Budget planning and heuristic models

An important development in municipal (and other) budgeting is the use of budget simulations. These simulations receive as their inputs (1) proposed policy variables and (2) parameters that are assumed to be fixed, and they produce as their outputs pro-forma financial statements and other information useful in budgeting. The simulations contain mathematical descriptions of the components of the urban system and they use these descriptions to calculate the variables appearing in the output as a function of the policy variables and parameters.

Although budget simulations may be used to predict the financial state of an urban system. their principal purpose is to facilitate sensitivity analyses. The purpose of this paper is to present a method, called heuristic modelling, for performing sensitivity calculations with budget simulations. The sensitivity estimate is an estimated first derivative of an output of the simulation with respect to a decision input. It is useful because it allows a decision maker to estimate after each simulation, the consequences of a marginal change in the input. Heuristic models make it possible to perform sensitivity analyses after one or a few simulations, thus giving insights into the sensitivity properties of the components of the simulation, and reducing the number of simulations needed to arrive at appropriate budget decisions.