, the need for direction is all the more present. He adds that it is almost impossible to predict the outcome of the pandemic on the economy. There are many uncertain variables, such as the reactions of households, companies and the government to events concerning the virus.

Economists have to compromise between alternating their forecasts and showing forecasts that are out-of-date. Both can damage their credibility: constantly changing predictions may seem as if the forecast is not certain or credible, but persisting that the weather will be very hot when it is freezing won’t add to your credibility either. However, discussing different scenario’s contributes to an understanding of the situation and is thus beneficial. Not giving any forecasts isn’t an option either, De Groot points out. Many parties base their choices upon predictions, so demand persists to exist.

Long-term forecasts

Strict measures damage consumers’ trust and thereby decrease the amount of expenditures. However, in the long-term strict measures empower the economy to recover at a faster rate. It’s important to keep long-term developments in mind as well when interpreting different scenarios. In times like these, Jacobs states it’s of great importance that we learn to live with bigger insecurities. Adapting models to rare phenomena such as a pandemic is unsustainable and therewith undesirable. Requiring economists to make accurate forecasts at all times is simply not realistic.

More information

The full article can be downloaded above.