Former steel worker Frank Behum speaks as he visits the former Bethlehem steel factory in Bethlehem, Pennsylvania on November 2, 2016.
In the heart of Pennsylvania, a key battleground state in the race for the US presidency, few people have yet to make their choice between Donald Trump and Hillary Clinton. Frank Behum, who spent 32 years working as an electrician for the giant Bethlehem Steel plant before retiring, said the choice was not ideal. Still, "You've got to pick the lesser of the two evils here."

 / AFP / EDUARDO MUNOZ ALVAREZ / TO GO WITH AFP STORY BY THOMAS URBAIN-"In Pennsylvania, few are undecided but many are turned off"        (Photo credit should read EDUARDO MUNOZ ALVAREZ/AFP/Getty Images)
© AFP

It has been a bumper week for economists writing about trade. Free Lunch on Thursday examined how much free trade has hurt US workers — the answer is that it has been good for most but bad for some. More strong analysis has been coming our way.

How does the US record on factory employment (chart below) stack up against Germany, typically held up as the gold standard of a manufacturing economy? Brad DeLong, in the essay we recommended on Thursday, argues that only the extent to which the manufacturing share of US employment fell more than in Germany (a small part of the total employment reduction) is excessive. So how did Germany fare?

Wolfgang Dauth, Sebastian Findeisen and Jens Südekum have just published a study of German manufacturing employment, which they contrast favourably with the US experience. While the 1990s were exceptional due to the job losses in the former East, in the 2000s Germany faced a similar global environment to the US. Their main claim is that China may have eroded US but not German manufacturing because of Germany’s trade surplus. That is plausible to a point — they document that German employment in export-facing manufacturing was stable throughout the 2000s. Industries exposed to import competition, however, saw large falls.

Is this really so different from the US? It’s true that almost 30 per cent of American factory jobs disappeared since about 2001, but the loss of one-tenth of German ones was also drastic, and equally bad for those thrown out of work. The same authors have previously documented that workers in import-competing manufacturing sectors do not find new jobs in the exporting sector. So while the magnitude of job loss is different, the overall picture is similar: concentrated losses that are not shared by a broader population made better off by trade openness.

Those concentrated losses can be large, as Dani Rodrik insists in a comment to DeLong’s essay. And in particular, Rodrik argues, they are much, much larger than the average gains from trade. Rodrik highlights two pieces of research on Nafta. One, by Lorenzo Caliendo and Fernando Parro, estimates the US welfare gain from freeing trade at just 0.08 per cent. The other (which we discussed on Thursday) estimates that high school dropouts in previously protected industries and living in places particularly dependent on them may have seen wages fall short by almost one-fifth of what would have been the case without Nafta. Rodrik’s point is that this is an awfully harsh loss for a marginal gain for the country at large.

Fair enough. But we should also compare like with like. If we turn to the magnitude of losses where they are concentrated, we should also look at the concentration of the productivity gains. If a 0.08 per cent gain sounds small (although it seems not to include any gains going to owners of capital, only workers’ price-adjusted income), it is because this is for the entire economy. But the effects of Nafta could only have come from the part of the economy that is traded with the US’s neighbours. That’s about one-third of total US trade, which in turn was only about one-tenth of the US economy before Nafta. So the 0.08 per cent gain derives entirely from at most a 3 per cent portion of the US economy. And it’s even more concentrated than that: Caliendo and Parro find that the bulk of gains came from the few industries that saw the biggest changes under Nafta, in particular electrical goods, textiles and auto production. These accounted for less than 20 per cent of US exports before Nafta — or less than 2 per cent of GDP.

What happened was that a few narrow manufacturing industries changed drastically under Nafta, and presumably when China’s trading role ballooned in the 2000s. These reaped large productivity gains, although they look small compared with the entire economy. In the process those who had benefited from protection suffered losses commensurate with the productivity improvement. Again, these were large, but small on average across a large economy.

At this point we should see that this is an instance of the old and fundamental issue in political economy, which Mancur Olsen analysed in the classic book The Logic of Collective Action. Many policies, in particular those that remove protection against market competition (whether foreign or domestic) give rise to concentrated losses but larger, though dispersed, gains. In most cases — certainly for domestic liberalisation, but also for issues such as pollution control — the approach political economists tend to take is to seek ways to address political and distributive challenges so as to build a consensus around realising the greater gains. What they rarely do is to say it’s not worth pursuing those gains in the first place.

Other readables

Numbers news

  • While UK consumers went on a spending spree after the Brexit referendum, eurozone consumers are squeezed by stagnant incomes.

To receive Martin Sandbu’s Free Lunch by email every workday, sign up here

Get alerts on US trade when a new story is published

Copyright The Financial Times Limited 2019. All rights reserved.
Reuse this content (opens in new window)

Follow the topics in this article