Matt Yglesias could be the Gilbert Arenas of libertarianism

If only he wanted to.  Matt writes:

I actually think I am pretty cynical about government.  I’ve learned a lot from my various libertarian friends, from my seminar with Robert Nozick, from libertarian blogs, etc. and I think public choice economics is a very important perspective.  The upshot of this is that, as a general matter, I’m considerably less enthusiastic about regulatory solutions to policy problems than are most liberals.

Sadly, though, the upshot of my libertarian-infused cynicism has mostly been to push me left of where I used to be on domestic policy issues.  It’s cynicism about government and the political process that, for example, has made me much more enthusiastic about labor unions and much more hostile to means-testing entitlements than I used to be.  If I believed that the deliberative democracy people weren’t naive fools, I’d be much more sanguine about various "third way" approaches to things.

Matt is probably the closest I will ever get to thinking I could be a Democrat.  But I am not sure what he is favoring in his post. 

One view is that a once-and-for-all change favoring labor unions would produce a stream of ongoing benefits greater than we could achieve through smaller piecemeal government interventions.  On empirical grounds I am skeptical of our ability to manipulate the union participation variable in a very useful way, and that is assuming I were to like labor unions more than I do (I do like them somewhat; I am not a union-basher, but I am not nearly as keen on unions as Matt.)  Union participation varies largely with whether "unionizable" sectors of the economy expand and contract.  Clearly we are headed away from labor unions as manufacturing shrinks as a percentage of gdp.

Another view, not excluding the first, is that Matt has abandoned Rawls’s "publicity condition."  That is, he is willing to advocate policies he knows to be bad, out of fear that they prevent a political tidal wave.  Means-testing Medicare, for instance, might lead the whole system to lose favor and collapse.  Therefore we shouldn’t means-test, even if the idea taken on its own terms has merit.  I don’t dismiss this possibility.

If Matt is willing to admit I am right about unions (I am pretty sure about that one), I am willing to call the other question a draw.  Deal?

Addendum: Here is Matt’s new book-to-be.