Two-Stage Budgeting: A Difficult Problem

Utility maximization with a weakly separable utility function requires a consumer create an optimal budget for each separable subgroup. We show that computational complexity of optimal budgeting is the maximum of an exponential in the number of alternatives and a quadratic in the number of budget increments. From a budget survey of undergraduates we show that an undergraduate procedural consumer can obtain a budget estimate from the experience of previous students and can monitor the flow of funds and can make adjustments at a minuscule fraction of the calculations needed for optimal budgeting.