Monday, July 20, 2009

Is Something Wrong with Certain Kinds of Trading?

Beautiful as France is, it's ski resorts are filled with pesky, queue-barging coquins with all their attendant bad manners, trampling upon one's skiis, and mayhem creation. It cannot help royally tick one off after civic-mindedly and patiently waiting in a queue and/or letting others take their deserved turn. As disturbing is to meander down to the hotel swimming pool after breakfast only to find that a contingent of Germans have staked out all of the pool chairs with their towels, never mind that they are off sightseeing on the tour bus, or being buffed and pedicured in the spa. Not dissimilar are the feats of the notorious Frigate Birds who routinely make their diets eponymously stealing the fish caught by pelicans, boobies or other more industrious birds.

What is their common thread? For one, they are reasonable analogies for a flavour of low-latency, high-frequency, order-sniffing, front-running strategies that parasitically intermediate real buyers and sellers, but only after excising a requisite toll that CUNY's Donefer (thanks to TD at Zerohedge for posting the presentation bullet points) estimates to be in the range of USD$15 to 25 billion. Of course not all short-term traders are predatory. and not all predatory traders are short term. However, given the estimated sized of frigatted profits, responsible civic-minded investors should at the very least ask some questions about its legality. desirability, as well as the enforceability of restrictions (not necessarily in that order) and the philosophically slippery slope and unforeseen externalities any restrictions might ultimately yield.

But before I have my say, allow me first to express my admiration at the undertaking. I admire it's cleverness, its technical feats, it's use of probability and game theory, and it's predation upon what is essentially others' doltishness, not having yet come to terms with technology or conceived of others' misuse . Having said, that, this still doesn't mean we - the royal "we" - should admire it or even encourage it, but we can still admit it IS clever - in the same way that the locals in Hong Kong would watch the Barings booth, and sell futures whenever the Barings phone rang on the floor, having learned that most of their index arb was long stock and short futures and that a knee-jerk sale was a reasonable bet for a scalp, or like the clever URL-nappers who registered and held hostage names like HOTELLS.COM and HOTELSS.COM and all other poorly spelled permutations, re-directing mis-hits to other places until the purveyors grudgingly bought back the misspelled real estate. Saints or sinners - what sayeth you? Or perhaps you nihilistically matters not in the least.

But really, how should we feel about it?? That USD$15 to 25 billion is coming from somewhere - probably your mutual fund, your defined-benefit pension fund, and possibly your own pocket via your P.A. While it's true that no one dies as a result of the clever gamester-ing, almost everyone is hurt, albeit small. Is this sufficient violation of the public interest to warrant action? Let's examine the boundaries: At one end, "front-running" of the contrapreneurial sort (i.e. on 100% information) is decidedly illegal in most civilized markets. At the other end, speculative buying on reasoned information or no information (like that peddled by CNBC) is wholly acceptable and encouraged. But predatory, parasitic order-sniffing strategies lay somewhere in-between though closer to Darth Vader than Alec Guinness, for one might aptly describe it as front-running on 75% probable information where said information is cleverly inferred from quasi-public, though admittedly obscure and difficult to tease-out information. This is not an easy case to find fault or guilt, even for the likes of Solomon. Call it legal-to-the-letter, but with substantive externalities negative to the public's interest.

But perhaps there are some redeeming qualities to the undertaking of parasitic trading based upon hidden order-revelation and algorithm-spoofing that shakes the electronic tree yielding the sought-after fruit? Perhaps they add liquidity - qualitiatively or quantitiatively? Errrr, methinks not. Their pings, probes, posts and cancels, add nothing to market quality or integrity. Their "firm" 100-share bid or offer is a quintessential poison chalice. You taste it at your peril and ultimately with regret.

One must look outside the marketplace for their "positives". They do spawn technological advances, and neaten the parks and libraries in Stonybrook, LI while increasing the municipal coffers in Chicago. It might be thought to generate tax revenue, though if its a zero-sum game, there is no net gain systemically. It has a Randian free-to-pursue cleverness-rewarded element, but even here, the conentious-philosopheresse (with lower-case "p") might begrudgingly yield on logical grounds of harming the public interest.

Might the main exchange itself have something civically useful to say about this? Sadly, the civic incentives of the exchange were jettisoned long ago, and are now somewhat shortsightedly intent upon solely maximizing turnover and volume, on behalf of its shareholders rather than pandering to the interests of such pedestrian constituents as the listed companies and their investors - irrespective of their primacy to game itself. Sympathy for the buccaneers runs deep, however - perhaps by tradition for specialists were themselves monopolists who systematically abused privileged information -however inimical this was to the community at large.

In the realm of The Trade , it's hard enough for bona-fide buyers and sellers (with horizons of longer than intra-day) to locate each other and maximize achieved results, or minimize slippage, without service providers themselves and all the minions privileged to transaction information front-running (whether large or small) without a dedicated army of thieves kitted-out with the highest-tech gear, and the shortest and quickest lines to the exchange tripping you up, knee-capping you, picking your pockets, at any and every opportunity. In this sense, they are not merely annoying gnats to be waved aside, nor even pesky mosquitoes, but nasty flesh-tearing financial midges or black-flies inflicting real damage. "Ouuch....f*&k that hurt!!"

Which begs the question: in the absence of pecuniary restrictions (message fees, fees for cancels? Tobin taxes? etc) or other yet-to-be-conceived outright heavy-handed rule-making, how can normal participants fight back, and snooker them in return. Hidden cameras watch the croupiers and box-men, keeping them honest. Surveillance exists on the few remaining pits, ostensibly though ineffectually to try to catch the covert shin-kicks, ear-pulls and information abuse that was rife anywhere rules or privileged information could possibly be capitalized into a dollars. But employ a series of probing orders, or a cute predictive algorithm upon a numerical sequence that does nearly the same thing (albeit with an element of risk and uncertainty) with the same intention, and Presto! one is sanitized and legitimate and courted by the very exchange whose constituents one preys upon. Hmmmm. Something doesn't seem quite right. They serve no purpose, even if purposefulness is over-rated.

