'Flash Boys' IEX stock exchange opens for business
latimes.com
latimes.com
They have been operating for some time already as an ATS, today they are now part of the RegNMS protected quote, meaning that up until today, brokers did not have to route orders to them.
it also allows them to have companies "list" on their exchange, though I don't know of any companies that are planning on listing with IEX currently.
They bring 4 unique things to the market
- their built in delay of 350 micro seconds on each incoming and out going message, ie they delay order's coming in and they delay fill notifications going out.
- they have a patented dynamic peg algo that will allow a user to place an order that the exchange will dynamically price. This order type was a bit contentious as the exchange doesn't delay quote changes when updating these orders, ie these orders are blessed in the sense that they don't obey the 350 microsecond delay.
https://www.iextrading.com/trading/dpeg/
- they aren't currently allowing co-location
- they don't participate in the maker taker model. They charge, I think, 9 cents on dark orders and nothing for lit orders.
Maybe the most interesting thing that IEX has brought to the US market is that the NYSE is now filing to add their own discretionary PEG order.
https://www.bloomberg.com/view/articles/2015-12-22/the-flash...
TL;DR:
> In other words, the objection goes, IEX uses the magic shoebox (and its router's ability to skip the shoebox) to do "latency arbitrage" and trade with stale quotes on other exchanges before traders on those exchanges can update their quotes, just like it accuses high-frequency traders of doing to its own investors. And, the objection goes, if it's unfair for high-frequency traders to do this to IEX's customers, then it's just as unfair for IEX's customers to do it to high-frequency traders.
[but maybe this is just leveling the playing field...]
That is a recipe for vilification.
Think of the wrath of the unemployed truck and taxi drivers that will be aimed at the evil computer programmers and engineers who perfect self-driving cars and trucks. The headlines will be amazing.
Where did you get your claim from? I can source mine, if you really need me to.
Is there a socially acceptable way to call a person an ignoramus on HN?
That may not be accurate, but that's why I personally feel uncomfortable with HFT.
Meanwhile, cost of trading for normal people like us has gone through the floor. And we're only really looking at the last 15 years when we think about trading costs, but even steeper reductions precede that, and it was all brought about by replacing human market makers --- who are crooked as a barrel of fishhooks --- with automated systems.
Respectfully --- I don't know you, and this isn't a personal comment --- but my guess is that you distrust HFT because you've been told to distrust it. If you do even the most superficial cui bono analysis, you'll see the people most interested in making you believe that are themselves major financial institutions, all of them far larger than the HFTs.
It wasn't HFTs that brought down the economy in '07-'08. It was their adversaries.
The industry standard for buying/selling mortgage-backed securities or collateralized debt obligations isn't HFT. It isn't even automated. It is one sweating jock yelling at another sweating jock over a recorded telephone call. You can see this dramatized in The Big Short, where to unwind the shorts that the "good guys" have made they have to get their own not-quite-sweating not-quite-jock played-by-Brad-Pitt to do the phone calls on their behalf.
Yes. No. Maybe. There's a massive push towards exchange trading the more liquid products: traders are expensive, and if you can get a robot to do basic inventory management and market making for you (even with a human in the loop), that's savings for a desk manager looking to cut costs in a highly straightened FI environment.
The highly distressed and/or exotic stuff that people are talking about in the Big Short are still slung by salespeople, with the connivance/approval of traders.
Some of them worked for firms as professionals, but many traded on retail platforms similar to what you'd see from Interactive Brokers or TT these days. Still you are right that the average investor never traded in this fashion. At worst, the people HFT put out of work were trading as a profitable hobby similar to playing poker online.
But human market makers colluded routinely, and in some cases famously.
My only point was that some people think all HFT is bad, and here we have a complaint that actually the situation is reversed and it was actually all human market making that was bad. I don't think either statement is completely true, and I was trying to highlight the contradiction because the human market makers and the HFTs are, in a lot of cases, the same people.
† https://www.bloomberg.com/view/articles/2013-11-06/cftc-sues...
I think it's harder to describe how HFT is bad. Most of the efforts in this thread are, we'll say, only loosely connected to reality.
