Sloppy Statistics Make for Sloppy Analyses
Here’s a letter to the New York Times. (For alerting me to Atkinson’s op-ed, I thank Jack Nicastro.)
Editor:
Robert Atkinson’s case for devaluing the dollar as a means of protecting America’s economy from China is stuffed with fallacies both factual and theoretical (“We Are in an Industrial War. China Is Starting to Win.” January 9).
A key factual error is his claim that “China’s incursions … have cost America millions of manufacturing jobs.” The evidence that he cites in support is in a paper that he published in 2018, in which he wrote that “from 1965 to 2000, manufacturing employment was more or less stable, hovering between 19.5 and 16.5 million jobs. But the 2000s saw the number of manufacturing jobs plummet by more than 5 million.”
These statistics are as sloppy as statistics get, as a growing workforce – America’s more than doubled between 1965 and 2000 – renders the absolute number of manufacturing jobs economically meaningless. A more meaningful figure is manufacturing employment as a share of total nonfarm employment. Had Atkinson examined this latter figure he’d have found that the percentage of workers employed in manufacturing peaked in 1944 and ever since has steadily fallen. Indeed, its average annual rate of decline from 2001 through today (2023) was, contrary to Atkinson’s story, slower – at 1.9% – than from 1965 through 2000, when it averaged 2.1%.*
A key theoretical error is Atkinson’s contention that Beijing’s subsidies of Chinese industries that export are strengthening China’s economy at the expense of America’s. Not only do these subsidies distort resource allocation in China, rendering that economy less productive than it would otherwise be. Also, to the extent that Beijing arranges to bestow on Americans (and other people outside of China) goods at prices below cost, Beijing enriches us at China’s expense. How, pray tell, are we – rather than the Chinese people themselves – being economically harmed by Beijing’s practice of compelling the Chinese people to work for us at a discount?
Sincerely,
Donald J. Boudreaux
Professor of Economics
and
Martha and Nelson Getchell Chair for the Study of Free Market Capitalism at the Mercatus Center
George Mason University
Fairfax, VA 22030* Calculated by me. I divided FRED Data on annual manufacturing employment by annual total nonfarm employment and did the calculations from the resulting annual figures.
The post Sloppy Statistics Make for Sloppy Analyses appeared first on Cafe Hayek.
