Alex Tabarrok has a very nice post on the morality of trade. You should read the whole thing. It is concise and artful. Here's a taste:

Peter and Jose [individuals who desire an exchange] presumptively are better off from trade otherwise they wouldn't trade so the individualist economist (the economist who takes Peter and Jose as the relevant moral community) will support free trade. The liberal internationalist will also support free trade because there is a strong argument from positive economics that trade increases total wealth (comparative advantage, specialization, competition etc.).

In between, we have the nationalist economist for whom it depends. The case for trade for the nationalist economist is pretty good - after all the individuals involved benefit and the world benefits - so the case is reasonably strong that Peter and Joe [a co-national of Peter] taken together will also benefit especially if we consider many trade pacts on some of which Joe benefits directly. Nevertheless, Rodrik is correct that when you exclude Jose it is possible to come up with examples where Joe's losses exceed Peter's gains.

I would argue, however, that economists are too quick to take the nation as the relevant moral community. It is quite possible, for example, for Peter to benefit from trade but for Peter's city to be harmed, for Peter's state to benefit but for his region to be harmed, for his country to benefit but for his continent to be harmed... Indeed, geography is not the only way we can define the moral community. Why not ask whether English speakers benefit from free trade or Christians or left handed people?

Here are a few responses:

1) How certain are we that "the world" benefits from free trade?

Alex writes, "The liberal internationalist will also support free trade because there is a strong argument from positive economics that trade increases total wealth (comparative advantage, specialization, competition etc.)." That may usually be true, but if the claim is that a liberal internationalist will always support free trade, it is false. The whole point of the current debate is to suggest caveats to that "strong argument from positive economics that trade increases total wealth". Those of us saying that "free trade" is not always good don't just mean that it is not always good for every country. There are circumstances in which it might reduce welfare for nearly everyone.

Here's an extreme example, but it does illustrate the point. Suppose a person who dislikes New York has enough money to purchase an atomic weapon and a delivery system. Suppose there is an actor (friend of dictator?) with the legal ability to sell a weapon, and who doesn't much mind what happens to New York. It is difficult to say precisely what the result to the world will be from this free exchange, but I think we can mostly agree that it will be bad, for almost everyone, although the two parties to the exchange might come out pleased. We, reasonably and uncontroversially, make an exception to "free trade" with respect to exceedingly dangerous weapons. But none of the standard models or strong arguments of economics addresses this case. There is no algorithm we can use to quantify, "yes, this trade is welfare increasing" or "no, this trade is bad" when, in the lingo, large externalities exist.

Externalities exist in more normal cases. Recent trade-skepticism in developed economies has largely to do with unbalanced trade. Persistently unbalanced trade involves creating debt, that is uncertain promises of providing goods and services in the future in exchange for goods and services today. Any trade involving debt creates the risk that the debt will not be satisfied, in which case the parties to the trade themselves (who may be more than two) lose in the aggregate. But more importantly, when debts are repudiated or means of payment devalued, there can be large negative externalities for parties uninvolved in the original exchange. Credit risk is much like pollution, and presents challenges to libertarianism and standard arguments for trade.

Standard arguments for trade are undermined by debt in yet another way. Most models of trade, from Ricardo on down, presume that economies will specialize in producing the tradable goods for which they have comparative advantage. We make cheese, they make chocolate, and the world benefits from more of both. Empirically, this assumption fails to hold, at least over the medium term. The "strong argument from positive economics" really has nothing, and certainly nothing good, to say about the case when they make chocolate and we make debt.

There are circumstances under which "the world" will not benefit from free trade, particularly credit-based free trade.

2) Are nation-states more relevant than other groupings as determinants of human welfare?

I dislike tribalisms, including nationalism. In Alex's story I'd be a "nationalist economist", but in reality I'm neither. Frankly, under this prickly skin I'm a utopian who'd prefer to see all borders disappear.

But we're not there yet. In this real world, doesn't the welfare of nations have more to do with well-being than ones city, handedness community, even religion or ethnicity? It sucks that Ohio is suffering economically, but the barriers preventing an Ohioans from moving to Silicon Valley are far less than those holding back a Bangladeshi. In the United States, there is not (and I hope there doesn't come to be) the level of ethnic cohesion and the sort of institutions that would allow for coercive policy at an ethnic level to be plausible and effective. If there were, we would have to deal with that.

Suppose left-handers and right-handers really did segregate, cohere, and belong to distinct political organizations with the power to compel compliance among members. Suppose the right-handers organized very carefully, and policed group behavior very strategically, in order to maximize the welfare of right-handers regardless of any effect on lefties. Would it necessarily be a good strategy for left-handers to ignore those facts, and pretend group identity doesn't matter? It might be, depending on the behavior of the cohesive righties. More probably it would not be. It is a game-theoretical matter that demands analysis, not an ideological or moralistic claim about what would be best.

Again, in my dreams, I would like to see nation-state politics become as hypothetical as handedness wars. But ignoring the actual relevance of nation-states, or failing to act strategically within some nation-states while others behave very strategically, is not a solution. Nation-states will not disappear just because we might will them to do so, and the interests of real people who live in nation-states will not be improved my merely pretending they don't matter.