Reasons to be a Census skeptic

Over the past week, the press was chock full of 2020-style headlines like “Census Bureau Confirms Pandemic Exodus from SF.” That’s because according to the Census Bureau, virtually every urban county in the U.S. (even urban counties in growing metros like Dallas and Atlanta) lost population between July 2020 and July 2021. But is the hype justified?

I suspect not, for a variety of reasons. First of all, Census Department estimates have, in recent years, tended to underestimate urban populations, at least in some cities. For example, in 2019 the Census estimated Manhattan’s population as 1.628 million, while the actual count of 2020 showed 1.694 million residents- an underestimation of over 65,000 people. The Census estimated Brooklyn’s population at 2.559 million, but the actual count showed 2.736 million- an underestimate of over 150,000. (On the other hand, the 2020 population count was actually a bit lower than the 2019 estimates for Washington and San Francisco).

Second, even the 2020 Census probably undercounted cities more than it undercounted suburbs. How do we know this? Because according to the Census Bureau itself, it undercounted Blacks by 3 percent and Hispanics by 5 percent, while slightly overcounting whites. These groups tend to be more urban than suburban (at least compared to whites) – so if the Census undercounted these groups, it probably undercounted urban population generally.

Third, the timing of the Census Bureau’s estimates does not quite make sense to me. By July 2021, rents had already began to rise in Manhattan; the low rents of February and March were already disappearing. This suggests that by July, population (and thus demand) was increasing.

Fourth, even if the Census Bureau’s population estimates were valid for the summer of 2021, they certainly aren’t valid any more. How do we know? It seems pretty obvious that in New York City, rents have skyrocketed to pre-COVID levels and beyond. According to streeteasy.com, rents in New York City bottomed out in January 2021, reached pre-COVID levels in December 2021, and have continued to rise. If rent is rising, it seems likely that demand is rising as well.