A few more notes. 1. Iron is non-magnetic at pressures above 10 Giga Pascals (where it enters an HCP-crystal-structure phase), so presumably the Earth's core is not ferromagnetic. (Sad, since if it were, that would help resolve the puzzle.) 2. The simple model I proposed could run in either (A) the direction where fluid flows up to equator & sinks at poles, or (B) the (opposite) rises-to-poles&sinks-at-equator direction. Only one of these directions yields a self-reinforcing dynamo. For the other direction, the "dynamo" effect should be of negative feedback! So, my hands got a workout applying "right hand rule" and suchlike to try to figure out which is the good direction -- I believe the answer is: Direction B yields self-reinforcing field! Contrary-sign to my original post - sorry. It is not obvious to me which of A,B the fluid inside the earth should prefer, although probably one of them is favored. 3. Higher pressure decreases electrical resistance in almost all materials. Higher temperature decreases it in semiconductive and insulating materials but increases it in metals. What is the electrical resistivity of the earth's interior? It might be that the resistivity varies with depth and is high at some particular particularly important depth(s).