Most folks here are focusing on Burry's comments regarding price-discovery. However there is another huge point: Liquidity risk. To understand his point, you have to know the gory details of how an ETF operates.
First: When you buy a ETF share for the S&P 500 (iShares, Vanguard etc), the share is not backed by all 500 S&P components. Virtually all the large-number component ETFs are using a sampling of shares to match the underlying index. (They would be buried by transaction fees otherwise.) The subsampling of the index is reasonably well-understood math, but relies on an assumption: That the buying and selling each component share will not be greatly affected by the ETF purchase or sale.
[Edit: I may be out of date - Some ETFs are full samples. Nevertheless, the bigger point that the ETF purchase/sale does not much affect the price stands.]
Second: The ETF uses a very clear process to keep the price of the ETF in equilibrium with the index it represents. Large players are allowed to go to the ETF adminstrator (say iShares) and turn in a bunch of the ETF shares, and iShares will transfer back the underlying components. So if the ETF price ever gets too cheap relative to the index, the big players will redeem the ETF share, and then sell the underlying shares they received, which results in a quick, nearly guaranteed profit.
Conversely, if the ETF price goes above the index, a large player will bring a basket of the underlying component shares to iShares, and iShares will give them corresponding ETF shares. So they buy the components cheap, sell the expensive ETF, again making a quick profit.
Now to what Burry is saying: Several components of the big indices are thinly traded compared to the amount of money in the index funds. In a large index drop, there will be disproportionate downward moves in those thinly traded shares: As large players will be redeeming the ETFs for underlying shares and then sell, these thinly traded stocks will drop further than you'd predict from the index. This will cause the index to drop further, which will cause more ETF shares to be redeemed, perpetuating the cycle.
I think his point should be better known than it currently is: The current wisdom that "you can't lose money in the stock market long-term" is reminiscent of "you can't lose money buying a house."
Yes, the liquidity risk seems the more interesting piece.
He seems to say that if you have trillions of dollars in ETFs, you should be seeing more volume in the shares in these indexes than we actually observe.
So some of this cash is going toward synthetics -- mathematical models that are supposed to mimic the underlying securities -- and not the actual stocks in the index.
In a general rout, the synthetics won't perform like you'd expect them to. Prices might eventually clear, but it's not "as good as cash" like many investors assume.
This feels like the most concise explanation of the underlying mechanics that I was intuiting from the article. Now the question becomes: how do I hedge out of this risk without going full day-trader?
> Now the question becomes: how do I hedge out of this risk without going full day-trader?
Here's anecdata from the past: I spent a lot of time during the housing bubble working on a similar strategy. I came to the conclusion that shorting the banks (with leverage!) would be a profitable way to make money.
It turns out that was a beautiful and correct strategy, up until the moment the SEC decided to ban shorting. We got out with profit, but it was stressful and certainly nothing life-altering as a consequence of the SEC action. (bastards!) If you have seen The Big Short, they had a similar problem: Because the CDOs stopped trading, Goldman and company unilaterally declared that there was no problem. Since there was no market, there was no mark-to-market. Burry and friends were able to wait it out, but you'll notice that Burry had to exercise some extraordinary clauses in the contract; Dealing with your investors after that must have been all kinds of fun.
The moral of the story is probably something like: When you are profiting from the system melting down, the system will invent new rules to impede your profit, so be prepared.
Edit: BTW, I don't think mark-to-market accounting has ever been fully restored, but it's been a while since I checked.
My guess is that would be more expensive than what the index fund / ETF would cost you even in a worst case scenario. Read my comment where over the last 10 years Vanguard has only deviated by 1%, even after factoring in trading costs. A typical active fund is going to charge .5% per year, minimum, which adds up to 10% over 10 years, assuming a 7% annual return. A more typical 1% expense ratio is going to be 20%. Active management is extremely expensive. But trading yourself, unless you have a million, is going to be expensive, too. IF you wanted to DIY, $5 a trade adds up if you're putting money into multiple stocks every month. $1000 divided into 10 stocks a month is going to be a 5% expense before you even get started.
A large cap index fund wouldn't have the same liquidity problems, but it has the problem that if this were to happen, funds might have to sell their large cap holdings to cover outflows, driving down those prices. At least it would reflect the NAV, though.
You could buy a closed-end mutual fund since it won't be balancing its holdings in the same way.
You could buy a small cap fund and short a large cap fund, betting that small cap shares will get distorted in the positive direction if there's a liquidity crunch.
I guess I'm somewhat less worried about people not picking stocks because hedge funds, and really, anyone greedy, will always try to do that, bringing some amount of pricing to the market.
There is a range of active - passive even within ETFs that are trying to perform like an index. An example of this in Canada is say Horizons vs Vanguard that's "fully" passive. You pay higher fees for someone else to do the hedging (= protection from having your money in a passively managed ETF that has underlying assets that don't have a high trading volume) for you.
> ..."you can't lose money in the stock market long-term"...
Put in a large enough number of years in that "long-term", and it is true. But many people don't have the time horizons of institutional investors, so what is "long-term" to an re-insurance company might be "lifetime" to an individual investor.
Now, if we could only resolve the principal-agent problem for institutional investors to the benefit of individual participants that make up the institution's backers...
First: When you buy a ETF share for the S&P 500 (iShares, Vanguard etc), the share is not backed by all 500 S&P components. Virtually all the large-number component ETFs are using a sampling of shares to match the underlying index. (They would be buried by transaction fees otherwise.) The subsampling of the index is reasonably well-understood math, but relies on an assumption: That the buying and selling each component share will not be greatly affected by the ETF purchase or sale.
[Edit: I may be out of date - Some ETFs are full samples. Nevertheless, the bigger point that the ETF purchase/sale does not much affect the price stands.]
Second: The ETF uses a very clear process to keep the price of the ETF in equilibrium with the index it represents. Large players are allowed to go to the ETF adminstrator (say iShares) and turn in a bunch of the ETF shares, and iShares will transfer back the underlying components. So if the ETF price ever gets too cheap relative to the index, the big players will redeem the ETF share, and then sell the underlying shares they received, which results in a quick, nearly guaranteed profit.
Conversely, if the ETF price goes above the index, a large player will bring a basket of the underlying component shares to iShares, and iShares will give them corresponding ETF shares. So they buy the components cheap, sell the expensive ETF, again making a quick profit.
Now to what Burry is saying: Several components of the big indices are thinly traded compared to the amount of money in the index funds. In a large index drop, there will be disproportionate downward moves in those thinly traded shares: As large players will be redeeming the ETFs for underlying shares and then sell, these thinly traded stocks will drop further than you'd predict from the index. This will cause the index to drop further, which will cause more ETF shares to be redeemed, perpetuating the cycle.
I think his point should be better known than it currently is: The current wisdom that "you can't lose money in the stock market long-term" is reminiscent of "you can't lose money buying a house."