Redistribution occurs whenever services are not directly connected with the revenues which fund the services. It is an efficiency concern in the private sector and an equity concern in the public sector. To address the question of redistribution, three basic issues must be addressed. First, the units of study must be determined. In this case, the geographic boundaries of every city and township as well as two towns define the units of study. Second, the amount of funds which are produced by each unit must be approximated. Third, the actual consumption of services must be estimated. The 1996 budget, 1990 census data, 1994 census estimates, and various reports from the County Sheriff, State Police, Equalization department, and other sources serve as the basis for these estimates. The budget outlines all County programs and their funding streams. The other reports supply data for estimates of local service delivery and tax payments by locality. Together, these sources are used to estimate the sum of services used and taxes paid by residents of sub-county units.

    All revenues were divided by their general nature (tax, intergovernmental transfer, department generated, etc.) and then by the specific source of the revenue (property tax, income tax, etc.). Determinations were made to attach the revenue with the most applicable data set. For example, most intergovernmental transfers are generated by the federal and state income taxes and state sales tax, so these transfers are attached to the income characteristics of the local jurisdiction, according to the 1990 census. Since sales taxes can not be attached to a specific data set at the local level, the income tax is used as an estimate of the sales tax burden. The income characteristics of residents therefore accounts for all revenues from intergovernmental transfers. Revenues which are generated by the department are usually from user charges. Since information about specific use is not available, these revenues are assumed to be produced by the general population. Other revenues which could not be linked with a specific data set are also assumed to be distributed according to the general population. These include the hotel/motel tax and inter-fund transfers.

    Appropriate assumptions were also made regarding the consumption of services and formulae were set up to reflect how the benefits are distributed. For example, the Health Department aims many programs at the lower income and poverty populations. Therefore, the poverty rate from the 1990 census serves as the best estimate for distribution of these benefits. Similarly, the Sheriff’s activity report is used to help determine how the benefits from the Sheriff’s law enforcement services are distributed. However, not every service can be linked to a specific data source. This is the case with County departments that provide services related to governance, such as the Clerk and Register of Deeds. Services from these departments are assumed to be equally distributed throughout the population.

    Once all expenditures are linked to a specific data set, the total amounts are divided among the municipalities according to the data characteristic. The expenses are then added at the municipal level to get an estimate of the total expenses consumed by the residents of each municipality. Revenues produced by the residents of each municipality are determined in a similar fashion. The total expenses are then subtracted from the estimated amount that each jurisdiction pays in department-generated revenues and various taxes on property and income. This total is examined on a per capita basis to determine which localities pay more on average in revenues than they receive in benefits and vice versa. Due to the per capita methodology, a negative outcome indicates that the municipality in aggregate receives more services than is paid in revenues.

    Certain adjustments were made in the data sets to correct for the anomalous nature of two jurisdictions. The City of East Lansing is one municipality which was given special consideration. Michigan State University is within the city limits of East Lansing and the census data includes these students. The low-income student population skewed East Lansing’s population to 50,320 and the number in poverty to 11,211. Yet students who reside on the M.S.U. campus receive many services, including police and health care, from the university without paying property tax or user fees for County services. Off campus students also receive very few County services and are affected little by County policies for the same reasons. For these reasons, the population of East Lansing was lowered in this analysis to 29,865 and the number in poverty was reduced to 1,041 to reflect the population characteristics of East Lansing rather than the actual population. On campus students were omitted entirely and off campus students were counted as 3/4 of a county resident. Since the census does not sort out the student population, estimates provided by the City of East Lansing were used. Census data for Delhi Township was used as an estimate for determining the number of East Lansing residents in the 18-24 age category whose incomes are below the poverty line. Actual 1990 census data were used for the remaining age categories.

    Mason was also given special consideration. Mason’s reported crime activity was abnormally high due to the fact that 85 percent of all summons in Mason are made to inmates at the county jail. This necessitated a change in the crime statistics to more accurately reflect actual crime incidence and Mason’s crime rate is 85 percent lower than the reported rate. The Sheriff’s Activity Report was also adjusted in a similar manner.

