The shield of the cable does nothing for the motor but lower the inductance of the cable slightly which would only matter for long cable runs. due to the square edges of the pwm out of the vfd, there is a voltage doubling resonant effect between the parasitic capacitance of the coils and the inductance of the cable. an rc filter at the motor can stop it, but it does nothing if it isn't a problem.
the common mode voltage delivered by the vfd looks like this
Attachment 410298
and you can't get rid of it without an LC filter. i made one from a standard line/load reactor, adding an extra mangetic path to it, you can then use 3 capacitors after the inductor, connected back to the negative (or positive) bus of the VFD.
Attachment 410300
you don't need to add the extra magnetic path to the reactor if you just want delta connected capacitors at the motor, and that will deliver sine wave phase to phase voltages, but that doesn't remove the common mode voltage from line to ground, which requires a 4 or 5 leg inductor. adding the extra magnetic path will increase the inductance of the choke which means the saturation current will decrease.. so keep that in mind.
anyhow i haven't seen anyone on this forum mention shorter bearing life without a shaft ground.