Still, the uproar outside rantings of ZeroHedge is decidedly muted. Exchanges make money. Brokers make money. Libertarians dislike any regulation or restriction. Perhaps regulation is simply not possible and pecuniary control is not desirable. But there remain weapons at the disposal of both perturbed traders, and the daring and the bold. Guerilla tactics. In simpler times, when Atari elicited a "wow" instead of a "huh?", this might have been achieved by leaving sizable but plausible limit, stop or MOC orders, then pulling them to see who and how much was leaning on them. Now more cunning is required. Torquemada-like fear and surprise. Spoof them back. Double-spoof. Triple spoof. In size. BOOOO!! Randomize. Maybe relax or eliminate all restrictions upon transaction etiquette, thereby allowing trade with oneself to paint the tape as required shake the parasites, perhaps leveling the playing field. Or merely to make it more like quick-sand. Declare all-out war so anything goes. Let - no, encourage the 'bots to fight and predate each other. "Greetings, Mr Anderson". Is this real or are we in the Matrix??! Hunt the hunter. "Kill the bear" as Anthony Hopkins rallied in The Edge. "What one man can do - another man can do!!" Heck, it's not about investing anymore (if ever it was) - its about winning the game, and as this time of The Quickening approaches, an algorithmic battle to the death, an epic battle cannot be far behind, leaving in its wake, RAIDs, bandwidth, over-educated Russians with no scrupples, recursive bloodshed and, then yes, SkyNet...

And all the while, the thoughtful genuine investors, will watch with morbid fascination while exiting such venues and making greater use of alternative auctions that somehow keep out the riff-raff, reveal less and more optimally perform the true role of the exchange - to raise capital for growing enterprises, to match true buyers and sellers in the secondary markets.


Gunther said...

If the big players get too much of the financial cake the small players stop playing after realizing the hard way they did not win.
As institutional investor(I am not one) I would want a broker who does not deal in the market to exclude the conflict of interest.
Over time the skimming of the market will be reduced, by most players leaving the unprofitable place or a reasonable amount of fairness enforced by a honest regulator.
Regulation seems to me the faster and less painful way.

Macro Man said...
This comment has been removed by the author.
Macro Man said...

I am not expert in the finer points of the microeconomics of equity trade execution, but in what way is the so-called "liquidity rebate" offered by the exchanges and ECNs not a kickback? Surely a simple solution would be to ban said rebates and, as the chap rom Themis suggests, enforce a minimum period for which any market order is valid...?

danps said...

How about starting up a government-run exchange with a handful of simple, clearly enforcable rules as an alternative for those of us not in the know (a public option, as it were). Let the Masters of the Universe play in their libertarian paradise to their hearts' content, and provide a safe haven those of us who want a more transparent system. Win/win, no?

Anonymous said...

Sheesh. Do people ever get taught the most basic elements of political philosophy? To be clear, libertarians do not condone pulling the wool over others eyes so to screw them amid business transactions. Full disclosure of this seems to be missing, and as such is a breach right off. Moreover, before you rail on “libertarian paradise” (even if doing so with a straw man), understand that the system as is, is anything but. It is heavily regulated, and does quite nicely at keeping the hurdles sufficiently high to dissuade any genuine free market competition.

As for your government-run solution, yes -- great idea. Why, we can even have a bunch of Goldman alum from the Treasury run it. That will guarantee it is fair. (From the frying pan into the fire.)

danps said...

If the current system is "heavily regulated, and does quite nicely at keeping the hurdles sufficiently high to dissuade any genuine free market competition" then let's throw that yoke off and allow the Goldmans of the world to pursue truly unfettered capitalism. Sounds like what they would consider paradise, so let them have at it. I have no problem with that. The rest of us should have a more understandable (simple, transparent) market to play in.

The problem with the current system - Goldman alum going through the revolving door to the government - is that there is too much discretion. The SEC can be intimidated or captured into not enforcing the regulations on the books (e.g. "We were concerned about generating volatility where there were large pre-existing open positions, and we wanted to start afresh with new regulation, not re-write history"). Right now the government is engaged in the fool's errand of convincing investors the system isn't rigged while allowing a select few favored players to rig it. I say split 'em apart - give the favored players free rein in a completely unregulated environment, then create a second one with a handful of enforcable rules that require no human intervention to enforce: Make it as vanilla as possible for those of us with less of a taste for swashbuckling free enterprise.

Provided I have just demonstrated an understanding of the most basic elements of political philosophy, what objection could you have to that?

dave said...

I favor a transaction tax, with a rate that varies inversely (and exponentially) with the duration of the position, fading away to near-zero as the position duration approaches a minute, and hitting 10% at about a tenth of a second, on its way to an asymptote at zero.

C said...

Can someone explain to me what the advantage is of matching trades in a time period smaller than a second. It seems to me that all these problems would be solved if the exchange simply said: we match trade flow every second or every five seconds. You get exactly 60 or 12 market prices a minute.

I just don't get how allowing distinct trades to take place every microsecond could be good for markets. It completely obviously fragments liquidity.

Anonymous said...

Two points:

1. I know a few high-frequency trading guys and, to the very limited extent that any of them has been willing to speak at all about what they do, it is apparent that a very large number of them got their start by designing the computer algorithms that are used to execute orders for regular people, and at least a few of them got their start by being friends with somebody else who had designed such an algorithm. Essentially, they are making their money by taking advantage of the flaws in the systems that they designed in order to front-run the people who are relying on those systems. Not that they intentionally built the flaws into the system, but that, having discovered the flaws, or at least suspecting where the flaws might lie, they chose to exploit the defects for their own profit rather than to fix them. Kind of like a designer of burglar alarms who discovers how to circumvent them and decides to use that knowledge to rob houses - except this is legal.

