When talking about welfare economics, two key theorems come to one’s mind:
1rst welfare theorem: All competitive equilibria are Pareto Optimal
2nd welfare theorem: All Pareto Optima are competitive equilibria.
In this paper, the author reaches some not so intuitive results by making some interesting assumptions on preferences. In his results, the first welfare theorem doesn’t hold under these assumptions; however, the second welfare theorem does. This means, that not all competitive equilibria are Pareto Optimal.
The author arrives to this results by proposing a definition of non-paternalist independence, where each individual respects the tastes of others, no matter what he thinks of them. However, he forms his judgment of their welfare, and whatever his opinion of the justice of the distribution. This means, the assumed preferences are interdependent, but non-malevolent.
Under that assumption, the author proves that non paternalist extended preferences accommodate the known tradeoff between the mean of the income distribution and its dispersion. This highlights the importance of the non-paternalist condition: it preserves the second theorem of welfare economics. Neither the first nor the second theorem survives the introduction of paternalist preferences. Since paternalists will prefer institutional arrangements, usually they have the effect of making relative prices different for different people. Then, market demand curves do not exist. They link this with competitive capitalism and market socialism. Since the second welfare theorem survived, an optimal point can always be reached by Lerner-Lange socialism, but perfectly competitive capitalism doesn’t offer the same guarantee.
I found this paper really interesting. We can observe how a simple change in the models assumptions ends up breaking the most fundamental results we usually observe. The best thing is the literature that came up after this results, which they mention in the article as well. Definitely a paper worth reading.