Hacker News Reader

Comment by _se | Hacker News Reader

"They" don't do anything - this is called a circuit breaker, and is automatically triggered. There are three breakers:

L1 - 7% down before 3:25pm - 15 minute halt

L2 - 13% down before 3:25pm - 15 minute halt

L3 - 20% down - halted for the remainder of the day

Only a single L1 and a single L2 breaker can occur in a single day, e.g. the market falling below 7%, rising, then falling again will not trigger a second L1 breaker, but falling to 7%, up to 5%, then down to 13% would trigger an L2.

FYI this is the kind of thing you have to know to become registered as a securities representative.

View on HN

To address OPs point, someone did do this. Someone (or more accurately, a lot of people) stepped in and introduced an artificial construct that constrains trading under certain conditions.

It's not intrinsically a "bad" thing per se and it was something already established long before the current set of conditions arose. But none the less, it's an artificial constraint introduced on trade systems that is likely beneficial.

To those who believe that all markets are rational and efficient, that interventions cause more harm than good, y, an enforced halt seems to be anti-capitalist.

But we are not rational actors. We can get into panics. Panics can stir more panic. Forced breaks allow for the market to reassess data for a few minutes without fear of loss for not acting immediately.

You would think, by the same arguments, that the stock market could always just trade at 15-minute intervals. Why not? But people go crazy when researchers (e.g. Eric Budish at U. Chicago) suggest lowering the frequency to milliseconds, let alone seconds or minutes.

There are different, additional, objections to that. For example, it would make life rather difficult for market makers, which would mean a lot of the liquidity would dry up, which means the spread would widen out, which means trading would be more expensive.

I think there is scope for designing market mechanisms which have the volatility-reduction effects of periodic auctions, but which still allow market makers to hedge. I hope people are working on those.

And then there are questions like: Should trading be convenient? Inexpensive?

> which means trading would be more expensive

And how is that a bad thing in and of itself? It would presumably knock out some of the ultra-low-margin HFT but so what? If it would have the effect of turning the stock market into less of a roulette table, with fewer gamblers compared to bona fide investors - isn’t that a good thing?

The circuit breakers also have an important function in relation to the role algorithmic traders play - it allows a bit more time for humans to step in and either switch off or tweak the algorithms if appropriate.

Do they have circuit breakers in the other direction?

No. This is just one of several places where the laws systematically favor the bulls.

Believe this is not a "law" but a "policy" Regardless, your sentiment is spot-on.

Its an enforcement policy directed by the SEC. Its effectively a law, you can’t opt out for instance.

Well, you can choose not to follow laws (but you'll face the consequences if caught!).

You are wrong about that. See T5, T6, and H10 halts.

T5: Single Stock Trading Pause in Effect Trading has been paused by NASDAQ due to a 10% or more price move in the security in a five-minute period.

T6: Halt - Extraordinary Market Activity Trading is halted when extraordinary market activity in the security is occurring; NASDAQ determines that such extraordinary market activity is likely to have a material effect on the market for that security; and 1) NASDAQ believes that such extraordinary market activity is caused by the misuse or malfunction of an electronic quotation, communication, reporting or execution system operated by or linked to NASDAQ; or 2) after consultation with either a national securities exchange trading the security on an unlisted trading privileges basis or a non-NASDAQ FINRA facility trading the security, NASDAQ believes such extraordinary market activity is caused by the misuse or malfunction of an electronic quotation, communication, reporting or execution system operated by or linked to such national securities exchange or non- NASDAQ FINRA facility.

H10: Halt - SEC Trading Suspension The Securities and Exchange Commission has suspended trading in this stock.

Those seem to me be to be directed more at price manipulation of individual stocks (eg pump and dump schemes) than restraining enthusiasm across the whole market. I'm open to correction on this (and will not demand a halt).

The T5/T6 are not really directional circuit breakers though. They are momentum breakers that don’t bias towards downward momentum.

Because my google-fu is so bad ... has a major exchange ever actually stopped trading because it gained too much?

I don't know if that's a good way to look at it. It's not just "bears" selling when you have a big drop. It's people who need the money for something and are worried about liquidity.

Trading halts may or may not fix that, but that's their purpose.

> It's people who need the money for something and are worried about liquidity.

Then they're doing it wrong. You shouldn't have money that you need immediately in the stock market. Yes, people will do that anyway, but I'm not sure we should cater to those people when the health of the markets as a whole are at stake (assuming circuit breakers are effective; up for debate).

