On December 12 BEA released comprehensive data on GDP by county (as well as metropolitan areas) for the years 2001 to 2018. The emphasis in the release was on the real growth estimates for 2018. While the interest in such figures is understandable, it should be understood that they are very dependent on the methodology BEA uses to produce them. Economic Census data allow one to construct plausible estimates of nominal GDP in local areas for those years; at other times BEA has to do considerable interpolation and extrapolation, though QCEW data allows for good higher frequency data on employee compensation at the local level. Furthermore, to construct real estimates, BEA assumes that output and input prices for an industry are uniform across the nation. If one thinks about smaller counties which might be dominated by a few establishments in one industry or sector, it is not implausible to believe that the particular production technology used in one, and prices and costs, may be considerably different than in other localities. One important instance where one might scratch one’s head about the real growth estimates is the report that the real output of New York County, New York, edged down in 2018, due to a contraction in financial output. Given the difficulties of calculating real financial output one might question that finding.

As noted, nominal GDP numbers for counties may be a bit firmer than the real figures. Unsurprisingly, nominal output is heavily concentrated in a handful of counties. The four largest (Los Angeles, New York, Cook, and Harris) accounted for more than 10 percent of national GDP in 2018; the 31 (this list includes DC as a county) with nominal GDP estimated to be $100 billion or more accounted for nearly one-third of the nation’s output. In contrast, the combined GDP of the nearly 1000 counties reported to have had nominal output in 2018 less than $50 million (the smallest was the $17.5 million of Petroleum County, Montana) was only about 1 percent of the national total.