Development of forecasting models for predicting disaster relief costs

To aid in the control of Federal spending and budget deficits, the United States government must develop a methodology for accurately estimating yearly budget expenses and reflecting those estimates in fiscal year budgets. Disaster relief provided by the Federal Emergency Management Agency (FEMA) is a Federal program for which the government does an inadequate job of predicting yearly funding needs. A method of forecasting yearly disaster relief funding for FEMA is proposed based on the development of models to predict the number of disasters that will occur in a year, the cost of each disaster occurrence, and the percentage of disaster costs that will be obligated in the year the disaster occurs and in following years. Disaster relief models are defined for hurricane, flood, fire, tornado, winter storm, typhoon, and miscellaneous disaster categories. These models are then used to predict Federal disaster relief costs for the fiscal year 1993. The 1993 predictions are validated against real 1943 disaster relief data collected from FEMA. The proposed method of forecasting disaster relief proved to be a good tool for predicting the occurrence of disasters and the fiscal costs of disasters in the years that they occurred. However, the method was found to be a poor predictor of disaster costs in the years following the disaster occurrence.