A short term forecast model is developed for use by the Baltimore FICB in estimating their new money requests two months in advance. New money requests are the difference between new loans made and paydown. New money requests were forecasted by estimating two separate equations, one for new loans made and one for paydown. Ordinary least squares was used to estimate the parameters. Results of the estimation are reported. The forecast model developed in the study accounts for 83.59 percent of the monthly variation within the database in new money requests. Tests of the model include Theil's inequality coefficient, turning point analysis, and forecasting beyond the database. The results of these tests indicate that the model's estimates closely track actual new money requests. Due to the presence of multicollinearity, the individual effects of the independent variables are not identified. Multicollinearity was not a concern, however, since the model is to be used exclusively for forecasting. In addition to the statistical results, a computerized forecasting program is developed. The program can be used to predict new money requests for the Baltimore FICB two months in advance. It incorporates the results of the research into an easy to use package requiring a minimum of user supplied input. / Master of Science