    Other adjustments were made to the 1996 budget’s data. Most notably, not all revenues and expenditures are included in this model. Of the $105,566,818 allocated in 1996, only $80,003,731 is incorporated into this analysis - just 75 percent of the total budget. Excluded from consideration are revenues from Sheriff contracts with Delhi, Stockbridge, and Webberville because the contracts can be linked to the police services provided to these localities. The level of Sheriff activity in these jurisdictions are controlled to reflect the contracted amounts. Revenues and expenditures from the Fair Board have also been omitted because the services are funded through user fees, thereby directly linking the revenue source with the service received. Also excluded are the operations of the Medical Care Facility. Due to the funding source and the nature of the service, it is virtually impossible to link the benefits and funding to the municipal level. Finally, the services of the Road Commission are eliminated because the funding of road services is closely linked to road use through gasoline and weight taxes.

    In this process, a series of variables were created to simplify the analysis. These variables, which appear below, begin with an "r" to depict a County revenue or an "e" to depict a County expenditure. These variables show the total revenues and expenditures for a local jurisdiction that is attributable to a population characteristic, such as poverty or crime activity. Some expenditures originate from multiple sources and some budget lines fund programs which are designed to address different population characteristics. For example, the Health Department has programs which are designed to assist under-privileged (poverty) populations. But many other programs are designed to control disease, from which the entire population benefits. Since the actual proportions of these can not be known with certainty, an informed estimate is used. The operations of the County court system, jail and related agencies as well as FIA operations were set up in a similar fashion.  (See the Appendix for a more complete discussion of the assumptions made for the distribution of the variables.)

    The model projects three outcomes which serve different comparative purposes at the local level. A table of the model results appears in the Appendix. The first outcome, per capita revenues minus expenditures (column 1), compares the estimates of the total amount of county funds which are produced by a jurisdiction with the services received by the residents of that jurisdiction. This outcome is useful because it gauges the municipalities which benefit most from the County’s role in a federalist system of government. A locality which generates a significant amount of revenues through taxes on income, residential property, and commercial property but whose residents do not receive a significant number of services would be a net payor.

    The second, per capita property revenues minus expenditures (column 2), removes revenues from the income tax, hotel/motel tax, inter-fund transfers, and department-generated funds. This outcome is useful because it only includes taxes levied on property - a revenue for which the County has full jurisdictional control. Revenues from the 911 and special transportation millages, debt service on drains, and delinquent taxes are included here. However, since benefits are received by residents but revenues come from both residential and commercial property, this outcome does not accurately gauge individual tax burden. A locality which has a high level of commercial property produces more revenues than a municipality which has little commercial property. If this outcome is misinterpreted, the residents of such a municipality would appear to bear a heavier tax burden than they actually do.

    The third outcome, per capita residential property minus expenditures (column 3), is the most accurate measure of residential tax burden. It includes only the revenues produced by residential property and the expenditures made in all County services. This outcome would be of most interest to residents who want a measure of the average amount of services received by residents of their locality compared to the average amount of property taxes paid by residents.

    After the three outcomes were determined, a different method was taken to measure cross-subsidization in order to confirm the results. This new approach is frequently used in micro economics to determine the most effective departments within a firm. This approach - the ratio method - is also the method utilized by the federal government to measure cross-subsidization among the states and geographic regions. It is in no way superior to the above per capita method, it is simply a different method used by some to examine the phenomenon of cross-subsidization.

    The major change in methodology from the above approach is that population is removed as a factor, thereby removing the resident as the receptacle for services. Instead, the recipient of services is assumed to be the municipalities themselves. Similarly, the producer of funds are assumed to be the municipalities themselves and not individuals or property value. The municipality is portrayed as a mechanism which produces outcomes (revenues) in exchange for inputs (expenditures on services) in the same way that a firm might portray an employee as a mechanism which produces services in exchange for wages. The factors used to estimate the level of services received within a municipality are the same as in the above per capita method. However, instead of subtracting the expenditures from the revenues and dividing by the population, the revenues are simply divided by the population.

    This approach yields three outcomes which mirror the outcomes of the primary methodology. The first, expenditures over revenues, divides the estimated expenditures made in a municipality by the estimated revenues received by County from the municipality. A result of zero would indicate that the municipality contributes some revenues but receives nothing from the County. An outcome of one indicates that the municipality receives the same amount that is contributed. The second outcome, expenditures over SEV revenues, examines the rate of return for every dollar paid in property taxes. The third outcome, expenditures over residential SEV revenues, examines the ratio of expenditures made on residents to the property taxes paid by residents. In most instances, the results from the ratio method are similar to the results of the per capita method. The areas in which the two differ were examined more thoroughly.