Home > finance, news, rant > High frequency trading: does it hurt the little guy?

High frequency trading: does it hurt the little guy?

August 8, 2012

I’ve already written about high frequency trading here, and I came out in favor of a transaction tax to slow that shit down a little bit. After all, the argument that liquidity is good so more liquidity is better only holds to a point – we don’t need infinite liquidity. It makes sense to actually have a small barrier to trade – you actually have to think it’s a good idea one way or another, otherwise you have no incentive not to do something dumb.

And as we’ve seen recently with Knight Capital, dumb things definitely are likely to happen.

It’s been interesting to see the media reaction. On the one hand, the Room for Debate over at the New York Times has a bunch of people discussing high frequency trading (HFT), and the most pro-HFT guy essentially says that the SEC should keep up technology-wise with these guys, and everything will be ok. That’s called living in a fantasy world.

More interesting to me was Felix Salmon’s post yesterday, where he rightly complained that, all too often, journalists dumb down and simplify reporting on these things, and then he proceeds to dumb down and simplify reporting on this thing.

Specifically, he complains that no “little guys” were hurt in Knight’s crash, even though the press is always looking for the little guy that gets hurt. [Side note: he also complains about the LIBOR manipulation not hurting municipalities, which is false, it did hurt them. He needs to understand that better before he dismisses it.]

But, if I’m not dreaming, Fidelity was one of the large customers of Knight that’s pulled out, and if I’m not unconscious, Fidelity manages quite a few of my many 401K accounts, as well as a huge proportion of the 401K accounts in this country. So it’s quite possible that my retirement money was part of that massive screw-up which is now owned by Goldman Sachs, not that I’ve been notified by Fidelity of any harm (but that’s another post).

As for small investors vs. little guys, there’s a difference. If you have enough money that you’re investing it through brokers, I personally don’t count you as small, even if you appear small to Goldman Sachs. So I’m not interested in whether the small investor was all that harmed by Knight’s meltdown, but I’m pretty sure the small investor was scared away by it.

But looking at the larger picture, I’d definitely say this is an indication of the outrageous complexity of the financial system, which most definitely is hurting the little guy, i.e. the taxpayer. This complexity is why we have the government guarantee in place, the Too-Big-and-Too-Complex-To-Fail banks and markets, and the little guy on the hook when things melt down. Moreover, there’s a direct line from that whole mess to the destruction of unions and pension programs, even if people don’t want to draw it.

So if you want to be myopic you can say that this was one firm, making one major blunder, and it’s self-contained and that firm is failing just like it should. But if you take a step back you see they were doing this as part of a larger culture of competition for speed and technology that they are so focused on, they threw risk to the wind in order to achieve a tiny edge over NYSE.

That laser focus on having a tiny edge really is the underlying story, and will continue to be, at the expense of risk, at the expense of our retirement funds trading for us, without regard to unnecessary complexity or, yes, the little guy, until our politicians and regulators grow some balls and put an end to it.

Categories: finance, news, rant
  1. August 8, 2012 at 11:53 am

    I just finished “Dark Pools: High-Speed Traders, A.I. Bandits, and the Threat to the Global Financial System “, by Scott Paterson


    There are many issues that this book brings to light which are relevant to this thread.

    Here are a few:
    1) ‘Order Types’ – REG NMS provides numerous ‘order types’ that give precedence in the queue to certain types of orders E.g. – submitted using ‘maker/taker’ pricing. This gives an advantage to HFT traders. I don’t know if there are any studies which show the impact of this ‘2 tier market’ on small retail investors. But the implication is that ‘the game is rigged’,

    2) ‘Turn it off Now!” – vs- ‘maintain a fair and orderly market’ During the flash crash, many of the top HFT firms hit the kill switch. This cause enormous liquidity to evaporate in seconds from the market. Knowing there are no specialists required to ‘maintain a fair and orderly market’ has to impact small investors. They know they’re not on a level playing field.

    3) Canceling 90% of orders submitted. The market is supposed to provide signals regarding
    supply and demand. This is one of the cornerstones in free markets (as opposed to managed markets – think USSR). The HFT orders are covering the ‘signal’ with noise.

    4) Painting the Tape – these patterns are apparent and part of the impact is to keep small investors out of the ‘shark tank’. Dark Pools were conceived the circumvent these issues.

    5) Colocation and $300 million cable which cuts 6 milli-seconds of transmission time.
    Does this 2-tiered playing field keep small retail investors out of the ‘casino’?


  2. August 8, 2012 at 1:08 pm

    To me, the vulnerabilities of high frequency trading is akin to the ridiculous situation that developed when Enron was (corruptly) allowed to turn the electric power grid into an arbitraged and thoroughly corruptible “marketplace”, rather than a regulated public commons utility. This was accomplished by computerized “wheeling”, the practice of rapidly switching huge supplies of power from one market where demand and pricing was low, regardless of the capabilities of the transmission infrastructure, to another market where demand was higher and the current pricing could be briefly arbitraged for outrageous profits – especially when corruption and bribery set the apparent demands.