Because bulls and bears aren’t equal. The laws should be favored towards not destroying the economy needlessly.

Because extremely bull markets are a strong indicator of a bubble the laws should be equal to prevent bubble bubble bubbling.

Are they? The market reaches all time highs almost every year, so would you say that we spend most of our time in bubbles?

It is actually expected that the market will typically go up, not down. Otherwise there would be no difference between investing and gambling.

It's a question of velocity - markets go up and down all the time... but speed breaks to prevent the market from going down to fast also need to prevent the reverse case - some sudden inexplicable and extreme price spiking.

Basically - we should protect against any really dramatic sudden changes as chances are that something is wrong.

In the last year the S&P grew ~22% prior to this recent drop, that's fine... and on a day swings in the low single digits might be fine. But if we woke up tomorrow and the S&P had recovered the pre-drop value and then grown 20% over that price - something weird is clearly going on and we should be worried.

This isn’t the economy

You can easily sell a market to zero, however it is much harder to buy a market to infinity.

* It's very difficult to sell a market to zero, but impossible to buy it to infinity.

Has any stock or financial instrument ever been valued at infinity? I don't know enough about finance to rule that out and it seems unlikely but at least possible.

For instance here is a toy example of unbounded value: Consider a market for a stock in which the only market participants are two bots with the behavior that they trade the stock back and forth at an asymptotically increasing price. The buys are backed by loans from a zero interest government bank. Since the bank is efficiently hedged to losses. Both sides of the trade owe the bank, the seller can also pay the bank the money the bank needs to back the buyers loan. Thus, the bank could allow both these loans to go to infinity causing the value of the asset to approach infinity.

Similar things have happened with flash loans in the cryptocurrency space.

At what point does a finite increase of the price turn it from a finite number into infinity?

No. Money is a scarce resource and modern algorithmic trading has limits to prevent another LTCM.

What is the price of a share that isn't listed for sale?

Emptying out all the asks is a thing that happens.

Why would you need them?

> Only a single L1 or L2 breaker can occur in a single day.

I don't follow. You'd hit 7% before 13%, so how would L2 ever execute?

Sorry - I meant that neither L1 nor L2 can trigger twice in a day, E.g. down 8%, up 2%, down 9% does not trigger a second L1 breaker.

> neither L1 nor L2 can trigger twice in a day

This phrasing makes sense. Thanks!

Especially relevant today since after hovering around -5% for most of the session after the first halt, right now we're at -6.9% and threatening to push below 7% again.

This means each execute once, not mutually exclusive.

I'm pretty sure these need to be hit consecutively.

As I understand GP: L1 and L2 can each fire only once per day. So if you hit L1, and after a brief rebound the market still goes down, you'll hit L2. If after that it still keeps falling, it'll get stuck at L3.

We hit 7% earlier today. If we hit 7% again today, nothing will happen. If we hit 13% we hit L2.

L1 hit today and there was a 15 minute halt this morning. If the S&P continues to plunge and hit -13% during today, there will be another 15 minute halt.

It's only a 15 minute halt. So if you hit 7% and trigger L1, trading is halted for 15 minutes, but then it resumes. If it drops down to 13% another 15 minute trigger happens.

But each of the L1 or L2 can only occur in a single day, not one or the other. You can hit L1, and then later hit L2, but you can't hit L1 twice.

Does that help clear things up?

I thought these circuit breakers actually existed to stop algo's from going haywire?

They have existed since the crash of 1929. There were no algorithms in place then. The goal was just to force human traders caught up in the moment to take a break and stop panicking.

I assume they have been tweaked since 1929, but they started with the crash back then.

Edit: someone else is claiming 1987 as the start. My memory says 1929. If this matters do your own research.

https://www.investopedia.com/terms/c/circuitbreaker.asp

Well something was put into place after 1929...

> Regulators put the first circuit breakers in place following the market crash of October 19th 1987, when the Dow Jones Industrial Average (DJIA) shed 508 points (22.6%) in a single day. The crash, which began in Hong Kong and soon affected markets worldwide, came to be known as Black Monday.

> In 1933, the U.S. Congress passed the Glass–Steagall Act mandating a separation between commercial banks, which take deposits and extend loans, and investment banks, which underwrite, issue, and distribute stocks, bonds, and other securities.

https://en.wikipedia.org/wiki/1933_Banking_Act

It was repealed in 1999.

https://en.wikipedia.org/wiki/Aftermath_of_the_repeal_of_the...

