Join 3,441 readers in helping fund MetaFilter (Hide)


Microsecond mercenaries
March 31, 2014 1:39 PM   Subscribe

It used to be that when his trading screens showed 10,000 shares of Intel offered at $22 a share, it meant that he could buy 10,000 shares of Intel for $22 a share. He had only to push a button. By the spring of 2007, however, when he pushed the button to complete a trade, the offers would vanish. In his seven years as a trader, he had always been able to look at the screens on his desk and see the stock market. Now the market as it appeared on his screens was an illusion.
In an excerpt/adaption of his new book Flash Boys: A Wall Street Revolt, Michael Lewis follows Brad Katsuyama from uncovering evidence of high-speed electronic front-running to the founding of the IEX exchange intended to discourage it. The Wolf Hunters of Wall Street (SLNYT).
posted by figurant (153 comments total) 57 users marked this as a favorite

 
I happened to see a TV segment about this (maybe 60 minutes?) and it was quite interesting.
posted by jepler at 1:49 PM on March 31 [1 favorite]


I just finished reading this, and it's absolutely fascinating.
posted by vibrotronica at 1:49 PM on March 31 [1 favorite]


And see Matt Levine:
In my alternative Michael Lewis story, the smart young whippersnappers build high-frequency trading firms that undercut big banks' gut-instinct-driven market making with tighter spreads and cheaper trading costs....The high-frequency traders make money more consistently than the old-school traders, but they also make less of it
posted by jpe at 1:50 PM on March 31 [8 favorites]


What really baffles me about this whole article is that I had no idea these Wall Street hotshots were so incompetent and/or clueless. "What's a millisecond?" - really? People who don't understand that information takes time to propagate along computer networks are designing proprietary trading systems?

(Great article.)
posted by RedOrGreen at 1:56 PM on March 31 [3 favorites]


Here's an interesting question: what makes Michael Lewis such a consistently compelling writer? What's he put in the water?
posted by shivohum at 1:58 PM on March 31 [7 favorites]


I love Michael Lewis's work in general and I'm very interested in this new book, but I find the opening paragraph impossible to believe:

"Michael Lewis writes that before he began working on “Flash Boys,” he had little interest in the stock market, “though, like most people, I enjoy watching it go boom and crash.” And like most people, he had an antiquated notion of what the stock market was. At the start of “Flash Boys,” his dazzling, troublemaking new work of reportorial storytelling, Mr. Lewis summons that sweet old image of a trading floor full of screaming brokers, slamming telephones and hysteria-inducing ticker tape. Picture that, he suggests. And then forget it, because it’s gone."

The dude started out as an investment bank bonds salesman in the 1980s, wrote about the experience for his first book which was a huge success , and then went on to be a financial journalist writing many articles and several further books on finance industry related issues. He became one of the most high-profile mainstream writers on finance and banking etc. Seriously, he's saying he's never had a big interest in the stock market before and had an old-fashioned view about it until he starting working on *this book*? Either this is really clumsy marketing spin from Lewis & his publisher or the NYT journalist messed something up in a weird way.
posted by Bwithh at 2:04 PM on March 31 [8 favorites]


I neglected to include the 60 Minutes segment in the FPP (I haven't watched it yet), but it's here, including a transcript.

My favorite detail of this is the 60 km coil of fiber-optic cable in a box at IEX designed to slow trading down enough to make HFT useless.
posted by figurant at 2:07 PM on March 31 [13 favorites]


You could find some on LinkedIn, but Wikipedia had never heard of them or their accomplishments.

yeah.... I think this book review must be written by a NYT cub reporter or intern or something
posted by Bwithh at 2:08 PM on March 31 [1 favorite]


Oh wait, no, Janet Maslin is a veteran literary/film critic and book reviewer for the NYT, since like 1977. Odd.
posted by Bwithh at 2:10 PM on March 31


Either this is really clumsy marketing spin from Lewis & his publisher or the NYT journalist messed something up in a weird way.

Arguably, bonds and subprime markets are not stock markets.
posted by pwnguin at 2:13 PM on March 31 [4 favorites]


The dude started out as an investment bank bonds salesman in the 1980s, wrote about the experience for his first book which was a huge success , and then went on to be a financial journalist writing many articles and several further books on finance industry related issues. He became one of the most high-profile mainstream writers on finance and banking etc. Seriously, he's saying he's never had a big interest in the stock market before and had an old-fashioned view about it until he starting working on *this book*? Either this is really clumsy marketing spin from Lewis & his publisher or the NYT journalist messed something up in a weird way.

There's probably some hype in here, but I actually think you could make a case that he's mostly ignored the stock market in his career--from the beginning, in Liar's Poker (I think this was his first work), he talks about the open disparagement the bond traders had for "equities in Dallas." The Big Short is also about people figuring out how to short housing, also not about the stock market. I can actually see how he spent relatively little time thinking about the (perceived) backwater of stock trading in favor of the more exciting worlds of bonds and derivatives.
posted by dsfan at 2:15 PM on March 31 [6 favorites]


Suffered through the 60 Minutes piece, which managed to go the entire time without tipping off to an actual stock trader whether their story had any substance to it. OP's posted paragraph finally makes clear that it does.
posted by ocschwar at 2:22 PM on March 31


And while the brokers often protested that there were no conflicts of interest inside their dark pools, all the dark pools exhibited the same strange property: A huge percentage of the customer orders sent into a dark pool were executed inside the pool. Even giant investors simply had to take it on faith that Goldman Sachs or Merrill Lynch acted in their interests, despite the obvious financial incentives not to do so. As Mike Gitlin of T. Rowe Price says: “It’s just very hard to prove that any broker dealer is routing the trades to someplace other than the place that is best for you. You couldn’t see what any given broker was doing.” If an investor as large as T. Rowe Price, which acted on behalf of millions of investors, had trouble obtaining the information it needed to determine if its brokers had acted in their interest, what chance did the little guy have?

Wow.

How is that even legal? it almost sounds like someone hired a bunch of very smart lawyers to tell them how to run a gigantic pyramid scheme without running afoul of any laws.

"I'm going to take your money, and i'm going to get you the best deal i can. but you aren't allowed to actually see where your money goes"

And it all runs on charisma and reputation of these banks being the "guys who know what they're doing" who are you know, on your side because you're a rich dude too.

I guess it's not as much of a ponzi scheme as a glory hole scheme but just... wow.

No wonder people hate wallstreet bankers.
posted by emptythought at 2:24 PM on March 31 [11 favorites]


you don't have to trade through on particular dark pool and there is data on where the other pools are trading at. Not to mention you don't have to buy at market.
posted by JPD at 2:26 PM on March 31 [4 favorites]


Further Analysis
posted by JPD at 2:31 PM on March 31 [6 favorites]


I really don't understand how this was not common knowledge among market makers and major players like fund managers. This has been in the NY Times multiple times. Years ago, when I was working for the company that automated one of the major exchanges and when they were looking for a new location outside of the city (after 911), we were all sworn to secrecy about that location because all these same people were trying to locate themselves as close to the facility as possible in hopes of gaining the shortest fiber runs. If I knew about it as a 20-something putz in HR, surely these multi-millionaire "geniuses" in charge of billions in assets knew all about this? I feel like this is a case of pleading ignorance after they were almost caught red-handed.

That being said, I'm really looking forward to reading this. I love everything by Michael Lewis I've read.
posted by nevercalm at 2:31 PM on March 31 [5 favorites]


Oh I buy that stock markets were not a special area of expertise for Michael Lewis but it's just odd that he's portrayed as a total newbie in that opening - really, even as just someone who reads the financial news, he didn't know about the mass computerization that replaced the open outcry trading floor?

Oh, and he's certainly done journalism on stock markets and new technology before - here's a piece about the impact of the Internet on stock trading from 2001
posted by Bwithh at 2:32 PM on March 31


It does seem from the article that investors were being actively deceived about the use of dark pools in preference to the investors explicit instruction. How is that legal?
posted by bonehead at 2:32 PM on March 31 [3 favorites]


you don't have to trade through on particular dark pool and there is data on where the other pools are trading at. Not to mention you don't have to buy at market.

The behavior described here, if described correctly, would work against limit orders as well as market orders. In fact, a market order in most venues is still a limit order, albeit one where the limit price is far from the market price.

So you see offers of 10,000 shares of Intel @22 , send a bid of 10,000@22.01 (*), and if the broker dealer you're routing through is corrupt, it or a croney will send an order of (say) 5000@22 before yours, getting your order filled with 5000@22 and 5000@22.01, and if they can pull the same stunt in the reverse direction, quickly rip you off 5$.

(*) If you really really really want 10,000 shares, you can set the limit price far higher, of course.
posted by ocschwar at 2:33 PM on March 31 [1 favorite]


It does seem from the article that investors were being actively deceived about the use of dark pools in preference to the investors explicit instruction. How is that legal?3

No one is trading on a dark pool unless they explicitly choose to. Most retail investors are getting filled internally anyway.
posted by JPD at 2:34 PM on March 31 [2 favorites]


So you see offers of 10,000 shares of Intel @22 , send a bid of 10,000@22.01 (*), and if the broker dealer you're routing through is corrupt, it or a croney will send an order of (say) 5000@22 before yours, getting your order filled with 5000@22 and 5000@22.01, and if they can pull the same stunt in the reverse direction, quickly rip you off 5$.


yeah - but I believe you'd see that you got filled above market if you are sophisticated enough to be using dark pools anyway.
posted by JPD at 2:36 PM on March 31 [1 favorite]


This article seems to use an excess of words to describe issuing speculative front run orders at one exchange based on orders publicly observed at another exchange.
posted by ceribus peribus at 2:36 PM on March 31


Here's an interesting question: what makes Michael Lewis such a consistently compelling writer? What's he put in the water?

