Lars Syll thinks [1] that economic systems aren't ergodic, which is a fine statement of a hypothesis (essentially the opposite of the assumption of ergodicity used to build thermodynamics -- and that I use on this blog when I assume ideal information transfer).