Some HFT is Good, Some HFT is Bad, But the Market Structure is Ugly

Cameron Smith wrote in Traders Magazine Online, today, a defense of HFT. He used an analogy that HFT blesses the market with its liquidity, and now people have the nerve to question it: the analogy is of a sailor saving a drowning boy, only to have the boy’s mother ask him where is the boy’s cap.

Our response:

We read Cameron Smith’s well-written editorial in Traders Magazine on March 17th, 2010 and it echoes many other viewpoints on this topic that we have also read in the past year extolling the virtues of high frequency trading (HFT). As a matter of fact, it echoes the same viewpoint that Mr. Smith made back in November 2009. We responded to his claims back then in our blog,, and we will do so again here.

The defenses of HFT are, and have always been,

–         HFT adds liquidity and reduces spreads

–         HFT aids in Price Discovery

–         HFT Lowers Transaction Costs for All

–         HFT is Market Making Evolved.

–         There are no studies that show HFT to be harmful.

The professionals who make these arguments are all lawyers, consultants, executives at ATS’s whose largest customers are HFT firms, and executives at exchanges whose largest customers are HFT firms. The defenders of HFT are defending a machine that made $21b last year, and expects to make multiples of that this year. We offer our rebuttal as agency traders. We offer our rebuttal as the voice for many, many buy side traders and investment firms.  We have been trading equities for clients on an agency basis for twenty years, and we have been doing this in a manner that always has embraced technology and change. In fact, we spent nearly a decade of our careers at Instinet Corporation. We have never been specialists, we have never been market makers making eighths and quarters, and we have never been SOES bandits working at Island or Tradebot.

I make the above point, as any reader should question the motives of any strong viewpoint on an issue. We run a high-touch agency trading firm that is small, and small by design. That is our only stake. We are not against any one type of trading or investing, and have no moral yardstick that we use to size up market participants. We are agnostic to day traders, value investors, stat-arb players, momentum shops, GARP players, psychic traders, and HFT firms. We do take issue with a market structure that has evolved into one which favors specific participants over others. OK. On to our rebuttal…

1)      While HFT adds liquidity and reduces spreads in the top 100 stocks, the other 95% of US equities have experienced wider spreads and more volatility. Why does every defender ignore that fact, and just lay down as gospel that HFT adds liquidity everywhere? We are happy that market players everywhere can trade Citigroup ($4), Fannie Mae ($1), Freddie Mac ($1), Bank of America ($17), Ford ($12), FAZ ($13), and countless other high volume large cap names and ETF’s, for penny spreads all day long. This is a good thing. If The Prince wants to sell out of his 200 million Citigroup shares, he can rest easy that the trade might take an hour or two to effect. But what about the small cap names? What about those names in which every posted display of liquidity results in 17 firms each penny jumping the original buyer or seller initially, and then each other subsequently, to create veritable volatile feeding frenzy events every time a player enters the picture? The predatory HFT at work in these lower market cap strata do not bless anyone with their “liquidity”, and their activity certainly does not minimize volatility. Traders will tell you that.

2)      Contrary to claims that HFT aids in price discovery, we again assert that HFT is divorced from price discovery in two meaningful ways:

  1. One just needs to look at what happens when there is a “keypunch error” (think DNDN or RMBS) to observe that HFT in those cases was not stabilizing, but destabilizing. The price swings that ensued hurt investors, and had absolutely nothing to do with the value of the security, or the firm the security represented.
  2. HFT, specifically the rebate driven kind, bought and sold Burlington Northern (BNI) at $60/share all day long (going home flat), and also bought and sold BNI at $70/share (going home flat) a few months later, and also bought BNI at $99/share all day long (going home flat) after Warren Buffet made his takeover announcement. At any and all price levels, HFT was happy to play, and agnostic to the price level. Does this sound like price discovery to you? They play agnostically at any and every price level.

3)      Does HFT lower transaction costs for all? Is this because retail trades for $10/trade with narrower spreads? Is the cost really $10? Why do large HFT firms pay the online retail brokers for their flow? They do so because they make money off of that flow. The cost then is not $10. Folks need to understand that. If you are reading this, please ask your online broker what they will charge you to execute your trades if you pick specific exchanges and ECN’s to route to instead of their “auto-router”.