The best you can say about sweaty men shouting is that if not corrupt, still inefficient. (To say nothing of spreads in eighths.) I think that's a claim that really is true for literally every trader.
You also could see who was on the other side of your trade. Savvy floor traders would use that information to their advantage. If a sharp broker traded with you, you could hedge more aggressively or speculate in the same direction. Most electronic markets are anonymous these days so there's less information leakage.
HFT is generally considered good for retail traders because spreads tend to be lower. You trade both more cheaply and more quickly.
However, it's generally not good for large institutions (which are more than just 'big evil hedge funds') because markets react very quickly to movements caused by this big firms. If they decide that something is priced wrong, they won't be able to make many trades taking advantage of that.
https://www.bloomberg.com/view/articles/2016-02-25/-flash-bo...
Market has a hard time reacting to option spreads when the long option is executed prematurely. I don't see why big players can't use them.
Also the market can't react too predictably. Because then the big player could just yank the market around and profit.
Some of this is limited by regulations on large holders / insiders.
Are you claiming this doesn't impact the profitability of those people with directional views?
See Vanguard on this: http://www.cnbc.com/2014/04/25/vanguard-chief-defends-high-f...
The story of automation in the financial markets is similar to the story of automation in many other businesses.
I may have that completely wrong though.
1) It reduces the bid/ask spread. There isn't just one price for a stock, there are two. The price at which you can buy and the price at which you can sell. The price at which you can sell is lower. So when you buy a share of stock you are immediately down a little bit. This amount is called the spread. By automating firms can reduce this spread which does the opposite of what your intuition told you. It will bring buy prices down slightly and sell prices up slightly.
2) It helps make sure that prices are as accurate and up to date as possible. When you go to buy a share of $GOOG you probably aren't really sure if it should cost 775.40 or 775.45 or 775.50. You just figure it's a good company and likely to go up in the future. But because there are all these firms working really hard and acting really fast you can be pretty confident that whatever price you buy at at any given time contains the total available knowledge currently available in the world about Google's future potential.
Here how it works... Imagine you want to BUY 10000 MSFT...
You send your order to exchange A, it does a partial fill for 1000 orders, and sends the remainder to exchanges b,c,d. An HFT firm sees your order to exchange A knows its not going to fill and sends its own orders to buy the liquidity on B,C,D and then sends sell orders at a higher price to B,C,D, your order fails to fill and you have to issue a new order at a higher price.
Since retailers will very likely never exceed the liquidity on a single exchange they'll never have any issue with HFT and will just experience increased liquidity and faster fills.
However, if you're a large dinosaur still sending huge orders now you'll need a group of suckers who want to trade only with you, enter IEX, and 'consumer' protection from HFT on their exchanges who now has a large pool of suckers to trade with.
Maybe, but that seems like a pretty risky strategy. A simpler and far less risky strategy that would look very similar (admittedly only if you're looking exclusively at orders on the book and not fills) would be for HFT market makers to cancel or reprice their existing resting orders on exchanges B, C, D in response to getting or seeing a large fill on A.
In the strategy described by the parent, in addition to having to cross the spread, the HFT firm would also be at the back of the line at the next price level (unless maybe they already have an order there? but no guarantee that it's the right size, or maybe they have multiple small orders and cancel whatever is in excess of the position..).
So I'm genuinely curious: is what the parent describes something that is really that commonly done? This is one of the things that made me highly skeptical of Flash Boys. It seemed to me they observed a phenomenon, came up with a single explanation for it and never even considered any other possibilities that didn't fit the chosen narrative.
Imagine your neighbor owned a Ferrari and you told him one day you were going to buy a gallon of milk at the store. "On sale for $3.99!" you say to him. Now imagine he sped past you on your way to the store and when you arrived there he had bought all the milk at $3.99 and was selling it in the parking lot for $4.01.
You wouldn't necessarily be ruined financially paying $4.01 instead of $3.99. The cost is negligible. But you would probably think he was kind of a jerk.
