G'day sticky667,
That's a great question and one with a pretty easy answer. The race engineer guys sometimes refer to them as "tube torsion members" in a pure mechanical sense and below is the conclusion from a white paper we published years ago on this subject.
---------------------------------------------------------------------
Conclusion
Tube torsion members can be used successfully as swaybars with some benefit in weight reduction.
As an example, 2 individual, rate equivalent (approximately), common sized generic bars 1100mm long with 300mm blades, one solid 24mm the other tube of 25.4 x 3.96mm (1” x 0.156”) gives a weight reduction of about 2.5kg with about 200g of that as unsprung mass per wheel.
However the downside to these is the increase in stress levels for equal OD or rate, and therefore reduced strength. Also the complications seen in manufacture swaybars from tubular medium, and the importance of geometric control under bending make these more difficult and expensive to make. The attachment points at the blades can also be a problem with these swaybar designs.
The additional complications and downsides of hollow bars seem to outweigh their advantage in overall weight and unsprung mass, which can be seen to only be marginal in street car application. Even OE manufacturers rarely use this type of design in their swaybars. However with the right conditions and setups, they can bring some gain in racing situations, which can require weight reductions to the gram.
------------------------------------------------------------------------------------
As you can see, back in 2003 when this was written there was not a lot of tubular product swaybars around and that's still the case. We do know that the frequency of failure through breakage is a lot higher than solid as well but that doesn’t really matter as much for racing at the very pointy end as they can program in a replacement strategy.
However, for OEM or stock it's not an option but on the other hand they can control a tighter tolerance between production and overall implemented function design, that is, how the bar works with the whole system. Having said that, part of the reason we got this paper done was because we were seeing a lot of hollow OEM bar failures in performance use or heavy duty use.
You can read the full post and attached copy of the original paper here at RaceDotCom.
Hope that explains it.
Cheers
Whiteline Jim
www.whiteline.com.au