Flash Crash Analysis of May 6 Stock Market Plunge 411
Jamie found an interesting site that has many charts and graphs about the strange May 6 stock market plunge and rebound. There's a lot of information to consume over there, but it does a pretty good job of showing high-frequency trading is getting to be a real problem.
Simple Fix (Score:2, Interesting)
Look at average frequency of trading for the two years leading up to the Flash Crash, and set that has the upper limit.
Summary... (Score:5, Interesting)
The problem with this is that since it has happened once, it *is* going to happen again in a slightly different way. Software glitches, fat fingering the keyboard, etc. are convenient excuses.
How is this a problem? (Score:4, Interesting)
Old way: 10,000 trades a day, every few months or years the market dips for a few months and rebounds, every several years the market enters a deep recession for years.
New way: 10,000 trades a second, every few weeks or months the market dips for a few minutes and then rebounds, every several years the market enters a deep recession for years.
The artifacts of trading become more temporally frequent and more temporally limited; the artifacts of real economy (growth-recession cycle) don't change.
Re:It should Flash Crash to about 5000 (Score:2, Interesting)
You're not too good at math are you?
1.5*6800 = 10200
Current market value: 10187
Yeah, I'd say he's decent at math, if you account for rounding errors.
Re:The next chaper of this story (Score:3, Interesting)
Those which were at a 60% discount or greater, yes. And it's serious bullshit.
ZeroHedge had a discussion on the Nanex report... (Score:5, Interesting)
...right here [zerohedge.com]. One commenter had some interesting things to say about "quote stuffing":
So it might not be the big advantage that Nanex sees it as.
Real problem (Score:5, Interesting)
If I'm going to gamble, I'm going to Vegas. At least there, you get free drinks as you piss your money away.
more sensible trading (Score:4, Interesting)
I have always thought that if you buy stocks that pay dividends you should have to keep them for a minimum of a year, and if you buy stocks that don't pay you should have to keep them for a minimum of 24 hours before being able to sell them. If you short stocks the sale should take 24 hours to take affect.
I mean really, what could happen in 24 hours that would honestly affect the value of a publicly traded company? Even the oil spill has taken weeks and months to lower BP's value, and really in the long run it is probably time to buy BP stock.
Re:It should Flash Crash to about 5000 (Score:1, Interesting)
Re:How is this a problem? (Score:3, Interesting)
ya if anything this was a strong supporter of high frequency trading. The market corrected before the vast majority of people were even aware there was a stock market dip. What caused the initial dip is well outside my area of expertise (since I don't follow the second to second activities of traders), but after it did happen, that it went back to be in line with it's roughly steady state seems like the system is work.
High frequency trading allows rapid price normalization between exchanges, which is good, and while the rate of the fluctuations is faster, that doesn't necessarily change the average value over time.
Re:HF Trading reduces spread, increases liquidity (Score:3, Interesting)
It is no accident that this story comes just as Wall Street reform bill it being released. This story is just a little lube - brace to be shafted as we see what the reform bill holds. Dug up some references for above:
http://www.amazon.com/Wall-Street-jungle-Richard-Ney/product-reviews/B00005X4AX/ref=dp_top_cm_cr_acr_txt?ie=UTF8&showViewpoints=1
http://books.google.com/books?id=5eICAAAAMBAJ&lpg=PA37&ots=JFKfDQyI8G&dq=%22richard%20ney%22%20conspiracy%20theory&pg=PA37#v=onepage&q&f=false
http://www.sec.gov/news/press/2004-42.htm
http://www.usatoday.com/money/markets/us/2005-12-06-nyse_x.htm
http://www.cfo.com/article.cfm/3903963?f=related
Re:How is this a problem? (Score:4, Interesting)
Well, I would suggest reading the article. Basically, there are two problems identified:
1) NYSE has a bug in their system which caused a crash.
2) High volume traders are effectively launching denial of service attacks through their otherwise unnecessarily high number of quotes per second ( as high as 5000 per second). The only reasoning besides stupidity is that they are trying to force their cometitors to burn though their cpu analysing the quotes, allowing the quoter to have a temporary advantage in analysis and trading.