A closer analogy would be that the neighbor is the owner of the local supermarket and you have recently announced that you were going to buy all the milk in the region. You go to one local supermarket (not his) first and buy out the entire stock of milk. On your way to the next one (his), he raises the price of milk to $4.20, knowing that your increase in demand is going to drive up the price everywhere and he doesn't want to be the idiot that sold you milk at $4.00/gal and will have to resupply at $4.20, losing 20 cents per gallon on that sale. When you get there your realize that the price of milk has gone up, and you yell and scream and stomp your feet, but you buy the milk anyways, because you have high demand for milk. The supermarket owner is not a jerk for responding to the increased demand for milk, but many people see it that way.
So maybe your a really rich guy who can afford a lot of milk, and so you lobby the government to restrict the ability of supermarket owners to talk to each other, maybe they have to wait a day or something. The supermarket owners are just going to respond by increasing the price of milk on average, because there are random milk thirsty people coming through every once in a while buying up all the milk, increasing the price, and they need to increase the milk premium so they can afford to resupply. They need to compensate for that risk. For that reason everybody loses out. Less people are going to buy milk due to increased prices, so that's bad for the store owner, and milk buyers are going to have to pay a higher price.
Same thing is going to happen at IEX. There is simply going to be a wider bid/ask spread.
That said, it should be mentioned that the "rich guy"/"institutional investor" includes various pension funds.
"In a letter urging the SEC to approve IEX as a full exchange, the Teacher Retirement System of Texas, a pension fund that manages more than $125 billion, suggested that trading through IEX could save the system millions of dollars a year."
What your described is more like real front-running: A customer places an order, the evil broker sees it and goes out in front to buy for himself then gives the customer a lower price.
Your buddy in the Ferrari reliably making money requires him to be able to properly predict. AND you're missing the other side of the trade! The poor shopkeep that priced at 399 when you were OK paying 401. Why should he lose out?
HFT market makers solve this.
The guy with the Ferrari would still outrun you and offer you a new price, with a margin just enough for them to be profitable, yet not substantially large as to talk you out of the deal altogether.
You should instead send 4 or 5 trucks to each location and buy the available milk, the guy in the Ferrari cannot outrun you because you're already there, and your order filled before he even got there.
Now imagine that you have at your finger tips a giant constantly updating database to the nearest millisecond about how much milk there is at every grocery store, but you still decide to take out a front page ad and announce your plans in advance, instead of breaking up your order into multiple parts and sending each to its own store. If you did this, everyone would laugh at you like they are at IEX and Capital Group.
In reality, this occurs i the stock market because the "national market" consists of something like 15 different exchanges in a big distributed system, so there are race conditions. HFT market makers trade on all of the exchanges, and adjust their prices based on trading demand seen from other exchanges. They spend a lot of money on fast networks between the exchanges, so that they can be/beat the proverbial ferarri. But synchronized trades a la tgemilk trucks or Katsayuma's "Thor" cannot be raced against.
IEX is intended to partially commoditize the Thor approach: take the advantage away from the grocers (HFT, market makers) and give it to the big milk buyers (institutions like hedge funds and pension funds). Retail trading is not affected one way or another.
The situation I'm describing is literally what happens in the real world. If a large trader wants to buy a big block of stock such that he can't fulfill the order on a single exchange he has to be pretty careful about how he executes the trade or the market will move against him. As soon as he purchases all the stock at the market price on a single exchange those selling stock on other exchanges will raise their prices.
It's very important to understand this. The price doesn't (generally) rise on other exchanges because someone swoops in and buys up all the supply. It rises because the people who were selling in the first place change their offers.
What you're describing is straight up hacking, and if there was as clear of a breach of the law like that, we wouldn't be having civil discussions about whether HFT is good or not.
Besides, almost nothing routes to Wall Street anymore. Especially not for stocks.
NASDAQ's trading platform is in Carteret, NJ. NYSE's trading platform is in Mahwah, NJ. Most HFT firms, if they're not already colocated in Carteret or Mahwah, are located in Secaucus, NJ.
https://www.amazon.com/Flash-Boys-Insiders-Perspective-High-...
Seriously, at one point, Lewis suggests that the trading station of some big trader is hacked. That just by typing numbers without submitting an order, stuff jumps. This should send huge red flags off on anyone that's even remotely familiar with anything similar to a computer. But it's another "see how rigged it all is?" anecdote blended in with his nonsense.
