This is an old question, but it might be possible to put the old saw to rest for good.
If you have a deSitter space, it can't be rotating--- deSitter space is unique. If you have a black hole in deSitter space, it can rotate (this is the deSitter Kerr solution recently discovered), but it is only one body rotation, the cosmological horizon can't rotate independently of the black hole horizon.
This might not be surprising, except that if you make the nonrotating deSitter black hole bigger and bigger, there is a point where the black hole and Cosmological horizon are symmetrical. In this case, you have two horizons. If you rotate one horizon, the other rotates in the opposite sense, so that only their relative rotation is meanigful. The two horizons are symmetric now, so you can't differentiate between their motions.
If you add matter in-between the two horizons, you will curve the universe inbetween, and if you put a lot of static dust in, you get to an Einstein static universe with two black holes at opposite ends. In this universe, the two horizons are clearly matter. So there is no boundary between matter and cosmological horizons, and it is a fair statement to equate all matter with some sort of horizon-object, so that the electron is like a little micrscopic black hole.
This is the point of view most consistent with string theory, since the strings in string theory are dual under strong-weak coupling dualities to objects which are clearly black holes in the classical limit, namely D-branes. The point of view that matter is the same as horizon puts Mach's principle to rest--- all motion is relative to distant "matter", either matter matter or horizon matter, which is also matter.
This statement is consistent with the holographic principle, and the holographic principle can be thought of as the ultimate in Mach's principle, since it says that all motion is relative to a distant holographic screen, so that the whole thing is moving relative to a distant horizon. This principle is more precise, more quantum, and more general than Mach's principle, and it is consistent with the solutions of GR in a space bounded by a horizon. I must say, however, that the deSitter formulation of string theory is not available right now, so that the full holographic principle is not completely known.