Balanced connections make use of a principle called common mode rejection. The idea is that the signal is split into two identical signals at the output, but phase inverted by 180 degrees. Any noise picked up by the cable will be in phase on the two signal wires. At the other end, the noise is phase cancelled by common mode rejection, and the differential signal is combined back into a single ended signal. It's a good thing for long cable runs, but usually not necessary for short ones.
I highly recommend you heed the advice above, and leave the effects processor out of the recording chain, unless it's a really high end unit. Even though you say you know what effect you want to use ahead of time, the fact that the entire signal will pass through the unit rather than just an aux send to it means that the audio quality of the efx unit will completely determine the audio quality of the vocal track. Besides, I've never set an efx on a vocal one time, and not tweaked it during the mix. Recording it first means you can't dial it in perfectly later.
If you use preamp - compressor, the balanced/unbalanced question is moot.
Set the efx unit up as an aux send and return using your soundcard ins and outs, or use software efx.
If you still want to keep it in the recording chain, pay close attention to gain stageing, and a keen ear on the noise floor.
Good luck