Why economic models have never been able to predict a downturn
Simon Jenkins (theguardian.com, 8 August) points to economists’ failure to forecast the 2008 financial crisis. He could have pointed to many more such forecasting failures, since there has never been a correct econometric forecast of a macroeconomic turning point. The question is, of course, why? Here is one answer.
The relevant macroeconomic theories all assume that no one ever trades or communicates directly with anyone else. What happens if people do trade directly, and generally communicate with one another? We know from a different kind of modelling that evidence-based assumptions of social interactions lead to episodes of volatility that cannot be forecast.
These models (which come under the rubric of complexity modelling) generate volatile outcomes of the sort classified by economists as “outliers”. These are dismissed by economists and economic forecasters as being one-off, externally generated events that are not part of normal economic processes. They have to dismiss these “outliers” because otherwise they cannot apply statistical theory to the data. The laws of statistical probability underlying econometrics are not relevant.
Economists claim that their theories and methods are tested by the accuracy of their predictions and not by the realism of their assumptions. In fact, more realistic assumptions about social and economic interaction lead to (and explain) the correct prediction that their models cannot forecast turning points such as the 2008 crash and the current economic crisis.
That is why economists have nothing useful to say about the current economic volatility, or specific proposals, such as those of Rishi Sunak and Liz Truss, to address the consequences of this volatility.Prof Scott MossChapel-en-le-Fr
Read more on theguardian.com
