The Urban Revival Is (Probably) Not Over

A few weeks ago, Richard Florida wrote in the New York Times that the "urban tide has crested." In support of this claim, he cites the Brookings Institution's compilation of data discussing 2016 Census estimates.

But if you look more closely at the data (at Table 1 in the Brookings article), it tells a more modest tale. First, it shows that the overwhelming majority of central cities are still growing: in 50-plus metro areas examined, only 13 central cities lost population. If you had told your average scholar in 1980 that 80 percent of central cities would be gaining population a few decades later, he or she probably would have suggested psychiatric help. In the 20 largest metro areas, central cities declined in only three—Chicago, plus the always-declining St. Louis and Detroit. Even long-suffering Snow Belt cities like Cincinnati and Philadelphia gained population.

So where's the decline? For five years in a row, Census estimates claimed that central cities actually grew faster than their suburbs—a rather unusual result if true, given that many suburbs have lots of undeveloped land. But in 2016, major cities grew by 0.82 percent in 2015-16, while their suburbs grew by 0.89 percent. Around this trivial difference, Florida spins a tale of urban decline.  

Moreover, the entire argument spins around highly questionable data: yearly Census population estimates. But these estimates have not always been very reliable. For example, the 2010 Census count showed that Atlanta had 120,000 fewer people than claimed by the 2009 Census estimate—about a 20 percent difference. Thus, there is no reason to believe that this year's Census estimates are accurate enough to show whether any city's population increased or decreased over the last year. 

At this point, I am tempted to declare that the urban revival is here to stay. But I am reluctant to do so for one reason: 2010 Census results often showed that the 2009 Census estimates overestimated urban population. Although Atlanta's overestimate was an extreme case, most central cities were less populous than the 2009 estimates suggested. If this pattern continues in the 2020 Census,* we may find that some cities that mid-decade data says are growing were actually declining, and that some other cities are growing less rapidly than many people now believe.

*Assuming the 2020 Census itself receives enough money to be reliable- something that is a bit uncertain right now.