The objective of this paper is to determine upper bounds of the potential savings that can be realized by the application of cash management optimization models. These upper bounds are found by simulation as the difference between the performance of a deterministic optimization model--which finds the optimal policy in hindsight--and the simulated performance of a hypothetical treasurer who uses simple heuristic cash management rules as informally practiced by many treasurers, based on prediction of random cash flows. The results of this analysis leave serious doubts as to profitability of cash management optimization models.

The purpose of this paper is to construct a model for the computation of an optimal cash balance for a bank, although it could be adapted to any organization. By a bank we mean to include both commercial banks and savings banks (mutual savings banks and savings and loan associations). One might also be able to adapt the model to an “international bank” such as the United States holdings of gold and foreign exchange.