≡ Menu

Stolper-Samuelson Supplies No Good Justification for Protectionism

In my latest column for AIER I challenge the assertion – often made by people who know just enough economics to fool themselves and others into thinking that they actually know a great deal about economics – that the Stolper-Samuelson theorem offers a theoretically satisfying justification for protectionism. Here are my concluding paragraphs:

When the U.S., with its relative abundance of high-skilled workers, begins to trade more freely with countries that have a relative abundance of low-skilled workers, the wages of low-skilled American workers will today indeed fall relative to those of high-skilled American workers. And as Stolper and Samuelson show, if conditions are just right, these low-skilled workers’ wages might even fall absolutely.

But falling relative and absolute wages of American workers with only low skills — or, put differently, rising relative and absolute wages of American workers with high skills — itself incites more workers than previously to enhance their skills. Thus, what in the short run is seen as a justification for protectionism — namely, the propping up of wages of low-skilled workers — is in the long run seen to be a cost of protectionism. After all, no country grows prosperous by dampening its citizens’ incentives to improve their skills as workers.

Intensified incentives for workers to further improve their skills is one longer-run benefit of freer trade for economies, such as that of the U.S., in which low-skilled labor is relatively scarce. In fact, workers’ incentives to transform themselves from low-skilled to high-skilled employees are even stronger in high-wage economies if the Stolper-Samuelson theorem applies in reality and freer trade causes the wages of low-skilled workers to fall not only relative to the wages of high-skilled workers but also absolutely.

One general conclusion from these considerations is that finding even a single plausible justification for protectionism is practically impossible.