I think those torque specs are what they reference to achieve rated collet runout. I had the same thought when I stumbled on a similar chart - that is a lot of gronk especially considering the length of wrench typically supplied. I did some informal tests just measuring runout on an accurate endmill shank. Seemed like anything reasonably tight enough to hold the tool seemed to give me pretty consistent runout reading, within tenths at least. All of this assumes you have a good quality collet holder & collets to begin with. If its a cheapo with 0.002" runout than no amount of wrenching is going to fix that.
Seating a tool in the quill is actually a different matter. It is pulling the 2 solid surfaces together axially to seat. There is no collet action (collapsing slits). Personally I'm of the opinion tight enough to not let go is good. Hard hammering to release cant be too good for the threads on the inserted tool nor the drawbar. Even though they pretty meaty & threads are strong, doing this repeatably must take some kind of toll. If the surfaces are clean & taper in good condition, it doesn't seem to take a lot of force.
I've always been curious, when people install those power drawbar units, do they set them to some kind of torque? And what kind of spring pressure to release when its un-tightened?