The Behavioral Implications of Aggregation and Delay on Budget Approval Decisions

An experiment with three treatments is used to evaluate the social efficiency of budgeting protocols induced by other-regarding preferences. In disaggregated superiors receive three individual proposals in iterated fashion, and must decide whether to accept each project without knowledge of the forthcoming proposals by the subordinate. In aggregated three projects are bundled together, and subordinates submit an aggregate proposal that is intended to cover the cost of all the projects combined. Superiors can accept all of the projects to which the proposal pertains, or none of them. In delayed subordinates submit individual proposals as in disaggregated, but superiors delay making a decision on any of them until they receive all three proposals. The results indicate that superiors are slightly more demanding in aggregated than in the other two treatments, and subordinates submit lower proposals for high-cost projects. However, superiors are not so demanding that the social benefit of aggregation is completely undone: the likelihood of project acceptance under aggregated is significantly greater than under disaggregated and delayed, especially for the highest cost projects. It therefore appears that the aggregation of budget requests into a single proposal to cover all costs is somewhat efficacious in increasing project acceptance. Interestingly, only when the superior is forced to consider an aggregated budget is social welfare improved. In this sense, aggregation may be thought of as a substitute for a commitment to view only the total budget proposed, and not its components. More generally our results provide a rationale for restricting the level of information in the evaluation of subordinates when superiors cannot commit to how they will use the information.