On the stability of the pressure method for the Navier-Stokes and Euler equations