2. What's going on with all of the banks (like Goldman Sachs, of course) that have earned zillions of dollars with these high-frequency trading programs? If there are defects in the algorithms that a bank is (in effect) selling to its customers, shouldn't the bank be fixing those algorithms for its customers' benefit instead of exploiting the defects for the bank's benefit? Why haven't they hired hundreds of programmers to fix the algorithms that they use for their customers and spent millions upon millions of dollars to create low-latency order systems located only microseconds away from the stock exchange for executing their customers' orders?

Charles Butler said...

Hey Cass,

You should consider that the reason the story doesn't get traction is precisely that ZH is its main proponent. He seems to think that GS is simultaneously:

1). Buying and holding huge numbers of stocks in order to support the market as an agent of the government;

2). Buying and holding huge numbers of stocks in order to later dump them on unsuspecting retail types;

2). Is the dominant, market direction neutral HFT player;

3). Is plain and simple front running order flow.

It's easier to imagine ways in which these strategies conflict with one another than the opposite. But that doesn't matter when your goal is to attribute all evils to one actor.

Your thesis needs some support from someone who is not playing to a lowest common denominator crowd.

You should look on Thursday as a good omen, btw.


Anonymous said...

Ms. Cassandra,

With your usual eloquence, you bring up some interesting points about high frequency trading. Being involved in the field myself, I have a few observations about your post and the comments.

1. High frequency traders (HFTs) typically net around 1 mil ($0.001) per share (including liquidity rebates and fees) -- this translates into a possible profit of around $2.5B dollars if every share traded on every exchange throughout the world were done by HFTs. I think the $15B - $25B numbers are wild ass guesses with no basis in reality. Or, these numbers include other types of trading that are not really high frequency.

2. I don't know of any flaws in the major exchange trading platforms. If someone has specific information about such a flaw or what you consider a flaw, I would be interested in hearing the details. And, so would the exchange. Dark pools, on the other hand, are a different beast entirely. Trade at your own risk. The owner of the pool will definitely be using the order flow to optimize their trading. But the public exchanges are as level a playing field as I can imagine.

3. The competition to add liquidity is fierce. If MSFT is trading with a one penny spread and is getting ready to tick up, to post the first bid at the new price level you have to get you order into force within a few hundred microseconds. And being among the first to post liquidity is necessary to be profitable. Analogously, you also would like to get your ask out of the way before the price ticks, hence the quick cancels. As far as I can tell, this does not disadvantage longer term traders. On the contrary, the price of MSFT will always be current and there will always be a competition to provide you with liquidity because if I don't someone else probably will.

4. Changing or eliminating liquidity rebates will not significantly change the composition of trading. This is borne out by both theory and empirical measurements between time spans with different liquidity rebates and fees and between exchanges with different liquidity rebates and fees. The reason is simple: As the liquidity rebate decreases, the average spread in the instrument increases by the the same amount offsetting the change. This happens because of the competition to provide the liquidity. With the rebate, the adders can be more aggressive and still make money. Without the rebate, they have to be slightly more conservative. Effecting the change you want in an open market with competition is a treacherous endeavor -- unintended consequences abound.

5. It is not obvious to me how printing individual trades fragments liquidity. For comparison, the Tokyo Stock Exchange does match trades at multi-second intervals because of technology issues. They have wider spreads than in the US and no more liquidity as far as I know.

I am open to some informed debate on the issues. For obvious reasons, I would like to correct some of the misconceptions out there about HFTs, but I am open to ideas about how their activities may negatively impact the market.

chad rich said...

Ok fine. I agree. Essentially, what the hell are the specialists doing now? They've been usurped.

But, you think the NYSE will allow my stack of cisco servers on the floor? If they did, I'd have no problem.

One thing I don't get is before long, the NYSE is gonna be broke from paying out these rebates. And, by the time they stop paying them the specialists will be gone. liquidity will be at a premium. Then, investors may end up losing just as much as they are now through the increased spread.

I don't know. But, it isn't fair. Because I can't put my servers on the floor of the NYSE (like I'd know what to do with them). But, that doesn't mean there aren't retail traders that wouldn't.

Bring back the nickel.

"Cassandra" said...

Anonymous 10:23 - These are interesting anecdotes, though to be clear, I mention no-where GS.

Dave- a transaction tax treats the symptom, but not the problem. But if you had to implement one, yours is neat (ignoring the measurement and book-keeping practicalities

C - you've made a wise observation, and the reason is the same one the casinos use to program higher payout frequencies (of smaller amounts) of the slots near the entrances and thoroughfares. This entices and primes the punters like a Pavlovian bell. Also, informed traders do better in continuous markets - picking off the innocents all the way down to the clearing or new equilibrium price. I wonder whether once a day is sufficient - though balancing the book at a clearing price may take some time as imbalance dissemination and real-investor response moves like treacle.

Charles - 'Tis a good omen indeed. Thanks for noticing. And you are of course correct that histrionics - in the end - minimizes attention. As for my thesis, I would suggest it relies not in the least upon ZH. He happened to have a posted a link to a paper that was the first I'd seen to quantify something that had been bugging me for years. The thing is: the natural world IS a dog-eat-dog place. That said, exchanges are man-made constructs, and so we can design them to minimize parasitic and uncivic behaviour and participation. I could swear to you that the airlines are getting good at this too: inquire about a flight too many times and they raise prices on spec, then lower them with some decay function. It's not dissimilar.

In the market when harassed by one, one gets the distinct urge to fuck them back, find the server and tear its knobs off, sever its cables use them to garrot it's Dr Tyrell, then launch it off a tall build. But its a machine, and Skynet will insure one is tracked down...

Charles Butler said...

I´m aware that you´ve been all over the matter for a long time. Didn´t mean to imply dependence on your part.

