In Michigan, only one known but widely cited COCS study has been performed, focusing on Scio Township in Washtenaw County. Initially, the analysis was prepared as a graduate student research project in the School of Natural Resources and Environment at the University of Michigan.[60] The study was later published by the Potawatomi Land Trust.[61] Like the studies that served as its model, the University of Michigan study deconstructed local tax revenues and government expenditures based on land use.

The study found that for every local dollar raised in Scio Township, residential property required $1.40 in local government service expenditures.[62] Farmland, in contrast, raised enough revenue to offset public expenditures. Commercial and industrial land had the largest positive fiscal impact on the township: for every dollar raised in revenue, the township spent only 26 cents.

In addition to the standard limitations of COCS studies, the Scio Township analysis demonstrated additional flaws and weaknesses. The reader, for example, is left with the implication that agriculture (as well as commercial and industrial property) pays for itself and should be encouraged over other land uses, particularly residential land uses. In fact, based on this analysis, one could conclude that Scio Township should encourage agricultural development to balance the negative fiscal impacts of residential development.

However, agriculture represents only 1.4% of the township’s total revenues from general taxes on land uses (see Chart 10). Agriculture generated only $203,532 in revenue[63] while commercial and industrial property generated $5 million and residential property generated $9 million.

Farmland Revenues Are Small Compared to Commercial and Residential Uses

Farmland does not produce the revenue per acre that either residential or commercial development does. Take the following possible scenario. Scio Township consists of 22,000 acres.[64] In 1985, 36.4% of Scio Township’s land was devoted to farming. This proportion has probably fallen although more recent estimates were not provided in the study. (Countywide, the amount of land in agricultural uses is about 41%.[65]) Suppose, hypothetically, agriculture now makes up 30% of the township’s land use (a 17.5% drop from 1985). This suggests that about 6,600 acres would still be used for farms. This also implies that agricultural uses generate about $30.84 per acre in revenues. If we assume all the remaining acreage is devoted to commercial, industrial, and residential uses, these land uses would generate on average at least $900 per acre.[66] If farmland produced the same revenue per acre as urbanized uses, Scio Township would need just 226 acres to generate the revenues currently derived from agricultural land. Clearly, the township cannot rely on agricultural land uses to generate significant revenues for local government services.

If local policymakers were basing their land use policy on the tax and spending impacts estimated in the Scio Township COCS study, they would want to discourage residential development and encourage agricultural, commercial, and industrial uses. In addition, if land use decisions were made purely on net fiscal benefits, Scio Township should reserve all of its land for commercial and industrial uses. This would, of course, require extensive commuting by workers from other parts of the county or neighboring counties, which, in turn, would ironically encourage sprawl-like development patterns in neighboring townships and nearby counties.

Scio Township, like most communities, would want to encourage a mix of land uses. Local land use policy should be designed to facilitate the synergies among varied land uses, not impede them. Unfortunately, a COCS study provides little insight into how this would be achieved.

COCS studies also fail to shed light on the appropriateness of different types of residential development. The push to preserve farmland and limit residential development is driven in part by the belief that low-density residential development is inefficient compared to high-density residential development. This means that COCS studies must break the "residential land" category down further than the broad classification that is typically used. Because they often do not do this, COCS studies provide little useful information about the cost of public services in higher density land developments.

A final weakness of the Scio Township study was its approach to school funding and spending. School districts are independent governmental units and have independent taxing authority. District boundaries do not necessarily conform to township or municipal boundaries. As economist Gary Wolfram pointed out in his critique of the Scio Township COCS study, the authors’ inclusion of schools to assess the revenue impacts on township government was inappropriate.[67]

The implications of including education in the Scio Township analysis are significant. Excluding schools, residential land uses generated $1.7 million in revenue and incurred just $857,800 in expenditures. Its expenditure-to-revenue ratio was therefore calculated at 0.49, suggesting that the township’s residential land has a positive net fiscal impact. But when schools were included in the equation, the scales were tipped against residential development.

While the Scio Township study made adjustments for the statewide school finance reform brought about by Proposal A in 1994, it failed to recognize that education is provided through a different government authority with independent financing and provision requirements. The state of Michigan, rather than local school districts, is now the primary financier of public education, changing significantly the cost and revenue environment in which schools operate. The failure to consider this fact skewed the study’s findings.