[pyrados]

This is an oversimplification. Density in fact makes public infrastructure more efficient and more predictable.

As Mason Gaffney notes in his article, "Containment Policies for Urban Sprawl"

"Consider water distribution. If demand doubles within a fixed service area by doubling density, we need simply expand all pipe diameters—and not by double, but by the square root of two, since cross-sections increase with the square of the radius. But if demand doubles by doubling the service area, at constant density, we must, (a) double our pipe mileage; (b) double the cross section of our old system at its base, and more than double it elsewhere, to transmit the extra load through to the new extension; (c) increase pressure at the system load center to maintain it at the fringes (especially if the new lands are higher); and (d) upgrade our pipe-joints to hold the extra pressure.

Actually those four simplest considerations understate the case a good deal. We should add the factor of peaking. The fewer customers on a given line, the higher is the usual ratio of peak demand to mean daily demand because there is less pooling of offsetting demand patterns, and more lawn sprinkling. There is also a factor of planning expansion.

"Containing urban sprawl" does not imply halting growth, but holding it inside compact increments, whose ultimate density is known in advance and will be reached quickly, saving utilities from the waste of under- or oversizing their lines in the face of uncertainty. Urban sprawl as known today not only reduces density but breeds extreme uncertainty of future density."