Re:How is this a problem? (Score:2, Interesting)
Another note, the billions of dollars "lost" during this flash were only felt by those that invest for the moment. The long term investor did not even notice this blip. IMHO if you play the investment game and are trading on the news of the microsecond you are very well aware that some news is FALSE. Your choice to act on information is your CHOICE. You either make some money or lose some money.
Re:HF Trading reduces spread, increases liquidity (Score:5, Interesting)
High Frequency Trading is _beneficial_ to the public markets at large, and why powerful interests keep blaming and attacking electronic trading as the root of all financial evils that befall us: http://www.tradersmagazine.com/news/high-frequency-trading-benefits-105365-1.html?zkPrintable=true [tradersmagazine.com]
High frequency trading is not beneficial when it shaves pennies and acts as an intermediary between a buy & a sell that would have executed anyways.
Your article discusses exactly that in the section called "Accelerating Price Discovery Benefits All Investors"
And they have the nerve to try and make it sound like a good thing. From your FA:
Some critics, however, still maintain that, while greater liquidity is valuable in theory, market participants with large orders (and the individual investors they often are representing) are nevertheless harmed by high frequency trading. In simple terms, these critics assert that when, for example, they try to purchase a large quantity of IBM shares, high frequency traders detect the buying interest and cause the price of IBM to rise before they can finish purchasing all the shares they desire.
The problem with their counter-argument is that it ignores the fact that the HFT is buying up all the shares you could have bought and then selling them to you for a higher price.
You, the original buyer get screwed and they, the original sellers, gain no benefit.
Then they try to stretch their asinine argument about trades that take milliseconds to execute,
into the realm of land sales which often take days if not weeks to arrange and execute.
Intellectual dishonesty at its finest.
Re:It should Flash Crash to about 5000 (Score:3, Interesting)
"Why do you believe that the companies should be worth 25% less than last year?"
Because they won't be profitable during the recession? Because investors need money for other purposes and have to sell shares?
There are lot of reasons. Stock market is not 100%-rational, but it's also not completely irrational.
Re:HF Trading reduces spread, increases liquidity (Score:2, Interesting)
Re:HF Trading reduces spread, increases liquidity (Score:3, Interesting)
Re:The next chaper of this story (Score:2, Interesting)
Zero Sum is NOT productive... (Score:5, Interesting)
All this HFT stuff is zero-sum, if someone makes $10 on HFT, someone else loses $10.
HFT is a market parasite at this point and, IMO, ALL quotes should have a randomly induced delay between 0 and 1 second (with the delay being DIFFERENT to different participants), to eliminate the advantage of high frequency trading.
Re:HF Trading reduces spread, increases liquidity (Score:5, Interesting)
As the FA points out, you are arguing in favor of inefficient markets. You want to be able to buy up all the shares without anybody being aware of the increased demand. HFT buys and sells in microseconds, at most minute time-frames - reducing the spread in the process.
I guess that depends on your definition of "inefficient."
Lets say you're at the supermarket.
You reach out your hand to take [product] off the shelf,
by the time you reach out to take another [product], the shelf is empty!
A HFT saw your first signal and then swept the shelf clean,
bought all of [product], and is offering to sell [product] to you at a markup.
Oh, and the HFT has done the same thing at every other supermarket you would visit.
Has the market been made more efficient?
Or is the HFT behaving like an anti-social asshole?
I'd say the answer to both questions is "yes,"
but that the needs of society outweighs the needs of the market,
which is why we have farking regulations in the first place.
You want us to believe that a few microseconds later your buy/sell opportunity at a few cents difference has vanished.
Uhhh... that's exactly what HFTers do.
That's exactly what the parent's article argues is a good thing.
Or do you have another explanation for why they need ultra low ping connections?
Re:HF Trading reduces spread, increases liquidity (Score:4, Interesting)
Efficiency in the sense that the buyer's order was too high (by a fraction of a cent)? and then the difference between the buy and the sell price doesn't go to the buyer or seller. That money just disappears from the transaction.
I guess if you worship at the shrine of "efficiency" then that kinda makes sense. It's the punishment for getting the price wrong. But at some point, at some level of detail, you might want to consider that the efficiency gains are inscrutable.
