The paradox of redistribution* refers to the fact that welfare states in which a greater proportion of spending goes to universal programs tend to be more redistributive than welfare states in which a greater proportion of spending goes to targeted programs. If programs are evaluated one by one, it might seem that targeted programs will in almost every instance make the poor better off (not counting programs that deliver public goods* like environmental protection, law enforcement and national defence). But the paradox of redistribution suggests that the poor will actually be better off when targeted programs are a relatively small proportion of overall spending. Effectively targeting the welfare state at the poor means giving up targeting in the design of many of the individual programs that collectively constitute the welfare state.
The paradox of redistribution only goes so far. It does not tell us that all programs ought to be universal rather than targeted. That conclusion would be absurd, given that some programs are needed to address special needs that not everyone has; it is not possible even in principle for such programs to be universal (unless they are conceived of as a form of insurance; regardless of this view’s merits, it is incompatible with the definition of universalism from which the paradox of redistribution arises). Nor does the paradox tell us which programs ought to be universal and which ought to be targeted. It does, however, mean that the efficiency of targeting in a particular case is not dispositive with respect to the wisdom of targeting in that case. If our standard of justice demands that we design the welfare state so that on the whole it provides the greatest benefit to the least advantaged, it may be prudent to design some programs according to other criteria.
This raises a problem for those who categorically reject means testing (i.e. income or wealth targeting) as a matter of principle rather than political prudence. If universality is a fundamental principle of justice, it is unjust to give priority to the poor. But as a matter of fact, the paradox of redistribution shows that welfare states that incorporate a greater degree of universality do give priority to the poor. If giving priority to the poor is unjust, then we should adopt a less redistributive welfare state. And the way to do this, apparently, is to make more programs targeted at the poor. Thus it appears that no feasible welfare state is consistent with the categorical rejection of income and wealth targeting as a matter of principle. Call this the paradox of the paradox of redistribution.