This is all true, and in particular, most Universities in my experience (based on a representative sample of directors of research, research support IT types and tame academics like myself at research intensive UK Universities) are incredibly bad at managing the inter-departmental billing. In my institution even for the big-ticket items like electron microscope, the technical support people spend more time chasing down bills than actually supporting users on the kit. Any cost-benefit analysis conducted by people who actually known how a lab works quickly shows the whole thing to be a complete waste of time and money.
Even for catching the idiot users who've broken something it can be difficult to make it pay. So you catch gormless post-doc whose just crashed a microscope stage into a sensitve detector, shorted out the HT and blown several boards in the back of the instrument. What happens - the HR people won't let you recover the costs from the salary (on the grounds that the post-doc would be destitute), the PI whose post-doc it was will refuse to pay because they've not been allowed to factor a charge for gormless post-docs into their grant application. Neither the host department nor the user's department will pay and will dispute liability (on the grounds that your system should have been interlocked against gormless users). You can ban said gormless post-doc from using the instrument again, but that's rather academic as right now nobody can use it all.
In more industrial/manufacturing evironments one can lock down the processes which reduces this sort of mayhem, but in academic research environments that's much harder. If you can keep the userbase small (10 users on any bit of kit) you can jsut about manage to sport the gormless ones before they do damafe, but when you go to larger userbases it just gets to be a hard problem.