In case you´re not an aficionado, they go in inverse order to the GC. The rottweilers don´t get released til about 4 PM. Cheer for Wiggins - this year´s long tail event.

Consumed with envy.

Anonymous said...


A couple of points of clarification:

1. Exchanges will not go broke from liquidity rebates because they charge takers a liquidity fee. Typically, the liquidity fee is $0.001 more than the liquidity rebate, so the exchange makes a small amount on every transaction. The open competition between exchanges has driven these fees to very small levels, but the exchanges will keep them large enough to remain profitable.

2. The NYSE and most other exchanges allow connections directly (i.e. an ethernet cable) to their trading systems. You do need to be a broker dealer or be sponsored by a broker dealer for this level of access. That is a minor hurdle and a reasonable control and monitoring requirement for direct access in my opinion. And any retail trader can access this technology indirectly by using a high performance retail broker (e.g. Lime).

3. You are correct that the specialist have been marginalized. They are only kept around for show and tell when NYSE is trying to convince a company to list with them. The specialist have been unnecessary for a while now and aren't really a big part of the competition when trading.

You seem to be very upset about the fairness of the current system. From my perspective, it is much more fair now than when the specialists used their privileged position to extract monopoly profits from outsized spreads. Volatility adjusted spreads as well as the related trading fees have been consistently driven down by the open competition that currently exists in the markets.

My belief is that the current openly competitive system, driven in large part by HFTs who go home nearly flat every night, allows you to get the best price knowable with minimal friction at any time during market hours. I am open to actual counter evidence to this assertion, but I haven't heard much yet.

If you want to pay a nickel to trade, I am sure there a plenty of brokers that would be willing to execute your trades.

"Cassandra" said...

Anon. said:
"From my perspective, it is much more fair now than when the specialists used their privileged position to extract monopoly profits from outsized spreads. Volatility adjusted spreads as well as the related trading fees have been consistently driven down by the open competition that currently exists in the markets.

You are not setting the comparative bar very high. The specialist system was not just a monopoly legalizing higher spreads, but a monopoly that de facto tolerated perhaps even encouraged front-running - whether directly or indirectly. You couldn't have designed a lamer system more conflicted system more derisory in respect of the varied constituents of the exchange.

The cost to trade is the cost to trade - particularly when ultimate trade size is well-outside the immediate book whether hidden or revealed. You can't conjure point-in-time liquidity where is doesn't exist, and those that do conjure it, obviously charge for it. But the HFT order-sniffers and front-runners are predatory/parasitic (meant to be descriptive NOT pejorative) in nature, and not liquidity providers, and as such I fail to see how further intermediation between the end risk buyer and the risk-seller whether natural, long-frame short, or short-term liquidity provider can possibly create value, enhance value or provide liquidity.

"Cassandra" said...

just to be clear, I'm not saying its illegal, just undesirable, and admit it's a slippery slope to restrict. From this point of view, I am looking for defenders to explain why it is or might be desirable and am open to actual counter evidence to this assertion, but I haven't heard much yet.

Anonymous said...


I agree with everything you wrote except the premise that HFT is predatory and/or parasitic. I will attempt to provide counter evidence to this assertion.

My understanding of HFT that adds liquidity is that the strategies are very analogous to the typical market-making strategies of humans in past eras. They post two-sided quotes and as their position varies away from zero (i.e. informed traders start taking liquidity more on one side than the other), they adjust their quotes to try to balance their position. They also adjust their quotes if the price looks like it is going to tick up or down. Of course, all this happens extremely quickly (response times in the hundreds of microseconds).

I submit that this is simply the open competition, modern version of old-school market making without the privileged position or other shenanigans and with less cost to all parties involved.

Let's analyze interacting with this HFT liquidity from the perspective both a small and large utility trader. Let's define the small utility trader as needing to execute orders that are less than the typical inside size and the large utility trader as needing to execute orders that are much larger than the typical inside size.

From the perspective of the small utility trader, interacting with the HFT liquidity is great because they can execute their trade immediately at the inside price with little cost and minimal spread. This is a definite improvement over the past.

From the perspective of the large utility trader, however, interacting with the HFT liquidity is not nearly as much fun. They can execute the first portion of their trade immediately at the inside price with little cost and minimal spread, but the market will very quickly adjust to a new equilibrium if their orders affect the balance of trades for the instrument.

I claim that this is exactly how the market must function. If you start buying enough of something, an efficient market will adjust the price upwards as the information about your buying filters to the participants. To the extent that this did not happen in the specialist universe, filling those large orders was being subsidized by other activities because you cannot fill large orders at the same spread you can fill small orders and still make a profit -- the large orders contain too much information.

I also claim that the current market configuration provides a good solution to filling these large orders. There are multiple players competing to fill the order and they will give you their best possible price allowing for only a few mils of profit for themselves. I trust open competition much more than a central authority when it comes to filling an order at a fair price. Admittedly, I don't have a direct measurement immediately at hand to demonstrate this efficiency.

I have attempted to give evidence that the current HFT that add liquidity in the market do indeed serve the market-making role of adding liquidity and that the open competition forces them to do this efficiently.

I agree that their is a continuum here and trying to regulate the micro-dynamics of trading is a slippery slop with a good chance for unintended consequences.

And, while you claim that order-sniffing, front-running, predatory and parasites may be descriptive terms and not pejorative, they are certainly not terms expressing a neutral point of view. For casual readers, using such language tends to drive the conversation away from fact-based discussion and i don't that is your objective.

Thanks for listening.

"Cassandra" said...

anon 10:07 - Thank you for your thoughtful and nuanced reply. I agree with almost all of what you say, with a few key exceptions. First, I am careful not to lump all HFT into the same cart as I specifically stated that not all HFT is predatory and not all predatory trade is HFT. Bona-fide market-making was precisely the reason for this.

