May

18

Reading the histrionics in the media (and seeing some comments on here of late) on HFT, I have decided to vent my frustration and throw my hat in the ring on the situation.

HFT is no more evil than any form of arbitrage. If you're fine with arbitrage (and I assume 100% of us are), it is my contention that you should, rationally, also be fine with HFT.

As far as I see it, HFT equally helps enforce the Law of One Price just as arbitrage does. What's more, both HFT *AND* vanilla arbitrage share the exact same rational premise. Both are seeking to drive prices from inefficient fragmentation to a place where no profit can be made via price discrepancy alone.

Allow me to explain: With vanilla arbitrage, profit comes from knowing two prices exist for the same thing (or cash flows) and taking advantage of that knowledge by buying the cheap price and selling it simultaneously at the expensive price before others do. With HFT it comes from knowing pretty much the exact same thing…"over on exchange x this stock is bid / offered at 23.05 / 23.07. Over on dark pool y, by virtue of my better infrastructure, I can see thousands of shares offered at 23.05…I will rip every share I can find at 23.05 on the dark pool, then turn round and offer them at 23.06 on the exchange in the assurance that lazy institutional VWAP algos will be forced to take it off me". Is this any different from being able to buy or sell the same quality and quantity of gold on two different exchanges at two different prices before anyone realises? There is no philosophical difference.

Think of trading in its ultimate philosophical and mechanical terms. Using the HFT vs arbitrage example, it doesn't require an enormous leap to see that both styles of trading rely on a qualitative difference in available information. Supernormal information is profit potential. Couple it with speed, and you shouldn't be surprised that someone consistently makes dollars….until they lose either their information or their speed advantage (or either one remains static while their advantage in the other one diminishes). This confirms *everything* we know in corporate finance: the whole notion of alpha and beta, CAPM, EMH - you name it - relies on the distinctions that information provides. Can't make supernormal profits without supernormal information…and so much the better if you can act faster on the same bit of information than anyone else. This is, after all, how information advantage works across myriad fields; Pheidippides' legacy came to be called Marathons, billions of dollars is spent each year on intelligence etc etc. Heck, even in its weakest form, those of you who pick stocks fundamentally look to do the same thing - you believe your information is better, so you buy or sell to the next mug who will take your price - who, in turn, thinks you are a mug for buying or selling at that price. How is that process philosophically different from HFT? And don't say that time is immaterial either….anyone who has ever lamented a bad fill from a broker when the market is moving away from them are guilty as charged.

I have no doubt that people railed against arbitrage in the early days. They probably railed against the firms that made money from it. There were probably enquiries and investigations and probes. The public likely lamented those who figured it out and called them cheats. Arbs doubtless caused panics and sparked runs in different instruments. And, just as with arbitrage, HFT will be shown to be benign because it is just one more method of sucking profit potential out of price inefficiencies. Markets benefit, long run, from having HFT. On that point, let us not conflate "efficient markets" with "profitable markets". It's obvious to all that traditional avenues of profitability are drying up very, very fast.Traditional sources of alpha have now very much become beta for the vast majority of players. That's not to say that alpha is dead - it's just tougher to find. One of the "easier" places to have found it is in latency technology - but this too shall pass. Arbitrage has more-or-less disappeared because people learned to spot the opportunities and cash in on them in ever decreasing time periods. So too with HFT. The first cracks are there to see. My point is that everyone screams about how unfair these programs are, but HFT and arbitrage aren't so much about investment skill as they are exercises in information theory, infrastructure design and speed; all elements of efficiency.

At this juncture, I'd like to draw a couple of quick comparisons. First up, sports have recently provided an example of increasing symmetry of information. Coaches and teams spend millions on film, statisticians, data analysis, equipment and so on. But it wasn't always this way. Did those teams garner an "unfair" advantage? Not many would say so. Was it unfair when athletes first figured out that a good pre-race meal of carbohydrates was better than steak and eggs? Not many would say so. The aftermath of stories like "Moneyball" or "Fortune's Formula" show that these little edges disappear in time as everyone else picks them up - and so it is with our arena. But there are always new edges to discover. Discovery requires creativity and persistence.

Secondly, HFT and flash trading are just another stage of evolution in the market. Biology teaches us that mutations / adaptations are not, by themselves, either good or bad. Rather it is the *environment* that determines whether or not an adaption becomes an advantage. And so it is with the market. Right now, the market environment allows for profitability in discovering sub millisecond price advantages across market venues. Those who have "mutated" to exploit this are profitable. Those who haven't will struggle to compete in that niche…unless the environment changes, or they find a different niche to compete in.

