This page has been archived and commenting is disabled.
Was HFT Responsible For Investors' Massive Dendreon Losses?
Yet more dusty footprints of High Frequency Trading's shadow domination of markets emerges, courtesy of Matt Goldstein's latest column "The victims of high-frequency trading." Matt focuses on the blatant example of a self-perpetuating myth, driven exclusively by reinforcing algorithms in the April 27 collapse of Dendreon stock, which dropped by 70% in 70 seconds. Bloomberg did a good summary of the events that cost many DNDN investors hundreds of thousands in losses, and which now, five months after the event, the regulators have still not disclosed any additional data to bring this presumably aberrant event to closure.
More than 3 million Dendreon shares changed hands as the
stock fell from $24.25 to as low as $7.50 in 70 seconds, before
trading was halted at 1:27 p.m. in New York, according to data
compiled by Bloomberg. In the first 20 seconds of the sell-off,
more than 260,000 shares were sold at the so-called bid price,
or about 60 percent of the total, showing investors were willing
to accept almost any offer to unload the stock.
Matt provides a different angle on the story: that from the rightful angle of the primary culprit of this dramatic and exacerbated drop in which market makers were supposed to step in... but did not, because they were simply a bunch of computer programs that had no mandate to provide liquidity.
There has been speculation that short sellers, traders who look to profit from the stock’s plunge, spread a rumor that Dendreon was going to report poor test results for its cancer-fighting drug. Others theorize that a broker incorrectly typed in an outsized sell order, which panicked others in the market.
But no matter what the precipitated the sell-off, it’s likely that high-frequency trading magnified it-given that these automated trading programs control more than half of the daily stock trading in the United States.
Some of the algorithmic programs that drive high-frequency trading desks are designed to spot an unusual trading trend — such as a sudden decline in a stock’s price-and jump on it. Other programs, meanwhile, are written to automatically cancel bids to buy fast-falling stocks in order to minimize losses.
“If an HFT guy steps away from a stock, that can drive it down,” says Joe Saluzzi, a co-founder of Themis Trading in Chatham, NJ and an outspoken critic of superfast computer-trading. “It’s not necessarily the shorts pressing a stock down, it’s also because of bids disappearing.”
It’s all perfectly logical from a trading perspective. But when these two strategies come together, it can create a vacuum-like force that allows a stock to plunge in a short span of time. This is the kind of unintended systemic shock to the markets that has got critics and even some advocates of high-frequency trading nervous.
Goldstein's cautionary conclusion confirms the same concerns that Zero Hedge has been voicing for months about potential risks associated with the reckless (and riskless) propagation of quantitive strategies, which can shut down on a moment's notice, creating one systemic Dendreon repeat for the entire market.
Yet it’s not clear securities regulators are sufficiently worried about the potential systemic risk posed by high-frequency trading. And that should worry everyone.
When regulators talk about high-frequency trading they often focus on seemingly obscure things like whether traders should be able to put their computers close to the stock exchange to maximize trading speed, or buy and sell shares through less-than-transparent “dark pools”. These are all important issues, to be sure.
Yet they pale when compared with the threat of a high-frequency trading program sparking a sudden and inexplicable sell-off in a stock.
That’s why it’s imperative for regulators to come clean with what, if anything, they know about the events that led to the April 28 debacle in Dendreon shares.
It’s been five months since that event, and investors are entitled to answers.
We agree.
»
- 6635 reads
- Printer-friendly version
- Send to friend
- advertisements -
Follow this example to its logical conclusion and you will arrive at the reason why Vladimir Putin is BY FAR the richest man in the world (please, anyone who disagrees is simply a fool). Pay attention people.
Friday morning. August 2000. Most traders already in the Hamptons. Bogus press release picked up by Bloomberg. Emulex drops from 113 to 43. Took about 5 minutes.
http://www.museumofhoaxes.com/hoax/Hoaxipedia/Emulex_Stockmarket_Hoax/
Here is one from your link over a 140 years old. The Civil War Gold Hoax. When they caught the guy he only got three months. Apparently the false rumor turned out to be true a few months later.
http://www.museumofhoaxes.com/hoax/archive/permalink/the_civil_war_gold_...
