[C]onsider three counterfactuals.

Counterfactual #1 is we somehow freeze trade with China at 1990 levels, when Autor, Dorn, and Hanson's investigation starts. How is this done? Through strict import quotas or high tariffs? What are the deadweight losses associated with this, and what is China's response? Presumably they would make a similar move, and it seems likely this kind of behavior would set off a global trade war. Tell me: would those communities that were subsequently made worse off by more trade with China be better off if we started a global trade war? Let's say I'll grant that it's possible that some communities who had the most harm from increased trade with China might still be net beneficiaries from this shutdown of China trade, but it's far from clear.

And this isn't even the most appropriate counterfactual. Instead, consider Counterfactual #2 where we shut down all trade in 1990 and went to full autarky. After all, if our claim is that the local economies are hurt by trade then turn trade off entirely should benefit them. But you'd be hard pressed to find an economist who thinks the move to full autarky would make anyone better off on net. Which means we agree that even these harmed communities are still net beneficiaries of free trade.

Of course, the full impacts of free trade accumulate over time as countries specialize further. So to understand free trade's true net effect you'd have to examine Counterfactual #3, which would examine the effects of going back in time and having the U.S. institute autarky in full way back in the 1700s and staying there forever. Not even the most anti-trade economist in the world would claim that this would make any part of the U.S. better off today.