a) the Curie temperature indicates the loss of magnetic order, not necessarily local moments. Withn the Stoner model of metallic ferromagnetism thereare no moments above TC. Nature appears to be more complex, though...
b) the thermodynamic answer to your 2nd question would be that the transition arises because the free energies of the two phases cross each others values as a function of T, the crossing point being the transition temperature... [I can't remember the mechanism, though, sorry]