60 Minutes lead story on Michael Lewis - Flash Boys

[oddballstocks]

I doubt HFT is happening to the types of stocks I buy, small illiquid ones, but I know of something worse.  That's the physical market makers who are buying the shares at $14 in little bits and pieces then offering them for sale at $14.50 in round lots.  Or buying the lots at $14 and selling at $14.50.  This is absolutely true because I know people who do this, that's the definition of their job.

 

I wish there was HFT for some of the stuff I am buying.  Then there'd be some liquidity, and bid ask spreads smaller than 30-40%.  I know market makers who sit on ask's in the 1000's for a $15 stock just so no one can enter an order online. 

 

As someone said earlier in the thread, if your investment thesis is wrecked by a penny or two then you need to step back and question what you're doing.

[Otsog]

Otsog, the quotes supplied in the Michael Lewis piece are even more confusing. They look at one side of a transaction, the trader who enters an order that doesn't get filled, and somehow end up at "front-running". What happened to the order? Is the HFT now taking your place as the investor? Or, more likely, is the HFT flipping the stock to an investor who is willing to pay more? In the latter case, that is just price discovery! What the people in the article are complaining about is the allocation of the fee for matching buyers and sellers. The bid ask spread is not a right. It is just the information provided given the existing order flow.

 

The HFT is taking your place as the investor. 

 

http://www.nytimes.com/2014/04/06/magazine/flash-boys-michael-lewis.html?hp&_r=0

[Otsog]

I think all the analogies in the thread so far have been beyond terrible so I might as well try one too, but only poker players will get it. 

 

The specific HFT practice that the 60 minutes and NYT pieces cover is akin to playing Rush Poker and landing in the button every single hand.  And they are paying FTP for that right. 

 

 

[ni-co]

To me this simply is a stock exchange malfunction. To larger institutions this feels like a tax and it adds up to tens of millions a year. It doesn't help to say that in the grand scale of things it's not that much money compared to their holdings or that they should be long term oriented etc. This doesn't change the fact that, in this case, HFT introduces a second middleman between buyer and seller beside the exchange itself (maybe even a third or fourth middleman – who knows?).

 

This makes trading less efficient, especially for large institutions like pension funds, mutual funds, ETFs etc. and now that they know about it, it will stop. Up to recently, the exchanges didn't have any incentive to solve this problem because larger institutions didn't know how the scheme worked. They simply noticed their trading expenses increasing every year. On the contrary, the exchanges had any reason to be silent about it because it multiplied their trading volume and they profited from every single trade.

 

Now that everybody knows about it, the exchanges have to do something about it or their clients will simply leave to create their own exchanges. It's too much money on the table for them not to do this.

 

By the way, the claim that HFT provides liquidity is a very weak argument in this case. Buyers and sellers are already there and the high frequency traders do nothing but putting themselves in the middle and making a transaction more expensive that would have happened anyway at a lower price. That's why those institutions feel ripped off. This has nothing to do with being against HFT in general or being against computerized trading. But this practice will stop no matter how much money went into shorter fibre lines and real estate next to stock exchanges.

[writser]

HFT introduces a second middleman between buyer and seller beside the exchange itself

 

No! HFT did not introduce the middleman. He has been forever: the market maker. Or day trader. Or arbitrageur.  Order books have never been sacred bibles where only long-term investors can enter orders. There are and always have been people trying to read the tape. People who think: "well if I can buy at this price I probably sell it over there", "that's a big sell order, let me adjust my quotation slightly lower" or "the market trends up, let me buy something". You make it sound as if HFT is suddenly a new class of assholes. They are not. Only the tools they use are different. Inter-exchange arbitrage is nothing new. These firms just took all the profits the old assholes made because they are far more efficient at their job. They also make transparent how much the middle men are actually making and that is why you suddenly hate them.

 

By the way, the claim that HFT provides liquidity is a very weak argument in this case. Buyers and sellers are already there and the high frequency traders do nothing but putting themselves in the middle and making a transaction more expensive that would have happened anyway at a lower price. That's why those institutions feel ripped off. This has nothing to do with being against HFT in general or being against computerized trading. But this practice will stop no matter how much money went into shorter fibre lines and real estate next to stock exchanges.

 

That doesn't make sense. If the buyers and sellers were already there the trade would, by definition, already have been executed. If the price changes either the buyer or the seller makes a profit. And if the trade happens at a lower price and a HFT firm is the buyer then nobody was willing to pay more - or the trade would've been executed at a better price in the first place! What you struggle to describe is called market making. If you add liquidity to the system (i.e. put a bid behind behind the best bid and have a coffee) then market makers make zero money. If you substract liquidity from the system (buy from the offer) them yes,  market makers will make money in the long run. It has been like that for centuries. If you don't like that: don't buy from the offer!

