A monthly updated revenue forecasting model for the state of Michigan. (Volumes I and II)

The monthly updated revenue forecasting (MURF) equations developed in this dissertation provide the State of Michigan with an improved means of both tracking revenue performance and refining estimates of revenue over the two-year budget forecast horizon.; Prior to the development of MURF, the econometric input into revenue forecasts was based solely on the output of a quarterly econometric model, SQEM, run by the Research Seminar in Quantitative Economics at the University of Michigan. The MURF equations use the forecasts of economic variables from this model.; However, the MURF equations for the most important taxes have a monthly frequency. This has both econometric and administrative advantages. From an econometric point of view the use of monthly data greatly expands the number of observations on the independent variables and so improves the accuracy with which the relationship between revenues and economic variables can be measured.; From an administrative point of view, the use of monthly forecasting models allows for an earlier recognition of trends in revenues which should allow for a smoother functioning of the budget process.; The MURF tax equations are shown to out-perform the SQEM tax equations in the 1987-1988 period. In addition the MURF equations generally outperform both individual autoregressive moving-average (ARMA) equations and a vector autoregressive (VAR) model of a group of these taxes.; MURF also contains some entirely new features. For the first time an econometric model is used to forecast ticket sales for the state lottery. In addition, a model is developed to predict income tax refunds. These two innovations add about {dollar}1.5 billion to the scope of econometrically forecast General Fund-General Purpose and School Aid Fund Revenues.