Which makes it all the more easier to get away with anything that sounds as sensational as this.
I've never quite understood the amount of HFT hate here. God's work it ain't, but it's a bunch of nerds using bleeding-edge technology and machine learning to break into and (dare I say) disrupt an industry whose avarice and sense of entitlement is perhaps unparalleled in modern history. It's tailor-made for this crowd, but somehow still gets a bad rap. I truly fail to see how that's less exciting than anything else anyone here gets worked up about.
In one corner, we have sweaty alpha male jocks [+]. In the other corner, we have geeks with computers. The geeks ran the table on the jocks primarily because the geeks can do math faster. The jocks complain that the geeks are cheating, because how can they be expected to out-math a computer and, also, isn't math just a little suspicious? And HN sides with the jocks?
[+] People might think I'm exaggerating. Videos exist of the daily life of a market maker. Here's a representative day at the office, circa 2000, in Chicago's open outcry options pit: https://www.youtube.com/watch?v=mvx3xM02iUs In stocks and bonds, you generally got to be sweaty in your own office (open plan, frequently) and yell at someone over the telephone rather than straight to their face.
So it's not so much that HN has sided with the jocks over the geeks. It's that (some portion of) HN dislikes both of them. And since the geeks killed the jocks the geeks are all there is left to dislike.
Part of this is also algorithm aversion (even here on HN) - watch how people are reacting to Tesla autopilot crashes. There is a lot of tin foil stuff about 'algos gone haywire', but I can tell you now that humans fat fingering in the market were both more common and more deadly.
You could say the same thing about ebola.
Which is not to say every single person is worse off just that this does not increase the long term value created by the companies who's profit makes the stock market a positive sum game.
www.investorsexchange.com
Hence they use the short form IEX.
I'm so happy Stack Overflow won that battle.
If technology exists, has been discovered, you cannot thwart it by trying to mask it by retarding it.
It think it is ironic that an exchange, IEX, founded with one of its aims being to level the playing field by calling out HFT as 'wrong' or 'unfair', and yet allow for their dynamic trades to bypass their 350 us delay is an example of we don't need it/we do need it.
People cry out about HFT being a game for the super players only, but out of it comes innovation - faster switches, well-honed programming tricks, hardware and software that eventually make their way to other uses. Sort of how the military and porn industry pushed a lot of tech we take for granted today.
I am not a trader, only an observer of technology. I do play with time series data processing, not networking, which is why I find HFT one of things I try to tap for tips and tricks. After all, a lot of very smart physicists and mathematicians being paid to work 40+ hours a week on it, does produce something worthy of study. A lot of optimizations that sometimes find use in other areas.
That was my comparison to the porn industry with online payment systems, and the military with DARPA projects and the early days of funding for what was the nascent Internet.
The crazy speculation that goes on in the minds of the 'quants' that work on these issues has no bounds, as brainstorming should before the realization stage. They are looking into quantum computing and using custom boards (FPGAs, ASICS and GPGPU) that find use eventually in things like bioinformatics, visualizations, big data processing and huge neural nets.
I program in J, and with Jd, which is similar to k in the APL family, you can process billion rows of data better than some of the systems built by the big money players in tech. k/qdb is mainly used in the financial sector, in which I have never worked.
My point was that it seemed kind of silly to put cable in a box to make a 'speed bump' to sort of negate something everyone already knows or has, and then still have to process their dynamic orders without the delay. I understand the intention, and I am not technical enough to suss it out, but it just seemed to go against common sense. That is all.
This talk presents compares discretization to standard continuous-time bidding, but doesn't go into a lot of detail about how it compares to IEX-style delays: https://simons.berkeley.edu/talks/eric-budish-2015-11-19
IEX delays almost all messages in and out by a small amount.
They don't delay messages for orders that their own router modifies, their discretionary peg order. This lets their peg orders update before any one, ie HFTs, can update their own resting orders on IEX or at other exchanges in the case of a fill at IEX.
It's important to note that even with the IEX design, speed is still very important as its still a price time priority queue for order placement and cancelling. An analogy would be if all 100 meter sprinters reaction times were delayed by 1 second from hearing the starters pistol. The fastest runner still wins, its just that all reactions times are delayed evenly.
