This doesn't answer your question directly, but Matt Levine has written some great articles, in particular, about index funds and their effect on the market.

Here's one: [1]

"Second: One of my little stock-market obsessions is that index funds free-ride on the work done by active investors. Someone needs to make decisions that allocate capital to businesses. A world in which everyone indexes, and in which no one thinks that active managers should be able to charge for their services, is a world that will spend too little time and effort on allocating capital to the right businesses. That's not the world we live in: A lot of people still actively work to allocate capital..."

1. https://www.bloomberg.com/view/articles/2015-07-07/can-you-r...

That the stock market is an efficient mechanism in allocating capital to the right businesses is a myth.

For one, the stock price of a company has no direct bearing on its capital. Only at IPO time or when a company is raising additional capital is the stock price relevant for the capital. E.g. Google's stock price has increased 15x since their IPO, but that has had no effect on their capital. Any other time besides the IPO and when raising additional capital, it is just money changing hands between stockholders, the company doesn't see any of that. Of course, the initial investors often only invest in the capital with the understanding that they will be able to easily sell their shares later. And there are some secondary concerns: employees might have stock or stock options, stockholders can influence the board and hence management, and too low a stock price might engender a hostile take-over.

Secondly, the market is frequently very wrong about pricing stocks. It was happily 'allocating capital' to internet stocks during the dot-com bubble, to financial stocks up to 2008,... A metaphor created by Benjamin Graham, and often repeated by Warren Buffett, is that of Mr. Market [1]: a manic-depressive fellow who has periods when he pays way too much for stocks at certain points and sells stocks at way too low prices at others.

That said, it mostly works. To a company it doesn't really matter all that much that the stock market undervalues your stock by 20-30% or more at some times (a good opportunity for share buybacks!), and overvalues it at other times (a good time to raise additional capital, or use your stock for takeover bids of undervalued companies). Of course, companies often do exactly the opposite: share buybacks when their stock is overvalued and takeovers overvalued companies when their own stock is undervalued...

[1] https://en.wikipedia.org/wiki/Mr._Market

> the stock price of a company has no direct bearing on its capital. Only at IPO time or when a company is raising additional capital is the stock price relevant for the capital. E.g. Google's stock price has increased 15x since their IPO, but that has had no effect on their capital. Any other time besides the IPO and when raising additional capital, it is just money changing hands between stockholders, the company doesn't see any of that.

But all the capital decisions are in the context of the stock price. If an outside group looks at investing, that will be in terms of current shares; if the company gets bought by another or does a merger, that transaction will be determined by the current share price.

> Secondly, the market is frequently very wrong about pricing stocks. It was happily 'allocating capital' to internet stocks during the dot-com bubble, to financial stocks up to 2008

Some of those internet stocks were worth far more than even their inflated valuations at the time (of course, many were worthless). The market price is always going to be a best-guess consensus estimate, sure, but a lot of these things are just inherently hard to figure out how much they're actually worth.

No, the decision is in the context of the market capitalization, not the stock price. If $10M company has a share price of $50/share, then it has 200,000 shares outstanding. If it has a price of $100/share, then it has 100,000 shares outstanding. A buyout, merger or acquisition is based on the total value of the company, not based on the smallest unit of ownership in that company.

And market cap alone is misleading without considering debt as well. But all that is technical nitpicking. The point is that the stock price is a measure that directly feeds into the overall value of the company that potential buyers etc. will consider.

This is an absolutely crucial point that is not discussed nearly enough in the context of passive investing. Thanks for the link, I didn't know Matt Levine had stated it so eloquently already.

A tangentially related point is that membership in e.g. the S&P500 imparts a big financial advantage to companies for a somewhat arbitrary reason, due to passive investing.

That advantage can become a liability when the market goes down.

Here's how it works.

An ETF or Fund like vanguard is a company that issues "coupons" and then buys and sells them (and various related administrative things, e.g. forwarding dividends while combining them).

So when you buy an ETF "share", what happens is that you buy a newly issued coupon from this company. This company gets notified, and as a result will put in market orders for these shares (while combining them in smart ways), and once it has bought the shares, issue the "share". (needless to say there's aggregation happening)

When you sell the reverse happens. You essentially request the company destroy the coupon. In response the company will sell shares. Once the shares are sold, the company will transfer that money (ie. whatever they got) to you.

So to answer your questions, in a flash crash scenario as an ETF owner you'll experience more lag in both cases. Ie. whether you're buying or selling the lag will add to your disadvantage compared to the rest of the market. So simplifying things, if a flash crash happens and you own an ETF or a fund you'll only be "allowed" to sell once the drop is over. If you try to buy at the bottom your order won't be filled for a while. Mind you this will be in the seconds range, or in particularly bad cases a few minutes.

