Short-term, population-based forecasting in the public sector

There has recently been a tremendous expansion of the range of problems to which the demographic perspective is applied. Development of a new population-based method to solve the problem of forecasting income assistance caseloads for the state of Washington represents yet another effort in which the demographic perspective helps solve two major public-policy problems: (1) providing accurate and useful forecasts of caseloads, and (2) creating a dynamic model with which to analyze alternative policy proposals. When forecasting or examining the caseload history, it is also common to look at these caseload levels as a time-series. A caseload grows and shrinks as time passes because new members enter the caseload from a population of potential clients while other members exit the caseload. Population-based forecasting, as reported here, is really quite a novel approach to forecasting public assistance caseloads. In most situations, simple extrapolations of past trends or econometric time-series models are used. Characteristics associated with entries and exits can be used to develop dynamic models of current and future caseload changes. For budgeting purposes, these models can be readily translated into average annual caseload levels and can be directly used to examine policy alternatives and programmatic options. Entry and exit rates and volumes can be related to historical, current, and anticipated changes in economic, social, and programmatic conditions to develop models of caseload behavior, and ultimately, forecasts of caseload levels that are used for budget development.