He uses reasonable and simple language to explain concepts to laymen (I know a TON of people that read Moneyball and The Blind Side with 0 interest in baseball/football) while at the same time he is also gifted at showing how this seemingly complicated thing he's talking about fits into the broader narrative of the story he's telling and tying it to a more human story.

For example, The Blind Side opens with the wrenching story of Joe Thiesmann's leg snapping to set up something: Why offensive linemen are so valuable and how the game has changed since the 1980s. Not exactly material for a heartwarming movie with Sandra Bullock, but it's wrapped very effectively in the story of why this enormous kid from the projects is suddenly a hot recruit. You have to know why an offensive tackle is so valuable to understand the human story, but you want to read the offensive tackle stuff because it's written in everyday, approachable language rather than the jargon of the trade.

Or compare something like Moneyball with another book on baseball statistics, Baseball Between the Numbers. BBTN is an interesting book about ways to use statistics in baseball, but is fairly dry and contains a lot of math problems. Moneyball is about the use of statistics in baseball but the dry statistical/educational bits about baseball are wrapped up in the story of how Billy Beane, Failed Baseball Player, Outsmarted Everyone Else.

I mean, the Moneyball reactions would be like some of the reactions in this thread. PFFFFT YOU'RE TELLING ME PROFESSIONAL BASEBALL SCOUTS AND OWNERS DIDN'T KNOW THAT GETTING ON BASE WAS IMPORTANT? COME ON! But watch any baseball game and listen to the commentators talk about RBIs and batting average and clogging up the basepaths and you'll find it's true.
posted by Ghostride The Whip at 2:37 PM on March 31 [14 favorites]


Michael Lewis probably read Joseph Campbell at some point in his life.
posted by b1tr0t at 2:39 PM on March 31 [1 favorite]


yeah - but I believe you'd see that you got filled above market if you are sophisticated enough to be using dark pools anyway.

If you can see an offer in the order book, so can everyone else in the market.

If somebody matches that offer before you do, that's life.

If somebody in cahoots with your broker matches the offer because they got tipped by the broker, that's a crime.

Getting one order filled above market is not a big deal. When that keeps happening, that's evidence of something fishy. (Might not be your broker. Might be your broker's broker. Or the exchange. Or a crooked sysadmin. Who knows?)
posted by ocschwar at 2:40 PM on March 31 [1 favorite]


I read a book about HFT a year or so ago, Dark Pools by Scott Patterson. I am kinda torn about this aspect of the economy. On the one hand, it seems like bringing in computers to do the job of matching buyers and sellers in the market should be a win. Open access to these systems also seems like a win. To the extent that front running and shady dealing happens, I suspect the computers make the situation worse, because of the huge prices people seem to be willing to pay for reduced latency. The rise of "algorithmic" trading is tied to this, and also introduces the fun potential for flash crashes. Claims that the algorithmic traders are providing liquidity have often proved to be false, usually the computers are the first ones aware of a problem and they respond by ceasing to trade, causing liquidity to disappear. The fact that the systems are opaque to see how trades are being carried out, and the SEC can't be bothered to regulate them, is the last nail in the coffin.

I don't really get why we need trades to happen every millisecond. Would it kill us if we just ruled that the market-servers would carry out trades once every 5 minutes?
posted by rustcrumb at 2:40 PM on March 31 [5 favorites]


I'm a big fan of Lewis, but IMO the best book on this topic has already been written, Dark Pools: The Rise of the Machine Traders and the Rigging of the U.S. Stock Market , 2 years ago.
posted by stbalbach at 2:41 PM on March 31 [5 favorites]


I had no idea these Wall Street hotshots were so incompetent and/or clueless. "What's a millisecond?" - really? People who don't understand that information takes time to propagate along computer networks are designing proprietary trading systems?

I know a couple of "professional" book "editors" who don't even know how to use the track-changes function in Microsoft freaking Word.
posted by Camofrog at 2:46 PM on March 31 [6 favorites]


This article seems to use an excess of words to describe issuing speculative front run orders at one exchange based on orders publicly observed at another exchange.

This was certainly my impression from the 60 Minutes piece last night as well. My professional background is in futures not equities, so I might be wrong and someone please correct me if I am. What ocschwar is describing is straight-up front-running, and it is illegal, but what the complaints here seem to be are a bit different: a buyer trying to buy more than the standing offer, so his order isn't completely filled. HFTs see this, and react more quickly by buying shares on other exchanges in the expectation that the large trader will route the remainder of his order to these exchanges (at a slightly higher price), so that the HFTs can sell back to him. To which my response is, so what? The offer price and volume was there, you chose not to buy at that price, so tough rocks.

By the way, for anyone who doubts Lewis's skill, the way he got 60 Minutes to turn David Einhorn into a hero of the little guy should be all the evidence you need (the very good Salmon piece JPD links mentioned it, but my jaw dropped seeing it last night).
posted by dsfan at 2:47 PM on March 31 [5 favorites]



I don't really get why we need trades to happen every millisecond. Would it kill us if we just ruled that the market-servers would carry out trades once every 5 minutes?


That's up to the market to decide. If you build a venue that only does periodic crossing auctions and nothing else, and every big investor decides that's the game they want to play, that's what will happen. 5 minutes? 1 minute? Whatever people want.

What ocschwar is describing is straight-up front-running, and it is illegal, but what the complaints here seem to be are a bit different: a buyer trying to buy more than the standing offer, so his order isn't completely filled. HFTs see this, and react more quickly by buying shares on other exchanges in the expectation that the large trader will route the remainder order to these exchanges (at a slightly higher price), so that the HFTs can sell back to him. To which my response is, so what? The offer price and volume was there, you chose not to buy at that price, so tough rocks.

The issue is what how to tell whether you're being subjected to front running or just not being nimble enough. If you can't tell, maybe you want to take your trades to another venue.
posted by ocschwar at 2:53 PM on March 31 [2 favorites]


Apparently the HFT brokers had set up a low-latency honey-pot exchange (BATS) to enable hiding of the true pricing. If I've got the scheme right, the customers were seeing orders reliably filled at the honey-pot, but not at the higher latency (to them) exchanges, where the high-speed trading was occurring.
posted by bonehead at 2:57 PM on March 31 [2 favorites]


I wonder what the impact of a transaction tax would be on the markets. Not huge, say 1% of every trade gets taxed.
posted by el io at 2:58 PM on March 31


We already have a transaction tax. It's called the Capital Gains Tax.
The current version only duns you when you gain money at the exchange.
The proposed new tax would hit you when you lose too.
posted by ocschwar at 3:00 PM on March 31 [4 favorites]


I wonder what would happen to this "industry" if the following conditions were required:

- Every order placed on any exchange have a minimum "time to live" of two seconds before it could be canceled

- Every firm using algorithms to place orders have the capital to execute on every order they have open at any particular instant
posted by de void at 3:08 PM on March 31 [1 favorite]


The pearl clutching from Lewis is a little much. Because I'm sure no one ever gamed order flow back in his day.
posted by 99_ at 3:09 PM on March 31 [6 favorites]


Apparently the HFT brokers had set up a low-latency honey-pot exchange (BATS) to enable hiding of the true pricing. If I've got the scheme right, the customers were seeing orders reliably filled at the honey-pot, but not at the higher latency (to them) exchanges, where the high-speed trading was occurring.

This part struck me as something that should at least be illegal in theory unless your a weasely ass too. It's like, letter of the law but not spirit kinda thing.

It's basically one way insider trading

This shit feels like formula one, but without any real regulation and also with the possibility to screw over a bunch of uninvolved people. There needs to be some regulating body to go "No, you can't abuse the wording of that law/regulation like that".
posted by emptythought at 3:11 PM on March 31


It’s a shameful vilification of the entire industry, says BATs O’Brien on Lewis’ HFTs allegations
The Holy Grail Of Trading Has Been Found: HFT Firm Reveals 1 Losing Trading Day In 1238 Days Of Trading
posted by the man of twists and turns at 3:13 PM on March 31 [1 favorite]


>I wonder what the impact of a transaction tax would be on the markets. Not huge, say 1% of every trade gets taxed.

That's actually breathtakingly huge, and would violently cutoff market liquidity and do profound damage to our financial infrastructure and integrity.
posted by jjmoney at 3:15 PM on March 31 [9 favorites]


That's up to the market to decide.

Technically, it could also be up to the government to decide. We can and routinely do decide that we want the market to be constituted in various ways (and letting it be a free-for-all is also a decision to have the market be constituted in a particular way).
posted by kenko at 3:18 PM on March 31 [8 favorites]


There's a fun Planet Money podcast that describes the birth of high-speed trading as invented by Thomas Peterffy. It's been a cat-and-mouse game since day one. The robot that constantly types on the Quotron machine is worth the price of admission.

