I'm guessing there's just more heat trapped in the Earth from its formation than suggested by formation models. A factor of two error would not surprise me because we're dealing with a fairly complicated system (with convection, multiple states, heterogeneous fluids, &c.).