The objectives of policy analysis

From the always-interesting Megan McArdle comes another post, "The Limits of Policy Analysis", that has stimulated two different lines of reflection, or maybe two and a half. Here's the half, first:

The substance of the post itself just provides a couple of illustrations of the unintended consequences that follow upon policy changes that were supposed to have certain intended consequences. "Does this mean that we should never do anything?" she asks, rhetorically, but doesn't answer. I will though: it means that we should do things we have a right to do with our own lives, for our own ends, taking the bad or unintended consequences with the good, and learning from the experience as best we can. But it also means that we should, at the very least, be both humble and very cautious about doing things using the power of the state to control other people's lives in order to bring about particular outcomes we deem beneficial for them. "Doing things" on that level should be limited as far as possible to adjusting state structures away from bringing about or facilitating outcomes for individuals, and more toward facilitating individual and voluntary processes by which people can bring about their own outcomes. Doing such things on a state level will no doubt still have their own unintended consequences, but learning from, and improving upon, such mistakes will at least keep policy makers focused on their proper and possible goals, rather than upon designing and guiding the lives of others.

The recognition of unintended consequences, she writes, is "a good reminder to be humble about what we say about policy" -- but, more than being humble, it should be a good reminder to keep our focus on the right level: on structure rather than outcome.