But something to note here too: most measures of historical returns do not look at liquidity either. I am in the UK, and I know there is research (I can't find it atm) showing that before 1970s, trading costs were significant. Historical returns rarely reflect that. Nor do they reflect the fact that most people before 1970 probably couldn't own any asset other than govt bonds (or that most banks were forced to own them too).
So I would say the issues with indexes are two-fold. First, they will fail if they are built on illiquid securities (recent example here is also Neil Woodford's implosion). Second, they are often predicated on historical returns that are, in any non-academic/practical sense, fictional.
Another reply is Horizon Kinetics...apart from GMO, they are the only investment manager whose letters I actually read. And on ETFs, they have written a lot.
https://horizonkinetics.com/commentary-type/under-the-hood-i...