Oh, and Peterffy also believes HFT should be reined in.
posted by JoeZydeco at 3:19 PM on March 31 [1 favorite]


This is far worse than the discussion of how really rich people lost a lot of money because Goldman or whoever sold them fraudulent mortgage derivatives.

this goes straight to the very underpinnings of the market and should be stopped.
posted by Ironmouth at 3:22 PM on March 31 [1 favorite]


The local stock exchange implemented a very specific alternative to a blanket transaction tax a couple of years ago, and it appears to have been much more successful than expected, even according to the exchange itself. (The purpose was to curb the worst excesses of HFT, but it apparently also helped liquidity compared to other stock exchanges, at least according to the media.)

A scientific paper got published a few weeks ago (it can be read here). It has a good explanation of the problems with HFT as well.
posted by Baron Humbert von Gikkingen at 3:23 PM on March 31 [4 favorites]


From the man of twists and turns first link:

"It’s not an exchange, it’s a dark pool. It doesn’t provide transparency at any part of the day. They spend a lot of time in this book demonizing HFTs and then they take them as customers. I’m not sure if they are good now but only if they use their exchange. I don’t understand how they are building a bullet proof system and using high-frequency traders."

So this guy is claiming that IEX is just another dark pool and that they are still engaging in HFT, when the actual technology and reality of how IEX works to defeat HFT is pretty clearly spelled out in the article. Forcing latency that defeats the ability of a HFT due to a forced slow down of the information is not a dark pool, and is pretty transparent if you ask me. It also categorically removes the advantages that BATS tries to provide, by allowing market monitors to utilize microsecond arbitrage. It removes the ability of HFT to utilize arbitrage, which is again, a GOOD thing for investors, just not for traders. BATS is an exchange sold to traders. IEX is an exchange sold to investors. Big difference, and something that O'Brien is going to regret when his proposed advantages get eaten. He also seems to think that the brokers are his clients, when in reality, he wouldn't exist if there were no investors to supply those brokers with trades to make in the first place. Very wrong priorities for someone whose livelyhood relies on the market remaining functional and solvent.

I think that this is one of the underlying arguments of the linked article in the OP, though, which is only kind of talked around when they mention the perverse incentives that led to the creation of prop shops and the functional realities of HFT. Of course, what it also led to was a massive increase in capitol investment by a lot of trading firms, in a massively escalating race to get as low latency as they possibly could. Maybe by removing the incentive, the brain drain that HFT has done to the computer science and programing industry might turn around (hey, a guy can dream, can't he?).
posted by daq at 3:28 PM on March 31 [2 favorites]


>>I wonder what the impact of a transaction tax would be on the markets. Not huge, say 1% of every trade gets taxed.

>That's actually breathtakingly huge, and would violently cutoff market liquidity and do profound damage to our financial infrastructure and integrity.

Okay, how about this one: a sales tax of 1% of every trade divided by the number of seconds you have held on to the stock you are selling. Hold on to the stock longer than, say, 100 seconds, and the sale is tax free.

Hold on to the stock only one millisecond, well, do the math....
posted by DreamerFi at 3:37 PM on March 31 [5 favorites]


I just wanted to drop in this link to a paper by the Norwegian central bank on HFT. I thought it was a great overview, especially the section describing the different types of HFT strategies.

Great to see constructive discussion on a finance related post on Mefi.
posted by pravit at 3:39 PM on March 31 [2 favorites]


I had no idea these Wall Street hotshots were so incompetent and/or clueless. "What's a millisecond?" - really? People who don't understand that information takes time to propagate along computer networks are designing proprietary trading systems?
The Wolf Hunters link mentions the image of traders as dumb jocks. I've taken service calls in a lot of types of businesses, and honestly, the trading firms I visited all really seemed to be living up to the stereotype. Far more than any other industry.
Of course, they were making 10 times as much money as me, so who's the dummy?
posted by MtDewd at 3:53 PM on March 31 [2 favorites]


de void: "I wonder what would happen to this "industry" if the following conditions were required:

- Every order placed on any exchange have a minimum "time to live" of two seconds before it could be canceled

- Every firm using algorithms to place orders have the capital to execute on every order they have open at any particular instant
"

To use the ancient internet analogy, Wall Street would interpret that as damage and route around it.
posted by chavenet at 3:55 PM on March 31 [1 favorite]


As a companion piece, Radiolab did a great piece on high frequency trading a while back and has some great discussions of how quickly things have sped up. It can be found here.
posted by craven_morhead at 3:55 PM on March 31


This article seems to use an excess of words to describe issuing speculative front run orders at one exchange based on orders publicly observed at another exchange.

You do understand that not everyone knows what that sentence means, right?
posted by Etrigan at 4:42 PM on March 31 [8 favorites]


We already have a transaction tax. It's called the Capital Gains Tax.

More than two-thirds of all stock equities are owned in tax-advantaged accounts such as pensions, endowments, IRAs and 401(k)s. They don't pay capital gains taxes on transactions.
posted by JackFlash at 4:50 PM on March 31 [8 favorites]


Saw the Sixty Minutes thing. They never said who exactly built the pipeline from New York to Chicago at fantastic cost, much less who the HFT actually are.

Investigative journalism indeed.

(Perhaps this is in the book?)

This article seems to use an excess of words to describe issuing speculative front run orders at one exchange based on orders publicly observed at another exchange.

Speculative? Technically, I suppose, but as a percentage, I wonder how many trades are cancelled because the price went up a fraction on the unwary would-be purchaser.
posted by IndigoJones at 4:53 PM on March 31 [1 favorite]


I just wanted to drop in this link to a paper by the Norwegian central bank on HFT. I thought it was a great overview, especially the section describing the different types of HFT strategies.

The simple proposal - if you have you more than 70x as many quotes than you have completed trades, you start paying additional fees. I just move these numbers from one place to another sometimes at several steps' remove so not particularly qualified to judge by IRL standards, but it looks good to me.

Something close to the endgame is neutrino sources and detectors located at the major trading centers to blast straight through the Earth and avoid the curvature. I'd imagine you'd get some freeish science out of the things on weekends.
posted by save alive nothing that breatheth at 4:57 PM on March 31 [5 favorites]


>I wonder what the impact of a transaction tax would be on the markets. Not huge, say 1% of every trade gets taxed.

That's actually breathtakingly huge, and would violently cutoff market liquidity and do profound damage to our financial infrastructure and integrity.


You're absolutely correct that 1% is far too much, but I suspect there exists a transaction tax rate, probably closer to 0.01%, that would remove the waste that HFT creates in the system and still retain functionally all of the utility and liquidity, particularly as viewed by market participants outside the richest 1%.

Although that Oslo paper linked above is an interesting alternative that specifically focuses on certain HFT strategies, while not creating a very modest tax on stock market participants in general (0.01% of transactions is on the order of a billion dollars tax on NYSE activity). I'm not sure if that is a feature or a bug.


The simple proposal - if you have you more than 70x as many quotes than you have completed trades, you start paying additional fees. I just move these numbers from one place to another sometimes at several steps' remove so not particularly qualified to judge by IRL standards, but it looks good to me.

There are a number of exemptions to the 70x as well, including things like adjusting bid and ask spreads - i.e. providing the liquidity that HFT parasites use as their figleaf - so these are very targeted for specific HFT strategies.
posted by Homeboy Trouble at 5:00 PM on March 31 [1 favorite]


I am a total layman to this stuff and I didn't understand the NYT article at all despite reading the entire thing :/

These guys created a separate stock exchange that automatically delays a signal going from one computer to another in order to make trading stocks more fair?
posted by gucci mane at 5:03 PM on March 31


What I love is that the epic past grift for the young female lead of the movie The Grifters was a group pretending to have the ability to do this, and using that carrot to bilk their marks out of their investment money. And now the HFT assholes have gone and actually done it.
posted by localroger at 5:16 PM on March 31 [8 favorites]


Goldman was an investment bank until the fall of 2008, when it became a commercial bank overnight in order to capitalize on federal bailout benefits, including virtually interest-free money from the Fed that it can use to speculate on the opaque ATS exchanges where markets are manipulated and controlled.

Unlike the NYSE, which is open only from 10 am to 4 pm EST daily, ATSs trade around the clock; and they are particularly busy when the NYSE is closed, when stocks are thinly traded and easily manipulated. Tyler Durden writes:
“[A]s the market keeps going up day in and day out, regardless of the deteriorating economic conditions, it is just these HFT’s that determine the overall market direction, usually without fundamental or technical reason. And based on a few lines of code, retail investors get suckered into a rising market that has nothing to do with green shoots or some Chinese firms buying a few hundred extra Intel servers: HFTs are merely perpetuating the same ponzi market mythology last seen in the Madoff case, but on a massively larger scale.”
HFT rigging helps explain how Goldman Sachs earned at least $100 million per day from its trading division, day after day, on 116 out of 194 trading days through the end of September 2009. It’s like taking candy from a baby, when you can see the other players’ cards.
The stock market is a casino, and the house is made up of extraordinarily powerful international banks. I'm sure everyone here would love to cheat at cards and make millions of dollars, but that is a game the banks reserve for themselves.
posted by deanklear at 5:29 PM on March 31 [8 favorites]


Has there been a flash crash that also hasn't had a subsequent flash recovery?
posted by garlic at 5:38 PM on March 31 [1 favorite]


What do you call parasites who feed on other parasites? Metaparasites?
posted by spitbull at 5:40 PM on March 31 [2 favorites]


These guys created a separate stock exchange that automatically delays a signal going from one computer to another in order to make trading stocks more fair?

