That's not how overclocking works.
First of all, where are you getting 40%? No chip overclocks its stated GHz speed by 40%. 3.5GHz -> 4.5GHz is common and that's barely over 20%.
Second of all, increased clock IS NOT a % increased speed. There are very few things where that applies so directly. For more common things like gaming you should be measuring in matters of FPS, where increased CPU by 25% might only see 10% increase in framerate.
Finally, the goal of OC should be to allow you to increase game settings without dropping your framerate to a point where you actually notice that there are frames per second. Eg, going from 100fps to 120fps seems like a 20% increase, but really you can't notice it and you're generating more heat. Going from 100fps to 100fps with higher settings actually is a good change.
Also, notice how stock clock i7's haven't gone up very much in rated GHz? That's because they're improving other things. Memory bandwidth, heat generation, OC performance, thread handling. It's not all about the GHz.
Hahahhaaha... way to n00b it up... ok let me break it down
4.8ghz is pretty much reachable on 90% of the chips on ivy and haswell, @ the 10% voltage boost @ 1.41v
4.8/3.5 = 37%.. so I rounded up, notice i said ~40%, because certain apps may even give you + 50%
Now with regard to How much this gives you... Well.. you can't count that by FPS alone.
As you've yourself stated, " increase in performance as the result of overclocking possess effects outside of increased frame rates.. "
well, you didnt' state it as elegantly as that, but that's what you meant..
SO.. then the increase in performance can not be measured directly in a complex program..
All we know CONCRETELY is that there will be on average ~ 40% more calculations possible at any given time...
so, I call THAT ~ 40% boost in performance..
this also does not betray my comparison to Future CPUs.. because I'm talking about Relative IPC gain.. which is consistent with my argument.. there is between 5-15% per generation @ 2 year intervals..