The best explanation that I have seen is that the number of occupants per residence has declined. The number of single occupant tenants has rapidly increased as a share of residents.
This is a story that could reconcile rising rents and house prices in what Kevin Erdmann calls the “closed-access” cities with a loss of population in those same cities. Although young singles may double up to save on rent, this is probably swamped by the decline in households with children. And there is a general, long-term trend of much more housing space per person in the U.S.
My guess is this:
1. In open-access cities, the result is many more people and much more space per person.
2. In closed-access cities, the result is slightly fewer people and only slightly more space per person.
This is true. In places like Manhattan, over the past century, total floor space has actually increased quite a bit, but population has declined significantly.
More recently, as the post-modern urbanization wave has developed, in the 90s there was an increase in the intensity of urban residential utilization. Then, the credit boom allowed Americans to claim more square footage nationally in the 2000s. That meant more space elsewhere, but in closed access cities, mathematically, it means de-population.
A richer society might also mean more space per person. Hence, depopulated closed access cities