When a customer makes a large order, say millions of dollars of shares, it splits up the order and sends it to several exchanges simultaneously. All of the exchanges currently are selling shares at the same price so you would expect to be able to buy all the shares you need at that same price.

But when an order arrives at the first exchange, the high frequency trader sees that a big purchase is in the works and races to all the other exchanges to buy up all of the shares available for sale at the current price milliseconds before the customer's orders arrive there. They then sell the shares back to the customer at a slightly higher price, profiting on the difference.

By calculating the time differences to the various exchanges and adding delays so that all orders arrive at the same instant, there is no opportunity for the high frequency traders to get see the order on one exchange and get to another exchange first.
posted by JackFlash at 5:44 PM on March 31 [11 favorites]


What do you call parasites who feed on other parasites?

Siphonaptera.
posted by localroger at 5:47 PM on March 31 [1 favorite]


Actually, JackFlash, that was the principle behind Thor, the original HFT-busting tool the OP protagonists developed before they understood exactly why it worked. IEX was meant to be an exchange immune to such shenanegans by design.
posted by localroger at 5:50 PM on March 31 [1 favorite]


What do you call parasites who feed on other parasites? Metaparasites?

Seriously, it seems weird to single out just one group on Wall Street and refer to them as parasites. Hello? It's Wall Street.
posted by indubitable at 6:01 PM on March 31 [2 favorites]


I'd be interested to hear an analysis of how front-running via HFT traders could influence other stock exchanges - and the whole global trading system for stocks. The article implies IEX is the only reliable exchange in New York because it compensates for geographic factors with a big ball of cable. Does this imply that all the world's electronic exchanges should adopt the same strategy?
posted by rongorongo at 6:01 PM on March 31


rongorongo, any exchange which adopts the anti-HFT strategy loses the business of those who benefit from the HFT trade, unless those folks are being pushed very hard indeed from the other side to use the fair exchange. OP makes it clear there has been a lot of resistance to that even when investors have demanded their trades go through IEX. It's clear that the investors, including major institutional players, like IEX but the middlemen with the dark pools and other insider gimmicks who have been benefitting from HFT will need to see enough benefit from being attractive to this huge pool of suddenly aware business to give up the perks they've been getting from the HFT profits.
posted by localroger at 6:07 PM on March 31 [1 favorite]


Aren't transaction taxes game-able by the rich to become essentially regressive?
posted by BrotherCaine at 6:23 PM on March 31


A 1% transaction tax is probably far too large and would induce a lot of friction in the market — it would basically increase the bid/ask spread on each security traded by 2% (one percent in each direction), which is more than it has been throughout the entire 20th century on NYSE blue chips.

However, I'm not sure that an 0.1% tax would be a terrible idea. Bid/ask spreads have only dropped to sub-0.1% levels in the last decade or two on liquid equities, and I don't know that it's necessarily a great thing or benefiting anyone other than the big players. (And the hell with them.) If overall trade friction were pushed back up to 1990 levels I think we could probably live with that.
posted by Kadin2048 at 6:59 PM on March 31 [3 favorites]


I've read the article, and think I understand it. I've read a bit about hft, but my exposure to it is limited. What does this do to the overall market? For example, could it be the insideous worm propping up an illogical bull market by constantly pushing prices higher? Or is it just the power to skim profit off the top but doesn't actually move the market?
posted by dejah420 at 8:32 PM on March 31


Maybe a dumb idea, but why couldn't an exchange set up a network of servers based across the world (maybe housed at other large exchanges, who would all be part of the network). When an trade comes in, it must be bounced off a random server somewhere in the world and then returned to the exchange before it is executed. That way, you would want to be fast, but you could never guarantee that you were faster than the time it takes to bounce the signal all the way around the globe. No tax, but a bit of leveling of the field.

Of course, I have no idea whether or not this is feasible.
posted by roquetuen at 8:37 PM on March 31


Actually, I should have phrased my question saying "why hasn't" rather than "why couldn't". The exchanges all have the infrastructure and it looks like a lot of people are pushing for a tax, which seems like it would be more negative than just leveling the playing field. Are HFTs really that influential across the world? Whether or not this is feasible to have coordination across exchanges seems…implausible.
posted by roquetuen at 8:47 PM on March 31


Another write up on Reuters, with a few good quotes not in the NYT article.

This is one of the key explinations:
"With the help of new hire Ronan Ryan, Katsuyama realized that his orders traveled along fiber optic lines and hit the closest exchange first, where high frequency traders would get a glimpse, and then use their speed advantage to beat him to the other 12 U.S. public exchanges and 45 private trading venues. HFT algorithms could then buy the shares Katsuyama wanted, and then sell them to him at a slightly higher price."

This is why it is called "front-running", in the most technical of definitions of the behavior, however, because it is done automatically by a computer, it doesn't fall afoul of the legal definition of front-running, which would mean collusion between two or more traders to push the stock price up by buying the stock themselves first, and then selling them at the new higher price to fill the client order.

This is the key, here:
"Katsuyama and Ryan created a system in which RBC would send its orders first to the exchange that was the furthest away, and last to the exchange that was closest, with the goal of arriving at all places nearly simultaneously, cutting out HFT."

If you can take away the advantage of having a faster, closer, lower latency connection to the other exchanges, you eliminate the risk of someone using arbitrage to make a profit. You also clean out the excess inefficiency caused by your order not completing at the price the market is supposedly showing you. It would be one thing if it were just a matter of someone in another trading firm clicking "buy" a few milliseconds before you. But when you, a human, are competing against a couple thousand servers that can send out "buy" orders between the time you enter the price you want and the time that you push "buy" on your workstation, the game, as they say, is up.

Last quote for emphasis:
""It almost felt like a sense of obligation to say we found a problem that is affecting millions and millions of people - people are blindly losing money they didn't even know they were entitled to. It's a hole in the bottom of the bucket," he said."

Yes, it is a hole in the bucket. It is also a lot like someone sitting and reading your request for a buy before it actually gets to the exchange, running in before you can place your trade request, and pumping up the prices, even just a few pennies, and then causing your order to fail because the price is now higher than you were willing to pay for the stock. Then you, the trader sitting at your trading desk has to find some other way to fill the order from your client, or be willing to lose money paying a higher price for the stock than you originally wanted to pay. It is a major distortion of the market price and a major distortion of market activity.

Some of the funniest quotes are coming from the heads of other exchanges, though. Many of them are saying that they need to allow HFT to increase the liquidity of the markets. What they aren't telling you is that they make money BECAUSE of HFT. Almost every exchange charges the brokers and traders to use their exchange for trades. The more trades that happen on their exchange, the more fees they collect from the traders and brokerages. And, of course, if half the volume on the exchange is due to HFT, that represents 50% of their revenue. It's really that simple. Now, some exchanges will pay you to use them, but only if you are a buyer, and some will do the opposite, and pay you to sell on their exchange. That's touched on in the original NYT article, "Why did one public exchange pay you to do something — sell shares, say — when another exchange charged you to do the same exact thing? " I think that right there tells you a lot about these other exchanges, and more importantly, just what kind of game has been played on Wall Street for decades. And, of course, all of it is perfectly legal, and "heavily regulated" by the SEC. The same SEC that couldn't seem to figure out what happened when the collapse of 2007-2008 cost real investors and American taxpayers billions of dollars.

Now, we can't say for certain that the SEC is full of idiots, but given that up until now, no one has bothered to do what IEX is doing, i.e. making a real effort to build an exchange where distorting influences like HFT and secondary transaction fees are not part of the incentive to trade on any particular market exchange.

I hope IEX is very successful. This is how to do transparency in the markets, and this is how you build something for the market to allow the market to do what it is supposed to do, which is a mechanism for channeling capital to productive enterprise.
posted by daq at 9:25 PM on March 31 [9 favorites]


JPD: No one is trading on a dark pool unless they explicitly choose to.

This quote seems to contradict that, but it's not entirely clear:
"The banks took different approaches to milking the value of their customers’ orders. All of them tended to send the orders first to their own dark pools before routing them out to the wider market. "
posted by IAmBroom at 9:57 PM on March 31 [1 favorite]


rustcrumb: I don't really get why we need trades to happen every millisecond. Would it kill us if we just ruled that the market-servers would carry out trades once every 5 minutes?

Sure, and while you're at it, let's reduce the maximum speed limit to 5 mph nationwide. It will save lives, reduce damage claims, and make the US a quieter place to live. Heck, some highway head-on collisions would be survivable!

Win-win, right?

