The question of public service cost and efficiency is an important one for debates over suburbanization. If low-density residential development were inefficient, an argument could be made for restricting it. But the empirical evidence on infrastructure costs is mixed. Even the SEMCOG study, which represents one of the more ambitious attempts to restrict development by imposing a particular urban form on Michigan residents, can claim only modest improvements over existing development trends. Moreover, the SEMCOG study is fundamentally flawed because it ignores the benefits of families living on larger lots.
The problem is that while some infrastructure costs decline as density increases, (e.g., street maintenance), other costs increase. Cities provide more than just one public service. As densities increase, cities tend to get larger and the level of general spending tends to rise (as do tax rates). Thus, while infrastructure costs may go down, administrative inefficiencies increase as cities get bigger and provide a broader array of non-infrastructure related programs such as housing and welfare. The net effect is an increase in general government costs.
The comments of Reid Ewing, a proponent of higher-density compact development, are worth repeating: "Having said all this, it turns out that density may not be the most important land-use variable after all. Density largely pays for itself, in the sense that developers pay for on-site infrastructure and successive property owners pay for public services through their property taxes."
As communities develop and commercial and industrial properties crop up around residential development, sufficient "cross-subsidization" of costs and benefits occurs to minimize negative fiscal impacts on local communities. In other words, surplus revenues from commercial and industrial land offset the fiscal drain of residential areas.