Batch auctions are a whole different animal, and to be honest, one I'm not very familiar with. Their greatest weakness is that they don't operate in a stand alone environment, ie the rest of the world continues to trade around them which eliminates most of their advantage.
As long as you have multiple exchanges that are not perfectly syncd (a distributed systems problem!) There will be time based arbitrage opportunities.
Canadian banker Brad Katsuyama notices prices for stocks he is buying change price almost the instant he places his order to purchase them.
Turns out that this happens when he places a larger order than can be filled at a single exchange. After some investigation he determines that HFT traders "see" his trade on the local NJ exchange and buy up stock at other (further away) exchanges before said banker can fill his order at the original (lower) price.
His solution: Send orders to multiple exchanges but delay the orders to the closest ones so orders arrive at exchanges more or less simultaneously thus defeating that particular HFT strategy. The coil of fiber is the method used to achieve that delay.
Human market makers at those exchanges are entering their quotes(by hand even). And Brad Katsuyama uses new electronic technology to crush them. Taking their entire inventories and moving the markets against them.
The human market makers fight back and make their computers do the market making for them and Brad is hurt because his technology is no longer the best way to trade.
This new exchange uses the long coil of fiber to enable all traders from around the world to benefit from that protection. I'm wondering if they really had to go to that extent to get the speed bump. But hey since shame has yet to force formally verified software on anyone's radar, the long coil will certainly do.
Edit: Oops. Maybe they can bypass the coil for certain orders. Well that doesn't make sense...
For those interested it looks like his next book will be about Behavioral Economics.
http://nymag.com/scienceofus/2016/06/michael-lewiss-next-boo...
Pro Tip: If a large financial institution tells you they are protecting you, they are bald faced lying to you and you are about to get a serious lesson in business.
The same idiots who tell you that IEX is a good thing will also sell you actively managed mutual funds that A) under perform vanguard and B) seriously under perform once you subtract fees.
Rough recollection of a couple of paragraphs in the book. Someone else please add if recalled incorrectly.
There are a lot of SEC comment letters claiming IEX's delay coil is an "intentional device"[1][2] but the SEC approved their exchange application because they consider the delay to be "de minimis." One area where the IEX delay differs from simple geographic latency is that IEX delays incoming orders, outgoing fills, and outgoing market data, but they do not delay market data from other exchanges. That difference makes it more difficult to "pick off" a stale order on IEX, but on the flip-side, it also means IEX orders get to free ride the price discovery happening on other markets without taking risk.
This delay also raises issues because under Reg NMS, brokers have to route orders to exchanges with the best price when trading through an entire price. Since IEX's data is lagged, it may be showing a quote that is no longer accessible. Brokers would face additional slippage/execution uncertainty if they need to route at this phantom quote and wait 700 microseconds to find out it wasn't there. Canada has exchanges with delays similar to IEX, but IIROC ruled that they are not "protected" quotes so brokers have discretion not to route there if they feel it would harm their execution quality.
Or find a way to route around the 30m cable...
Alternatively if you want to lower your risk (which also lowers your expected reward) you may invest half in REIT ETFs and the other half in an uncorrelated sector. Think of what sector goes up when real estate goes down.
DISCLAIMERS of course: This is not professional advice. Invest at your own risk.
It might even be better to think about an investment vehicle more closely correlated with housing costs. The main Vanguard REIT is mostly commercial property I believe which could be somewhat uncorrelated with residential property.
If you can stomach a bit of risk you might consider buying something like Vanguard's Total Bond Market ETF.
Edit: WTF, downvoter: that's entirely what his question comes down to. If he's cool with it, then maybe something riskier might be warranted. If not, then 'sitting in the bank' might be ok.
If your edit, minus the wtf, had been in the original comment, then probably no one would have downvoted.
This or relevant CDs would be your best bet.
Move your emergency fund there, too, if you like.
These are probably the best "riskless" places to keep the funds. If you can stomach more risk, you can certainly do better.
It's easy to become resigned to finance being an insiders game. But the efficient allocation of capital is to important to leave to "the professionals."
Flash Boys: Not So Fast
https://www.amazon.com/Flash-Boys-Insiders-Perspective-High-...