Streetwise Professor

July 3, 2011

Quick Thoughts on HFT

Filed under: Derivatives,Economics,Exchanges,Financial crisis,Regulation — The Professor @ 6:51 pm

Deus ex Machiatto has an interesting post on high frequency trading.   His basic take:

I would argue that HFT has reduced trade sizes and decreased liquidity/increased costs for block trades, especially combined with the move to trading the VWAP rather than brokers taking on blocks as a risk trade, but (3) really gives me pause for thought. Is it true?

. . . .

My recipe for HFT is not to withdraw it, but to reduce the impact of its speed by (1) requiring all market participants to trade on a central order book – no dark pools and (2) requiring all quotes to be good for a minimum of half a second. This would affect real trading activity very little while completely wrecking the high frequency strategies that generate flash crash risk.

I am skeptical that requiring trading on a central order book (CLOB) will eliminate flash crashes.  Note that flash crashes have occurred on futures markets with CLOBs.  The eMini on 6 May 2010 is the most notable example, but even in markets without linkages to other markets–such as cocoa and natural gas–(a) have CLOBs, and (b) have experienced flash crashes.  I’m also skeptical that dark pools have anything to do with flash crashes, and am also not convinced that the dark pool structure–which as DEM notes, has effectively supplanted the block market for facilitating large trades–is detrimental to the interests of those who want to trade in size.  After all, a block positioner could still attempt to compete for the business of those who want to trade in blocks: the fact that dark pools have become the venue of choice suggests they outcompete the old way of trading blocks on cost.

That said, I have long been critical of the SEC rules that protect only the top of the book rather than the depth of book.  Top-of-book protection provides a way for HFTs to make money by exploiting fragmentation of liquidity.  Creating a real CLOB entails a whole set of issues (see the linked article above for a brief discussion), but these can be avoided through the creation of a virtual CLOB that protects the entire books of the various liquidity pools.

A related issue is flash quoting, in which the operator of a matching system gives those supplying liquidity on it a first chance to execute orders before they are routed to other venues.  This undermines time priority, which also precludes the creation of a virtual CLOB by integrating the books of multiple venues.  This can also affects the incentives to supply liquidity.

DEM also links to a presentation that notes that HFTs make money by collecting liquidity rebates.  He further quotes a Bank of Canada study:

Some HFT participants to overload exchanges with trade messaging activity; use their technological advantage to position themselves in front of incoming order flow, making it more difficult for participants to transact at posted prices; or withdraw activity during periods of pricing turbulence.

Both these issues relate to the subject of the very first substantive SWP post: exchange pricing.  That is, the prices that electronic exchanges charge their customers.  I wondered then, and I wonder now, why exchanges don’t charge users for causing congestion on their systems.  If quote stuffing and overloading is problematic, exchanges could penalize it through how they charge for message traffic.  You’d think they have the appropriate incentive to do so.  If it is parasitic on the customers who generate revenues for exchanges (because it is undertaken by those whom exchanges typically pay rebates–see below–and harms the people whom the exchanges charge positive prices), the exchanges would have a strong incentive to devise pricing strategies or quoting rules that would tend to deter this conduct.

Exchange pricing is also relevant for rebates.  It makes sense for exchanges to subsidize some activity (providing quotes) that increases the revenues they can realize from other activity (taking liquidity).  Again, exchanges should have the right incentives to balance the costs and benefits and come up with pricing schemes that encourage the right amount of liquidity supply.

It would certainly be odd, I should add, to want to discourage this kind of pricing if one of your concerns is the disappearance of liquidity.

And related to the disappearance of liquidity, I agree that contributes to crash risk–in both old school floor markets and newfangled electronic ones.  But restrictions on quoting–e.g., DEM’s suggestion that quotes be good for one second–could well be  counterproductive in that regard.  The current evidence suggests that order floor became much more toxic on 6 May 2010, and that’s why HFTs stopped supplying liquidity.  Forcing them to keep their quotes good for longer than they would choose to on their own in such circumstances makes it more likely they will not quote at all.  That is, establishing minimum times that quotes must be in effect is likely to accelerate the wholesale withdrawal of quotes altogether during circumstances in which flash crashes are most likely.

I think there are many issues implicated with HFT–not the least of which being whether there is even an approximate match between the social and private benefits of this activity.  Many of the issues associated with it, however, I think are only peripherally related to flash crashes.  HFT may have a role in flash crashes–though exactly what is hard to pin down, and it may also have a role in making them flashier, i.e., shorter.  But many of the controversies involving HFT don’t appear to be connected to flash crashes.  As in many issues involving financial infrastructure these days, many issues are getting mashed together.  This makes coherent analysis hard, and coherent responses even harder.

Print Friendly, PDF & Email


  1. […] Some thoughts from my friend on HFT.  He and I disagree on this.  I would eliminate dark pools and have everything trade on an arbitragable CLOB.  Been saying that for years and it has not much to do with competition, but risk/reward.  I also would eliminate the penny for stocks and go to nickels.  You’d get bigger bids/offers, thicker books and more spread opportunities. […]

    Pingback by Breakfast Links | Points and Figures — July 4, 2011 @ 9:10 pm

  2. […] Streetwise Professor responded to my post on HFT with a number of good points. I am skeptical that requiring trading on […]

    Pingback by Deus Ex Macchiato » Nuancing HFT — July 5, 2011 @ 5:20 pm

  3. […] – On HFT and block trading. […]

    Pingback by FT Alphaville » Further reading — July 6, 2011 @ 1:09 am

  4. A few comments, if I may.

    “I wondered then, and I wonder now, why exchanges don’t charge users for causing congestion on their systems.”

    I believe this is due to the competitive environment. The first exchange that tries to impose realistic limits will be branded a Luddite by the other exchanges and watch their business shrink. Collectively, however, the exchanges would welcome a governing mechanism (minimum quote-life and/or cancellation fee) to reign in the out of control explosion of quote traffic.

    “Forcing them to keep their quotes good for longer than they would choose to on their own in such circumstances makes it more likely they will not quote at all. That is, establishing minimum times that quotes must be in effect is likely to accelerate the wholesale withdrawal of quotes altogether during circumstances in which flash crashes are most likely.”

    They already withdraw their quotes in a millisecond or less whenever the slightest market imbalance is detected. The only ones left standing in front the train are those who are impaired by things like the speed of light.

    There has to be some minimum time that a quote should be honored and/or a cost associated with canceling it. Many feeds have a tag with each quote signaling that it is auto-executable. How can a quote be auto-executable if it expires before it reaches the computer processing it? Today, it is common to see dozens if not hundreds of micro-burst traffic events where a significant number of quotes have already expired before leaving the exchange network!

    So where does it end? Is 1,000 quotes/sec unrealistic? How about 10,000 ? How about 1 million quotes/sec per stock? What? You say, that’s just silly? It was last year we transitioned from the era of silly into the era of the absurd.

    On July 5, 2011, CQS increased traffic capacity from 750,000/sec to 1 million/sec. That same afternoon, traffic rates spiked and saturated the lines on a millisecond basis. Similarly, OPRA regularly saturates at a capacity of 4 million/sec. We estimate the limits need to be at least 3 times higher for both in order to avoid saturation. At these rates of growth, there will no longer be a diversity of participants with accurate market data. And the regulators will have no hope of ever piecing together what happened in the next disaster.

    Eric Hunsader

    Comment by Eric Hunsader — July 7, 2011 @ 5:27 am

RSS feed for comments on this post. TrackBack URI

Leave a comment

Powered by WordPress