I remember when I was installing the (stiff) set of IPD bars on my PV. I did the rear bar first, and took it for a quick spin with that on and no front bar. Lol, it almost *was* a spin, it was pretty prone to oversteer. More fun than I need in my life.
Put the front bar on and now it is very nicely balanced between under and over steer, you can coax it into either one.
Roll bars do two things, one is the more noticeable (if you're not going to and past the limit of adhesion) - simply making the car lean less in corners.
The second involves the roll stiffness balance between the front and rear, which is a combination of the spring rates (and shock valving in a quick transition) and the roll bar stiffness. Generally speaking - the end with the greater roll resistance/stiffness will be the end that breaks loose first. Again, in very general terms, the reason Volvo generally tended to put front bars on and not always rears is to ensure that the cars had a fair amount of understeer from the factory - understeer being what is safer for drivers on public roads. The tail coming around in a corner on an unsuspecting driver is a very bad thing, it's not great on a suspecting driver either. The car just plowing straight ahead scrubbing the front wheels is a lot safer for all involved.
If you're not worried about which end will break loose, or if you don't care that there's a fair amount of understeer, just go with a nice stiff front bar. Getting the balance closer will allow you to go around corners a little faster, but only if you're hustling it at 9/10ths or more, at or close to the limit of adhesion.
Also, I don't think you can compare the thicknesses of the front and rear bars directly, because on a 240 at least, they operate in different manners. The front acts directly between the lower control arms and the front subframe, the rear just ties the two lower control arms together. You'd have to ponder the geometry between the two methods to determine what the diameters really mean in relation to each other.
I don't think IPD did any sort of complicated maths like that, they probably just tried out a bunch of different combinations and drove the test car around a track, and figured out which combination leads to a nice balance with just a small amount of understeer.