My first venture into the professional world 25 years ago shared an office with Petterfy and Timber Hill. His luminary financial partner had written a lovely thoughtful working-paper meditating upon disciplined electronic market-making that anticipated all of what we currently see (and then some) excepting the predatory shenanigans. What was (and remains fascinating to me) is that position acquisition in market-making in one instrument creates natural opportunities for market-making in many others, and linking these and consequential risk management is the holy grail. This is true of cash and derivative markets, as it is in related instruments that some would term chinese hedges. After 25 years it's amazing to me that many MMs still haven't cracked this nut to its full potential. Anyway, I wholly approved then, and continue to believe a good MM algorithm with attendant risk mgmt will always be better, fairer and more efficient than Vinnie and Tony, perhaps not as profitable (if it isn't abusing the structure), but less likely to spend time in Allenwood or pay fines to regulators who disapprove of such practices.

But as you know I am talking of "the other HFT", the one that is not first providing liquidity but demanding it on the basis knowledge garnered from it's privileged position, clever order-sniffing and real-order location gymnastics, or even clever inferential data analysis, in order to supply it thereafter. This IS parasitic. This adds nothing. This games the system. This gets in the way of bona-fide buyers and sellers and unnecessarily taxes and undesirably intermediates. I am still waiting for the rationalization (probably Randian) that says it IS good, or desirable, or noble.

jck said...

"the other HFT" you don't like merely replicates what the "locals" have been doing on futures exchange floors for decades, there it paid to have the body of Chicago Bears and now it pays to have (electronic) speed. As there are multiple trading platforms with different latency and therefore different bid/ask, the HFTs keep the various markets in line, it's not parasitic, it's basic arbing, low risk and low pay and it's better and cheaper than the old "specialist" system.

"Cassandra" said...

jck - seem to suggest its benign because its less bad than the structure that was very bad. You also dismiss what it as "basic arbing" but I think most cognoscenti would beg to differ. Finally you sort of contradict yourself by saying it's like the floor (a single market where rife illegal front running is/was done by secret signal, knowing looks and shin-kicks from those with privileged order information to those who could profit from it for a slice), but then you go on to say this is NOT what they are doing, since they are just cleverly sitting in the middle of different venues. I do not have a problem with inter-market arbitrage for shame on the broker who is not working all venues. But front-running across venues is still parasitic front-running, however sliced.

jck said...

Cassandra, there is a simple fact of life: if you want to trade in a market it will cost you, whether it is specialists in a funny jackets or HFT bots. I see HFT as benign and normal given the multiplicity of trading venues AND the fact that they all have different technical specs and latency, so you must expect prices to be different and arbing to occur. For the front running bit,I am sure it existed on futures exchange floor but much harder to do electronically because with e-trading you also get e-audit is not illegal to try to guess where the hidden orders are and that's not front running in my book. Problem for old fashioned asset managers, life is a bit harder and they generally lack the skills to operate in these markets, hence the whining...

"Cassandra" said...

I understand the simple fact o tradng life...really! But structure and participants actions can make it more or less expensive to transact. I think the public interest is served where it costs less (both expliitly and implicitly).

I realize it is not illegal to guess which way someone is going, or will go, though it is illegal to use material non-public information to do the same. As I suggest in the post, which is not a whine, it is grey where some are using their privileged positions to "guess", or where they are gathering the obscure information through ways in which the venues had not intended, nor logical structural construction had not contemplated, the result being the "guess" (again as I suggest in the post) is closer to being materially informed than "a guess". I am not suggesting it is illegal. I am saying that Dr Tyrell should consider this when he goes back to the drawing board on exchange structure v.5.2 - something he can do in Blade Runner II because prodigal-son Roy (and this is my new plot) killed a likeness of Dr Tyrell, and the real Dr Tyrell lives!

Eric Hirschberg said...

I just throw in the comment that the Hong Kong futures predators were trading off the Merrill Lynch booth and not the Barings Booth. They actually went as far as to suspend trading until we told them what it was we were doing, and then proceeded to game us. I think gaming real orders adds no benefit to price discovery and therefore is a detriment to capital markets in the whole. That said, I don't see a clear way to delineate the intent of an order. I believe that the answer lies in using dark pools, but I'm not sure that that cannot be gamed by manipulating the reference price, in what would become an illiquid exchange mechanism (assuming most orders went dark). Another solution would be a quasi continuous market comprised of a series of discreet interval auctions. Sort of a digital solution to an analogue problem. I am a proponent of this solution as it acts to re consolidate liquidity to a single pool mechanism (think 2 minute blind auctions)

My 3 cents

Anonymous said...


I am unsure specifically what you mean by delineating the intent of an order? What possible intents are there for an order you want filled? I would claim that the intent is to get filled at a favorable price relative to the price at some future time.

I don't think dark pools facilitate price discovery at all. In fact, they tend to fragment liquidity. Not to mention that the owner of the dark pool typically allows their proprietary trading group to trade against the pool (and it may not be completely dark for them). That seems like a terrible solution. The playing field is not at all level or auditable. The dark pools are popping up because it is more profitable than sending those same customer orders to the public exchanges. We consider those dangerous waters in which to tread -- and our business is to be the shark in the water.

The Tokyo Stock Exchange has quasi continuous order matching (for technological reasons). The spreads and fees are higher than in the U.S. and there is less liquidity overall. And to boot, the HFTs that are there have a higher profit margin on their TSE trading than their US trading. This is probably due to less competition because of the higher barrier to entry and non-standard matching algorithm.

If you want to consolidate liquidity and make it easier to trade large size without moving the market (at the expense of making it more costly to trade small size), you need to widen the minimum tick size and use a pro-rated matching algorithm. An example of this are the Eurodollar contracts on the CME. There is very large inside size. The pro-rated matching levels the speed playing field by matching all orders at the same price level in a pro-rated fashion relative to their size. This takes away the advantage of being ultra-fast at posting your order at new price level while at the same time encouraging you to post larger size.

