Ilya Somin reads my piece on Illinois and has some very cogent thoughts on how libertarians should handle the problem of people who rely on government programs:
This issue is a common challenge facing libertarians or indeed anyone who wants abolish a longstanding government program: what do about the interests of people who depend on that program, sometimes to the point of "planning their lives around it."
My own view is somewhere in between McArdle's and Henderson's, but closer to the latter. It is indeed true that there are people who depend on government programs that should should be abolished, and in some cases it would be unjust to simply cut them off immediately. But that doesn't mean we should simply leave the programs in place forever. There is a wide range of options in between going cold turkey and feeding the addiction indefinitely. For example, we could lower program benefits without eliminating them completely. We can also continue paying current beneficiaries, but gradually eliminate the program for future ones (who generally rely on the program less). Another option might be to give some or all of the beneficiaries a lump sum "severance payment" to cushion them through the transition they face.
There are also some relevant moral distinctions to be made between different classes of program beneficiaries. There is a difference between an ordinary citizen who relies on a program he or she had no hand in creating and a politically sophisticated, powerful interest group that aggressively lobbied for its establishment and perpetuation. That's the distinction between, say, a low-income old lady who lives off a Social Security check and big agribusiness interests that lobby for massive farm subsidies and then whine about having become dependent on them.
It's also worth noting that people who rely on government programs, even those who are not politically sophisticated, are aware that such programs are subject to change by future legislation, and that they accept the risk of such change when they decide to participate in the program. That doesn't mean they have no legitimate reliance interests. But it does diminish the extent to which the rest of us are morally bound to fully compensate them for their losses if the program is cut or abolished.
Finally, reliance interests cut both ways. If some people have "planned their lives" around existing government programs, others have planned their lives around existing tax rates. Doubling state income tax rates imposes considerable harm on the latter. That doesn't mean that Illinois is morally required to never raise them. But it does mean that any reliance interest rationale for increasing taxes in order to pay for programs must take account of reliance interests on the other side.
I think we have to specify that we're talking about the case of Illinois, which needs to cut billions out of its budget right now, and not about long-term strategies for the libertarian state. I would, for example, be very pleased to turn Social Security, over the long-term, into a means-tested program that keeps the unlucky out of destitution, rather than this absurd attempt of middle-class Americans to get rich by picking their own pockets. But if some libertarian were to argue that starting tomorrow, we should terminate the program and put those who really need it on welfare or disability, I would say no. The current beneficiaries relied on the program's promises to plan their retirement; you cannot simply say "haha just kidding after they've been retired for ten years.
Now, if circumstances are sufficiently dire--say, it's a choice between middle-class seniors, and kids born with severe disabilities--then I would endorse such a change in the program. But that's not quite where Illinois is. The state's income tax isn't particularly high, and it has a lot of social service programs that people rely on, as well as a lot of contracts that would be very costly and disruptive to break. They need to close their budget deficit now, not thirty years from now; that means some tax hikes are simply necessary.