Here’s a letter to a pro-free-trade correspondent – one of many who wrote to alert me to the extant objections to my recent New York Times op-ed on trade and job churn:
Mr. Dan Shevlin
Thanks. I’m indeed aware of David Autor’s  (and Kevin Drum’s ) objections to my recent New York Times op-ed  on trade and job losses. I’ll likely soon post a detailed response on my blog. But let me here offer a whiff of what will be a principal part of my response.
Autor and Drum insist that the 2.4 million American jobs that Autor and his co-authors  identify as being destroyed by trade with China – and that I mention in my op-ed – is a net figure and, hence, evidence against the validity of my point that we should not worry about the particular jobs destroyed by trade because ordinary job churn in America routinely destroys far more jobs than are destroyed by Americans’ trade with non-Americans. That is, Autor and Drum claim that, as a result of Americans’ trade with the Chinese, there are today in America, contrary to my argument, on net 2.4 million fewer jobs. But in fact the nationwide employment changes in America estimated by Autor, et al., are only inferred from industry and regional changes in employment. This inference, I believe, is unjustified. I don’t doubt that employment in U.S. industries and regions that are heavily exposed to Chinese imports is on net lower as a result of these imports. Yet contrary to Autor, et al.s’ claim, this fact implies neither a flaw in classic trade theory nor that overall U.S. employment is lower.
First, more U.S. imports allow Americans to export more as foreigners spend their additional dollar earnings here. But these additional exports need not come from the regions most affected by imports. Also, because imports reduce domestic prices and, thus, increase Americans’ disposable incomes, imports allow each American to buy more goods and services not only from nearby producers but also from producers in regions of America distant from his or her home. Finally, jobs in the U.S. are created also because more U.S. imports permit more foreign investment in the U.S. – and such foreign investment might well occur in regions other than those most exposed to imports. The resulting employment gains are ones that classic trade theory demands be counted against any employment losses caused by imports. Yet by focusing only on the industries and regions directly affected by Chinese imports (including industries upstream and downstream from the directly affected American firms), Autor and his co-authors miss these offsetting employment gains.* Being so dispersed and coming through countless different channels – many of which are several steps removed from trade with China – a large number of these created jobs are practically impossible to detect with econometric tools. This fact, however, does not make these jobs less real.
A closing observation to strengthen the above conclusion: the core point of my NYT op-ed is that trade plays no special role in destroying particular jobs. If Autor, et al.s’ findings can be validly extrapolated to conclude that job destruction caused today by trade does indeed reduce net employment nationwide, then it is likely also that job destruction from any other source should have the same negative consequence. As my op-ed reports, the number of particular jobs destroyed, on average, each month today in the U.S. is 1.75 million, only a small fraction of which are destroyed by imports. Therefore, if Autor, et al., are correct that labor-market imperfections and other forces are now at work causing particular job losses to translate into net, long-run job losses nationwide, it is a mystery why we do not see the much larger number of particular job losses caused by innovation and by changes in consumer preferences having nothing to do with trade causing far more massive net job losses nationwide. Yet we see no such thing – thus casting doubt on the claim that increased international trade somehow, alone among all the countless sources of job churn, results in net job loss.
Donald J. Boudreaux
Professor of Economics
Martha and Nelson Getchell Chair for the Study of Free Market Capitalism at the Mercatus Center
George Mason University
Fairfax, VA 22030
* Here’s a key passage from Autor, et al. (although their paper should be read in full): “The national industry estimates exclude reallocation and aggregate demand effects, whereas the CZ [commuting zones – that is, regional] estimates exclude the national component of these two effects, as well as nonlocal input-output linkages.” [page 228]