I am trying to understand Michael Burry's cautions regarding ETFs (https://www.bloomberg.com/news/articles/2019-08-28/the-big-short-s-michael-burry-sees-a-bubble-in-passive-investing). He seems to be making two related arguments:

1. we are in a bubble. He thinks ETFs are causing things to be overvalued because passive investing is shoving money into stocks based on their appearance in an index rather than their underlying fundamentals.

2. when the "bubble" bursts, the massive volume of activity flowing into ETFs is going to exacerbate the pain of a market downturn.

I am primarily interested in better understanding this 2nd point. My question is how, specifically, would/could ETFs act as an accelerant in a market downturn?

Here's what I've come up with so far:

In the case of an ETF that tracks an index like the S&P or NASDAQ, my understanding of ETFs is that shares are created when an Authorized Participant (AP) borrows shares of stock for the companies in the index that is to be tracked. In the example on Investopedia they say that often these shares are borrowed from large holders like Pension Funds. Those shares (the creation unit) are passed to the ETF sponsor and held in trust. The EFT sponsor then created shares of the ETF which the AP can sell on the open market.

Here is the part where I'm a little fuzzy and would like some input: I assume that part of Burry's argument about ETFs is that if the price of the stocks making up the ETF start to decline that the original owner of those shares (in this case the Pension Fund that loaned the shares to the AP) might want to sell. In this case they need to get those shares back from the AP. So the AP needs to buy the ETF shares on the market, give those ETF shares to the ETF sponsor who will return to the AP the underlying shares of individual stocks. The AP then returns those stocks to their owners who dumps them on the market.

I think that part of what Burry is saying is that this action is probably fine for ETFs tracking high volume stocks like Microsoft or Apple. But it could be really bad for indexes like the Russell 2000 that tracks some stocks that don't have large daily trading volumes.

To be really long-winded, consider a hypothetical. Suppose there is company that trades around 1 million shares/day. This company becomes part of an index. Somebody decides to create an ETF to track that index. APs now go out and acquire say 250,000 shares of that stock to form a creation unit for the ETF. Demand for the ETF causes prices for the stock to rise beyond what would be expected given the financial fundamentals of the company. The EFT effectively takes 250,000 shares out of circulation (because they have to sit in trust to back the EFT), so average daily volume of the stock is around 750,000/day. Now the market turns, prices start declining, the owner of the stock shares wants to dump his/her shares. So you now get this sudden influx of big blocks of shares from the ETF (1/3rd of the average daily trading volume) adding to the existing selling pressure.

Is this story consistent with others' understanding of (a) how ETFs work and/or (b) Burry's claims that ETFs will cause the pain of a market downturn to be magnified?

For reference I have identified the following SE threads as related to my question (although I did not find the answers I was looking for in them):

Index funds and finite supply of stock

Do physical ETFs possess the same problems as CDOs?