What? Our regulators haven't done anything about obvious crimes? Unthinkable, I tell you!
Didn't this stuff cause the 1987 crash?
Derek
Check the drop of ELAN on July 29-30 --- After hours
I remember Emulex... the company put out its own press release and by the closing bell everything was back to normal... and by the end of the week the perpetrator was arrested... things can work quickly when certain individuals want something to get done...
When the profit (most would say crime) wasn't directed by and beneficial to the "made" boys of Wall Street, it's truly breath taking how quickly the SEC, FBI etc are on the job protecting Wall St......er....Americans from those bad bad people.
I lost thousands on that event since my stop was triggered. My broker offered some lame excuse for a reason and no recourse.
It really sucks but I feel you, stop orders afford no protection from moves of this type. The vanilla stop order puts in a market order to sell once a threshold has been crossed...but if the stock is falling like a rock the fill price can be substantially lower than the stop. One can also use the stop-limit which puts in a limit order to sell at the threshold price once crossed...but if bids dissappear then your order will never get filled and you'll be riding that thing down to the bottom unless you can adjust course in a matter of seconds.
Individual Investors 0 - Algos 1
don't even mention the market-open gaps... up or down. stops be damned - no safe haven to be found.
i *really* hate it when it gaps down, triggers our stops, then recovers, after we've been shaken out... just kills ya.
Just like the Fed sees all the banks' books and thus can game the markets, don't you think GS and friends (with a little help from the ultimate friend the Fed) might have HFT software algos to take advantage of stops, which are usually set at certain technical levels?
Regardless of your thinking about how or even if technical analysis works, if enough people follow it, it works simply because enough people follow (and act on) it. Which means GS and friends are on it.
Two of my favorite "v-shaped" movements rescued by the algos this summer were BAC on the day the SEC settlement was announced and PharMerica when the FDA supposedly raided them.
Steep multi-percent drops off a cliff which were buoyed just as quickly and recovered to a higher price than before the fall...all within maybe a half hour.
Comprehensive detailed names named report on organized hedge fund criminal conspiracy to destroy DNDN, HFT just one part of the MO.
Michael Milken, 60,000 Deaths, and the Story of Dendreon (Chapter 1 of 15)at: http://www.deepcapture.com/michael-milken-60000-deaths-and-the-story-of-...
Slightly OT, but also on this subject: Taibbi has another great piece in this month's Rolling Stone (no web version yet). Even more evidence that our regulators are asleep at the wheel...
They bust these erroneous trades anyway do they not? I have had many trades busted on me where I had bids and offers away from the market by too much. I understand the point is the lack of liquidity but if the trades are busted is there really hundreds of thousands lost by investors?
The DNDN trades stood.
They did not bust the trade and I tried multiple times to get it reversed and they said it was 'legit'.
None of the dndn trades were busted. Which was the real shady part.
Copnsidering that only 118 million shares of SPY were traded today, could it be next? All they have to do is stop bidding. The rest will be history.
> ... But it’s a bitter pill to swallow when you consider that Dendreon shares quickly rebounded from their pre-crash level after the company reported generally positive test results and trading resumed
Damn those HFT machines! They drove the price of the stock back up to non-rumor level prices as soon as the rumours were debunked.
I fail to see the issue here WRT computers. The SEC should investigate if someone shorted and spread a rumour of a failed drug trial. Whether it was computers or J6P selling after that rumour is not relevant.
well well well; look who crawl out of his GS subsidized PR hole. welcome back !
Cheeky is en fuego.
Wow. Great argument. The average IQ on this site just dropped about 50 points.
If you're going to dis an article, at least use some logic. Or wry humor. Or both.
But to try and slip the premise by us that since the price returned to where it started, in effect saying "move along, nothing to see here, all is good and nothing illegal happened," is just plain stoooooopid.
That was not my point. There was of course something to see.... but I don't understand why anyone is focusing on the computers.
Rather - humans and computers on a baseless rumour together moved a stock price, and then when the rumour was debased, they move the price back.... so the culprit is whomever created the initial panic with a false rumour.
