I don't understand your reasoning. If you put a big wide offset on the thing, the car will sag lower. The spring rate will effectively decrease. "Saying" that you are figuring it to the ball joint may make "the math" happy, but the real world effect is it will be softer
What you are doing effectively, is putting a pipe extension (wheel offset) on the end of your breaker bar (LCA)
Ok, so apparently I need to clean this up some.
So first off, this isn't "my reasoning". There is a standard way to calculate the wheel rate for a torsion bar, and it does not include anything beyond the lower ball joint in a Mopar style suspension. That's what the OP wanted to know, basically we're talking about the advertised wheel rate for that particular torsion bar on an A-body. That's the standard way to calculate the advertised rate.
Second, "the math" doesn't need to be "happy". You can calculate anything you want with the math, it doesn't care. It's not "the math's" fault that the standard doesn't include anything beyond the lower ball joint, that's just the standard. The math can calculate the rate at any point you want, if you want to work through the math.
And if you look at the disclaimers on the Sway-Away web page I linked with the torsion bar rate calculator, you'll start to understand why there's a standard. Because the
actual wheel rate at any given moment depends on a whole bunch of things, and it changes every time the LCA moves up or down. The
physics only cares about the horizontal plane for calculating the wheel rate. So, it's not even the length of the LCA that matters, it's only the horizontal component of the distance between the lower ball joint and the torsion bar. That means if you raise or lower the LCA from the point where the ball joint is in horizontal alignment from the torsion bar hex, the wheel rate changes. If you raise the ball joint 2", then the LCA length becomes the hypotenuse of a triangle, and all that matters for the calculation is the horizontal leg. So ride height has an effect on wheel rate, and so does the amount of suspension travel. But it's a small effect. I've calculated the difference over an A-body suspension travel range, you're talking about a few pounds per inch in either direction. Why is the ball joint used? Well, all of this stuff moves in separate arcs. The lower ball joint, the upper ball joint, the spot on the spindles where the wheel bearings ride, the hypothetical center of the axle. All of those parts are in constant motion, and the angle of each part is a factor in the horizontal distance. So if you really want to be technical, you need to consider the entire range of suspension travel, the angle of the spindle and ball joints over the entire range, and then work out the horizontal distance over that entire range. And that will give you the range of wheel rates for a given car, at a given height, with a particular torsion bar and wheel combination. There's probably a suspension program out there that can do it if you input all the suspension points.
Or, you could just use the standard. Because the difference over that range is fairly small, and there's really nothing you can do about it. There will always be a range for the wheel rate if you're calculating it exactly over the suspension travel of the car. And unless you're going to change the suspension geometry or range of travel you're not going to alter that range a whole lot, just control the overall rate with your torsion bar choice.
You could also use the torsional constant for the bar, because that's not changing. It would be inch-pounds/degree, or some form of force/radians. That doesn't change. But the industry standard is to use the wheel rate, and the standard is to calculate it using the length of the LCA between the center of torsion bar to the center of the ball joint, assuming that the LCA is horizontal (parallel to the ground).