Partisans of the market are inclined to see greater efficiency in the minimization of public expenditure and the provision of as many goods as possible through private contract. Thomas Schelling has even suggested that poor people should be allowed to patronize cheaper airlines and airports that offer a lower level of safety than rich people require—since it would be worth the savings to them. In money terms, after all, rich people value their lives much more than poor people do, not because they value their lives more, but because they value money less. Often, however, there is no satisfactory individualized substitute for public goods—not only those of the minimal nightwatchman state, but others as well. Rich people can band together in restricted private communities where the streets are clean and the landscaping and security perfect, but this is not enough, even for them. They also want to be able to live and work in safe and attractive cities with diverse populations. Leaving everything to the market will in certain respects leave everyone worse off than they could otherwise be.
It may be that the justification of public provision not for purposes of redistribution but rather to provide public goods for reasons of efficiency—goods that benefit everyone—can be extended to cover a great deal. The classic public goods are defense, domestic security, the legal system, environmental protection, and public health. But there may be important aesthetic, social, and cultural goods that cannot be supplied privately. If we can ensure a decent level of education for all, independently of their ability to pay, the result will be a society that is much better for everyone to live in, and economically better for almost everyone, than a society with high levels of illiteracy and innumeracy. A considerable support for universal education by the haves, even with a minimal tax contribution from the have-nots, will produce on balance a result that is advantageous for the haves as well as the have-nots, in both social and economic terms.
Similar things could be said of support for the performing arts in order to foster a creative cultural environment, support for scientific and scholarly research, and so forth. finally [sic], there is also a case, based on this type of efficiency consideration, for traditional social welfare policies guaranteeing a decent minimum standard of living, or decent minimum earnings, for everyone in the society. Such programs are usually regarded as redistributive, but the alternative to a decent social minimum is a society with real poverty, which often results in higher rates of crime, drug addiction, and single motherhood, all of which impose their own costs not only on the poor but on everyone else, too. To be grim about it, the cost of subsidizing wages for unskilled labor to make them sufficient to support a family might well be balanced by savings in the costs of prisons and law enforcement that such a change would produce, not to mention the value for everyone of the change in the social environment.
Again, such programs would not be redistributive in the usual sense of benefiting some at the expense of others. The poor would benefit, but only to the extent that the rich would also. The size of the benefit to the poor would depend on what would equalize marginal benefit to the rich from among competing categories of expenditure—how much the well-to-do could contribute before alternative uses of their money, including private consumption, would be more valuable to them—the diametrical opposite of Rawls's difference principle.
The reduction of social and economic inequality is in this way seen as a public good, paid for according to its monetary value to different individual taxpayers. This case differs from that of national defense, for example, in that it makes no sense to tax the poor for some of the cost of raising their spendable income. But it is still driven by efficiency, not fairness—a direct appeal to the interests of each, with no sacrifice being imposed on anyone. There are obvious political advantages in portraying social welfare policies in this way, but that does not mean there is nothing in it.
(Liam Murphy and Thomas Nagel, "Taxes, Redistribution, and Public Provision," Philosophy & Public Affairs 30 [winter 2001]: 53-71, at 63-5 [footnotes omitted])