    The grid system was designed and developed over time to accommodate reasonable, gradual shifting of loading, carefully accounting for the limits of the infrastructure linkages to keep within human response times in cases of unexpected imbalances or mishaps. With corrupted techno-wheeling for dollars, this human override response disappeared, and disasters loomed as ongoing threats as the idiotic task of building out the linkages to such insane load switching was obviously impossible as well as stupid. Fortunately, corruption destroyed Enron before wheeling destroyed the grid, but very little was learned as far as limiting the application of wild-eyed arbitrage schemes to inappropriate structures.

    High frequency trading is taking advantage of extremely transient arbitrage opportunities in a manner totally unavailable to an individual human being placing an individual ask-bid trade, whether a billionaire overqualified-investor or a sucker amateur trying to do better than the losing bets of his/her 401K. If the transaction costs forced on the human ask-bid traders were also suffered by the hordes of digital servers filling up Greenwich Village floorspace (displacing human beings), unforeseen technological/coding/corruption vulnerabilities could probably come back under human-scale emergency control far better than artificial means such as circuit-breakers or ex-post lockdown adjustments that shaft everybody and their human trust.


  3. HFTer
    August 8, 2012 at 2:27 pm

    I still don’t understand why don’t see this as an example of a “good thing”: someone took too much risk, lost their(!) shirt, got bought by other private companies. There is no “systemic” collapse or government bailouts. We have another example of how things can break, and exchanges can put in more safeguards to prevent this from happening again (e.g. tighten price circuit breakers, add volume circuit breakers, etc.) There will probably be an SEC/FINRA investigation/fine, with tightened risk controls across the industry.
    The US equity market is probably the “best” market there is (if you know of a better one – please tell me!) — in terms of regulations (protections for the “little guy”), liquidity, transparency, trading cost, everything I can think of. The fact that it functions so well normally is what makes it so familiar and “understandable” to “little guys”, and what magnifies every little hiccup. The only way I would see this one as a scary example would be “See what happens in our best, most transparent market? Can you imagine what happens in all the other markets?”
    To address some of your points, I was not convinced that the small investors lost any money in this event, including your Fidelity example. Fidelity “pulled out” probably means that they stopped trading with Knight and traded with other brokers (what would you do if your broker might go bankrupt before your trade clears?) — how does that affect your 401k? The trades were executed on Knight’s behalf (not its customers), otherwise it would be impossible for Knight to get into a large position. (It is also possible that they took the loss to protect their clients — the little guys — from the errors in their trading algo.) The only way I can see your retirement money participating in the screw up was if your manager decided to trade one of the 140 stocks that morning, and took the wrong side of the trade. (“Check this out, the stock is up 20% on no news, and 100x normal volume — I should probably buy a ton!”) If that’s the case — fire your fund manager and invest in an index fund.
    As for larger transaction taxes (SEC already collects a fee for every sale), they will certainly not remove speculation (my proof: round-trip real estate transactions costs are almost 10%). They might improve some aspects of market microstructure; that’s an empirical question and no one ever treats it as such. Luckily, it looks like some European countries are setting up this experiment, so we can take a look at results in a couple of years — I just wish we had a better idea of what we want our markets to do, not just “screw those HF guys!”. Additionally, my worry is a sea of exemptions that any such tax would necessitate (“you have to pay x in tax if you trade equities, but y if you are delta-hedging your options positions”), giving another advantage to bigger institutions.
    I think a very good way of addressing the issues in algorithmic trading is algorithmic: it should be easy for exchanges to identify undesirable scenarios and disable / overcharge the offending algo. I believe this is happening more and more (e.g., see the recent introduction of extra charges for high cancellation rate), and these incidents will probably speed up the adoption of such safety measures.


  4. Bindicap
    August 9, 2012 at 1:21 am

    I like your math posts, but this is one of those that I really don’t get. It’s such a strange contrast to your proof post.

    It’s hard for me to extract the core of your reasoning here without making it sound dopey. You reject that little guys weren’t really hurt by Knight’s screw up, but your argument is really nebulous — something about Fidelity re-routing order flow at the time, so maybe some dollars connected to managing your 401k accounts somehow flowed into Knight’s error portfolio which got sold off to a dealer in clean up. And this causes some kind of harm…. Huh?

    Then your other argument seems to be that buying and selling equities rapidly and how a participant can screw that up is an indicator of the outrageous complexity of the system. And that hurts the taxpayer. And there is a direct line to destroying unions and pensions. I mean — there is more than a missing lemma in that chain of reasoning.

    Knight rolled out something new without adequate testing or control and their shareholders are suffering for it. And it’s a bright warning to everyone else in the space.

    I do think there are some interesting issues in connection with the rise of faster trading. The main concern I see with HFT would seem to be if it has externalities — costs imposed on outsiders that don’t factor into its own economics and hence activities that are bad for the system continue while the outsiders bear the cost. The resources of the exchange would seem the main point where this would bear, but in fact what I see is exchanges want to promote more trading activity and see the associated technology and infrastructure costs as making economic sense. When exchange resources get abused, it gets fixed.