However, this simply extracts a tax on the small size (which could happily trade at a smaller spread) to subsidize the large size having less market impact. I don't see that as necessarily fair or beneficial for the market. The question becomes how to tune the minimum tick size to maximize your desired liquidity characteristics.

In Europe, for example, they have different tick sizes for different price ranges which a crude knob to try to balance the liquidity equation.


I want to respond, but I don't have time right this second. One question that will help me though: could you more specifically characterize this "other HFT"? Are these the take liquidity HFTs? Or, the arbist talked about by jck? I am still a little fuzzy on the exact behavior that you think is predatory or parasitic.

Thanks to all for the input.

Anonymous said...

Nice discussion! I'll jump in with two feet.

The comments above seem to stem from some Judeo Christian (or maybe Anglo Saxon?) sense of "fair" that would be laughed at by Darwinian Nature (with a capital N). These "egregious" expenses have directly make it possible for someone to sit at a terminal and get a decent fill trading 200 names at the click of a button without having to hire a roomful of expensive traders to spend all day working orders. Evolution in action. The fact that some floor trading monopolies made outsized profits caused smarter participants to use developing technology to steal their business. Sucked along by Moores law, this is now a death march to ultimate efficiency or a single surviving monopoly player. The survivor(s) will in turn be at the mercy of someone cleverer changing the game on them yet again. Just like Nature or Andrew Lo's Adaptive Market Hypothesis.

Current profits are not as risk free as people seem to believe.
You must get your probabilities correct, filter outliers, deal with hardware issues, software bugs, etc etc as alluded to above. In any case, costs are probably far below the noise level of prices given by some kind of Graham and Dodd fair value calculation with all attendant assumptions. In fact there are some optimization problems where adding a bit of noise to a system produces a better result. Perhaps this is one of them.

On balance, I would say that, like the parasites in the human gut, HFT serves a useful purpose.
I find it interesting that there is a lot of noise about this now while no one complains about other structural inefficiencies such as limited market hours or quarterly cycles which probably cause bigger dislocations with e.g. uneconomic behaviour into the close.

"Cassandra" said...

Anonymous -11:15

In Darwinian nature with a capital N, the hordes would be looting and pissing all over your servers, doing unspeakable things to the wife, and kids, squatting in your waterside home, unless of course your private army was able to keep them at bay. The rule of law, and more importantly, its incremental improvement, and its ability to adapt to the evolving times, has its basic advantages and contribute immensely towards human progress. I accept your point that yours is hard work and not failsafe. So is safecracking, cat-burglarly, cocaine importation and sex in poublic toilets with George Michael. This has absolutely nothing to do with whether the the undertaking has utility (or negative utility) to society and warrants pecuniary restraint. This is the discussion, and while I possess my suspicions on the subject at the moment, I reserve judgement for it may be that the slope is too slippery and one cannot do anything about it without unpleasant unintended consequences.

I agree with you that there is a reasonable amount of other uneconomic behaviour too, and greater transparency all around (e.g. sunshine time&sales of end-user purchase and sales around quarter-ends would go a long way towards shaming such behaviour.

Anonymous said...

The leech produces an anesthetic to numb the flesh as it attaches, thereby avoiding detection. The true nature of the current financial system is in the process of being detected by the "good" people. The next government, brought forth by an acute anemia may rip the sucker off and step on it.

Mencius Moldbug said...

anon and Cassandra,

Denninger describes the "other HFT" quite clearly in this post. Clear, of course, does not mean correct.

The question is whether HFT players can, as Denninger describes and Cass suggests, hack the trading system to determine limit prices that should be invisible to all traders, eg with "exploits" such as systematically issuing and then cancelling trades. If so, it's a scam. If not, it's not.

I for one would find the response of any HFT defender to the Denninger post quite illuminating.

And why can't these markets simply use a central limit-order book? Simple, optimal, unhackable. At least to my eye.

Anonymous said...

Mencius and Cassandra,

And why can't these markets simply use a central limit-order book? Simple, optimal, unhackable. At least to my eye.

I agree and I fully support forthright, centralized matching engines (e.g. like used by the CME). Not having to keep up with the latest innovation in special order types allows one to focus more energy on competing to provide the best price. I think many (but certainly not all) HFTs would agree.

Admittedly, I may be biased because that plays to our strength. We have never tried to gain an advantage by using special order types or a privileged position because we always assumed that those were short term opportunities and as outsiders (not anywhere close to Wall Street) we would loose at that game.

I believe the only reason that the CME can operate in their current mode, however, is that they have a somewhat protected market position given that you cannot trade their products on any other exchange and they currently have the liquidity. This is in sharp contrast to the equity markets both in the US and Europe where there is a fierce competition between exchanges to attract liquidity. One of the competitive mechanisms for the exchange is to differentiate themselves by offering unique order types that make it easier or more advantageous in some way to use their exchange to execute your orders. Hence, the proliferation of special order types (there is really an unwieldy number of permutations).

So, how about this flash order?

For one, the flash order was only introduced into the market a few months ago. Thus, flash orders clearly cannot be responsible for anything that happened (or didn't happen) in the market until quite recently. Just wanted to make sure everyone is aware of the possible scope of the influence from this order type (i.e. GS didn't make $5B last quarter by using flash orders).

My understanding is obviously limited to what I know (the Rumsfeld unknown unknowns caveat), but I do not understand how flash orders can be used to illuminate hidden liquidity any differently that regular orders can be used to do the same thing. You can always send a small order to test the waters. It is always possible that others who are more clever understand how to harness the flash orders to their advantage -- but the currently available public information about how they are being used to illuminate hidden liquidity does not make sense to us. I am open to changing my mind given more information.

One thing I do know is that flash orders allow less sophisticated market makers (most) compete with more sophisticated market makers (e.g. GetCo) that use inter-market sweep orders. I suspect that the people on the losing end of this innovation are trying to stir up enough controversy to get the flash orders banned. They have some Congresspersons talking about it, so they must be having some success.

