25 Apr 2014

Challenging Landsburg on HFT

Banking, Krugman, Shameless Self-Promotion, Steve Landsburg 33 Comments

You guys thought I was nuts reading my analysis of the Bundy standoff? Well now I’m publicly accusing Steve Landsburg of a basic math mistake. But I have a lot of life insurance, remember. An excerpt of the part where we’re still chums:

Reading guys like Paul Krugman rail against the socially useless financial sector, Landsburg raised an excellent question: Haven’t Keynesians like Krugman for years been lecturing us on the desirability of even pointless investment spending when the economy is stuck in a “liquidity trap”? And yet, I don’t remember Krugman or any other Keynesian thanking their lucky stars that traders had built a $300 million fiber optic cable between Chicago and New York just when–according to their own analysis–there were large multipliers on investment spending. The project actually started in March 2009 and “soon had 125 construction crews working at once.” The timing couldn’t have been more perfect. Talk about stimulus!

33 Responses to “Challenging Landsburg on HFT”

  1. Silas Barta says:

    I would be interested in developing an intuition of the social benefit of these faster trades. Does the economy really perform better in some sense when it can absorb information .001 seconds faster than otherwise?

    • Bob Murphy says:

      Man you’re making the same mistake (or not) as Steve! If billions of new bits of information are falling into various people’s laps every minute, then yeah I think a population of 6 billion people can measurably benefit if each bit of information is processed (somehow) .001 seconds faster than before.

      Look, does email really help all that much? What’s the benefit to society of getting an email somewhere in 3 seconds, rather than calling somebody up when that takes 8 seconds? You save 5 seconds on sending someone a message, is that worth the gajillion we spend on the internet? etc.

      • Silas Barta says:

        My argument is different from Landsburg’s.

        Landsburg: if you make these assumptions about the social value of the speedup, here’s what you get.

        Me: wait, how do we get a positive social value for the speedup in the first place.

        We both suspect, though, that faster trades at this point just transfer wealth between traders rather than create it.

        Yes, it’s great (welfare enhancing) when telecom improves email speed from 8 to 3 seconds or even .2 to .1 seconds. But I draw the line (in the sense of needing a different-in-kind explanation and new intuition) at humans’ inability to respond faster than a certain time.

        Why do stocks have value? Because they grant voting and dividend rights. Yet neither of these benefit from the price being reflected in a microsecond rather than a millisecond. Hence why I would need a proof-of-concept scenario in which such a change is welfare enhancing.

        FWIW the best explanation I’ve heard so far for HFT returns is that they’re an artifact of enforced price granularity, which spurs these tournament-like competitions for “first to process”; and that, if it were possible to trade at any price, we would just see more decimal places in trades rather than such intends speed competition.

  2. Steven Landsburg says:

    Response posted at the site you linked to, and awaiting moderation….

    • Bob Murphy says:

      Hmm I don’t think I have permission to approve it. Let me email the guy…

    • Bob Murphy says:

      I would pay 1/8 of a penny to see your comment now, rather than waiting till he approves it.

    • Keshav Srinivasan says:

      Steve, why don’t you post your response in this thread? This is the comment section that usually has more back-and-forth about Bob’s Mises Canada posts.

    • Major_Freedom says:

      Steve, you wrote:

      “Spread Networks recently spent $300 million to build a fiberoptic cable that will let Wall Street traders shave .003 seconds off their execution times.”

      “What’s the social value of that cable? If you can shave .003 seconds off the time it takes to execute a trade, how much good have you done the world?”

      Wait, who exactly do the individuals in exchange need to convince in order for a purported exchange to result in gains? Isn’t it themselves only?

      Sounds like you’re saying people other than the exchangers have to give their opinion before that exchange is “valuable”.

      But value is subjectively determined, meaning an exchange does produce valuable outcomes if the exchange parties believe it so! It is valuable because the exchangers, not you, not me, not “society”, believe it so.

      I do believe you are taking your inability to understand why others would trade the way they do, and arrogating it to the status of “social value”, where because you don’t value it, it cannot really be valuable even to the exchange parties.

      This is what Krugman said about the internet back in 1998:

      “The growth of the Internet will slow drastically, as the flaw in “Metcalfe’s law”–which states that the number of potential connections in a network is proportional to the square of the number of participants–becomes apparent: most people have nothing to say to each other! By 2005 or so, it will become clear that the Internet’s impact on the economy has been no greater than the fax machine’s.”