 

@Otsog: it's not an analogy, it's a completely identical situation. You can have an opinion about it but markets have been functioning like that for centuries. I understand your sentiment, I even somewhat agree with it but you have to understand that your arguments are not specifically targetting HFT trading. You just want to forbid arbitrage and allow exchange access only for long-term investors. That's a very nice notion of how an ideal capital market should look like but guess what: we have have FREE markets. Things don't work that way.

[ni-co]

HFT introduces a second middleman between buyer and seller beside the exchange itself

 

No! HFT did not introduce the middleman. He has been forever: the market maker. Or day trader. Or arbitrageur.  Order books have never been sacred bibles where only long-term investors can enter orders. There are and always have been people trying to read the tape. People who think: "well if I can buy at this price I probably sell it over there", "that's a big sell order, let me adjust my quotation slightly lower" or "the market trends up, let me buy something". You make it sound as if HFT is suddenly a new class of assholes. They are not. Only the tools they use are different. Inter-exchange arbitrage is nothing new. These firms just took all the profits the old assholes made because they are far more efficient at their job. They also make transparent how much the middle men are actually making and that is why you suddenly hate them.

 

By the way, the claim that HFT provides liquidity is a very weak argument in this case. Buyers and sellers are already there and the high frequency traders do nothing but putting themselves in the middle and making a transaction more expensive that would have happened anyway at a lower price. That's why those institutions feel ripped off. This has nothing to do with being against HFT in general or being against computerized trading. But this practice will stop no matter how much money went into shorter fibre lines and real estate next to stock exchanges.

 

That doesn't make sense. If the buyers and sellers were already there the trade would, by definition, already have been executed.

 

I know perfectly well what arbitrage and market making are and I get the impression you purposefully "misunderstand" what I'm trying to say. Let's say there are three sellers offering 10,000 shares each at $100 at three different exchanges. I say: "Fine, buy 30,000 at $100". I get a fill at the first exchange at $100 and the ask price ticks up at the other two exchanges to $101 because high frequency traders saw my order coming, bought the shares at $100 and now offer them to me for $101. I'm now left with two options: 1. Pay $101, 2. Not buying the rest of the shares.

 

Now tell me: How does this help the market place? There were three sellers of a total of 30,000 shares asking for $100; there was one buyer bidding $100 for 30,000 shares, but, in the end, the (full) trade didn't happen at $100 simply because the buyer's order didn't reach the exchanges at the same time. At least, that's how Katsuyama describes this phenomenon and it sounds perfectly reasonable to me. What they did to counter it is delaying their orders to the nearer exchanges so that their orders hit all three exchanges at the same time – and their fill rates went up to a 100% (at $100). That's why I say it's an exchange malfunction.

 

What on earth has this to do with market making or arbitrage? How does this provide any additional liquidity? This is simply front running orders. These traders wouldn't have bought these shares were it not for my order – that's what differentiates them from market makers, that's why they don't provide additional liquidity. That's my whole point.

[writser]

I'll try to explain where your example goes wrong. The moment you buy 10k shares on the first exchange, nobody, except for you, knows that you sent more buy order to other exchanges. That is private information until your orders actually hit the other exchanges. You make an implicit assumption that HFT firms know that you try to buy 30k shares in total. They don't. That would be frontrunning, that would be illegal and if they had that information and acted upon it they should be shut down.

 

But because you have to buy more shares in this example, the HFT firms look like geniuses. If all you had to do was buy 10k shares they look like idiots, buying 20k shares @ $20 while nobody is interested in them. They don't know beforehand what scenario will unfold. The only information they have is that somebody bought the entire offer on a single exchange. However, the alternative outcome in which the HFT firm is left with 20k shares on his book is never brought up in the 60 minutes show. Your example is a subset of possible outcomes in which the HFT firms gets lucky.

 

What HFT firms are doing in your example is speculating based on public information: 10k shares traded at exchange 1 (public information) so I buy 20k shares at exchanges 2 and 3 (speculation). In your example the HFT firm will probably make money because you use a predictable, blantantly obvious strategy to buy the maximum number of shares in the shortest possible timeframe. But it is not a guaranteed way of making money and it has NOTHING to do with frontrunning. They're just trying to predict what you have to do and are preying on your psychological fears (the shareprice is running away from me!) and hope that you instantly buy shares from them. Now you may think that that is despicable but people have been speculating, trying to predict and trying to scalp in stock markets for centuries. I say: nothing wrong with that. The more participants the merrier.

