One thing that has always troubled me against the question of urban growth boundaries around cities. It seems to me that the regions where growth boundaries are most necessary and least likely to be harmful are precisely the regions where they are least likely to be tried.
How so? Because in a slow growth, Buffalo/Cleveland type region, the costs from growth boundaries are low. Due to slack real estate demand, it is unlikely that growth boundaries in such a region will lead to inflated
housing prices. And because the city may die without growth boundaries, the benefits of growth boundaries are high.
By contrast, in a growing region such as Portland or Seattle, there are enough affluent people for city and suburb alike, so even without growth boundaries the core city will be moderately prosperous (at least compared to most Rust Belt cities). Thus, the major benefit of growth boundaries(i.e. preserving the core key and its older suburbs as decent places to live*) are smaller in a growing region.
And where thousands of people are moving into the region every year, there is a fairly significant risk of growth boundaries causing housing price inflation. Thus, the costs of growth boundaries are higher in a Seattle or Portland than in a Buffalo or a Cleveland.
And yet it is precisely the fast growth regions where growth boundaries and similar experiments are most feasible politically. Why?
*See my article on growth boundaries (available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=816885 ) which points out that Portland's growth boundaries have been more successful in improving the core city than in solving other problems such as traffic congestion.