      Everyone: The lesson here is that we have to be careful to not put too much weight on our inability to understand why other people would trade the way they do. For what we don’t currently understand, may eventually pan out n the future along with innovations and complimentary capital accumulation, to result in subsequent outcomes that we may end up valuing very highly.

      Many people thought the first cars were just silly rich man’s toys. “What is the social value in these “automobiles”?” etc.

      Ignorance masquerading as profound insight is what actually grounds the drug war. “What is the social value in people trading cocaine for money?”….let’s use force against them, because clearly they not able to understand “true”, i.e. “social” value.

      My heuristic is to always give the benefit of the doubt to the exchange parties. Clearly they are finding value in what they are receiving. I don’t arrogate my judgments over theirs like mine is representative of “society” and thus more legitimate.

      • Harold says:

        MF: I think I see an error in your argument. Landsburg acknowledges the private value, which is what you are talking about, and that private value is generally in synch with social value. If I grow an orange I slightly reduce the value of other oranges, but that loss is exactly balanced by the gain to my customers. He is proposing that making a faster connection does not slightly reduce the value of the second fastest cable, but devastates it, effectively transferring wealth rather than creating wealth.

        He may be wrong, but I think your argument is the wrong one to challenge his point.

  3. Transformer says:

    Lansburg says :

    “we should expect there to be more invested in such projects than they return in social value. “.

    The calculations you refer to are the “social value” , Landsburg thinks the private value exceeds that. It doesn’t make any difference if you are talking 1 transactions or all the transactions in the world.

    Your about global v US transactions seems valid though.

    • Major_Freedom says:

      All of what is purported to be a “social value” is in fact private value. For not everyone has the same value of it, if at all.

      • Transformer says:

        Well, I don’t really understand .Landsburg’s argument either.

        I was just pointing out that the argument wasn’t that the social value of the new link was less than its cost, which is the argument Bob was criticizing.

  4. Transformer says:

    Your POINT about global v US transactions seems valid though.

  5. Steven Landsburg says:

    Reposted from the other site:

    Bob: You’re assuming that all the private benefits of this fiberoptic cable are captured by Spread. I simply asked myself “If I were a trader in this market, how much would I be willing to pay to speed up one of my transactions by .003 seconds”? Admittedly, at this point, I’m just taking a wild-ass guess — but it seems very plausible to me that the answer is more than 1/8 of a cent. Part of that private benefit flows to Spread (which accounts for the value of the cable), perhaps some is captured in quasi-rents by traders, perhaps some is passed on to consumers.

    You can argue that my 1/8 of a cent is an overestimate, but the fact that it’s not all showing up in the price of the cable is, I think, not a convincing proof of that.

    As to the policy recommendation, why did you assume it was addressed to the government rather than to the exchanges?

    • Bob Murphy says:


      I may do a whole follow-up post, but to quickly respond:

      1) I realized (long) after I posted that I had misunderstood your argument. But, if you care, you might re-read your post and see how you kept referring to the social value of “the cable” and I think twice referred to the $300 million figure. So I really thought you were trying to show that the total social benefits were lower than the total use of resources, which is $300 million.

      2) Now that I grasp your argument, my answer is: “Darting in and out of trades, HFT firms make just fractions of a penny per trade, but the sheer speed and volume of their trading activity allows those that are successful to make significant profits.” From this Reuters article.

      3) I never said you were relying on government! Really, go look! And my logic still stands! (But I got lucky, I thought you were…)

  6. Cosmo Kramer says:

    Krugman rails against a monster Keynesians created.

  7. Tel says:

    The first question is how to measure “social benefit” since Austrian economics provides no system for aggregating preferences into a global metric.

    I seems to me (from my subjective point of view) that one of the useful things a stock market does is price discovery, and another useful thing it does is provide a meeting place for potential investors and people who want to borrow money in order to float their business idea. In the first case (price discovery) the market is reacting to information from many sources, but HFT for the most part does not bring in new information, it merely transports existing information from one city to another a little bit quicker. I do accept Bob’s point that many transactions are required to make a price settle, but I think it is more subtle than that.

    Someone who might be a random member of the public might see something important (e.g. maybe he thinks Facebook has been hacked) so this person reacts by speculating on the market (e.g. short-sell Facebook shares) and the HFT front-runs the trade and gets a small profit out of it, which is essentially zero-sum so the speculator is now that small amount behind. Presuming the speculator knows about HFT he has choices to allow himself to be front-run, or to pay an HFT trading firm to do the trade faster than anyone else, or maybe just decides not to speculate at all. Any time the HFT discourages any person from partaking in the market it is doing damage to the available information input that the market can use.

