Aggregate expectations under the stable laws

Following Lucas, expectations have become central to macroeconomic theory. Empirical implementations generally start from tendency surveys, where respondents indicate the expected direction of change. Carlson and Parkin, and others, turn this into quantitative measures by assuming aggregate expectations followed a normal distribution. These show signs of irrationality. However, the Central Limit Theorem requires only that such a distribution lie in the class of ‘stable probability laws’. Indeed, the Lucas information assumptions, and evidence on individual survey responses, argue against the limiting case of normality. Experiments on European Business Surveys show that substitution of skewed stable distributions can eliminate symptoms of irrationality previously found in survey-based expectations measures.