4)      HFT is not the same as market making. It is not “market making evolved” that is no different than the specialist taking his cut. The specialists had an obligation to maintain a fair and orderly market. They organized blocks, and aided blocks with their firm’s capital. Many days they made nickels, and some days they lost dollars. Net/net they made money for sure. But they were there in bad tapes. HFT is not the same thing. If the conditions are not right for profit for them, they turn off. In fact, the very Fed FX study that Mr. Cameron mentioned in his commentary also demonstrated that HFT disappeared  for minutes before and after market-moving announcements (like a Fed Rate Day), and as a result volatility spiked and spreads widened. Good prop trading for sure! Not market making though.

5)      HFT defenders have in the past challenged its critics to produce studies that demonstrate harm to long term investors. At that time it was challenging to quantify the ill effects, as HFT practitioners don’t exactly offer up their books and records for the task. This is one reason why we think that the SEC idea to put an HFT order tag on all HFT orders would be so useful in that regard. Studies can then be easily done to prove the benefits or detriments of this market participant responsible for 70% of all orders, and 60% of the volume. We also would like to point out that more recently, many studies have emerged that have highlighted some of the negative aspects of HFT, and its detrimental effect to investors. These studies are quantifying adverse selection in dark pools frequented by HFT. We urge your readers to check out studies by Jeffries (A Report on Information Arbitrage & Its Impact on Execution Quality), and more recently QSG (QSG Study Identifies Impact of Predatory High Frequency Trading on Institutional Equity Managers). More studies will come as well. One should be careful for what they wish for!

HFT is a very big bucket that catches many types of trading. We don’t for the most part question HFT morality, or legality. Stat/arb for example is a stabilizing and market benefitting type of trading. We do question a market structure that has allowed predatory HFT to flourish. The predatory trading that picks off dark pools using a plethora of tools (actionable IOI’s for example), and amped up with co-located speed is an issue in our opinion. But make no mistake: it is a dwarf issue relative to the fact that for-profit exchanges, focused on next quarter’s profits, cater to HFT firms at the expense of others. Exchanges rolled out Flash order types, admitting that they were wrong and unfair. Think about that.  Exchanges sell co-location space to HFT firms, looking for micro/nano/pico second speed advantages, so that they can beat slower orders to the quote. Why is the SIP, the public quote that powers the vast majority of institutional buy-side algorithms, operating at a much slower speed than those who are paying to co-locate? This is a market structure issue much more than an HFT issue.

We also question the effect HFT is having to asset prices in general. This is a “big picture” hang-up we have. If there are hundreds of different styles of market players (value, GARP, momentum, index, HFT, day-trader, etc), and no one style is completely overpowering, then we feel comfortable that market prices are reflecting underlying values (i.e. S&P500 = 1150). Let us ask you this, however. If you are looking to buy a diamond, and you go to a diamond district where one player controls 70% of the transactions, do you feel comfortable that the $10,000 price tag on the ring you want to buy is an accurate reflection of its value? Would you feel more comfortable if you were buying a ring with a $10,000 price tag in a market place with hundreds of players, none of them that overpowering? We would.

Our current market structure, contrary to Mr. Smith’s opinion, is not illustrative of a fair and transparent market. Mom and Pop, be they located in the financial district in Boston, in Des Moines, or at their condo in Pompano, should not have to pay tens of millions of dollars per year in order to take the 1,000 shares offered at $17.30. If an HFT firm can hit the enter key faster than my mom in Pompano, then they should get the stock. But if my mom is quicker to hit the enter key, then she should get the stock, and not an HFT firm who paid co-location rent to have their server at the exchange, which will see her buy order (through many predatory means) and beat it to the offer with its warp speed. And with regard to analogies, instead of the one that has HFT saving a drowning boy only to hear his mother complain of the whereabouts of his cap, we like the one where someone relieves himself on your shoes, and then tells you it is raining.