 

The problem is that you look at a whole series of transactions in hindsight and pick the easiest explanation to support one specific outcome (Looks like they know about my orders! They must be frontrunning me! Evil computers!), instead of the hard explanation that explains all possible outcomes (They use sophisticated hard- and software to predict what I am trying to do. They were actually right this time. If that happens constantly maybe I should adapt my strategy).

[ni-co]

I know it's unbelievable. Read the book – Lewis explains it. They lure brokers into placing orders at exchanges where they get the information first and can execute their orders on the other exchanges before the large orders hit them. They are able to recognize from which broker and even from which firm orders come. They are able to figure out from whom the order comes from the time delay and route an order takes to hit one exchange after another. It's not that they know it 100% but they figure it out and the odds are very much in their favor – it's very, very educated guessing.

 

It has nothing to do with "luck". The more data they feed into their algorithms the better they know from whom specific orders come and at what price. When they figure out that T. Row Price buys 100 shares of KO at BATS they can be pretty sure that this is only a partial order and they buy KO on NYSE and NASDAQ and then sell it at a higher price.

 

How do you explain that slowing down orders so that they hit all exchanges at the same time works? This is not a tinfoil hat conspiracy theory. RBC actually sold this system to large institutions when they had figured out how to circumvent this scheme by delaying the orders precisely.

 

Nobody says that's illegal but it's front running nonetheless.

[Packer16]

This is probably the reason IB trading desks can make profits most days.  How else do you explain that all these banks all of sudden are profitable most days trading?  Did they all of sudden become trading wizards?

 

Packer

[APG12]

Nobody says that's illegal but it's front running nonetheless.

 

This is not true and I wish people would quit saying it.

 

Are management fees legalized theft? Give me a break! Front running is FRAUDULENT.

[ni-co]

Nobody says that's illegal but it's front running nonetheless.

 

This is not true and I wish people would quit saying it.

 

Are management fees legalized theft? Give me a break! Front running is FRAUDULENT.

 

Front running is not fraudulent when you don't pay the front runner to execute orders for you or advise you in any way. It's not illegal to figure out that you want to buy more shares and snap them up right before your nose. That's not to say it shouldn't be illegal but it surely isn't.

[APG12]

Front running is not fraudulent when you don't pay the front runner to execute orders for you or advise you in any way. It's not illegal to figure out that you want to buy more shares and snap them up right before your nose. That's not to say it shouldn't be illegal but it surely isn't.

 

OK, but why do people keep equivocating the fraudulent and the non-fraudulent? I'm glad to see you are recognizing the difference but I see this everywhere! HFT is not fraudulent.

 

There are all sorts of fees and expenses I pay in life that I'd prefer not to pay. Or transactions that take place that I'd prefer didn't. The standard in a free society by which we decide what is allowed and what isn't does not consist of the following: What does Mary Jo White think about this? How does the 'little man' feel? Which would I personally prefer? What will make the market less volatile? What will increase 'efficiency'? etc.

 

The standard is: Are the transactions voluntary? Again, HFT is not fraudulent. Use limit orders.

[morningstar]

I know it's unbelievable. Read the book – Lewis explains it. They lure brokers into placing orders at exchanges where they get the information first and can execute their orders on the other exchanges before the large orders hit them. They are able to recognize from which broker and even from which firm orders come. They are able to figure out from whom the order comes from the time delay and route an order takes to hit one exchange after another. It's not that they know it 100% but they figure it out and the odds are very much in their favor – it's very, very educated guessing.

 

It has nothing to do with "luck". The more data they feed into their algorithms the better they know from whom specific orders come and at what price. When they figure out that T. Row Price buys 100 shares of KO at BATS they can be pretty sure that this is only a partial order and they buy KO on NYSE and NASDAQ and then sell it at a higher price.

 

How do you explain that slowing down orders so that they hit all exchanges at the same time works? This is not a tinfoil hat conspiracy theory. RBC actually sold this system to large institutions when they had figured out how to circumvent this scheme by delaying the orders precisely.

 

Nobody says that's illegal but it's front running nonetheless.

 

Of course slowing down orders so they hit all exchanges at the same time works. Why would you think that represents a flaw in the system? Previously, the broker was doing the following steps: #1 - clearly indicate to players on the NYSE a strong interest in purchasing stock (through activity on another exchange), #2 - deliver an order for stock to the NYSE. Naturally, it turned out better for RBC to do #2 before #1. That's the sign of a healthy interaction across exchanges.