Seems like that kind of thing would be implemented on the algorithm side, not the market.

You'd hope so, but why not have protection on the other end as well?

It probably doesn't make sense for an individual actor to stop trading when all the assets it holds are tanking. Everyone's selling to cut their short-term losses. Crashes aren't always feedback loops.

No one would self-impose the limit on themselves, because you could profit if you trade while everyone else isn't trading.

You do not want to trust your market's health to some random trading firm's coding skills and goodwill, just like you wouldn't open up a public webservice without some basic rate-limiting.

They exist to stop any participant, human or electronic, from going haywire.

"Let's all take a deep breath, y'all."

> Only a single L1 or L2 breaker can occur in a single day.

Then why does L2 exist? It seems redundant, L1 would get triggered before L2, and only one can occur in a single day, so why have L2 at all? What am I missing here?

Edit: Also, anyone who downvoted me for posting the same question as someone at the same minute, you know what to do.

See my other response to a similar comment - I will edit the original.

Makes sense, thanks.

No matter how much you didn't deserve downmodding, complaining about them only invites more. You have a better chance of recovery by not bringing them up, and the News Guidelines ask that you don't.

Please don't comment about the voting on comments. It never does any good, and it makes boring reading.

https://news.ycombinator.com/newsguidelines.html

Your paragraph was a lot more annoying than his sentence.

It's true that such comments are annoying, but they're necessary as a feedback mechanism to regulate the site. Otherwise the site guidelines would have negligible effect.

They're more tedious to write than to read, if that helps at all.

https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...

Okay thats fine, but who implements/decides these circuit breakers ? And what purpose to they serve only to limit a mass sell off ?

Nice point about only 2 daily. But still seems crazy.

AFAIK people can still trade in private, just not on the NYSE.

So don't worry, it's only us regular guys that get screwed. Big firms can still contact each other to make deals.

How exactly are you getting personally screwed by a 15-minute stop in trading?

> So don't worry, it's only us regular guys that get screwed. Big firms can still contact each other to make deals.

And when everyone is selling, who do you think is buying? No one is moving much of anything for those 15 minutes

How do you sell stock if no one is buying? If people are selling, others are buying.

Market makers are buying - their job is to always buy or sell stock from everybody. They (generally computers, but humans traditionally) will always buy your stocks, or sell you stocks. Their algorithm is simple: buy for $.10 (or some other tiny number) less than you sell - if the amount of stock owned is too low raise the price, if the amount is too much lower the price. They pretty much always make money in the long run.

For agreeing to do this (and having the capital to do it) the Market Maker for a stock typically secures certain benefits from the market in respect of that stock.

For a very popular stock on a typical day the market maker isn't really necessary. Your trades would absolutely execute immediately based on positions other people wanted.

When your stock is more thinly traded, or when things are a bit frantic the market maker is your saviour. When everybody and their dog is selling, the market maker will buy anyway.

Under some circumstances market makers can signal they intended to cease to make a market for specific stocks. When the market makers exit, all hobbyists should make sure they are gone too. Once there is no market maker for the stock you're holding, you will need somebody else to actually take the other side of your trades. "Prices" without a market maker are just a guess, there may be nobody actually promising to take your stock at any price, even if the last trade was for $1.40 your stock might be not sell even at 14¢. This makes for an exciting space in which to gamble with money you can afford to lose if you really know what you're doing, otherwise it's just a way to throw money away.

Having worked for one of the (smaller) big firms, and had a chance to watch from the sidelines and seen how days like this work: No, these circuit breakers don't screw the regular guys. They discourage the regular guys from screwing themselves.

There are a lot of firms whose core business strategy is to keep a level head and take advantage of people who panic and (over)react on days like this. They get damn rich doing it, too.

> They discourage the regular guys from screwing themselves.

If you're a little guy who believes that, say, 2019ncov is about to tear the world a new asshole, that's probably a decision you'd like to make for yourself.

I don't doubt what you're saying, but it's a matter of perspective. Sometimes the "panic" is the correct reaction. We're sitting on top of a perfect storm which is shaping up to be a massive potential black swan. And with the current circuit breakers, trading is only interrupted for something similar in concept to "the 99%".

