^ I'm not sure but I suspect you are right, or at minimum, larger/more frequent increases in dosages are required and this has substantial drawbacks in terms of increased physical dependency/abstinence syndrome and prolonging Post-Acute Withdrawal Syndrome to durations beyond what is seen in any other substances.
I have spoken with people who have remained on stable, low-doses of benzodiazepines for prolonged periods of time (several years or more) but they seem to be the exception and while in the past, physicians viewed this class of drug as rather benign, few are willing to maintain patients on them for prolonged periods anymore. It's also disputable whether those who stay on low-doses for years are still getting therapeutic benefit from them.
Most long-term benzo users I've known, spoken with or read about here seem to require substantial dose increases over the years and if they try to come off them, experience symptoms worse than what initially lead to the first prescription (and for very long periods). This would seem to lend some evidence to your theory that eventually they lose all (or a substantial amount) of therapuetic efficacy but users still experiencing the anxiety (or worse as a result of their prolonged use) keep trying to increase the dose to feel better.