Mutually Consistent Revealed Preference Demand Predictions

Revealed preference restrictions are increasingly used to predict demand behavior at new budgets of interest and as shape restrictions in nonparametric estimation exercises. However, the restrictions imposed are not sufficient for rationality when predictions are made at multiple budgets. I highlight the non-convexities in the set of predictions that arise when making multiple predictions. I develop a mixed integer programming characterization of the problem that can be used to impose rationality on multiple predictions. The approach is applied to the UK Family Expenditure Survey to recover rational demand predictions with substantially reduced computational resources compared to known alternatives.