The WSJ's "Numbers Guy" recently wrote about the problems with the "city livability" rankings published by various magazines and others. Readers love rankings so magazines and various groups will keep doing this. But urban economists model the spatial equilibria among cities as involving adjustments in housing (land) markets and labor markets (see Glaeser and Gotllieb's 2006 paper). The simple result involves a four-way partition: high- or low-amenity cities offer high nominal or high real wages depending on where housing prices end up. New York is a high-amenity place with high nominal wages, but low real wages because equilibrium housing costs are high. Were there fewer amenities (Houston?), spatial equilibrium indicates lower housing prices.
The model leaves out differences in land use regulation, but these too should be endogenous. The other problem is that the metro areas involved are too large, diverse and complex to easily fit the model. Averages mislead. There are usually wildly contrasting neighborhoods in each metro that defy the overall characterization. This is the major reason that ranking the places that are the home of often millions cannot easily be done.