And at some level (which we might have already reached) the efficiency gains might be outweighed by what would act as a kind of transaction tax paid by market participants to the HFT machine.
This is not a research but ad for API developer (Score:3, Interesting)
The guys have some skin in the game. From their website: Nanex is the creator and developer of NxCore, a streaming whole market datafeed.
Their recommendations are moot too:
> 1. Quote and trade data must be time stamped by the exchanges at the time it is generated. This will ensure delays can be detected by everyone.
This data is already timestamped by all US exchanges. Data delays are mainly due to line delay and application delays that can be measured quite acccurately. Not many people do it properly though. Many does not look at it at all.
For individual long-term investor (5+ years) it does not matter at all anyway.
>2. Quote-stuffing should be banned.
Exchanges have already requirements for designated liquidity providers to quote at NBBO for X% of time (where generally > 90%). Presumably these providers receive some benefits for doing so (lower rates etc.) Do you think it is reasonable to force other people to buy and sell at a certain price? What if retail investor wants to sell his 100 IBM shares at $110 when market is $100?
> 3. Add a simple 50 millisecond quote expiration rule: a quote must remain active until it is executed or 50ms elapses.
Why exclude Alaska/Hawaiii? What if US citizen wants to trade from Baghdad Green zone and it takes 500ms to get signal there via sattelite link? Should we change thange the rules to accomodate that? Or should we allow people to compete by improving their systems?
Logically flawed (Score:3, Interesting)
The flash crash (and high-frequency trading in general) is really only symptomatic of a deeper underlying problem - the modern stock market has no fundamental reason to exist. When the concept of stocks originated, it was a way to own part of a company. Companies paid dividends, and so if they did well then they sent you (the investor) a check in the mail. In that way, stocks could be thought of as investments where the payoff was receiving profits in the mail.
But the stock market has changed into something different and really bizarre. Everybody knows that a company's profits and stability are what drive its stock prices. But why is that? Although a few stocks pay dividends, these days most don't. And to the common investor who holds a small percentage of overall shares, I don't think there's any easy way to _force_ dividends out of a company you own stock in. In a theoretical sense I could buy up enough shares of the company to force them to pay me dividends, but that's not something the average investor can realistically achieve.
That means that the the only payoff possible from my stocks is the money I could make by selling them. This is really strange if you think about it. If I'm never going to see significant dividends from Google, so their financial success or failure should have no underlying reason to affect stock prices. If the ONLY thing stocks are good for is selling them to somebody else, then they have no intrinsic underlying value. They don't pay dividends. I can't take them over to Google HQ and say "here's my share of the company, I'd like to take this office furniture now."
It's like the stock market has changed from a way to invest in companies and share their profits to some strange cult where everybody's drinking the kool-aid and the only people winning are brokerages. People put their retirements, their life savings, into something that has no intrinsic value whatsoever. It seems like the market is essentially dependent on an having ever-increasing influx of new buyers, like a sort of giant distributed pyramid scheme. To me, it's not a question of _if_ the stock market will collapse completely but more a question of _when_. Nothing logically inconsistent can endure forever.
I do question the source though (Score:4, Interesting)
I'd never heard of Zero Hedge. Well it appears the guy who run it (at least according to Wikipedia) is banned from working at hedge fund companies for insider trading. That leads me to question the neutrality of his position. Seems like he's the kind of guy who wants any unfair advantage and as such may argue to keep those in place.
What's the problem anyway? (Score:3, Interesting)
Not to cast scorn on what was obviously a very scary thing for many a day-trader, but for a long term institutional investor, this was barely a blip on the radar. REAL crashes happen for a reason, and you can almost always see them coming, at least in hindsight. Looking back at 2008 and 1929, there's a very clear pattern that indicates exactly what is wrong, and how the likely result came to be. They still can't even figure out what happened this time, and a day later, it's hard to tell why it really matters anyway.
In a REAL crash, there will be an underlying reason. Prior to the crash, a lot of smart people, investing over the long term, will recognize the signs that the stock market is about to become a bad investment, and will either stop investing in it, or it appears to be cost-effective, sell off some of their more risky holdings, and re-invest in something more recession proof. This is what starts the decline (which happens over a period of months). Then suddenly, one day, when everyone really gets all panicy, it'll start dropping more rapidly, the media gets in on the fun, etc.