It's always easy to regulate the other guy. Especially if you don't really get what's going on there.
posted by IAmBroom at 10:02 PM on March 31 [5 favorites]


The takeaway for anyone who finds "excess of words to describe issuing speculative front run orders at one exchange based on orders publicly observed at another exchange" a little opaque is this:

Wall Street has always profited most and best from the vig. This is simply the latest incarnation.
posted by 99_ at 10:04 PM on March 31


el io: I wonder what the impact of a transaction tax would be on the markets. Not huge, say 1% of every trade gets taxed.

If you mean a 1% tax on every trade is "not huge"... the long-term gain of the market is only 9%. Would you say a new tax that cuts your earnings by 11% is "not huge"?

ocschwar: We already have a transaction tax. It's called the Capital Gains Tax.
The current version only duns you when you gain money at the exchange.
The proposed new tax would hit you when you lose too.


Wrong. We have a profit-based tax. It isn't per-transaction at all; selling 1,000 shares at 10% profit is exactly the same tax burden as selling 1 share 1,000 times at 10% profit.

A transaction tax would be 1,000 higher on the latter example.
posted by IAmBroom at 10:08 PM on March 31 [1 favorite]


daq: Now, we can't say for certain that the SEC is full of idiots, but given that up until now, no one has bothered to do what IEX is doing, i.e. making a real effort to build an exchange where distorting influences like HFT and secondary transaction fees are not part of the incentive to trade on any particular market exchange.

I'm not saying the SEC isn't full of idiots, but to be fair: Brad Katsuyama hired the best and the brightest he could find, both in and outside the industry, and was given permission to lose $10,000 A DAY in order to decypher this mess for his employers. That's considerably more firepower than government regulatory agencies can summon.
posted by IAmBroom at 10:19 PM on March 31 [4 favorites]


If you mean a 1% tax on every trade is "not huge"... the long-term gain of the market is only 9%. Would you say a new tax that cuts your earnings by 11% is "not huge"?

Just a few decades ago capital gains were taxed at up to 40% and the economy did fine. It says something about how low tax rates have become in recent history that a small transaction tax is now unthinkable to many.
posted by JackFlash at 11:36 PM on March 31 [6 favorites]


Why are the buy and sell orders not encrypted while in transit between broker and exchange?
posted by rongorongo at 2:17 AM on April 1


I have no idea whether the orders actually are encrypted, but the problem is that the orders are public once they hit the exchange: they have to be. The important thing to know is that there are only so many shares for sale at a particular price on each exchange; if you want to minimise the price on a big order you have to break it up and buy the cheapest bundle for sale on Exchange A, the cheapest bundle for sale on Exchange B, the two cheapest bundles for sale on Exchange C, and so forth.

Katsuyama's problem was that his orders would hit one exchange quickly, but take longer to get to the other exchanges. In contrast, his competitors had quick access to all the exchanges. If Katsuyama placed an order that was split among several exchanges his competitors would observe the transaction and guess that he might be going to buy the same stock on the other exchanges, too. They'd get in ahead of him and buy the cheapest bundles for sale on those exchanges, and put them back for sale at a slightly higher price. When Katsuyama's order reached those exchanges he'd be forced to buy the shares at that higher price, and his competitors would profit on the difference.
posted by Joe in Australia at 3:48 AM on April 1 [3 favorites]


his competitors would profit on the difference.

Here's the thing that's being missed - the profit from that<less than the profit an old school market maker would have made. So net-net - "frontrunning" ends up being cheaper than trading pre-HFT.

Not to mention marketmakers were also "frontrunning"
posted by JPD at 6:59 AM on April 1 [2 favorites]


Just a few decades ago capital gains were taxed at up to 40% and the economy did fine.

When was this?
posted by leopard at 7:02 AM on April 1


Can you explain that JPD? The article makes it appear that front-running added costs to the institutional buyer, not reduced them.
posted by bonehead at 7:03 AM on April 1


Yeh - Dark Pools and HFT replaced traditional market makers. The spreads you traded at back in the day (that halycon time with thick necked men in garish sport coats yelling from pits on an exchange) were much much higher than the spreads to you see today. The reduction in costs from narrower spreads is much greater than the loss from "front running" by HFT.

Indeed most of the guys who are the most stridently anti-HFT are people who saw the profitability of their own businesses impaired by this shrinkage in spreads.

LT Cap Gains were in the very high 30's in the late 70's.
posted by JPD at 7:06 AM on April 1 [2 favorites]


Thanks. However, it is true now, is it not, that HFT is itself an inefficiency? If buyers can use careful ordering timing, like Thor, they can get those lower prices themselves and remove another layer of arbitrage, right?
posted by bonehead at 7:21 AM on April 1


Arbitrage is a response to market inefficiency, after all.
posted by bonehead at 7:23 AM on April 1 [1 favorite]


The reduction in costs from narrower spreads is much greater than the loss from "front running" by HFT.

This is always a claim made without substantiation. Typically it is stated that spreads are much lower than they were 20 years ago, but almost all of that change was due to decimalization, not HFT.

Before 2001, stocks were quoted with a minimum dimension of 1/16 of a dollar, or 6.25 cents. After 2001, stocks were quoted to the nearest cent. That is a more than 500% reduction in spreads due to decimalization, not HFT. Since that time, even the best case savings you could attribute to HFT, which didn't really get started until the mid-2000s, is a tiny fraction of a cent.

Don't believe the HFT propaganda that they are saving people lots of money. The savings on spreads was due to the decimalization policy change, not HFT.
posted by JackFlash at 7:28 AM on April 1 [1 favorite]


yes it is - but again lede buried - HFT is slowly arbitraging away their own profitability. Look at a marginal player who really relies on capital rather than an advantage in skill and you'll see their returns have decline precipitously. A friend of mine who was an HFT trader but sold in '09 and claims that marginal players need very high VIXs to earn a decent ROE, whereas 5 years ago you could earn a 15% ROE with a VIX below 15.

It'd be nice if trading were frictionless but someone has to make a market - and they probably deserve to get paid something for it.
posted by JPD at 7:28 AM on April 1 [3 favorites]


yes decimalization was the big bang, but spreads have continued to decline since then. There is a meaningful reduction in the profitability of the market making and cash equities biz over the last few years.
posted by JPD at 7:35 AM on April 1 [2 favorites]


There is a meaningful reduction in the profitability of the market making.

Meaningful to the market maker. Maybe. Meaningful to the investor? Not so much.
posted by JackFlash at 7:40 AM on April 1 [1 favorite]


Why are the buy and sell orders not encrypted while in transit between broker and exchange?

The frontrunning issues aren't caused by any wiretapping shenanigans; just by reading the public stream of orders (and trades) at an exchange at one location, making an educated guess about what that party would have sent to another exchange, and then winning the electronic race to get their order into that other exchange ahead of the ones they predicted are coming. (On preview: as explained above; sorry to repeat)

Orders are already compressed in transit (in order to slightly increase bandwidth) and security is probably lax enough to let someone impersonate a false order from someone else, however there is a very thorough confirmation and booking process that will detect and reject false orders within a few minutes after they happen. This actually gets tested every so often when a bank/fund/HFT shop sends garbled, buggy orders because of technical screwups, and the banks have gotten so good at this kind of validation that they first check to see whether they would have made profit on an inauthentic order before rejecting it.
posted by ceribus peribus at 7:45 AM on April 1 [1 favorite]


Meaningful to the investor? Not so much.

Well - that depends how much you trade.

The reality of it is that for buy and hold investors who turn their portfolio little if at all this whole thing is a lot of smoke, not much fire.

Sort of the irony - the IEX guys are well intended but the entire equity trading world exists as a machine to convince people to trade more than they ought to.
posted by JPD at 7:49 AM on April 1 [4 favorites]


Front-running is a crime. It requires you to trade on non-public information. But these large institutional traders are shouting out on one market that they want to sell a million shares. So everyone in that market knows their order, and can react to it. Before computers, you'd have people on telephones doing the same sort of thing. That's a lot of shares, enough that the market is going to move based on selling those shares. And if you don't want to move the market, it's a dumb thing to do. So if you're making a dumb trade, you may lose some money on that dumb trade. Institutional traders reacting to high frequency traders by splitting their orders up, or using algorithmic trading to dump a million shares over a day isn't an onorous thing for them to do, it's just smart trading.
posted by garlic at 8:00 AM on April 1 [2 favorites]


It'd be nice if trading were frictionless but someone has to make a market - and they probably deserve to get paid something for it.

Well sure, but middlemen have a history of transiency as buyers and sellers (and technology) improves. There's nothing wrong with making a market function, but when their "value" depends on a technical detail that's easily fixed once understood....

I think the outrage comes from the sense that this was a market exploit which needlessly added a layer of buyers and sellers. It could easily have been fixed had it been understood properly as high-speed trading got going. The implications of high speed trades weren't, of course, as detailed in the article, and the HFT crowd developed organically to fill that gap as a result.
“I hate them a lot less than before we started,” Katsuyama said. “This is not their fault. I think most of them have just rationalized that the market is creating the inefficiencies, and they are just capitalizing on them. Really it’s brilliant what they have done within the bounds of the regulation. They are much less of a villain than I thought. The system has let down the investor.”
The system, the people who dropped the ball were the big investors and buyers, who trusted the crappy software they had, as detailed at the beginning of the article. They feel tricked now (and they were, from a certain view point), but they were largely deceived by their own ignorance and arrogance.

