Ever since the rise of rational expectations economics in the 1960s, the underlying assumption of rational expectations has both been defended for being realistic and been attacked for being unrealistic. The concept of rational expectations starts from the idea that individuals should not make systematic mistakes. Agents are not stupid, they learn from their mistakes, and draw intelligent inferences about the future from what is happening around them. While the adaptive expectations hypothesis had the disturbing implication that it allowed individuals to make systematic forecasting errors period after period, the rational expectations hypothesis asserted that people learned from their mistakes. It was based on the idea that guesses about the future must be correct on average if individuals are to remain satisfied with their mechanism of expectations formation. Rational people would take all available information into account and then discount that information into the future. If errors followed a pattern, they held information that could be used to make more accurate forecasts. The resulting predictions might still be wrong, but what mattered was that the errors would be random. People with rational expectations did still make mistakes, but not the same ones each time. Individuals could differ from one another in their expectations and still be rational if they were using different information. But when all these individual expectations were added together, errors tended to cancel out – producing an aggregate view of the future that reflected all the available information.

While some economists embraced rational expectations because it allowed a more realistic interpretation of expectations formation, others have argued that, despite the rational expectations school's insistence that expectations should be endogenous, it has dodged the crucial question of of how expectations are actually formed (Arrow 1978; DeCanio 1979; Friedman 1979; Pesaran and Smith 1992).