Sofa: A simulation program for predicting and evaluating the policy effects of grants-in-aid

This paper motivates and describes a computer simulation program to predict the general equilibrium effects of grants-in-aid to local governments on the tax and spending decisions of these governments. The SOFA (Simulation of Fiscal Assistance) program explicitly incorporates the effects of family relocation (the "Tiebout" adjustment process) and the availability of competitive private market providers (e.g. private schools, hospitals, recreation facilities) on the levels and distribution of local public services. SOFA can simulate the effects of the four main forms of grants-in-aid: (1) lump-sum transfers, (2) matching aid, (3) base equalizing aid, and (4) tax relief aid. In addition to the individual city predictions on taxes and spending SOFA provides the following summary statistics: (1) mean level of spending, (2) coefficient of variation of spending, (3) values for a utilitarian, Rawlsian, and egalitarian objective functions, (4) the central government tax rate needed to finance aid, and (5) the percent of families in the region who prefer the private provision of the "public" service. A copy of the program and a program users manual containing a detailed derivation of the model as well as user's instructions are available from the authors upon request.