More bank applied results in less yaw needed from the flight mechanics point of view, the automatic consequence of which being a reduced rudder drag.
I think that you might be missing the point. Bank directly affects slip angles which affects yawing moments (something about further effect of controls, as I recall ? - bank with aileron and you see the nose turn to the low side - severe adverse yaw cases notwithstanding). By picking up some useful yawing moment associated with the bank angle, you are reducing the proportion of the total yawing moment which needs to be found via the rudder input. This lets you put some more rudder in to reduce the Vmca a bit further which is the aim of the exercise.
Conversely, find yourself back around the real world Vmca on the day (and we're talking only a few knots) AND you don't put the bank in, then you might get a surprise from the aircraft's response ....
Surely the 5 deg mark is a good one not to exceed though.
For sure, this is a certification limit to put a fence around the OEM innovation paddock. However, it is a Vmca thing, not a performance consideration.