His words, not mine (well, I am paraphrasing). But I do appreciate his willingness to reconsider. We should all be so forthcoming.
The question at issue may seem technical, but it is of considerable policy importance. It is a standard result in public economics that the efficiency cost of a small tax or tariff is small, but rises rapidly (with the square of the tax or tariff). A direct implication is that when tariff barriers are low--as they are currently (average tariffs in the U.S. are around 2 percent)--reducing them further does not create a whole lot of aggregate gains. But the losses suffered by those producers or workers who are adversely affected do not get correspondingly smaller. A 10% reduction in wages or prices hurts as much, regardless of what the initial level of tariffs is. Therefore, the ratio of distributional "costs" to net gains tends to rise as the remaining trade barriers get smaller. This is where I think we are in trade.
It is of course a much different ball game with labor mobility, since the existing barriers are so high to begin with.