 

Of course I'd like to be able to buy an unlimited number of shares and not have my previous, visible activity factor into the market price. Buffet and other big stock buyers would like this even more! (Is it "front running" to buy WFC on the basis I suspect Buffet will likely buy some in the future - or perhaps has an order in transit right now?) You don't have a right to this, however.

[20ppy]

http://www.nytimes.com/2014/04/06/magazine/flash-boys-michael-lewis.html?hp&_r=0

 

To set up an Exchange, doesn't one need to get regulatory approvals? Or anyone can run an exchange business?

 

 

[writser]

How do you explain that slowing down orders so that they hit all exchanges at the same time works? This is not a tinfoil hat conspiracy theory. RBC actually sold this system to large institutions when they had figured out how to circumvent this scheme by delaying the orders precisely.

 

How do you explain what? It is completely logical, if your orders end up at exchanges within a shorter timespan it is harder to react to their respective executions. What do you think is at work here, magic? So now RBC is selling a trading system that can send orders to multiple exchanges in a shorter timespan. They're using their computers and proprietary software to be less exploitable by other market participants. Just like the HFT firms ..

[Guest 50centdollars]

This is probably the reason IB trading desks can make profits most days.  How else do you explain that all these banks all of sudden are profitable most days trading?  Did they all of sudden become trading wizards?

 

Packer

 

Exactly. Some of these investment banks don't even have 1 losing day. How can you have 365 days of no losses from trading?

[bobp]

"Virtu Financial Inc.,....... disclosing that it had earned money every day but one in the last five years."

 

They're frontrunning. Edit:That's of course just my opinion in case their lawyers are reading.

 

This is probably the reason IB trading desks can make profits most days.  How else do you explain that all these banks all of sudden are profitable most days trading?  Did they all of sudden become trading wizards?

 

Packer

 

Exactly. Some of these investment banks don't even have 1 losing day. How can you have 365 days of no losses from trading?

[rkbabang]

I know it's unbelievable. Read the book – Lewis explains it. They lure brokers into placing orders at exchanges where they get the information first and can execute their orders on the other exchanges before the large orders hit them. They are able to recognize from which broker and even from which firm orders come. They are able to figure out from whom the order comes from the time delay and route an order takes to hit one exchange after another. It's not that they know it 100% but they figure it out and the odds are very much in their favor – it's very, very educated guessing.

 

It has nothing to do with "luck". The more data they feed into their algorithms the better they know from whom specific orders come and at what price. When they figure out that T. Row Price buys 100 shares of KO at BATS they can be pretty sure that this is only a partial order and they buy KO on NYSE and NASDAQ and then sell it at a higher price.

 

How do you explain that slowing down orders so that they hit all exchanges at the same time works? This is not a tinfoil hat conspiracy theory. RBC actually sold this system to large institutions when they had figured out how to circumvent this scheme by delaying the orders precisely.

 

Nobody says that's illegal but it's front running nonetheless.

 

Of course slowing down orders so they hit all exchanges at the same time works. Why would you think that represents a flaw in the system? Previously, the broker was doing the following steps: #1 - clearly indicate to players on the NYSE a strong interest in purchasing stock (through activity on another exchange), #2 - deliver an order for stock to the NYSE. Naturally, it turned out better for RBC to do #2 before #1. That's the sign of a healthy interaction across exchanges.

 

Of course I'd like to be able to buy an unlimited number of shares and not have my previous, visible activity factor into the market price. Buffet and other big stock buyers would like this even more! (Is it "front running" to buy WFC on the basis I suspect Buffet will likely buy some in the future - or perhaps has an order in transit right now?) You don't have a right to this, however.

 

You now have no right to suspect?  You have no right to attempt to figure out the intentions of others?  This is going beyond ridiculous.

 

 

[benhacker]

I think this topic is strangely controversial.

 

The issue seems to stem from expectations.

 

Namely:

 

When a large fund investor sees offers at several exchanges at a given price, they believe that amount (across all exchanges) is "liquidity" for them.  This belief seems patently wrong, as if they forget we don't have "one exchange".  Reading the Lewis piece I was honestly confused at how big traders could be confused by this...

 

The mere fact of having multiple venues for trading means there WILL be arbitrage between venues.  Why else would they exist unless pricing, execution, fees, etc are differing between them.

 

It seems like for much of the last decade, managers of many large mutual funds have been thinking the bid / ask liquidity was equal to the sum of all bid / ask NBBO pricing at ALL EXCHANGES.  When it was really only the sum of the bid / ask liquidity at the largest single exchange (and that was only true if that exchange was direct routed the order...).