    The other common argument is that somehow market dynamics are made worse by faster technology, and we need to control speed and computation to promote better dynamics. I have seen various things about that, with nothing convincing. Mostly it’s unfounded accusations about front-running or volatility. It looks like the old guard fighting against a new generation and trying to preserve a structure that favors their older technology as they can’t keep up.

    Of course the old guard would love for regulators to come in and shut down the new guys taking over their jobs and doing them more efficiently. And they’ll talk nonstop to move opinion in support. Doesn’t mean it’s right.

    It could be interesting to see some finer points and productive ideas discussed. It’s not an area I know well. But I’m a bit doubtful we are headed there with this start.


    • August 9, 2012 at 6:33 am

      Hi Bindicap,

      I’m willing to admit my writing may have been bad, but I don’t think my reasoning was completely off. Let me try to explain, and you can tell me if you think I am wrong with the facts.

      First, with Fidelity: they have enormous amounts of “little people” money, which they trade opaquely. The fees people pay Fidelity don’t even include bad trading prices, just management fees, so bad prices are invisible to the average retirement fund contributor (me). So if Fidelity is using someone who screws up their order and gives them a bad deal, there’s no way I’ll ever find out. So there’s no way I’d know if the Knight screwup actually did have an effect on my money. But I do know that Fidelity refuses to use Knight anymore, so I can merely suspect that something bad happened.

      Second, I no longer buy the idea that markets should be as complicated as traders and quants want them to be, no matter how hard they are to regulate or understand. I know how behind the SEC is, and all the other regulators, and I know that won’t change. So the only way the situation will become reasonable is if the markets get simpler. It’s not going to collapse the markets to make them simpler, it’s just going to give people less incentive to spend millions of dollars getting colocated to exchanges, which is ok for the productivity of the country.

      Third, yes, there is a direct line from the credit crisis, to the ensuing bailout, to the federal deficit problems, to local deficit problems, to the pensions being cut. It’s not hard to follow, especially if you follow timelines and politics.

      You already understood the basic idea, though, which is that there are externalities not only to Knight’s mistakes but to the culture of HFT itself.

      Does that make sense?


  5. AnotherHFTer
    August 11, 2012 at 12:47 pm

    Re: bad trade prices.

    Fidelity (and Knight Capital as their broker) is required by SEC regulations to execute the trades at the National Best Bid and Offer (NBBO). The idea is to prevent this exact type of front-running client’s orders you describe. This parasitic behaviour was “legal” in the past, in the days of calling-your-broker-on-the-phone-asking-for-a-quote.

    When Fidelity sends its order flow to Knight, it gets the NBBO price or better (probably sub-penny improvement; because Knight thinks it can make money using this volume). It also gets a payment from Knight (again, probably because Knight uses this volume for its market-making businees). You can acutally see the Sec Reg 606 report regarding order routing for Fidelity here: http://personal.fidelity.com/products/trading/Fidelity_Services/Service_Commitment.shtml .

    The point I’m trying to make is that “little people” (aka retail investors) get the NBBO price when they sends their order. Which is probably what you expect when you send an order and have no better idea what the price should be.

    Claiming that HFT is screwing retail by sticking to the narrow aspect of trade prices is a weak argument, IMHO. The question of systematic (in-)stability due to HFT activity is an important question, which deserves a serious academic/media treatment (i.e., not just paranoidic ZeroHedge posts and hysteric Nanex plots).

    In my (deeply) biased perspective, the traditional industry still seems like a bigger hazard (SEC slaps on the wrist of IBs who defraud their clients, CEOs who get away with just about everything, and other extremely old tricks in new disguises).


  6. MacCiarogain
    August 27, 2012 at 9:03 am

    The idea that the little guy didn’t get hurt is laughable. Aside from the potenial for stops to be triggered, (you can argue the merits of stops as an aside topic…fact is, retail investors use them) if I’m an investor in Knight Capital (big or small investor) I just got taken to the cleaners. Of course there are risks to investing in any company, and that’s part of the game, but for Felix to claim the small guy didn’t get hurt is outrageous. How many retail holders of Knight even realized the types of risks involved in the business model of the company? How many realized that a seemingly strong comapny could crush its capital by buying 7mm in stock per second? The structure of the market has gone beyond ridiculous. Buying and selling by the milli or microsecond does not benefit the investor, and if you truly believe otherwise, you need to put down the coolaid, and think things through more closely. So you can sell 1000 BAC at a penny (or less) spread…you realize that spread has only narrowed because the liquidity provider can trade for no profit and still make money on the rebate? Liquidity provision was never an issue in these stocks. Get over the fact that penny spreads are the norm for already-liquid names, and take a look at the broader market…where often liquidity hasn’t changed.


  1. August 10, 2012 at 10:52 pm
  2. August 11, 2012 at 1:15 am
  3. August 11, 2012 at 4:12 am
  4. August 11, 2012 at 10:10 pm
Comments are closed.
%d bloggers like this: