In relation to my previous post on trade adjustment assistance, a quick quote from knzn on efficiency of trade. He notes:

The traditional Ricardian theory of gains from trade applies in the case of certainty: those who benefit from liberalized trade will have sufficient benefits that they could (theoretically) compensate the losers and still end up better off, so trade is Kaldor-Hicks efficient ex post. But in a world with uncertainty (and with incomplete insurance markets), liberalization of trade can be Pareto inefficient ex ante.

In normal-speak that is: just because trade can make everyone better off if things are efficient doesn’t mean that it will since our world is often inefficient. An important point.

His full post is here, including a link to the paper that brought about his post.