Megan McCardle sums up the libertarianism approach to governing:
“It won’t work” is the easiest prediction to get right; almost nothing does. The thought process that tells you something probably won’t work is not always a good way to figure out what will, even if you were right for the right reasons, as I agree lots of people were. That’s why libertarians have a great track record at predicting which government programs will fail (almost all of them) and a lousy track record at designing ones that do work.
But libertarians are generally smart enough to realize that we aren’t qualified to design a government program. (No one is, libertarian or not.) The prediction guides the desire not to plan. There are government plans, of course, unless one moves beyond a minarchist approach into anarcho-capitalism. And they will most likely not work as planned. We are smart but we are not the next Nostradamus. We remember that our first objective is the maximization of individual liberty. Thus, we seek to limit government to the barest necessity, to wreck as little as possible.