Sarah's insight is correct. A New York Times article from March 6 of this year describes work by Glaeser and Gyourko:
"In 2000, Glaeser took a sabbatical from Harvard and began to spend a few days a week in Philadelphia working with Joseph Gyourko, a real-estate economist at the Wharton School of the University of Pennsylvania. Glaeser had already been thinking about the relationship between housing and urban poverty when one day he and Gyourko began to discuss why cities like Philadelphia and Detroit — places with poor future prospects, both economists believed — weren't doing even worse in terms of population. Why didn't everyone leave, Gyourko wondered, and go to a place like Charlotte, N.C., that had a fast-growing economy? This question addresses a puzzle of urban economics. Cities (think of Las Vegas or Phoenix) can grow at a very fast rate, exploding overnight with businesses and residents. Some can increase in population by 50 or even 60 percent in a decade. But cities lose their residents very slowly and almost never at a pace of more than 10 percent in a decade. What's more, when cities grow, they expand significantly in population, but housing prices tend to rise slowly; even as Las Vegas grew by leaps and bounds in the 1990's, for instance, the average home there cost well under $200,000. When cities decline, however, the trends get flipped around. Population diminishes slowly, but housing prices tend to drop markedly.
Glaeser and Gyourko determined that the durable nature of housing itself explains this phenomenon. People can flee, but houses can take a century or more to finally fall to pieces. "These places still exist," Glaeser says of Detroit and St. Louis, "because the housing is permanent. And if you want to understand why they're poor, it's actually also in part because the housing is permanent." For Glaeser, this is the story not only of these two places but also of Buffalo, Baltimore, Cleveland, Philadelphia and Pittsburgh — the powerhouse cities of America in 1950 that consistently and inexorably lost population over the next 50 years. It is not just that there were poor people and the jobs left and the poor people were stuck there. "Thousands of poor come to Detroit each year and live in places that are cheaper than any other place to live in part because they've got durable housing still around," Glaeser says. The net population of Detroit usually decreases each year, in other words, but the city still attracts plenty of people drawn by its extreme affordability. As Gyourko points out, in the year 2000 the median house price in Philadelphia was $59,700; in Detroit, it was $63,600. Those prices are well below the actual construction costs of the homes. "To build them new, it would cost at least $80,000," Gyourko says, "so there's no builder who would build those today. And as long as those houses remain, the people remain."
The resulting paper, "Urban Decline and Durable Housing," caused a stir among urban economists even before its publication last year. (It was initially circulated with a subtitle along the lines of "Why Does Anyone Still Live in Detroit?" until the authors, thinking it politically insensitive, removed it.) In addition to illuminating some of the forces shaping our poorest cities, the research proved to Glaeser that it is impossible to think about urban economies without thinking of urban buildings at the same time. "
Indeed, old affordable housing almost certainly kept New Orleans'population from declining more rapidly than it did. The Glaeser-Gyourko work actually is part of a long tradition of urban economics work on "filtering." Ed Olsen at the University of Virginia wrote a paper in the early 1970s called "A Competitive Theory of the Housing Market," that predicted that in general the poor would live in the center of cities, for the simple reason that they would live where the old housing was. The poor actually tend to pay more for land rent than higher income people do, because (1) they can't afford to pay for transportation, and thus need to live near jobs and services and (2) because the old housing is on the most valuable (centrally located) real estate (it doesn't look valuable because the housing stock is so poor).
This phenomenon also explains why New Orleans population will likely not return to anything like what it was a little over a year ago. The "affordable" housing cannot be replicated, and so a principal reason many people stayed in New Orleans is gone forever.