Across all indicators, too! Oil price, US markets, international markets, t-notes, gold price, and a bunch I'm probably not aware of because I'm not a professional investor. China just shut it's economy down for two weeks. Long term outlook is rightfully poor.

> Sometimes the "panic" is the correct reaction.

Panic is almost never the correct reaction. Deciding that there's going to be a downturn and you should prepare yourself for it financially is one thing, but under what circumstances would it be optimal for you as an individual to panic-sell?

In my mind, panic is the thing you do when you realize that you haven't prepared. That you don't have enough resources for an extended downturn, that you're financially over-leveraged, and that you are in danger of losing your home and being unable to feed yourself and your family. There are TONS of people who are experiencing this right now in China and Italy, and many others of us who will be experiencing it the next few weeks in the rest of the world.

But panic also implies that you don't have time to fix that, and there's nothing you can do as an individual to change it. If that's the case, you probably don't have a lot of investments in the stock market anyway. Or if you do, and you're over-leveraged in the markets because you were gambling with your money instead of investing with it, then yes, you're panic-selling right now.

Panic sell is the correct thing when your doctor gives you a short time to live. Since your money will be worthless no matter what happens you may as well get it all now and spend it on whatever can buy a moment happiness.

There is an exception if your religion lets you take your stock with you. I don't believe in one, but I guess if you want to.

You can take that position after we all wait 15 min and take a breath.

If you’re a little guy that truly believed that, you wouldn’t have waited until the panic to happen to sell then.

But that exactly is one of those fallacies. If 2019ncov "tears the world a new ..." then your shares are going to be worth about the same as bills in your hand: nothing. Nobody will seriously risk their health for money.

And what's the rational thing to do in such a case? Let's say you're at a poker game. And you have a deal: you lose, you get shot, you win, you get your winnings. What is the rational thing to do?

You should go all-in. It is one of the very few cases it is actually rational to do that.

Nope, trading stops everywhere. Even Canada stops trading when the US has a market-wide halt.

The idea that the little guys largely have a chance actively trading in a fast moving market like today’s where they are gonna be crushed by algos is ridiculous to begin with.

If anything this is helps the little guy, by not completely crushing their stock value, and hurts the big guys who can outlast huge swings (something little guys cannot).

https://www.nasdaq.com/articles/market-wide-circuit-breakers...

Really it's just saying "the market is now closed". But you can still sell those securities on other markets or direct to someone who wants to buy or sell.

Mass sell offs tend to create unorderly markets, which is not beneficial for anyone.

The concept was introduced in US equities after the ‘87 crash, but was only consistently implemented for NYSE-listed stocks. In ‘13 these were made consistent and market wide (thus MWCB), set against a widely published value of the S&P (so that the control was predictable; thus how it executed today).

FYI, there are also bidirectional halts that exist intraday (Limit-Up/Limit-Down) that serve a similar purpose and control rapid, uncontrolled movements in individual stocks. These are also defined in exchange regulation and are well defined so that they are predictable.

> Mass sell offs tend to create unorderly markets, which is not beneficial for anyone.

It's beneficial for people who want to buy stocks cheaply or if we truly believe that markets are about price discovery.

> markets are about price discovery

Exactly, and when you have an unorderly market, price discovery becomes problematic.

It’s the same reason the single stock LULD bands exist (which put the brakes on both rapid downward and rapid upward movement). Stopping for 15 minutes (or 5 in the case of a LULD pause) is not detrimental to the process of establishing orderly price discovery.

In the event a stock is going to keep rising or falling due to legitimate changes in valuation it will continue to do so (look at NASDAQ’s halts page today to see stocks that have hit their bands multiple times).

> if we truly believe that markets are about price discovery

I think you can discover the price tomorrow.

> we truly believe that markets are about price discovery.

this is not incompatible with the believe that prices over a single day (or hour) can be dangerously noisy.

> Mass sell offs tend to create unorderly markets

So are we saying the market will be perpetual because it's not allowed to fail ?

At what point does the https://en.wikipedia.org/wiki/Pareto_efficient not apply ?

If it's manipulated, the efficient seems moot.

Pareto efficiency does not apply, because it is a theoretical construct that assumes perfectly rational actors. It's useful for thinking about the market, but not an actual law.

(For a more humorous statement: https://www.youtube.com/watch?v=oap6_U8-HvI)

The exchange does. Exchanges are companies too.

Reply on news.ycombinator.com