We'd be paying ~45 PP, 160 BR, 80 SR, and 8 years of berth time for a total of +8 C, +8 S and +8 D.
If this was JUST about the resource costs and berth time then it might be worth considering, yeah. The above, versus, for example 2 Centaur-A which would cost 160 BR, 120 SR, 4 years berth time, and crew for +6 C, S,and D, but also some additional P, H, L and there just being some more hulls to shuffle around in general.
Not that huge of a difference, and might be worth some extended debate.
But what makes the Constellation refit such an incredible waste and perfect sample of Sunk Cost Fallacy is the PP-cost. For those 45 PP we could just as easily go and request a new 1mt berth and an increase of our budget, and then simply use those to keep churning out Centaur-As on a constant basis.
Getting a 1mt berth (10 pp) and budget increase (30-35 pp) means that we can, for the same amount of PP and berth time, get 4 Centaur-As (total; 12 C, 12 S, 10 H, 12 L, 12 P, 12 D) and 160 BR. On top of the 160 BR and 80 SR saved from not refitting the Constellation.
There simply is no scenario where the Constellation refit actually gets us ahead when compared to the option of simply shoving them into a quiet corner of our territory and then ignoring them until we can replace them with something better.
The thing is that the Constellations are already good enough for rear area garrison duty, and keeping them in that role makes their combat upgrade almost completely irrelevant, meaning we'd effectively pay 45 PP plus refit cost and time for +8 S and +8 D. Which makes the refit an even worse choice by a rather large margin.