So, perhaps we should not blame the HFT exploiters for be amoral (not immoral) in an amoral market, but nor need we feel any sympathy for their passing. We should be celebrating this in my view: by ending their arbitrage, the market gets a bit more efficient and thus things get cheaper for everyone, rather than a few getting richer. Anyway the HFT folks had what, 5 years of free money? They've got their reward.
posted by bonehead at 8:58 AM on April 1 [2 favorites]


But this ignores the fact that everyone has always complained about their execution. Its not like pre HFT guys thought they weren't getting screwed. I mean I feel like that's what crappy PMs and mediocre execution guys spend half their work day complaining about.

High speed trading disintermediated out a different set of inefficiencies and is doing that created a different set of things that could be exploited. I mean that's bad, but its better than it used to be - and those exploits are being arbitraged away.

I think the article and general opinion vastly overstates the degree to which most sophisticated large investors feel like an aggrieved party.
posted by JPD at 9:06 AM on April 1 [1 favorite]


I feel like saying "Don't hate the journalists for selling the story", but I've been on the receiving end of that too many times myself. It is the way the world works though.
posted by bonehead at 9:24 AM on April 1


My advice to people reading and viewing these stories is always the same.

1. consider a subject on which you are a relative authority. The times that subject has been in the news, especially in a sensationalist context, how accurately were things reported? (in my experience, usually very poorly, especially with more technical things)

2. In this particular case, there is a level of friction the market is willing to bear. Capitalism almost guarantees that someone will pocket that friction. The people who are complaining the loudest are the people who used to profit from that friction, and no longer do. If you go back a little ways, they were silent as the people they were taking money from complained for similar reasons (see the documentary "Floored" for examples of this - it's about the change that happened on the chicago mercantile exchange when people started trading on terminals instead of the floor)
posted by RustyBrooks at 9:33 AM on April 1 [7 favorites]


On that note, the Canadian financial media is being as smug and parochial as one would expect: Canadian says ‘moral compass’ led him to solve unfair gaming of stock markets by high-frequency traders
posted by bonehead at 10:09 AM on April 1


I did love the line in 60 minutes about how the guy was "conformist even for a canadian"
posted by RustyBrooks at 10:24 AM on April 1


Unlike the NYSE, which is open only from 10 am to 4 pm EST daily

Wait...what's that bell at 9:30 then? Can I start sleeping in half an hour every day?
posted by malocchio at 10:59 AM on April 1 [1 favorite]


Yeah, that's a good example of reporting accuracy right there.
posted by ceribus peribus at 11:15 AM on April 1


> [A tax on securities transactions] would violently cutoff market liquidity and do profound damage to our financial infrastructure and integrity.

This is complete fear-mongering, I can say in a professional capacity.

There is some level of tax that would cause these issues - let's say 1% of notional (face) value.

And there is also a level of tax so low it wouldn't change anything - say, 0.000000001% of notional value ($1 on $1 trillion).

The effects change continuously as we raise that tax variable from 1E-9 to 1E-2. There is some level that will have only marginal effects on liquidity, and both generate revenue and deter HFT.

I'd also say that this word "liquidity" that people wave around is two different things - one of which has great value to your average person or even market maker, and one of which has basically no value to your average person.

There's long-term liquidity, the ability to get out of (or into) a security "fairly quickly" at a price "close" to its "actual value". And then there's short-term liquidity, the ability to "instantly" transact with a price "very close" to its actual value.

An investor is holding securities for some finite period to make profit based on fundamentals. As long as an investor can execute a trade that goes off in, say, within one minute and whose price is within the price bounds in that minute, they should be completely happy. Yes, in a crash there might be some actual losses in that minute, but on average for each investor that loses, another one wins, and these will be limited (oh, and in crashes, HFTs don't provide one whit of liquidity anyway, so the point is moot).

An HFT trader is not an investor, but someone seeking to profit from technical mispricings in the sub-second range. As such, they are not providing any liquidity at any interval over a second - HFT traders are by no stretch of the imagination "market makers".

I'm strongly, strongly in favor of a tax of financial transactions - though a small one. You might want to make some sort of adjustment for heavily leveraged securities like options or futures where the notional is much larger than the size of the transaction, but then again perhaps not: Wall Street's constant strategy of "selling volatility" - going short against market volatility over the long term - is a bad habit it can't seem to break and perhaps discouraging it might result in stronger markets.

A particularly neat way to structure it would be to make the tax variable, depending on how long the security is held. We already have one case of this in the capital gains tax.

So we might be talking about 1E-7 (1 in ten million tax) for positions that were held for over a minute, that rises to 1E-4 (1 in ten thousand tax) for holdings for less than one second.
posted by lupus_yonderboy at 12:05 PM on April 1 [5 favorites]


see the documentary "Floored"

I hadn't heard of this...sounds pretty interesting.

*checks IMDB*

Oh god, it features Rat-faced Santelli. Dunno if I can stand it unless I know he's meeting some Fargo-type gruesome death in the end.
posted by malocchio at 12:06 PM on April 1


> Capitalism almost guarantees that someone will pocket that friction.

Not even capitalism - the "Law of Conservation of Money" guarantees that the friction has to go somewhere.

(The "Law of Conservation of Money" doesn't really exist - governments and even banks make money appear and disappear all the time - but in these microtransactions, the sum of the inputs has to equal the sum of the outputs, so the friction has to go somewhere, even if you're Marxist!)
posted by lupus_yonderboy at 12:13 PM on April 1 [1 favorite]



> [A tax on securities transactions] would violently cutoff market liquidity and do profound damage to our financial infrastructure and integrity.

This is complete fear-mongering, I can say in a professional capacity.

There is some level of tax that would cause these issues - let's say 1% of notional (face) value.


The person you're quoting was literally responding to--and quoted--a plan of 1% of value.
posted by dsfan at 12:15 PM on April 1


> The person you're quoting was literally responding to--and quoted--a plan of 1% of value.

Yes, I read that perfectly clearly, which is why I picked that 1% number.

The correct response to that plan is not, "Any such plan is doomed to failure," but, "This is a good idea, but you need to turn that number down."
posted by lupus_yonderboy at 12:22 PM on April 1


Let's at least discuss this in context. It's normally referred to as a Financial Transaction Tax, or a Stamp Tax. There was a fair bit of interest in imposing one in Europe a few years ago. Typical rates run between 0.05% to 0.1%.
posted by bonehead at 12:26 PM on April 1


Yes, I read that perfectly clearly, which is why I picked that 1% number.

Look, it's no skin off my back, but I think the clear inference from what you wrote is that jjmoney is engaging in "complete fear-mongering" despite the fact that you seem to agree that 1% is too large and jjmoney said absolutely nothing about the desirability of a smaller tax.
posted by dsfan at 12:30 PM on April 1


He is engaging in fear mongering.

Someone made a perfectly reasonable proposal except that, probably because he doesn't know the field well, he got one number wrong.

The response does not say, "Taxation is possible, but your number is too high" - it says, "Your proposal will wreak havoc," without explaining why or highlighting the specific error. It implies that the brokenness of the proposal is in the tax part and not in the 1% part.

If the response had started off, "Some sort of tax is possible, this isn't it," then I'd have no trouble with it.
posted by lupus_yonderboy at 12:36 PM on April 1


What's the problem a transaction tax solves?
posted by JPD at 12:37 PM on April 1


He is engaging in fear mongering.

Someone made a perfectly reasonable proposal except that, probably because he doesn't know the field well, he got one number wrong.

The response does not say, "Taxation is possible, but your number is too high" - it says, "Your proposal will wreak havoc," without explaining why or highlighting the specific error. It implies that the brokenness of the proposal is in the tax part and not in the 1% part.


This is madness. The person literally started the comment with "That's actually breathtakingly huge." That obviously refers to the 1% part.
posted by dsfan at 12:53 PM on April 1 [2 favorites]


What's the problem a transaction tax solves?

In my opinion, going back to the argument that there is a maximum friction the market can bear, it will decrease the amount of money HFTs get of the frictional value, and increase the amount, I suppose, that the government gets. I am very skeptical of any argument that this would actually reduce the overall cost people pay to trade.

(It may or may not drive HFTs out of business - people in this thread and overall have a specific picture of HFTs that I think actually only describes a fraction of the actual HFT strategies out there)
posted by RustyBrooks at 12:53 PM on April 1 [1 favorite]


Y'know, the original exchange was this:

>I wonder what the impact of a transaction tax would be on the markets. Not huge, say 1% of every trade gets taxed.

That's actually breathtakingly huge, and would violently cutoff market liquidity and do profound damage to our financial infrastructure and integrity.


And lupus_yonderboy paraphrased it to this:

> [A tax on securities transactions] would violently cutoff market liquidity and do profound damage to our financial infrastructure and integrity.

Sorry, but I don't think that's faithful summary of the original comments.
posted by malocchio at 12:56 PM on April 1 [3 favorites]


Katsuyama realized that his orders traveled along fiber optic lines and hit the closest exchange first, where high frequency traders would get a glimpse, and then use their speed advantage to beat him to the other 12 U.S. public exchanges and 45 private trading venues. HFT algorithms could then buy the shares Katsuyama wanted, and then sell them to him at a slightly higher price.