Now - if one of you believe that the computers sparked this rumour, then I see your point about blaming this issue on HFT (but suggest you go back on your meds).
Otherwise, the issue here is a plain and simple human initiated stock manipulation.
History is rife with examples that predate the invention of the computer (let alone using them to trade) with movements of that magnitude in equity pricing... and single (or small number) product bio-tech companies in development or testing are always going to be the whipping boy of rumours, regardless of who is trading them.
If anyone here believes that a human specialist quoting in fractions would have maintained an orderly book for this stock in the wake of that rumour, I suggest you are wrong, and that you read Edward Chanellor's excellent book Devil Take the Hindmost, which chronicles several hundred years of speculative market activity.
Oddly and on a slightly different thread - I doubt that retail investors short this stock who lost their money on a floated rumour of a successful drug test would have be worthy of an article... since we are only supposed to have sympathy for long only investors.
The point is that the computers did not provide liquidity as touted by the trumpeteers....and in fact contributed to excessive downside. Your argument is like saying since Jesus Christ says take my body and blood...he is a cannibal.
70 seconds? By 2020 they can do that kind of a dump in 28 seconds flat.
Very detailed story about DNDN
Cramsters right in the thick of it a usual
See that job he did with CIT a couple weeks ago?
Boooooooooooooyaaaaaaaaah!!
I think Jimbos going to get something real sweet to front run soon
Booooyaaaah 2
http://www.deepcapture.com/michael-milken-60000-deaths-and-the-story-of-...
Waldo
Someone warned about this an hour before...
http://messages.finance.yahoo.com/Stocks_%28A_to_Z%29/Stocks_D/threadvie...
I remember the day well as I owned 17k shares, but fortunately, no stops in place. This was a planned bear raid perhaps facilated by HFT. Our regulators are captured.
So the HF market makers are bad because they might not voluntarily supply liquidity. But, it's also bad that they are providing 70% of the liquidity during the average day. Thus they should both trade more and trade less. Crystal clear argument.
Hey, is that a windmill over there?
please don't try to reason with the mob - just pick up a pitchfork and get in line.
Check out Patrick O'Byrne, Mark Mitchell, Bagley et al. on the Dendreon story. http://www.deepcapture.com/
DNDN dropped because its a highly speculative pharmaceutical company and a case can be made that it is tremendously overpriced. A few hedge funds were bidding it up at the time and if they decide not to hold it up or dump their shares then thats what happens. IT HAD NOTHING TO DO WITH HFT!!!!!! I knew quite a few traders that put in bids because we thought it had dropped too much but it doesnt matter if there is a huge seller. STOP FUCKING BLAMING HFT FOR EVERYTHING. DNDN has NEVER had a profitable year EVER. You play with fire you may get burned.
“Yet they pale when compared with the threat of a high-frequency trading program sparking a sudden and inexplicable sell-off in a stock” … A STOCK … what if the whole market has simultaneous sell off precipitated by some external event or black swan sighting?
As newly created 'program trading' was a contributing factor to October 19, 1987, newly created HFT could inflict instant death … Beware 10/26/2009
"Others theorize that a broker incorrectly typed in an outsized sell order, which panicked others in the market." wait - so the risk IS that traders fuck up and fat-finger outsized orders - NOT that Skynet (HFT) becomes self aware - right? right... i thought so...
You guys are really fucking annoying; if you don't like this place, and with all the evidence provided, and with the ( i hope ) legislation in place, still make praise of HFT, then why the fuck do you even bother with us un-educated non-HFT peasants. Why don't you just fuck off to Jim Cramer's web site and scream booooooooyah all day long ?
I am not convinced that HFT is a problem. All the evidence? Are you kidding me? Provide one documented example of the supposedly rampant HFT community ripping Joe Q Public off over and over. I'm not interested in a wet dream of "the machine bids a penny higher, then gets filled, then takes all the offers, then gets taken on the offer, then gets short on the offer, then hits all the bids, then gets filled on the bid...repeat." That example, which is the only example I've seen spouted off, really, is so unrealistic as to be laughable. Whoever thinks that's the way stocks trade, that a machine can magically get itself hit on the bid and taken on the offer, has NO IDEA what they are talking about. Explain to me the exact mechanism by which they are cheating. The EXACT mechanism. I trust you have vast knowledge relating to inter ECN routing practices and execution.
