If a government wants promote some value, it could do so directly or indirectly. The distinction is best illustrated by an analogy. Suppose you think government’s one and only job is to promote good art. (No one thinks that, I know.) This alone doesn’t tell you what government should do. Setting goals doesn’t automatically tell you what the best means are to achieve those goals.
A government might attempt to promote art directly, by creating art schools, offering subsidies and prizes to artists and museums, and buying art, or indirectly, by providing a basic institutional framework (such as the rule of law, a well-functioning property rights regime, open borders and other institutions that encourage the exchange of ideas) that in turn induce people to produce good art. If you care about art, it’s still an open question whether a direct, indirect, or mixed strategy best produces it.
The same goes for social justice and other welfarist principles. Let’s say you believe that the basic structure of society ought to be arranged such that everyone (as much as possible) has enough to lead a fully human life. That doesn’t automatically commit you to any particular set of institutions. Under any particular set of conditions, it might turn out, as an empirical matter, that a extensive, heavily redistributive welfare state realizes that principle. Or might turn out that this welfare state does not, and regime that maximizes growth realizes the principle instead.
Some libertarians (though not all) object to the idea of social justice because they misunderstand what it means to live by a principle of justice. They think that a commitment to social justice logically entails a commitment to socialism. Because they are hostile to socialism, they are thus hostile to social justice. But they’re wrong.
For example, Rawls’s Difference Principle holds that inequalities in people’s holdings should be arranged such that they are to the maximum benefit of the representative member of the least advantaged (contributing) group. If you advocate the Difference Principle, this doesn’t thereby commit you to believing that every government ought to have a Department/Bureau/Ministry of the Difference Principle. (It doesn’t commit you to believing there should be a government at all, actually.) It doesn’t commit you to holding that a society should be engaged in continual, heavy redistributive taxation and transfer schemes.
What role the Difference Principles calls for government to play is an open, partly empirical question. The Difference Principle is not meant to be a law, or even something enshrined in a political constitution. It is instead a principle by which you assess what the constitution of a society should be. It might turn out that the Difference Principle could be satisfied in an anarcho-capitalist or anarcho-communist society. It might turn out that the principle requires heavy redistribution, heavy economic regulation, and a social democratic state. It might turn out that the principle can be satisfied any number of ways. Rawls himself favored a regime that would create conditions that minimized the need for redistribution by (he believed) ensuring at the outset that people would be set on a path to prosper without redistribution. He didn’t believe that government should just rearrange incomes year after year.
One commentator on an earlier thread, and some people in other blogs, have already assumed that in advocating social justice, I must therefore be in favor of continued extensive interference with markets. But a commitment to social justice does not logically entail a commitment to having government pursue justice through a heavy-handed, direct strategy.