| abstract | - There is debate as to how beneficial a bought share can be to a company who has it sold back within a second. The question has to be asked not from the simulation of one investor, but of a whole system of investors behaving as such. HFT companies tend to be very in-house about their money. (Jane Street accepts investment only through employees.) If HFT is giving a net profit, it might not be giving a net profit in a fair and uniform manner. It may be concentrating capital towards only those who can get a supercomputer connected to the stock market with fiber optic cable. Regulation requires micromanagement of laws to react whenever a loophole is found, as profits begin to show up immediately (Allows for more experiments, thus causing a more volatile/innovative field.) While high frequency trading may give a net benefit to the economy, it is nevertheless a highly ununiform and finite benefit HFTers claim to provide liquidity. Liquid assets attract investors because they are assets which can quickly liquidate into money, as they will not prove so detrimental to the seller's value (Hence money is the most liquid of assets.) This is a property HFT thrives on: Immediate monetization of trades. The liquidity might exist, but if a non-HFT investor looks to profit from it, they may be surprised to find that as soon as shares begin to be sold away from the investment, the HFT supplied liquidity gets sold. HFT value provides liquidity because it's a fluid investment By ununiform, I mean HFT is imposing a winner/loser relationship. By finite, I mean that HFT isn't winning because of HFT. HFT provides liquidity, but of such a liquidity that it flows about the market being almost everywhere and nowhere at the same time. Take out classical investments, HFT will evaporate. Take out HFT, classical investments will remain HFT appears to be parasitic, but to what degree does every other investment show similar properties? (Does HFT act as a second layer, working off of investments? Take away non-HFT trading, does HFT survive? Vice versa does) Nash equilibrium: Parties are in Nash equilibrium if they are using the ultimate strategy with respect to the other. This does not imply a cumulative optimal, only optimal worst case. By disallowing some strategy which forces others into a destructive strategy, fair strategies can exist with safety. The issue is that Braess's paradox can manifest itself in attempts to enable superior choice, but accounting for it by mandating the strategies of investors implies disabling competitive trading. Trading is a part of capitalism, which bases itself on the self organizing properties of competition Implies HFT is highly automated. While the speed trading occurs at excludes human intervention, firms working with HFT must thoroughly analyse the data to maintain a competitive edge "Infinium taps into the CME Group's computers ... via dedicated fiber-optic lines capable of transmitting up to 5,000 orders per second with a lag time of no more than 10 milliseconds" Data with respect to the 10ms times: My ISP, electronicbox, advertises fast ping times. It takes me 20ms to roundtrip ping them. Outliers can raise up to 70ms. The 10ms was an upper limit "It is our goal to democratize high frequency trading" Issue: This is highly software focused. HFT relies heavily on hardware, which is not readily affordable; and location, which is not readily changed Seems to imply HFT is good because it encourages the market to adapt and protect itself against HFT. A problem which solves itself is still a problem. Alternative: If the market reflects a fractal, fortifying the system against a high speed low noise environment rapidly protects market models from over specializing with respect to current trends Mandelbrot opined that the market looked the same all the way down. Perhaps HFT is therefore lucrative because it is able to exploit the market as if it had no noise. Alternatively, there might be some form of noise which is compensated for, as implied by fullfinance.com, which HFT makes to profit on. This is analogous with the story of the bank scam where fractions of pennies were stolen from the rounding off of interest. By leeching onto the market's negligible, and self correcting, deviations before the slower market can, HFT might be seen to be licking the leftover scraps of what the classical market would clean in a fell swoop later on. Allowing deviations to accumulate makes dealing with them cheaper in ratio, so long as someone else does call them first at a smaller profit Quote stuffing is the offer of deals which won't be accepted, and are used by HFT firms to add artificial noise for competition (The producer of noise can more easily cancel it out in analysis) Quote stuffing baits others to make mistakes. Happens in a few millisecond runs Quote stuffing caused latency which caused margins of errors which at normal resolutions are indescernible. HFT uses a lot of bandwidth, which is detrimental to everyone since the exchange's bandwidth is a shared resource Some will defend quote stuffing as an emergent behavior. It's difficult to regulate what can be claimed as an emergent behavior, because it becomes easier to claim innocence with every new layer of indirection Strogatz, S. (2008, December). Steven Strogatz on sync [Video file]. Retrieved from Strogatz suggests that the most mundane swarms can evolve self organizing sychronized systems. HFT is a mass of small transactions, each reacting to the slight of every other. Similar to his bridge example, the liquidity market might pick up a similar volatility held by Nash equilibrium. Nash equilibrium resembles entropy with how it often ends in evenly distributed lukewarm rot. If everything is the same, progress stalls. Systems must work to not foster imitation L'interview de Benoît Mandelbrot par Annie Kahn (17.10.09,16h17) pour Le Monde (édition 18.19.09) Mandelbrot argues that current mathematical models of economics do not take into account fast pace change, being developed while the market was simpler in the 1900s. He is a proponent that complexity is to be dealt with through infinite recursion. His interactions with bankers imply that so long as some profit is being made alongside HFT, nobody cares to adapt. If the math with which post mortems occur is wrong, there will be too many failures attributed to bad luck Preliminary Findings Regarding the Market Events of May 6th 2010. Report of the Staffs of the CFTC and SEC to the Joint Advisory Committee on Emerging Regulatory Issues. Published May 18th, 2010 Underlines that the temporal nature of the flash crash implies liquidity failure. In other words, liquidity is not behaving as models suggest LONDON From Saturday's Globe and Mail Published Friday, Nov. 26, 2010 5:29PM EST "The Lunch, The revenge of Gordon Nixon" by Eric Reguly Nixon mentions that assets are easy to acquire. Independently, they don't matter. In terms of assets, debt doesn't matter: If I loan someone a dollar, I've lost a dollar for an asset worth a dollar. The only loss is liquidity, since I can't turn that debt into money as fast as I can turn money into money. HFT supplied liquidity is liquidity which the investor keeps. If the thing being invested in wants the liquidity, they must require the investment remain alive for a time. When the market gets unstable, a time when investments are required (Hence why the US government put out the stimulus package,) HFT investments back out. Any markets relying on HFT investments then crash, only after the HFTers get out safe. Since the market was becoming unstable, complete blame can't be made on HFT. But HFT serves to be a positive feedback on instability. It happens too fast, not giving those who want to take the long road and come out clean enough time to stabilize. Nixon's daughter works at Goldman Sachs, a big player in HFT, so he probably knows enough to not play with them if he isn't playing their game. Liquidity is only good if it's only liquid for you. He's also not a fan of American banking, which is an HFT hotspot Regulations must be defined in a way to adapt with a changing market. HFT is too fast for regulations to be amended, so they're axiomatic nature shines, which then requires an analysis of how impacting things like Gödel's incompleteness theorem are, for which a fair analysis is beyond the scope of this meditation on HFT. So I'll end it at that: Further thought requires too much thought, and a topic of such complexity might be best removed from the market for reasons of simplicity alone
|
