Here’s an e-mail sent to me by Jerry Heavin, which I share, in full, with his kind permission (link added):
I argue regularly with my friends about free trade and they are often in the Krugman camp advocating for some redistribution to the “losers “. I counter if that is legitimate then the losers (really all consumers), when prices are higher due to lack of international competition, should be compensated as well. So the protected workers who have higher salaries should compensate my 82 year old mother who lives on a small social security check and is buying from Walmart and the Dollar Store and thus will be forced to pay higher prices, lowering her standard of living. Of course my sister and I will make sure her living standard is fine, maybe they should compensate us instead.
Jerry’s point is excellent. As I read it, his point is not quite the same as, or goes beyond, the point that Steve Landsburg famously makes that producers who enjoy artificially high incomes due to protectionist policies owe these ill-gotten gains to consumers who were forced to pay the higher prices that fund those gains. The significance of Jerry’s point is that it reveals the importance of appropriately identifying a baseline from which to judge a policy’s effect.
When Mr. Krugman insists that the “conventional case” for free trade rests on the assumption that government will* redistribute income from the ‘winners’ to the ‘losers’ – and, therefore, that the case for free trade does not apply in reality when government is unwilling to carry out such ‘redistribution’ – he implies that the pre-trade-liberalization incomes and economic positions of people within the country are the appropriate ones against which whatever changes caused by trade liberalization are to be judged. People whose incomes rise or whose economic positions otherwise improve as a direct result of trade liberalization are ‘winners,’ while people whose incomes fall or whose economic positions worsen as a direct result of trade liberalization are ‘losers.’
But why start the welfare analysis with the existing, state-imposed trade restrictions already in place? Such restrictions aren’t natural; they were consciously imposed sometime in the past. And when they were imposed they caused some people to lose even as they caused other people to ‘win.’ Was there, at the time these trade restrictions were imposed, state-engineered redistribution of incomes from these winners to the losers – a redistribution that ensured that the losers, after the redistribution occurred, were at least no worse off after the imposition of the restrictions than they were before the imposition? No. At least in all actual cases of which I’m aware (which are mainly those in the U.S. and in Britain), government has not attempted any such targeted redistribution. Moreover, I never hear in any discussions of pending trade restrictions today that such restrictions are justified only if the state stands ready to compensate the losers from the gains enjoyed by the winners.
Overlook the reality that, when trade barriers are raised, such redistribution is not only practically or politically impossible, it’s also theoretically impossible. The reason is that the total gains to the winners are, in such cases, less than are the total losses suffered by the losers. This reality is ethically irrelevant to Mr. Krugman’s recent argument against trade liberalization, for his recent argument rests on the ethical proposition that the state should not change trade policy whenever any such change will cause uncompensated losses to innocent people.
So if existing tariffs and other trade restrictions were imposed without any state-directed redistribution to those who lost as a result of such tariffs and restrictions, why is today’s distribution of incomes – the one that exists with such tariffs and restrictions in place – the appropriate baseline for making welfare assessments of changes in trade policy? I don’t see that it is an appropriate baseline. I don’t see any ethical (and certainly I see no economic) reason why the unlikelihood of state-directed redistribution should be a barrier to free trade if the unlikelihood (indeed, the impossibility) of state-directed redistribution is not, and never has been, a barrier to protectionism.
* In Mr. Krugman’s essay to which I refer – the one linked above – he writes that
the conventional case for trade liberalization relies on the assertion that the government could redistribute income to ensure that everyone wins — but we now have an ideology utterly opposed to such redistribution in full control of one party, and with blocking power against anything but a minor move in that direction by the other.
Notice the wording (emphasis added): “government could.” Someone might use this clever wording to say that all that Mr. Krugman is here doing is referencing, by implication, the famous Kaldor-Hicks welfare criterion  that says that an economic or policy change is welfare-enhancing if in principle the winners could compensate the losers and have, after such compensation occurs, at least some people still better off with no one being worse off than before the economic or policy change. (Importantly, the Kaldor-Hicks welfare criterion doesn’t require actual compensation of losers from the gains of winners.) But Mr. Krugman is clearly doing more than reminding readers that Kaldor-Hicks is an often-used welfare criterion for justifying trade liberalization, for in the very same sentence he complains that government today will not engage in such actual redistribution – thereby implying that, on his reading, the “conventional case for trade liberalization relies on the assertion that the government” will “redistribute income to ensure that everyone wins.
UPDATE: Fernando Tesón tells me on Facebook that Gary Becker  made the above point. I’m not surprised. I don’t, however, know where Becker did so. If anyone knows, please do tell me so that I can mention it and, if possible, also supply a link.