How accurate are the economists' predictions

Economic forecastsWhy economic forecasts are constantly wrong

In a crisis that no one could have foreseen, people long for a glimpse into the near future. The economic researchers provide it: The "economic wise men" forecast a 2.8 percentage point lower gross domestic product (GDP) for 2020, the leading economic research institutes a 4.2 percent drop in GDP. The Munich ifo institute expects up to minus 20 percent. But what added value do the many predictions have if they are imprecise and differ so greatly? After the crisis, most will inevitably not show up.

In fact, the forecasts of economists often go wrong, especially in times of crisis: The British Economist calculated for the years 2000 to 2017 that the most important economic forecasts in years of upswing were 0.6 percentage points off, in recession years 1.8 percentage points [1]. The statistician Nassim Taleb writes in his book The black swan even “we just can't predict”. For many laypeople, the economic forecasts are likely to appear almost arbitrary.

But that's not true. The forecasts are precise for a while and provide politicians and entrepreneurs with important information for their decisions. The different results are partly based on different assumptions and therefore even offer the readers added value, because they answer the question: What if?

In the current situation this means: what if the lockdown ended on April 19th? What if it lasted longer? The duration determines the added value of entire industries. The gastronomy, tourism and some other industries are currently not generating any sales. If the lockdown lasted longer than five weeks, their income would be all the longer. In other industries such as the auto industry, parts and workers for production are missing. Of course, the different assumptions lead to different forecast results.

Regardless of the economic situation, economists use indicators that give a good indication of future value creation today. For the frequently cited Ifo Business Climate Index, for example, the Ifo Institute asks thousands of companies about their plans and expectations for the next three or six months with regard to production, demand and the order situation. Purchasing managers' indices provide similar information about the production situation in the future. Various providers form these from surveys of purchasing managers and managing directors. They comment on the procurement of raw materials and preliminary products. If they buy less material, this is the first indication that economic activity will soon decline. In the past, the mood barometers were closely related to the actual development of GDP: If the values ​​of the indices fall, GDP usually falls in the following months too; if they rise, GDP rises. Economists therefore assume that the relationship will also apply in the future. The indicators are a standard instrument for almost all economic forecasts.

How is it then that different economic researchers come to such different forecasts despite similar assumptions about the lockdown and similar instruments?

One explanation is that assumptions change very quickly, especially in times of crisis. A forecast is always based on the status quo. However, if the government assumes higher unemployment based on a forecast and then decides to support companies with political measures such as state aid and short-time work benefits, this will also have an impact on future economic performance. So the economic researchers also have to change the forecast of GDP. Conditions change daily, especially in times of crisis.

This is why economic indicators do not have the same meaning in times of crisis as they do in more normal times. Too many factors can change in the short term, too many entrepreneurs change their minds. So there are fewer clues with which to compare the development than at normal times. Instead of relying on the statements of the models, the economic researchers have to interpret their statements more strongly and find new indicators with whose interpretation they have less experience. The result: the forecasts are more uncertain and their statements differ more widely.

So forecasts cannot be accurate; They cannot look far into the future either. The coronavirus pandemic could not be predicted using various economic indicators. But that is not the task of economic forecasts either. They are supposed to show how the economy would develop if everything stayed at the status quo and show scenarios of how things could turn out differently. So we can only hope that at the end of the year it will become apparent that many of the current forecasts were wrong.