This is because I agree with the intuition that it is morally right that "we" (royal "we") have a responsibility to help the worst-off, the badly treated, and the trotted upon. We really *should* have welfare-like institutions.
I call them "welfare-like institutions" rather than just welfare for three good reasons. For one, we can't conclude from the fact that we have a moral responsibility to help the badly-off that it is some precise and specific institution that should do this. It's an open question whether the government is the right institution that will allow us to put our moral responsibility into practice. It just doesn't follow.
For two, moral responsibility implies a choice. There is a difference between a legal responsibility and a moral responsibility. Setting up a governmental welfare institution is to generate a legal responsibility, and not necessarily a moral one. We fail to act on our moral responsibilities if we are made to act in some way for fear of some evil befalling us. Like being sent to jail. Or being fined. Or getting grounded. A child is not acting morally when she shares her toys with her brother because she is under the threat of being grounded if she does not share. She acts morally when she chooses to share for the reason that she thinks it is the morally right thing to do.
The same is true with welfare. We do not act morally when we pay our taxes when we do so out of fear of being punished. We might if we do so out of a sense of moral responsibility.
And for three, it might turn out that non-governmental institutions will do the trick. This is an empirical question, and needs empirical studies to work. Maybe people will deny that this is true. Maybe charities and voluntary, non-governmental institutions will not be sufficient. Maybe. But we need to at least answer the empirical question, since governmental institutions come at a significant expense compared to equivalent voluntary institutions.
It's interesting to point out that Rawls himself, in the preface to the second edition of his A Theory of Justice (somewhere around pages 11 - 13), writes that he did not intend to justify the welfare state. To make sense of this claim we have to take the above three points into consideration. I take it that Rawls had something like the three things above in mind when he made this claim.
So that's one reason for thinking that a Rawlsian, or welfare liberal, libertarian is not entirely implausible. Here's another reason. This reason should appeal to ordinary welfare liberals from the left, and it is a reason that appeals to me.
Suppose the following is true. Suppose Bob, a welfare liberal, strongly supports a welfare state. His theory of justice demands the inclusion of a welfare state. Suppose he also virulently objects to a warfare state. His theory of justice says that war, in the overwhelming majority of cases, is morally illegitimate. Suppose further that Bob ranks his opposition to war higher than his support of government-run welfare. We might say that he considers the benefits of a welfare state at, say, 60, but the negative harm of a warfare state at -90.
I think Bob describes a great many Americans and Canadians who consider themselves on the left of the spectrum. We should help one another, even make people help one another, these sorts of people would say, but we should not engage in wars.
Follow me so far? Okay, here's the other important assumption I will ask you to grant for the sake of argument. Suppose that it is true that the governmental institutions we set up for the purposes of a welfare state have a 70 per cent likelihood of also engaging in war. Suppose, just for the sake of argument, that governmental institutions are like that, for some reason.
Bob, the welfare liberal, makes the following calculation.
"If we set up a government, we can expect a benefit of 60. There is, however, a 70% chance of this government engaging in war, which would yield a harm of -90. If we multiply the probability of engaging in war by the harm of a war, we get -63. Subtracting the expected harm from the expected benefit yields a net expected benefit of -3."Bob wants a welfare state. He is a genuine welfare liberal. But he really doesn't want war. He sings to himself, "War. Huh. What is it good for? Absolutely nothing!" And he believes the lyrics. But his calculation has led him to the conclusion that he should expect a harm if we set up a welfare state, even though Bob wants that. What should Bob conclude? He should conclude that, in practice, we shouldn't have a welfare state (because it is likely to engage in war) even though, in theory, we should have a welfare state.
"If only we could have a welfare state without a warfare state," he laments to himself. "But I'm not a naive welfare liberal, I take empirical facts seriously, and I am honest about what I expect."
All I am trying to show is the possibility of a welfare liberal libertarian. I am making no claims about the status of the numbers used. If you adjust the numbers, you might conclude something else. Suppose you think the benefit of a welfare state is 70, rather than 60, for instance, keeping the probability of war the same. That would mean a net benefit of seven to our society. In that case, the welfare liberal should support government-run welfare, even though he can expect wars.
Whether or not any particular welfare liberal should fight for a libertarian society will depend on her beliefs about the probabilities, and her rankings of the relative benefits and harms of a welfare state as compared with a warfare state.
On what basis should she ground her probability claims? One approach would be to take a historical sample of liberal welfare states, and see how many of them also engaged in the sorts of wars that she strongly opposes. My (strong) suspicion is that the probability will be higher than 70 per cent.
I have abstracted away one other component of the argument for simplicity's sake, but I should at least mention it since it will strengthen my claim. The benefit that we have assumed comes without a probability assesment. But the benefit is not guaranteed. Welfare states make mistakes. Sometimes rich people get welfare. Sometimes poor people fall through the cracks. No welfare state is perfect. Thus, we must also downgrade the expected benefit of a welfare state by appeal to a probability of actually getting the benefit. That probability will be less than "unity" (as they say), or 100%.
There will be a less-than 100% probability of the benefit (welfare), just as there is a less-than 100% probability of the harm (war). It would be neat to put together a welfare liberal libertarian calculator, so that you could input your own preferred probabilities, and your own preferred relative benefit and harm assessments. I don't have the skills for this. But I wonder just how many welfare liberals would actually turn out to be libertarians, for non-traditional (or non-standard) reasons.
You would be, morally speaking, a welfare liberal. You would be, politically speaking, a libertarian.