Am I asserting with authority that I know defensive metrics to be very accurate? I hope not. What I am asserting is that simply coming up with a story of why defense doesn't vary much does not in any way convince me (nor should it convince anybody, in my view) that the data must be wrong. If there is no burden of proof required, one can easily come up with any number of plausible hypothesis regarding the amount of variability in defensive performance.
My only point real point is that if I have to move forward and either accept or reject the data, as one must do if he/she is to come up with a system of holistic performance measurement, then I think UZR provides us better information than simply ignoring defense or applying some arbitrary defensive performance number based on our gut intuition, ad hoc adjustment to UZR, or any other approach I've seen put forth.
Hopefully, over time, defensive data will get better. But I've yet to see anybody put forth an approach to measuring defense that actually produces a better estimate. I have no more confidence in the accuracy/reliability of a "WAR" measure that ignores defense than one that incorporates data that might not be as reliable as we would wish it to be.