Why economic models have never been able to predict a downturn | Letter

Dismissing instances of direct trade as outliers' that don't fit statistical theory is a major failing, writes Prof Scott Moss

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.