This is great. I have two requests (for Tyler or anyone who may have some answers):
1. Could you address the idea of preference preferences? It strikes me that psychoactive drugs alter one’s utility function. Can that be modeled, or does that put you in non-linear-dynamical-system territory that doesn’t admit functions as we know them?
2. Am I being overly simplistic in this stripped-down political model? It seems to me that ideal welfare would measure your lifetime productivity at birth and then compensate you to the extent that it falls below $X. Obviously this comes at some cost, but it seems to me that a consequence of Arrow’s impossibility theorem is that there’s no political solution to choosing X (and how the cost of X would be distributed).
Thanks to Tyler and all who can shed some light!
When it comes to meta-preferences, it is usually acceptable to assume convergence through a fixed point theorem. So whether I have a preference for a preference for a preference, etc. comes to a coherent finish and we can speak meaningfully of meta-preferences. That said, I don’t see why we should always favor the meta-preference over the (“baser”) preference. “Wanting to listen to Top 40” is not obviously a worse preference than “wishing I enjoyed opera more.”
The second question refers to a longstanding debate in optimum taxation theory. There are a few reasons why we don’t grant transfers contingent on ability alone. First, we may find the outcomes of privately accepted gambles unacceptable and wish to grant aid to the losers. Second, there is sometimes a higher social rate of return from investing more money in the skilled. A final question is why we make so many transfers near or at the end of life and so few, relatively speaking, at the beginning of life. Obviously the elderly vote at a high rate but I don’t think it is just that. There is a more general intuition that the elderly deserve the aid and the young do not to an equal extent. That intuition is a big obstacle to fiscal balance.