The May 6 crash, on the other hand, had less to support it. Sure, there was that whole mess in Greece, but they're hardly holding up all the world markets anyway. One possibility is that someone purchased a huge amount of stock at significantly below market price. Since the ticker prices we are all intimately familiar with are little more than the last price something sold at, if enough consecutive trades go through at that same price, it will have the effect of suddenly dropping the market price for everyone.
For your normal investor who ISN'T hitting reload every 10 seconds, they aren't going to notice this. However, the computers will. And everyone who has put in automatic sell orders will get triggered, and those shares will also sell automatically at whatever cost someone will pay for them. Seeing a sudden sell-off, any automatic purchasing might also be temporarily put on hold, increasing the downward trend.
Now, some savvy investors, both long-term and short-term, realizing there is no sensible reason for the selloff, will see this as a rare buying opportunity, and will rapidly start purchasing shares in huge quantities. This will quickly cause the market to correct itself back upwards, until it gets close to where it was supposed to be. The problem solved itself. The REAL problem was investors who put in automatic sell orders, as if the stock in a stable company is going to plummet so rapidly that they won't have a chance to escape before the company goes bankrupt. Check some of the stock graphs for Enron and Worldcom. Those companies both went to pot practically overnight, but even then, you had months to watch the stock price fall, and could sell off at any time you wanted. Automatic sell-offs are just asking for trouble, and are just an excuse to not pay attention to your own money.
-Restil
Waste of time (Score:3, Interesting)
The "analysis" fails to account for what was happening in the currency markets - specifically the USD/JPY market that day. The events in the Forex market preceeded the equities market all day - specifically there was a huge drop in the US dollar (vs Japanese Yen) ten whole minutes BEFORE the S&P plunged. Looking only at the stock market will never let you understand what happened - it was a crisis of confidence not in equities, but rather in the US dollar as a whole.
Pretty easy to get rid of it (Score:2, Interesting)
There's a real easy solution to get rid of flash trading and excessive speculation and get the stock market back to investing...a sales tax on stock sales. I see no reason stocks should be exempt from other "products".
All those big houses use computerized trading to game the system, remember this story?
http://www.guardian.co.uk/business/2009/jul/06/golman-sachs-computer-codes-stolen [guardian.co.uk]
Then some fed prosecutor (I forget who know) let it out that this was bad because "in the wrong hands" the code could "manipulate the markets".
Well, in ANY hands that means it could and was designed to
"manipulate the markets". These too big to fail places get a license to steal, to skim off billions, and when that isn't enough, they still get bailed out, loot gets stolen from everyone else and handed to them. Then they can take that loot and buy bonds and government paper of assorted kinds, get even more money, put everyone else into "debt" to them.
And having their boys in and out of the Fed and Treasury and Sec..naw, that isn't a conflict of interest...
Re:HF Trading reduces spread, increases liquidity (Score:3, Interesting)
No, they both offer value. The difference is in value offered.
The buy-low/sell-high strategy broker takes a risk. At the instant he buys, he is not sure if he will be selling later. However, by buying at that moment, he gives cash to the seller, cash that would not have been available then. The seller benefits: he gets cash now, not later. And, in return for the risk of maybe not finding the right buyer later, the broker takes his cut - which is the reward for added value: the increased liquidity (seller got his money earlier and could use it).
The HFT brokers don't take risks. The way they operate is that they already know there's a low seller and high buyer - it's just that they didn't had the time to make the trade yet. He buys low from the seller and sells high to the buyer immediately, within milliseconds. And therein lies the quantitative difference: the buyers and seller are already on the market and are on the way to meet each other, when the huckster intercepts each and make him an offer right there.
In essence, the guy above is right: the liquidity has been improved - the seller got his money earlier. All of 5 milli seconds earlier. And, in exchange for this wondrous risk (read: 0, since the "HFT bandits" already knows a buyer) and service (5 millisecond more of money in your pocket! Wonders!), he gets a tiny bit of profit.