I tell you the truth, friends: the day that HFT becomes unprofitable for institutions is the day they begin offering it to retail. Everyone on this list is smart enough to recognise this phenomena. Retail will get their crack at HFT - but it will only be once all the fat has gone. Regulators: let the market do its self-correcting thing to sort out this problem. The tax system of every G20 nation should ably demonstrate that the more stupid rules abound, the more creative people get at working around them. If you want to really change things, change the environment. Hong Kong is a great example of this.

The right response for every participant (especially myself) is to get better, not to whinge about how "unfair" a system is that is patently not unfair at its most general level. Be more competitive. Take a hint if you're consistently losing and figure out why. Do better homework. Throw out the tired old "do this" books, and replace them with books on scientific method and practice. Learn new fields. Develop new specialties. Don't be satisfied with the linear, but go hard after the non-linear. Focus your energy on your own game, rather than worrying about what GS or the Sage or the Palindrome or the woman on a winning streak down the desk from you are doing.

You want to beat the machines? Do what the machines can't do - think for yourself.

Sam Humbert gives his point of view:

It's hard to know whether Mr White is limited by his knowledge of the field, or has an axe to grind. In any case, the core issue is stepping ahead via subpennying — bidding 41.0301 to step ahead of a 41.03 bid — which is permitted for some market participants, but not others.

Before writing further on the topic, Mr White would do well to read deeply on the subject, and attend one of the excellent high-frequency conferences currently on offer. The one sponsored by Wall St & Tech magazine May 11 in NYC was very good, and there are others upcoming.
 


Comments

Name

Email

Website

Speak your mind

5 Comments so far

  1. Andre Wallin on May 18, 2010 12:30 pm

    agree but it is a lonely path and the psychological challenge of doing what nobody else is doing…very few are capable.

    those of us who are, work in complete isolation and are forced to become something more or less than human. more or less is a matter of perspective. you have to be anti matter in a the middle of a blasting siren of matter. if you are distracted for even a second by the sirens you get taken in and immediately destroyed in the market.

  2. rj on May 18, 2010 12:56 pm

    A masterpiece of vague arm-waving.

  3. Sam Humbert on May 19, 2010 8:46 am

    It’s hard to know whether Mr White is limited by his knowledge of the field, or has an axe to grind. In any case, the core issue is stepping ahead via subpennying — bidding 41.0301 to step ahead of a 41.03 bid — which is permitted for some market participants, but not others.

    Before writing further on the topic, Mr White would do well to read deeply on the subject, and attend one of the excellent high-frequency conferences currently on offer. The one sponsored by Wall St & Tech magazine May 11 in NYC was very good, and there are others upcoming.

  4. Nick White on May 19, 2010 5:45 pm

    Neither limited in my knowledge of the field, nor do I have an axe to grind. My goal was to be accessible and draw out the *core* philosophical issue, rather than engage in the minutiae which will be foreign to all but those who actually have worked in HFT; but whatever warms your coffee in the morning, Sam.

    At any rate, I don’t think your point does anything to illuminate, further or correct the substance of what I was actually saying - the core “arbitrage” argument still applies to your subpennying comment. But I wouldn’t be so rude as to suggest you brush up on your rhetoric :-)

    Good luck with the conferences.

  5. Sam Humbert on May 20, 2010 2:33 pm

    My mother always told me explain something three times, then give up. But I’ll limit myself to two times.

    Some market participants (who, is left to the reader as an exercise) are allowed to frontrun public limit orders by economically meaningless amounts. If you want to call this “arbitrage”, then it’s “regulatory arbitrage”. It has nothing to do with trading “before anyone realises” or with “information or speed advantage.”

    And subpenny frontrunning isn’t “minutiae” — it’s the fountainhead of profits.

    Imagine you’re making a 41.03-41.04 market, and that others (but not you) are permitted to step ahead of you — 41.0301 bid or 41.0399 offer — whenever market conditions (say, an up/downtick in SP, or in a related stock) indicate it’s advantageous. The result is you absorb nearly perfect adverse selection. Whenever you’d want to be filled, instead the fill goes to the participant who stepped ahead by .0001. But whenever you wouldn’t want to be filled, you are. It’s the digital version of the ol’ “locals ahead of you” in the futures pits.

    If either everyone, or no one, were permitted to trade subpennies, then true time/price priority would exist, and I’d agree with your essay. But if you aren’t willing to address subpennying, where’s the beef? You’re discussing an important subject but ignoring the gorilla in the room.

Archives

Resources & Links

Search