Okay, on second thought, something here is bothering me.

Two sides of a triangle are always longer than the third, by trivial proof. So if routes to exchanges were straight lines[1] and information traveled at constant speeds[2], there would be no way that the HFT guy could read your order at exchange A and then get in line ahead of you on the same order at any other exchange B. Clearly [1] and/or [2] is wrong.

But this is saying that Katsuyama's [Time to exchange B - Time to exchange A] is longer than the HFT guy's [Time from exchange A (to know about the trade) + Processing time + Time to exchange B].

Notice the signs: time difference vs. time sum. Either Katsuyama had totally incompetent network engineers, or someone somewhere was getting a fat payoff to screw with him, tap him at the source, or something else nefarious. This is screwy.
posted by RedOrGreen at 1:12 PM on April 1


The HFT shops invest heavily in very high speed data links; think dedicated microwave relay versus Verizon fiber optic. It's only a few tens of microseconds faster, but it's enough.
posted by ceribus peribus at 1:21 PM on April 1 [1 favorite]


Also, it's the HFT shops that push the hardest for GPU and FPGA accelerated processing, to minimize that term in the equation. From a tech perspective, it's very interesting work.
posted by ceribus peribus at 1:24 PM on April 1



Notice the signs: time difference vs. time sum. Either Katsuyama had totally incompetent network engineers, or someone somewhere was getting a fat payoff to screw with him, tap him at the source, or something else nefarious. This is screwy.


Not totally incompetent. Just that his employer didn't want to spend the money keeping up with the state of the art in setting up a trading network. If it makes more sense for such a firm to shift their order flow to the new IEX exchange than it is to redo their networking set up, then that is what they should do.
posted by ocschwar at 1:56 PM on April 1


Has anyone posted this interchange between Katsuyama & O'Brien yet? I found it riveting, though I haven't been able to watch the whole 20 minutes yet.

http://blogs.marketwatch.com/thetell/2014/04/01/epic-debate-on-high-frequency-trading-between-michael-lewis-brad-katsuyama-and-william-obrien/
posted by small_ruminant at 3:22 PM on April 1 [3 favorites]


Not even capitalism - the "Law of Conservation of Money" guarantees that the friction has to go somewhere.

+1 for drawing a control volume around the transaction and formulating a conservation law :-)
posted by indubitable at 4:02 PM on April 1


Two sides of a triangle are always longer than the third, by trivial proof

Equilateral triangle
posted by thelonius at 4:33 PM on April 1


> Equilateral triangle

Sorry, guess that wasn't clear enough. I should have said, the SUM of the lengths of two sides of a triangle is always greater than the third side.

In this case, the two sides are K to HFT shop and HFT to xchg B, which sum to more than K to B. As ceribus paribus said, it boils down to quality of links.
posted by RedOrGreen at 6:08 PM on April 1


From the second link:

Taking advantage of loopholes in some well-meaning regulation introduced in the mid-2000s,

I would like to know more about this "well-meaning regulation" introduced in the mid/late Bush Administration. Which turned out to be exploitable due to network lags.

Network lags changed after 9/11 because a lot of electronic exchanges were relocated to various sites in New Jersey. A few years later, some "well-meaning regulation" was introduced, during the Bush Administration, remember them, somehow containing "loopholes" that let the distribution of lags be exploited for very reliable, very large profits.

This article explains a huge amount very clearly. Additional reporting on these regulations would be very welcome.
posted by kadonoishi at 11:34 PM on April 1


RedOrGreen: What really baffles me about this whole article is that I had no idea these Wall Street hotshots were so incompetent and/or clueless. "What's a millisecond?" - really? People who don't understand that information takes time to propagate along computer networks are designing proprietary trading systems?
The story points out that there is on Wall Street a fashion, a cultural norm, of rampant macho egotistic self-aggrandizement, projection of phony confidence and competence, and Master-of-the-Universe posturing. Dick-swinging, basically.

It's not hard to imagine that this might manifest itself in, among other ways, 1) exaggerating personal jock-ness as opposed to nerd-ness and 2) exaggerating the conceptual difficulty of any nerdy obstacles the jock/storyteller managed to overcome.

So a Wall Street audience may automatically fill in a back-story for "What's a millisecond?" that goes something like this: "Even though I too spend all day slapping peons around with my enormous penis, my can-do attitude and infinite capacity for spending all day at the office (aka 'hard work') make it possible for me to do nerd stuff when I really have to. That nerd stuff turns out to not be such a big deal after all, AMIRITE?" The "What's a millisecond?" framing may be a calculated.

Of course, I haven't met the guy, and I don't want to suggest he's automatically a horrible jerk. But he may be embedded in a horrible jerk culture and find that it rubs off some. In any case I don't think "What's a millisecond?" should be taken as an accurate account of his state of mind; it's storytelling.

I haven't finished the article yet but it's looking like a great read.
posted by Western Infidels at 7:40 AM on April 2


I think perhaps its better to describe it as one culture being superseded by another rather than talk about dudes slapping one another around with cocks.

I know lots of ex-D-1 athletes who are sellside traders. My friend the HFT is the sort of person who formerly shared an office with a Fields Medal winner. Before computerization aggression and relationships made one good at that job. Post computerization those skills were devalued.

Trust me - very few old school sell-side traders even know WTF the Fields Medal is, nor were they pretending to. Most of them have spent the last few years figuring out how to deal with making a lot less money TBH.
posted by JPD at 7:52 AM on April 2 [1 favorite]


There are lots of ex-physicists and ex-EEs working for hedge funds and brokers and the like. The scientific literacy of some of the traders (particularly the "quants") might surprise you. A couple of my theoretical physics grad school classmates went that way.

In any case, the time delays are more about end-to-end processing power and signal amplification/processing on the transmission switches (how many, what's the clarity of the lines/length between repeaters, etc...) than directly about distance factors. Theoretical signal transmission speeds are not the limiting factors, as can be easily seen by the time map in the NYT piece: some of the physically closer destinations to the initiation point have higher latencies than further ones. Network processing speeds are the fundamental limitations they are wrestling with. Distance factors are still a minor component of that.
posted by bonehead at 7:58 AM on April 2


JPD: Trust me - very few old school sell-side traders even know WTF the Fields Medal is, nor were they pretending to. Most of them have spent the last few years figuring out how to deal with making a lot less money TBH.
Okay, but Ryan (the "What's a millisecond?" guy) isn't an old-school trader, he's a 30-something IT geek who's also quoted as saying "[technology is] pretty captivating, when you take the nerdiness out of it." I read that as a geek pretending to be less geeky than he is, the better to fit in to a doubly-alien culture (he's a recent immigrant, too).
posted by Western Infidels at 8:03 AM on April 2


And, because this makes my teeth hurt, the time scales under discussion are microseconds, a thousand of which pass in a single millisecond.
posted by bonehead at 8:05 AM on April 2


(When we're talking about network latencies. Exchange latencies are in the milliseconds).
posted by bonehead at 8:07 AM on April 2


There are lots of ex-physicists and ex-EEs working for hedge funds and brokers and the like. The scientific literacy of some of the traders (particularly the "quants") might surprise you. A couple of my theoretical physics grad school classmates went that way.

Of yeah - but most of these folks weren't historically in client facing roles. Also there were always shops filled with guys like that. But execution guys at fundamental shops and sell-side sales traders were never like this. Changed some in the last five years. Actually I've heard the hardest job to find is a new role for a 40's ish buy-side execution trader. Time has just passed them by.
posted by JPD at 8:12 AM on April 2


As you said above, I'd bet that most of the ex-science and math types in the story are behind the HFT desks. The NYT piece is, as you say, about the less-literate banks catching up to the reality where a bit of this kind of knowledge matters.
posted by bonehead at 8:25 AM on April 2


Actually I've heard the hardest job to find is a new role for a 40's ish buy-side execution trader.

I really didn't need to read that this morning. But yeah...

Time to refocus my efforts on backstabbing and poisoned coffee.
posted by malocchio at 8:27 AM on April 2


makes you feel better the guy who told me that said that my job was the second toughest to find.
posted by JPD at 9:02 AM on April 2 [1 favorite]


Kind of, yeah!
posted by malocchio at 12:58 PM on April 2


Will we ever hear from someone who knew about this, designed the system that took advantage of it? Or do the HFT algorithms "just work", without their designers actually understanding the full picture?
posted by So You're Saying These Are Pants? at 2:43 PM on April 2


Most of the people who knew how the HFT algorithms worked are probably under indefinite NDA's and made enough boatloads of money to be uninterested in violating same.
posted by localroger at 3:32 PM on April 2 [1 favorite]


From the description in the article, there may not be very much actual cleverness in the algorithms. In fact, I suspect there isn't: complicated algorithms are slow, and HFT must be fast. Instead, they mostly rely on special treatment: being physically closer (or co-located) with the exchanges, as well as maximising their processing speed.