Let's look closely at the scenario painted in the NYT and other articles. There is news overnight. Some traders calculate BRCM's new equilibrium price to be $26.40 while the opening price is $26.15. Just for the sake of this discussion, let's assume that these traders are correct about the new equilibrium price. The market opens ... what should happen to a random retail trader who want's to buy or sell BRCM at 09:34:15 EST?

... continued ...

Anonymous said...

... continued ...

Scenario One: The market is omniscient and the price of BRCM instantly adjusts to the new equilibrium of $26.40. The retail trader who transacts at 09:34:15 will get the new fair price (less the $0.01 spread) for buying or selling. The traders won't make any money and neither will the HFTs. The only drawback is that I hear omniscience is hard.

Scenario Two: The market has a competitive ecosystem of traders and HFTs. The traders start buying BRCM. Before they can accumulate very many shares, however, the HFTs sense that there is upward pressure on the stock and start moving the price upwards until the buying and selling pressure are balanced. If this process is as dramatic as stated in the NYT article, by 09:34:15 BRCM will be trading at it's new equilibrium and the retail trader will again get a fair price whether buying or selling. The traders and HFTs compete for and split the profit gained by moving BRCM to it's equilibrium price, but the picking are slim because the price adjusts very quickly and it is impossible to take anything but a small position during this window.

Scenario Three: The market has traders but HFTs have been banned to slow down trading. The traders start buying BRCM and they are able to accumulate a large position before pushing the price up to $26.40. This takes most of the morning. For the retail trader, they will transact at $26.17 which is not the knowable equilibrium price and a real bummer if selling. If buying, your lucky day. The traders make good money driving BRCM to equilibrium on the backs of the retail sellers who sold to them all morning at something less than the equilibrium price.

Clearly, the market would be well served if we could implement scenario one. My current claim is that the current competitive market landscape is as close to scenario one as the market has every been. Not to say that there are not imperfections, but like democracy, the current system is the worst except for all the others we have thought of or tried.

My take on the flash order is that one set of HFTs are using it against another set of HFTs and they are unhappy. I have insufficient information to say that it is being used in the manner described by the linked article. The impact to the retail trader is unmeasurable -- that is they are not getting any better or worse prices. So, I won't be upset if they eliminate the flash order or for that matter any of the other special order types. I just don't think that will improve the functioning of the market.

As an aside, one thing that continuously frustrates me is that people very often only look at one side of the trade. And when they lead off with something like there were more buyers than sellers today .... there is always the same number of buyers and sellers. The NYT article laments that buyers paid $1.4MM more for those shares because of the HFTs. Yeah, well, sellers got $1.4MM more for those shares because HFTs were in the market. The traders were trying to utilize their information advantage over the other market participants to make money. The HFTs dispersed that information into the market very quickly once it was detected neutralizing the information asymmetry. That seems like a good thing for the retail trader and the efficiency of the market.

Sorry for the long post.

And, check the fine-grain price chart for BRCM. The NYT article makes it sound like the price jumped right out of the gate to $26.40. It took six minutes. That is an eternity.

kristiina said...

Seems like the story's getting traction.

This discussion has bothered me somehow. Strange expressions like it's a dog eat dog world crop up. This looks like cognitive bias to me. Of course, Hobbes is firmly in this camp, but there are others, like Kant, who are firmly in opposite camp. Jury's still out in this matter, and i don't know if they ever will be in...Meanwhile, if it looks like dog eat dog, maybe look at other animals. I manage to see often also the wonder of moral law within that Kant values as high as the starry skies above.

The venerable mr Darwin keeps popping up too. In a stunted version, though. The man himself says: "Our difficulty in regard to sexual selection lies in understanding how it is that the males which conquer other males, or those which prove the most attractive to the females, leave a greater number of offspring to inherit their superiority than the beaten and less attractive males". Sabotaging other males is just half of the evolution story. The other half is to be attractive to the other sex. Rapists seldom are. I'd say dog eat dog happens to those wo can't get the attention of bitches.

As to parasites, they will always exist. Safe haven is a dream of those who are complacent, lazy and fat. Of course, that doesn't mean nothing can (or should) be done. The conversation seems to have been partly practical - what to do - but the moral part just keeps coming up. I think monkeys eat certain plants to keep internal parasites in check. Isn't that part of daily routine to keep track who is eating you and keep them in check? I think tao te ching says something like: What is a bad man but a good man's work?

"Cassandra" said...

Kristiina - I read PWs piece in the times, and I am sympathetic with his sentiments - that there is an elevated probability of negative externalities resulting from a flawed market structure and mechanism that encourages what are - in essence - pursuits undesirable in BOTH moral and technical realms. You are adept in distilling these two separate threads.

While I may eschew technology is some spheres of my life, I am most certainly not a luddite in respect of markets and investment. But I do believe that just because "one can" doesn't mean that "one should". Technology and evolution of a flawed market structure (partly out of parochial self-interest and partly out of incremental inertia) have seemingly run ahead of The Market Polity's ability to examine and define what is good and desirable for the larger constituency.

Like your comments, I am more interested in entertaining the philosophical questions, and also find the cognitive bias disturbing. I am also unsatisfied with the Darwinian point of departure. Please drop me an email offline re: this discussion

Anonymous said...

Interested in hearing a response to anon @8:17 and 8:21.

Anonymous said...

Our tribal evolution has left us with a desire for fairness and equity while a different thread of behavioral evolution working within the same brain has led us to undermine the very rules of our cooperation. Cheating and scamming must be a successful trait or it wouldn't be so ubiquitous. But in these days I don't see monetary success of the Wall St. kind being translated into large families. It is more likely to be translated into cocaine and booze. But perhaps many of these tycoons are successful in scamming women as well, having many illegitimate children. Since there really are no rules, except those established by immutable laws of thermodynamics, those that procreate the most and avoid the evil of their pathological overlords will be most successful.

kristiina said...

