I just had a dream about economics (NB: the word "dream" means that what follows is not real). It turned out that Fischer Black wrote a short, secret manuscript shortly before his death and that manuscript was just dug up and published in an obscure portfolio journal. In the piece Black expresses new yet serious reservations about the idea of arbitrage. He built a simple model in which there are four kinds of arbitrage but each serves as a substitute for investment in human capital. The more arbitrage an economy engages in, the worse off it is in the long run. For various second best reasons, most of all related to the concept of superfluous assets and "spanning," the arbitrage did not bring the usual welfare gains from equalizing prices across different markets.
It is rare that my dreams are so detailed or for that matter so analytical.