There's a theme in the FPP that I've encountered before: some people are mysteriously rich and successful, and others wonder how they do it. Are they extremely clever? No. Are they well-connected? No. Do they know some sort of secret? No. Instead, the answer is that they are crooks or liars. That is, either:
  1. They aren't actually rich, just good at self-promotion;
  2. They're working off inherited wealth, and aren't actually successful;
  3. They are literally stealing or extorting money from other people;
  4. They are rentiers, who profit from being in a privileged position.
The HFT crowd are the last: their deals with exchanges mean that they learn about other people's trades before the information is generally available, and they can place their own orders before other people can react. There's a certain amount of cleverness in knowing how to use that information; and there's probably a decent amount of cleverness in setting up the computer systems to take advantage of it; but I can't see anything that couldn't be easily duplicated. Remember, Brad Katsuyama's employers were losing money because they didn't understand the speed of light.
posted by Joe in Australia at 5:24 PM on April 2


Remember, Brad Katsuyama's employers were losing money because they didn't understand the speed of light.

Last fall, Traders may have gotten last week’s Fed news 7 milliseconds early
Markets swung rapidly on the 2 p.m. announcement last Wednesday, with stocks, bonds, and the price of gold all skyrocketing. Somebody placed massive orders for gold futures contracts betting on exactly that outcome within a millisecond or two of 2 p.m. that day -- before the seven milliseconds had passed that would allow the transmission of the information from the Fed's "lock-up" of media organizations who get an early look at the data and the arrival of that information at Chicago's futures markets (that's the time it takes the data to travel at the speed of light. A millisecond is a thousandth of a second). CNBC's Eamon Javers, citing market analysis firm Nanex, estimates that $600 million in assets could have changed hands in that fleeting moment.
posted by the man of twists and turns at 6:20 PM on April 2 [1 favorite]


naked capitalism: Michael Lewis’ Repeat Omission: No Crimes Were Committed
posted by the man of twists and turns at 6:49 PM on April 2


Also from Naked Capitalism, a really good argument about the problems with HFT: Yes, Virginia, HFT and Liquidity are Not All They Are Cracked Up to Be
The best indication of the level of aggressive algorithmic strategies is provided in a study of the practice of “Quote Stuffing.”34 This is a tactic in which HFTs flood an exchange or other transaction-matching venue with quotes to buy or sell in order to slow down the venue’s processing times. HFTs do this so that aggressive tactics can be implemented without intervention from other traders. They also employ Quote Stuffing to slow down one exchange so that price differences between that exchange and another that is operating normally can be exploited.35 Therefore, when quote stuffing occurs, it is highly likely that an aggressive tactic is underway. The study reaches the following conclusion:

We find that quote stuffing is pervasive with several hundred events occurring each trading day and it impacts over 74% of US listed equities. Our results suggest that, in periods of intense quoting activity, stocks experience decreased liquidity, higher trading costs and increased short term volatility.

Quote stuffing operations have been estimated to “consume 97% of computer system resources that the whole market has to bear.
My highlight. OK, this isn't a free market; this is a squeeze play that should be criminalised.
posted by Joe in Australia at 9:46 PM on April 2 [1 favorite]


a bumper crop of links from Omnivore: The new financial industry
posted by the man of twists and turns at 8:17 PM on April 3 [1 favorite]


Mark Cuban gives us The Idiot's Guide To High-Frequency Trading
posted by the man of twists and turns at 7:12 AM on April 4


He's something of a rebuttal: Michael Lewis: Shilling for the Buy Side
posted by chrchr at 11:15 AM on April 4 [1 favorite]


For instance, they bought 10 million shares of Citigroup, then trading at roughly $4 per share, and saved $29,000 — or less than 0.1 percent of the total price. “That was the invisible tax,” Park says. It sounded small until you realized that the average daily volume in the U.S. stock market was $225 billion. The same tax rate applied to that sum came to nearly $160 million a day.

Really? This whole thread, and no one's made the Superman III / Office Space reference?

Also, I'm a little surprised this is news... For years now, there have been stories about trading firms laying their own fiber optic cable, paying deals to be co-located, laying a special cable between New York and London. Like... why would they do that except to try to run ahead of other traders?

But this is saying that Katsuyama's [Time to exchange B - Time to exchange A] is longer than the HFT guy's [Time from exchange A (to know about the trade) + Processing time + Time to exchange B]

Notice the signs: time difference vs. time sum. Either Katsuyama had totally incompetent network engineers, or someone somewhere was getting a fat payoff to screw with him, tap him at the source, or something else nefarious. This is screwy.


Well the HFT guy is probably paying to be Co-Located at the exchange, which is something only an algorithmic trader would pay for. So "Time from exchange A" is extremely low. The HFT guy also probably has a dedicated link between exchanges. The algorithm is very simple (ie, buy what that guy's buying), and I bet in terms of processing power it doesn't take much longer than it would take to route packets between a single network link.

As mentioned before, one thing to keep in mind is it's not all about speed of light. Differences in signal transmission are only part of the story - every switch or hop between networks adds significant latency. It's not a matter of Katsuyama's network engineers being incompetent - as I understand it, the algorithmic traders have built a private network to connect the exchanges. If you're relying on Verizon or some other provider for your connectivity, you're just never going to beat them.

So, put another way, lets say there are 8 network hops between Katsuyama and A, and 3 hops between him and B (maybe they have the same ISP). The HFT end runner has a dedicated link between A and B, so they only have two hops, plus one for processing time. And of course, the HFT algorithm can stack the deck in their favor by DDOSing exchange A with quote requests
posted by heathkit at 5:03 PM on April 4


Also the HFT guy could install a key sniffer or some other kind of malware at the trader's end, and know about the trade before the request is actually sent. That's probably not widespread, but I bet it's happened a bunch. I'm sure there's some kind of equivalent not-actually-illegal version of that which is commonly used, though.
posted by heathkit at 5:07 PM on April 4


Ok I should probably read the whole article before posting.

Inside BATS, high-frequency-trading firms were waiting for news that they could use to trade on the other exchanges. BATS, unsurprisingly, had been created by high-frequency traders.

So they literally created a honey pot exchange to catch all the trades coming out of Manhattan, then built a private network to disseminate that information to the other exchanges in under half a millisecond. That's pretty much the not-actually-illegal equivalent of using a key logger that I mentioned.

Also, it's astounding to me how all the technical details in this article would be common knowledge for any decent CS undergrad who had taken a networking class.

Also I'll stop posting now, sorry
posted by heathkit at 5:22 PM on April 4


Why Should “Flash Boys” Shock Us?
posted by homunculus at 11:16 AM on April 5


High Frequency Trading: Threat or Menace?
As Schwall looks further back into the past — he ends up spending several work days at his local public library on Staten Island — he discovers that this was just part of a long cycle going back to the 1800s:
The entire history of Wall Street was the story of scandals, it now seemed to him, linked together tail to trunk like circus elephants. Every systemic market injustice arose from some loophole in a regulation created to correct some prior injustice.
When he got back to work and described his findings to his colleagues, Schwall’s conclusion was that “U.S. financial markets had always been either corrupt or about to be corrupted.”
posted by the man of twists and turns at 10:43 AM on April 7


Most of the people who knew how the HFT algorithms worked are probably under indefinite NDA's and made enough boatloads of money to be uninterested in violating same.


I've done HFT. I'm under NDA, but I've not violated it yet and will not violate it with the comment below:

Some HFT algos work in the fashion shown in this story. That does not mean the HFT people are breaking the law. But it does mean the retail investors can't tell the difference between being pre-empted by HFT and being subjected to illegal front running. That's very, very bad. The exchanges have a duty to prevent both impropriety and the appearance thereof.

But recall this story starts with a trader seeing a bunch of offers for a stock and a price and deciding to match them with bids. Most HFT shops are the people filling the board with those offers to begin with.
posted by ocschwar at 8:22 PM on April 9 [1 favorite]


-The Real Problem with High-Frequency Trading
-Let's stop talking about HFT for a little while
-Matt Taibbi Extended Interview
posted by kliuless at 2:38 PM on April 11


The Infrastructural Mystery at the End of Michael Lewis's Flash Boys
posted by homunculus at 3:25 PM on April 13


JackFlash: Just a few decades ago capital gains were taxed at up to 40% and the economy did fine. It says something about how low tax rates have become in recent history that a small transaction tax is now unthinkable to many.

That's a ridiculous claim. Every successful businessman in the history of Ever has been concerned with changes in profit.

Mind if I take a mere 11% of your own income away from you? It's not like you'll notice it.
posted by IAmBroom at 9:38 AM on April 15


Every successful businessman in the history of Ever

These aren't businessmen. They don't make anything, and it's hard to find a single tangible benefit they provide to society-- if anything they are a burden on the economy rather than a driver of it.

Money changing shouldn't be a massive, highly lucrative industry. It sucks resources away from productive sectors. Taxes on the financial industry should be greatly increased.
posted by cell divide at 10:32 AM on April 15 [2 favorites]


It sucks money away from productive sectors.
That's not wrong at all and I'm not disagreeing, but another way to look at it is that it figures out what sectors are productive and directs money to them.
posted by chrchr at 8:17 PM on April 15


In other news: U.S. Set to Bring Criminal Charges Against Banking Giants
posted by homunculus at 4:57 PM on April 30


other news as in unrelated to high frequency trading, dark pools, and the post?
posted by garlic at 6:47 PM on April 30 [1 favorite]


« Older You say you just learned of the passing of comedia...  |  Six games that really get suic... Newer »


This thread has been archived and is closed to new comments