You're right, there's no hard evidence. Google keeps its secrets well, making it difficult to prove these claims definitively. I'm speaking from experience, but there is a way to observe this phenomenon yourself if you have a website:
1. Register on Google Search Console
2. Go to the page indexing section
3. Look at the rows "Discovered - currently not indexed" and "Crawled - currently not indexed"
For example, my own site has a two-digit number of URLs in both categories. These are blog posts Google simply doesn't want to index for reasons unknown.
I have access to Google Search Console data for over 100 websites, and most/all of them have the same issues. This includes sites (like my own) that rank well for certain keywords and receive traffic.
