An article in yesterday's Advanced Trading magazine, written by Pipeline executives Fred Federspiel and Alfred Berkeley, which was supposed to extol the virtues of HFT (or of the Pipeline product offering specifically, we were a little confused on that issue), ended up doing anything but, and in fact confirmed many of the concerns voiced with regard to high frequency trading in the blogosphere and in other venues.
While the positive spin on HFT presented in the article is no surprise - basically a recap of what all other proponents chime in with, it provides liquidity, it must be good, etc. (and later on, we will dissect this argument more closely), it is the negatives that AT presented, that were very surprising.
First, the focus is on what the authors call "rebate harvesting algorithms" or rebate seekers as they are more popularly known:
Some high frequency strategies "the so-called rebate harvesting algorithms" do much more than simply harvest rebates. Many rely on very short-term alpha predictions to create profitable trading opportunities, and the best of them wring every timing advantage possible from the markets. When the alpha is strong enough, they will cross the spread to take advantage of short term fluctuations. Operators of these strategies will co-locate at the exchanges or ECNs in an attempt to be the first to act on any signal to enter, or pull back from the market. The extreme measures they deploy to time their orders results in adverse selection losses for the institutional orders on the other side: the institutions trade more slowly when they should have traded quickly, and they trade too quickly when they should have held back.
The key issue here is that so called liquidity providers are in fact doing much more than just the altruistic act of collecting a third of a cent on any limit order, especially when matched with a client's own market order. This is especially true when, in the case of Goldman Sachs for example, they provide their clients almost exclusively with a market order option following VWAP for large block trades. All the while, Goldman as principal will eagerly wait on the other side of the trade, not only collecting rebates for every single trade initiated by its own clients via REDI and other market order routers, but will potentially exacerbate the "adverse selection losses for institutional order on the other side", i.e., its very own clients. Indeed, the architecture of the market is one which reinforces in a positive feedback loop adverse selection for those who are unable to unwilling to mask their participation under the guise of liquidity providers. It is in this light that the NYSE's SLP program has to be much more carefully evaluated, as even Pipeline admits there are many agendas driving "liquidity providers", of which providing (transient) liquidity being the least of.
Continuing with AT's observations:
Other high frequency trading approaches "a class of stat-arb strategies sometimes called information arbitrage" look over longer timeframes in an attempt to detect asymmetries in trading interests, and then profit by trading before institutions have a chance to finish their orders. These intra-day timing tactics have been called "front running" or "penny jumping"; they directly generate market impact losses for institutions.
And there you have it - HFT's direct and mandated involvement in what is explicitly front-running of various trading interests, as a function of information traffic speed and asymmetries. And this is not merely Flash orders - this is the whole market landscape: the underlying premise of today's HFT participants is to promote riskless (or as close to as possible) trading for those who are embedded within the market topology and have been given the green light by exchanges to "front run" or "penny jump" - call it however you want. Maybe it is time the SEC provided its own semantic definition of this phenomenon.
Another purported benefit of HFT:
Interestingly enough, the evidence shows that institutional trading costs, taken in total, have remained remarkably constant during the transition to high frequency trading. The implicit components of transaction cost " adverse selection losses, and direct market impact losses " have indeed been driven up, but the commission component has decreased in measure.
Ironically, Zero Hedge analyzed the explicit costs associated with HFT - namely implementation shortfall (slippage) and commission costs, and Pipeline's claim unfortunately seems to be far and away from the facts on this one. As we highlighted in our post "The Cost of High Frequency Trading", while commission costs have indeed decline, over the past 18 months IS costs have increased by almost 50%! The actual data, courtesy of ITG, below:
Claiming a 50% increase in HFT related costs is "remarkably constant" is a little skewed. Of course, if Pipeline can provide us with supplementary data that refutes ITG's facts, we would be happy to present it.
And the conclusion of the article: how should traders proceed in this market (assuming they do not wish to subscribe to Pipeline's product):
Exceptional traders can harvest that high frequency liquidity, while limiting adverse selection losses through sophisticated control of trade timing, and limit their impact losses through obfuscation of trading interests. An ongoing focus on state-of-the-art techniques for limiting adverse selection losses and direct market impact losses will put institutional traders, and the millions of individual savers they represent, back in the driver's seat.
Zero Hedge agrees that institutional traders, ("and the millions of individual savers", although based on Bernanke's adverse treatment of the last category from a macro economic perspective with his consistent push to devalue the dollar, it is likely that in no time America will have negative savings rates once again) will only benefit from the ongoing evaluation of all the various interrelated issues that comprise HFT, that have gone unanalyzed for far too long.
And in conclusion, returning to the topic of HFT providing liquidity, Reuters' Matt Goldstein had some prophetic words in this context: "I’m not impressed with the securities industry’s main defense of computer-driven high-frequency trading, which essentially is that all this lightning-fast trading provides liquidity and better prices for investors."
Matt has the right idea:
It’s a hard argument to swallow when you consider that many high-frequency trading programs are simply engaged in trading the same stock thousands of times a day in less than penny increments. Now maybe all those rapid-fire automated trades are getting better prices for some investors. But when a broker excessively buys and sells securities to generate higher commissions, it’s called churning, and that can result in an investor lawsuit or a regulatory sanction.
Indeed, when fast-fingered day traders were doing much the same thing as today’s high-frequency traders — albeit without the benefit of a sophisticated algorithmic program to guide them — Wall Street’s biggest firms were quick to dismiss them as either amateurs or rogues who were causing unnecessary volatility in the price of tech stocks.
And Matt's conclusion is the one that all proponents of HFT realize is the right one, yet are unwilling to put in front and center, as it goes to the very heart of the landed interests in the debate over high frequency trading:
If the main purpose of all that extra liquidity is to simply make fat profits for high-frequency traders at Goldman Sachs, UBS, GETCO, Citadel Investment Group and Interactive Brokers, that’s liquidity the markets can do without.
Zero Hedge will continue working with the proper channels to continue exposing the hypocritical charade that is HFT, and unmasking each and every unsubstantiatable defense of the vampyric technology that these days accounts for 70% of stock volume, whose sole purpose of generating 60-80% reutrns with a 5 Sharpe ratio for a very select few.