In the US, there is also regulation that allows funds to pause redemptions. So if you own a fund that fears it may be significantly affected by a market drop, it can then block your money (regardless of what a contract you have with them says) for a period of up to months. Given what has historically happened, for small funds this means if there is a large drop, they will block your money making things worse (but somewhat avoiding feedback in the market that would cause individual share crashes). You will lose something like 20-80% of your capital if this happens. The smaller a fund the more likely this is to happen.

So an ETF should only be used for amounts of money that are truly too small to buy individual shares, something under maybe $10k. For everything else you should put in the work to buy the individual shares. If you don't do this, yes there are costs that will be imposed upon you in adverse scenarios.

this isn't how ETFs work at all

have a read about the ETF redemption mechanism[1]: it is almost the exact opposite to the way mutual funds work (which is what you have described)

their behaviour at times of stress is correspondingly different to that of a mutual fund

http://www.investopedia.com/articles/mutualfund/05/062705.as...

Almost.

When you buy an ETF, you don't buy a newly issued cupon. You buy it from another market participant on an exchange - hence Exchange traded fund. What you're describing is closer to classic mutual fund. Each ETF will have "Authorized Participants" who make sure that the ETF mirrors the underlying assets.

I'm leaving that part out, because it obfuscates things. In one interpretation "Authorized Participants" on the exchanges keep the number of available "coupons"/shares at a fixed number.

So the market depth for any ETF should be constant, regardless of the value of the underlying assets (assuming non-extreme values and outside of crashes or rapid movements).

And yes, to be more exact, the market depth for an ETF should be a function of the amount of trading occuring in that ETF, not so much a constant. In any reasonable timeframe it should be constant.

Great explanation! Thanks!

>> So to answer your questions, in a flash crash scenario as an ETF owner you'll experience more lag in both cases

So the lag is due to inability to quickly buy/sell the percentages that were allocated for different stocks to build a share, right?

And if I understand correctly, this means if people keep their emergency funds in ETFs, in a crisis, they probably won't be able to access those funds quickly enough to get it at list price (because the price will move because of the lag).

Authorized participants (APs) exist for all ETFs. their purpose is to make sure that the ETF mirrors the underlying assets, and to provide liquidity.

Consider a ETF that consists of a single stock. IF the stock drops, the AP will lower its Bid/offer spread, and force the ETF down to the same level as the stock. But as this is a reaction to the stock falling, there is going to be a lag between the stock falling, and the ETF falling. This doesn't mean that you can sell the ETF at a higher price, necessarily.

A lag of milliseconds. Arbitrage is largely automated for this.

If everyone sells then the market will go down as it does now. It doesn't make much difference if that happens via a fund holding or not.

You can't sell without having someone buying.

Hence why the price goes down until someone does.

Yeah it was slightly loose speech. When most people put in sell orders the price drops till someone steps in to buy, if only the market maker, so the number of shares sold and bought match.

Another effect is over-valuing of stocks with sufficiently high caps to enter the index being tracked.

(A similar already occurs with huge institutional investors (e.g. retirement fund managers) that have too much money to invest to worry about small caps - Warren Bufett also has this first-world problem. Yet another similar effect occurs through bigger caps getting more publicity, so more people invest in them, disproportionately driving up the price.)

The result is smaller caps become more attractive (to small investors). Which results in greater investment in them - but only up to the point allowed by the above effects. So the undervaluing persists.

Kinda similar to tiny startups finding tiny markets attractive, that aren't worth it for incumbent to pursue.

> will everyone try to sell the same set of funds and will crash the funds themselves?

The ETF is backed by stocks in other companies, so if one person tries to sell someone else should be willing to buy, so long as the sale price * the number of shares is below the collective value of all of the shares held by the ETF.

Collective ownership of companies is a more interesting question. If everyone owns fractions of every company, what are the incentives for good governance at any one company? If it goes bankrupt, it's only a tiny fraction of everyone's portfolio.

Matt Levine has covered the latter beat for quite some time.

> If everyone owns fractions of every company, what are the incentives for good governance at any one company? If it goes bankrupt, it's only a tiny fraction of everyone's portfolio.

And all the other companies in the same market segment will then gain a share of the customers and income of the now bankrupt company, so there is little to no effect overall.

Of course this is exactly why some people choose to invest in markets instead of individual companies, because that has less risk.

Of course this is also why other people choose to invest in companies they assessed are way ahead of their market, because that is more profitable.

Active traders will have an outsized impact on the movement of stocks. Which could allow for fun games in the derivatives market. E.g. force a stock up and short it in a derivative. This kind of activity was much easier in the past so its not unprecedented.