Speaking of the laws of thermodynamics, this is my favourite: Newton's third law - for every action there is an equal and opposite reaction. Hit the world with goodness, and it will hit back with - what? Trying to be really good will bring out some nasty things. Or that's how i understand it. In this day and age, knowing my Goethe and Bulgakov and Jung, i cannot afford to get too attached to good or evil.

So, build a wall to protect what's yours, but remember that you can never know wether the good is inside or outside the wall.

"Cassandra" said...

Anonymous 8:41 - These are important observations. But isn't cooperation, the unnatural state, rather than cheating/scamming? While cooperation has seemingly aided mankind's vault into its current realm, it sits in healthy tension juxtaposed to the less salubrious darker forces of unadulerated self-interest.

I am no granola-head, but I suspect there are things related to this dichotomy that are yet to be discovered - vibration, harmony, energy etc. - poorly if at all articulated using the current paucity of language at our disposal, which I am reticent to dismiss. Despite the apparent natural order which demeans - no, abnegates - meaning, I remain open (and admittedly hopeful) to the possibility this is merely lack of knowledge, understanding, vision, perspective which may accrue if the species manages to survive long enough.

PaxAmericana said...

"Still, the uproar outside rantings of ZeroHedge is decidedly muted. Exchanges make money. Brokers make money."

Uproar is not necessarily the most important sentiment. I remember how average Japanese changed their perspective regarding stocks after the bubble burst. Eventually, many or most people came to the conclusion that things were rigged by insiders. This could happen to Wall Street, and it would be silent, it's just that all trades would be conducted by the sharks who run the game.

Anonymous said...

Cooperation is the preferred manner in which higher order organisms eat lower order organisms - cellular cooperation. The mind's logic is built upon language and its cousins in the mathematical realm and in combination they are not up to the task of describing reality experienced by humans. The logic of thermodynamics, systems, neural systems, molecular biology (I was almost going to put economics in here) is unapproachable by 99.99% of earth's human population and they have by default or coercion adopted a simpler explanation.

Technological growth and organization is no great departure from its organic counterpart, rather a reiteration of a theme, conducted unconsciously without serious consideration of the ramifications. The rational part of the mind, Dr. Jekyll, exists to rationalize the actions of the nefarious Mr. Hyde (there are many) operating in the background in satisfaction of our appetites. The religious common folk try to beat down Mr. Hyde, keep him in the dark in return for a most irrational reward of everlasting life while many on Wall St. celebrate his wild and unclean spirit and the immediate rewards he bestows.

Of course the prisons are full of Mr. Hydes, each from unique circumstances and all layers of society. Imagine a revolving door policy with ex-cons being hired into local police departments and city governments. Just imagine if the Hydes of Wall Street could accomplish that with the SEC and federal government.

kristiina said...

There are so many stories we can weave about what we see. Social Darwinism, Jekyll&Hyde, no rules, laws of thermodynamics, all this is stories we tell ourselves about ourselves. In the newspaper today there was a story about a 77-year old man who has done a lot of research on petrographs. He says he noticed, when teaching children, that it is tremendously hard for people to see, because we see only what is in our brains. But that must be cast off before we can really see.

It may be that our world is ruled by immutable laws. But if so, we do not yet know them. To paint a picture where we are at the mercy of our meanest traits may be truth. It may also be a clever way to manufacture helplessness and victimhood - those sturdy pillars of perpetuating evil. It may make one feel better when back-stabbing someone to say that the laws of thermodynamics/god/devil or whatever made me do it. How would it feel like to say I did it? What if people really started to own their choices and decisions?

Anonymous said...

Eventually the scientific proofs and explanations will be brushed under the rug in favor of more parochial and accessible explanations of reality. In my home state of Kentucky we have a creationism museum that hosts an impressive display of immutable laws.

Time and energy consuming erudition will be lost as fabricated explanations find suitable habitat in the unplowed minds of the masses whose main objective is to feel good, not discover reality's truths.

Most men and women cannot distinguish themselves from the "Mr. Hydes" that animate their actions daily. The small bit of rationality they possess will find adequate excuses for their subconsciously derived actions.

Anonymous said...

Technological advantages, legal or illegal, are soon vanquished once the tipping point of "participants" jump into the "game" and eventually cuts the pie into so many pieces that the profit potential vanishes.

The dot com craze was a perfect example of the pie getting cut into so many pieces that most participants starved and went under.

If GS has a quant trading system that is able to "liberate funds" from other trades, you can bet that others already have or will follow with their own.

GS, like the Federal Reserve Bank, is a quasi private/government operation. They are dirty as hell and everyone’s seems to know that. They fund a lot of political big shots, so there are very few with a backbone or bullet proof vest big enough to bring the bums to justice.

Jim said...

I agree with the commenter above who would like to see more acknowledgement of and response to Anon 8:17 and 8:21.

That poster's point seems to be that the HFTs are making the market more efficient. *Efficient does not mean everybody wins.* If the current price is below the equilibrium price, then a move towards efficiency means *buyers lose* and *sellers win.* If the current price is above equilibrium, then vice versa.

Lots of the complains, as the anonymous poster notes, seem to involve only looking at one side of the trade.

An HFT that sniffs out a large buy or sell order and buys or sells ahead of it is hurting that particular trader but moving the market more towards the efficient price and helping other traders. Really the buyer/seller competing against the HFT is angry that he couldn't fleece some other trader by buying/selling all of his large volume at a price that didn't reflect the fact that this large volume was there to be traded. I'm sorry for him that he couldn't profit by trading at price that didn't reflect his own volume, but who says he has a right to? Why don't other traders have just as much 'right' to try and figure out if his volume is out there and trade accordingly, thus making the price more, not less, accurate?

Anonymous said...


where r u?