For decades following World War II, America’s urban crisis was one of decline and population loss—problems that persist in some US cities. But the 1990 Census showed that an important set of cities had begun to gain population over the previous decade (and some neighborhoods had begun to attract new residents even earlier). Crisis became renaissance, and these cities came to experience an entirely different set of problems. Today, we observe the divergent fates of American cities: some are becoming extremely costly while others continue to struggle with the problems of abandonment; some grow at a rapid pace while others resist new development. Broadly speaking, we can classify US cities into three types in terms of their housing cost dynamics. First, some cities continue to have shrinking populations, so the existing supply of housing is large compared to the quantity demanded and housing is often quite inexpensive.

Housing markets and labor markets—conceptually, the same thing in most cases—exist at the scale of metropolitan regions. Because people within a metropolitan area can easily live in one city but work in another, it’s not possible to bring down the cost of housing in one city without bringing it down in the metro region. But as we will see, the decisions that affect housing costs are not made at the metropolitan scale, they are made at the scale of individual cities. So, it is usually correct to speak about the housing policy choices of cities, even when the outcomes of those policy choices will be manifest at the metropolitan scale. I will try to be clear about scale throughout this discussion.