In my attempt to referee the conflict between Jim Manzi of the Manhattan Institute and National Affairs and Atul Gawande of Harvard Medical School and the New Yorker over what percentage of the American workforce worked on farms in 1900, one thing I have found out so far is that neither one of them went back to the actual census data. Mr. Manzi, who wrote that in 1900 "about one-third" of the labor force was in agriculture, tells me that he relied on a 2007 report from the Federal Reserve Bank of Cleveland. Dr. Gawande, who wrote that in 1900 farming accounted for "almost half the American workforce," sourced that claim, according to an email from his research assistant to me, to a 2005 paper from the US Department of Agriculture. I went back to the actual U.S. Census of 1900, and the special report "Occupations at the Twelfth Census," which indicates that in 1900, the percent of the workforce devoted to agricultural pursuits in 1900 was 35.6. The information is table XXI on page lxxxvi (page 92 of the 1062-page PDF.) Is 35.6% "about one-third," as Mr. Manzi would have it? Or is it "almost half," as Dr. Gawande would have it? Seems like a fine question for those vaunted fact-checkers at the New Yorker, or for readers here to decide for themselves and voice their opinions in the comments section. As I noted in the earlier post on this topic, this is more than just an obscure dispute over historical statistics, because the White House is using the New Yorker article to make the case that passing a health care overhaul will lead to cost reductions. If the big reductions in the farm workforce took place before the implementation of the government farm-productivity-increasing efforts touted by Dr. Gawande, it undercuts that argument for passing the health care bill.
Table XXI from the Report "Occupations at the Twelfth Census"