    We could also look at price stability, I think there is evidence enough that HFT makes prices more volatile, which in turn makes it more difficult for people to judge whether a price swing represents real information or just some HFT activity probing to trigger stop-loss orders or something like that. We would need information in the real world about how people change their economic activity in response to information from prices. Suppose I’m a developer who might be sketching out ways to write a Facebook game, but then I see the Facebook shares crashing badly, maybe I’m going to put more effort into Android development instead. I would suggest that no one makes real-world business decisions in 0.001 of a second. They are going to look at the typical price trend over several days or weeks.

    Let’s look at the market encouraging new investment. Do the HFT guys bring more investment money to the table? I would argue the answer is no, they will discourage the small investor (e.g. the “mum and dad” share buyers) who feel they are not getting a good deal because they don’t have access to a fast fiber for their trades.

    Finally, since all these discussions go the same place, I would also argue that there is no need for government intervention. If a more efficient market can be designed to exclude HFT (or actually it would need to destroy high frequency information completely) then sooner or later someone will start a market that does this, and they can compete like anyone else. The free market can solve this problem, just like it solves other problems.

  8. Benson says:

    A market solution to HFT
    Interactive Brokers to offer direct routing to ‘Flash Boys’ dark pool IEX

    • Tel says:

      I’ve heard a lot about IEX but I have yet to hear an explanation of what they offer that is different. Exactly how do they go about preventing front-running?

  9. Bogart says:

    The only think wrong here are the estimates of profits. There is no computer YET to beat all humans in gambling. The fact is that once known, and rumors travel fast on Wall Street then you can change HFT to HFL (High Frequency Losses). The only scenario given to show how investors get hurt by HFT is that everything is in a state and then one big order comes in and wham the HFT computers swoop in and do the same transaction then force the helpless investor to pay them. This is silly, these markets have millions of trades going on all of the time. I keep reading about small or single order investors being beaten into transactions. But are these HFT systems really that intelligent? What if a high volume brokerage sells 10,000 shares first before buying 100,000? That is free money for the buyers as all of the HFT computers jump on the first transaction when the bigger one is right behind? What if the high volume brokerage does this randomly then HFT will break even?

    • Tel says:

      That is free money for the buyers as all of the HFT computers jump on the first transaction when the bigger one is right behind?

      I’m pretty sure the HFT can see the big transaction in the pipeline and have plenty of time to prepare for it. That’s what annoys people.

  10. Bogart says:

    As for the social value of HFT, The other respondents here correctly point out that this is not relevant. It is the value of HFT to the entrepreneurs doing it that matters. Now there is another interested party here and that is the owner of the exchange. And if HFT has no social value as you say then entrepreneurs will open their own exchange and prohibit this behavior and pull customers away from the ones that do allow it.

    Unfortunately the exchanges are even more regulated than the rest of the grossly over regulated securities business. So entrepreneurs are kept out and small investors pay the price.

    • Tel says:

      I think the exchanges will gradually find ways to block HFT because most of their customers prefer it that way. I agree that government regulations don’t help the market process… governments are always behind, and cannot be trusted to be fair.

      • Matt M (Dude Where's My Freedom) says:

        Given how tightly regulated exchanges already are, this seems unlikely to me.

        If an exchange just decides on its own to block HFT, the HFT people will sue them or appeal to the SEC or whatever it is they do for redress. They will claim they are unfairly being discriminated against or some such thing.

        Ultimately, it will be decided either within some bureaucratic agency or by the court system, but, as always, the government decision will be all encompassing. There will be no nuance. There will be no “each exchange can decide what is best for itself and its customers.” The government will simply decree that HFT either *must* be allowed or *must* be banned entirely. That’s how regulatory bodies operate.

  11. Kevin Donoghue says:

    “Private spending that is wholly or partially wasteful is also a good thing, unless it somehow stores up trouble for the future.” — Paul Krugman.

    • Bogart says:

      How does Paul Krugman know what is wasteful and what isn’t? What theory is he using that could give him the information to determine this? Murray Rothbard has such a theory but I doubt Krugman would subscribe to it.

      • Matt M (Dude Where's My Freedom) says:

        Right, and to get into the semantics of it all, if the spending is still “good” then it ceases to be “wasteful” by any meaningful definition of the term.

        IF societal wealth somehow DID increase by paying a bunch of people to dig holes and fill them back in, then that would no longer be a “wasteful” project. It would be a useful and valuable endeavor.

      • skylien says:

        What would interests me is “, unless it somehow stores up trouble for the future.”

        How could he ever know when this is the case?

Leave a Reply