— I'm with Tyler on this one:
So that is a significant reason why I am not a [modern] liberal. I prefer high growth, minimum domestic transfers, and a higher rate of immigration. Growth plus resource mobility is the best anti-poverty strategy we are likely to find. And this recipe is closer to classical liberalism than to modern liberalism. I might also add that the United States, through immigration, satisfies the Rawlsian formula better than does Western Europe.
Regarding the “Rawlsian formula,” the difference principle, it's worth noting that one of Rawls's many idealizing assumptions in Theory of Justice and Political Liberalism is a “closed society” assumption. That is, we are to think about the principles to govern the basic structure of society on the assumption that there is no immigration or emigration, or even trade between the citizens of separate states. So, on Rawls's own terms, immigration can't possibly satisfy the difference principle, since the difference principle is formulated to serve a state that, ex hypothesi, has no immigration. (And neither can international trade.)
That said, the “closed society” assumption is just wild, even as a helpful simplification. If the “fact of reasonable pluralism,” that is, the fact that under a liberal democracy the “burdens of judgment” will lead to a plurality of conceptions of the good, requires a major revamping of Rawls's system, then surely the “fact of national porousness,” that is, that fact that under any genuinely liberal state, people and goods will cross boundaries, requires a similar revamping–especially considering the fact that traffic of people and goods across boundaries tends to be good for everyone involved. Furthermore, if our natural talents are arbitrary from a moral point of view, and we do not strictly deserve the fruits of our contingent abilities, then surely state boundaries are arbitrary from a moral point of view, and citizens of states do not deserve benefits simply in virtue of having been born a citizen of one state rather than another. I don't think the closed society assumption can be justified either in substantive terms–it is not in reflective equilibrium with our considered judgments about the moral arbitrariness of borders, the importance of exit to justice, and the inevitability of international trade in liberal regimes–or in pragmatic theory-building terms–it seriously distorts more than it helpfully simplifies.
A better Rawlsianism would dispense with the closed society assumption. One very dumb way of trying to solve this problem is just to retain the assumption, but simply move the boundaries out until everyone on earth is included. The smart way to solve this problem is to recognize the moral importance of various levels of association, and work out a theory of justice assuming a structure of nested and overlapping jurisdictions. (Rawls's centralist/nationalist assumptions are also out of reflective equilibrium with our considered judgments about the moral and political importance of local governance and voluntary associations.) In this sort of cosmopolitan theory of justice, the nation state loses its preeminence as the subject of political philosophy, and becomes but one jurisdiction among others whose legitimate powers are a function of its role relative to other associations and jurisdictions in meeting the requirements of justice.