I know there is nothing obviously wrong with what I did. It's just that there has been so much inexplicable errors and behaviors as shown above that I thought it might be worth reporting. As usual, I won't hold my breath regarding follow-up. NI is not bound by any of our comments, suggestions and even less requests!
I looked into this issue more and there have been multiple CARs filed for this issue as well as other empty cluster references causing errors. The most recent CAR specific to this issue was set at a higher importance than the others so I would say that R&D knows about it and is out there working on it. The error is genuine, but when it is displayed it is flawed. Fixing a non-related error should not cause this error to go away.
What Matthew Kelton suggested at the top of page 2 is the correct fix to this problem, you will need to populate those clusters.
Anthony, thanks for the clarification, but why should I bother if my code functions as is? I was merely reporting an oddity, and rediscovered an old bug/feature in the process... I doubt I will find the motivation to look throughout my code for so-called empty cluster references without a guaranty that I will find them all.