Is it rational to have rational expectations?

In economics in situations where there is uncertainty one has to attribute some attitude to handling this uncertainty to individuals. The original idea was to assume that "people do not make systematic mistakes" for which Muth coined the term "rational expectations". This was replaced by a much more formal vision which suggested that people fully understand how the economy evolves. In this paper I will argue that the foundations of the "rational expectations" hypothesis which has underpinned most recent modern macroeconomic models, which have, in turn, been used for policymaking, are unsound. From a philosophical point of view the idea that agents can not only forecast the evolution of their environment but also their own evolution seems unjustified. If agents do manage to coordinate on specific expectations these may not be rational and can be self fulfilling. From an econometric point of view, when there are changes in the evolution of the economy, it is not rational to satisfy the hypothesis. Furthermore the empirical and experimental evidence suggests that we should abandon this route. Copyright Springer-Verlag Berlin Heidelberg 2014