"you guys" ??? who's that Cheeky - anyone who disagrees with you or doesn't write "AWESOME ONE CHEEKY!" or "CHEEKY IS ON FIRE!" in response to your comments?
I am confident that the comments i leave on ZH posts will make anyone who takes 30 seconds to read them and THINK about them smarter and more educated on the subjects at hand. And that's important, because the mob mentality overrides the common sense, and it gets dangerous.
The point of my comment above was just to point out ANOTHER Inconsistency - as ZH has another post up yesterday about how there is so much risk that algo's can go crazy and have an error impacting stock prices, when AGAIN, this post explains that it's the humans that have errors - not the computers.
I know all of "you guys" hate HFT - but to use DNDN as an example is patently absurd.
KD +10000!!!
</echo_chamber_stupidity_sarcasm>
Had computer troubles when this was posted yesterday.
It is not absurd to blame a 70 second drop on HFT and as a defender of HFT I will say that this is what happens if you don't have the proper speedbumps.
Forget any illusion that floor brokers on the NYSE would step in and provide liquidity in the middle of a plunge. Not how they got those houses in the Hamptons. A post '87 crash investigation found that they did not provide liquidity depsite the mandate.
Equity futures used to have a 2% downside limit. Then you had a trading halt and expanded the limit. SPY made theirs 5%. Exchanges got the CFTC to change the limits to match @5%.
When thing get faster you need better speed bumps. HFT should wise-up and get on board. You are much more likely to get price limit changes than you are to eliminate HFT all together.
DNDN was coming out with news the next day, the stock had had a massive ramp since early April. Perhaps all of the market makers who had to go short to make a market in it on the way up decided they had better cover their arses in case the thing really took off. I watch my stuff like a hawk and use mental stops. Figured it was a head fake and didn't sell. It paid off that day.
Forget about them doing anything. At the same time HFT is on hiring spree. See the solicitation I got from the recruiter. Read carefully the words bleeding, latency etc...
Last time I posted, Marla decided (I think) not to post, hence reposting it...
You've received this email because the skills in your resume matched our search criteria for a Senior C++ Developer in our database. It is possible that you may not be best suited for this particular position, but we have multiple positions available in all areas and levels of IT where you may be interested and better suited.
Please respond if you are qualified, available, interested, planning to make a change, or know of a friend who might have the required qualifications and interest in working with us as we have an excellent referral program.
Below is the job description to which I am referring for a current job opening located in NYC (midtown)
Client: Global Investment Bank
Location: Midtown (NYC)
Type: Contract (corp or w-2)
Duration: 12 mo+
Rate: Pay range from $650/day to $725/day based on corp-to-corp
Status: Immediate phone/onsite interviews
Job Description:
This senior software developer position with Enterprise Application Infrastructure (EAI) group of Brokerage includes responsibilities for developing and supporting core C++ application toolkits and components for use throughout the business.
The focus of the role is the development of the C++ libraries which are used to build applications across the firm, the firms next generation C++ infrastructure, as well as further developing the functionality of our existing systems
The successful candidate should have performance optimization experience, in-depth knowledge of low-latency networking and multicore scalability.
The products developed by EAI are used across the firm. This means that the role requires a developer with strong technical skills and the experience to understand the implications of their design decisions and good understanding of current best practices in application architecture, software design and code development.
The main responsibilities of the role include:
- Design and development of C++ libraries in collaboration with global team.
- Maintenance and support of existing C++ libraries.
- Identification of current technology gaps and potential solutions.
- Identifying the 'bleeding-edge' technologies that will form the basis of the firms next-generation systems.
-Interact with internal clients to on component requirements
Skills Required:
Extensive knowledge of C++ on Linux and/or Windows.
- Deep understanding of multi-threading and multicore architecture as well as low-latency networking issues.
- Track record of multi-tier application design and development in C++.
- Experience building re-usable code.
- Strong problem solving and analytical skills; Broad interest in technology.
-Strong communication skills
Skills Desired:
Experience with tuning-up OS kernel is a plus.