 

If I see someone trade LUKOY in US at $X, and at the same time the bid in UK for the underlying hasn't moved to $X adjusted for forex, am I an evil HFT if I take advantage?  If someone is buying IRE in the US and not buying BKIR in UK and a price discrepancy happens, shouldn't there be players that arb the exchanges to make prices efficient?  If not, should I be able to do it?

 

If you want to place an order to buy 100,000 shares of C @ $40.00, and you see 100,000 on "offer" but across 7 exchanges, and you only fill 30,000 at $40.00 and the rest of $40.02 (or $40.10, or $40.25).  Should you be surprised by this?

 

I'm honestly amazed that anyone who has traded in markets for more than 5 years should be.

 

The issue of FRAGMENTATION OF EXCHANGES is an issue I think should be discussed by the SEC.  The nature and special treatment or market makers should also be clarified and justified.  But the issues raised in the Lewis piece about HFT being bad seem strange.  There are many areas where "HFT" generally probably should be banned (quote jamming / infrastructure attacks, if there really is front running (I don't know), etc) - but unless I missed it, they weren't discussed (and may not be right to capture under the heading "HFT", I don't know).

 

I think if this discussion raised the debate so that the regulatory environment of the exchanges becomes more rational, that would be good... I have no idea why so many stock exchanges exist in the US and ZERO bond exchanges for example.  Demonizing HFT seems too easy, but I think Lewis (from what I've seen so far) let his narrative make more out of the story here than even he would normally do.

 

Writster, thanks for (what appears to me) bringing rational points into the discussion.

 

Something about "really fast" and "mysterious" makes HFT sound so bad.  If you just call them a market maker it wouldn't make the headlines.

 

Ben

[yadayada]

i read somewhere a while back that michael burry also had something to do with this and was collecting funds just for that. Anyone have more info on that?

[fareastwarriors]

http://video.cnbc.com/gallery/?video=3000263240

 

Lewis on CNBC

[ni-co]

@benhacker, @writser

Are you disagreeing with anything Lewis has to say in this video?

 

http://video.cnbc.com/gallery/?video=3000263240

 

Lewis on CNBC

 

As far as I understood your posts: You are not. It's much more about the perception or the tone ("stock market rigged") than about the facts. Most of this discussion seems to me to be a straw man argument around whether HFT is legal or "real" front running.

 

What I'd love to hear your opinion on: Do you agree that this is a (solvable) problem or do you say that this is how an efficient market should work?

[constructive]

I doubt HFT is happening to the types of stocks I buy, small illiquid ones, but I know of something worse.  That's the physical market makers who are buying the shares at $14 in little bits and pieces then offering them for sale at $14.50 in round lots.  Or buying the lots at $14 and selling at $14.50.  This is absolutely true because I know people who do this, that's the definition of their job.

 

I wish there was HFT for some of the stuff I am buying.  Then there'd be some liquidity, and bid ask spreads smaller than 30-40%.  I know market makers who sit on ask's in the 1000's for a $15 stock just so no one can enter an order online. 

 

As someone said earlier in the thread, if your investment thesis is wrecked by a penny or two then you need to step back and question what you're doing.

 

I've thought about this recently - it would be pretty cool to create a market making program for thinly traded securities. It would not need to be that technologically advanced since there is less competition in small nonscalable strategies. I think it would need to incorporate some value bias (quick human screening to set the parameters for what direction and how large it can trade).

[Otsog]

The issue of FRAGMENTATION OF EXCHANGES is an issue I think should be discussed by the SEC.  The nature and special treatment or market makers should also be clarified and justified.  But the issues raised in the Lewis piece about HFT being bad seem strange.  There are many areas where "HFT" generally probably should be banned (quote jamming / infrastructure attacks, if there really is front running (I don't know), etc) - but unless I missed it, they weren't discussed (and may not be right to capture under the heading "HFT", I don't know).

 

The market information asymmetry and obliqueness is being driven by the HFT firms.  Lewis may have taken artistic liberties with the delivery, but he is factual about the mechanical operations, I have absolutely no qualms with his vilification of HFT.  They are parasitic entities that are lowering the efficiency of public markets. 

 

 

 

[dcollon]

CNBC discussion on HFT with Lewis and others.  Thanks to ValueWalk

 

http://www.valuewalk.com/2014/04/katsuyama-vs-obrien-heated-hft-debate-full-video/?utm_source=mailchimp&utm_medium=email&utm_campaign=EMAIL_DAILY&utm_content=quick_link