Mar

30

Speed Traders Make Peanuts in Profits From Economic Data Plays - Bloomberg News Item

Vic Niederhoffer:

Excellent academic paper with many fascinating facts, e.g. the human reaction time is 200 ms. vic

Paolo Pezzutti adds:

It is called:

"Do High Frequency Traders Need to be Regulated? Evidence from Algorithmic Trading on Macro News" by Tarun Chordia, T. Clifton Green, and Badrinath Kottimukkalur

The link:

http://www.bus.emory.edu/cgreen/docs/Chordia,Green,Kottimukkalur_WP2015.pdf [38 page pdf]

Burgess Humbert comments:

Taking the paper at face value, and Ceteris Paribus, the elimination of excess profit down to a 'utility' rate of return is a natural phenomenon in a field whose technological advance has resembled an arms race of late.

Unfortunately - ( and this does not denigrate the thrust of the paper or its authors ) - a call to ex colleagues in the field just now led to them falling off their chairs laughing at the idea that the marginal returns are ( or are beginning to ) decline.

As to the other part of the paper about the need for regulations etc. Let me answer this way -

Execution via High Frequency execution has both improved and decimated 'liquidity'. ( Here , I refer exclusively to spot FX, commodity futures, long end interest rate futures and stock index futures )

It has improved liquidity for transacting small parcels that are small enough to be executed within the first two levels of the bid/offer depth order book. I find the improvement stunning - particularly FX.

In terms of dealing in size - well, just a year ago, one could call a counterparty and get a stunning bid and or offer in say 75 mio GBP/ USD Spot FX. No more…. a combination of regulatory change and HFT execution has worsened fills by about 0.00015 in this amount. ( please let's not even start with discussion about using market maker provided execution 'algorithms'…. )

A real life example to put some meat on the table ; The last time I sold 50 mio GBP USD the rate going in was 1.4987/89… the following 'fill profile' is typical of a market order of this size nowadays-

3@87

17@86

30@84

Worse case fill at 86 when dealing with senior professionals just a year ago.

There are ways to improve on this, but I wanted to demonstrate how HFT has both improved and decimated liquidity at the same time.

Andrew Goodwin is skeptical:

What is the cost to every resting limit order in every correlated asset? Think of all assets including options, stocks, futures and derivatives or combos. Isn't is a fool's errand to look at profit from S&P 500 futures to generate the conclusion that HFT does not make much money on news breaks?

Burgess Humbert agrees:

Indeed. Let's call it 30 billion a year with a Sharpe of something approaching infinity.

stef says:

The real problems with HFT are the rules and how they protect the manipulators

See "Direct vs SIP Data Feed" http://www.nanex.net/aqck2/4599.html

Burgess Humbert adds:

There are indeed some very nasty yet very legal order types. The description of some of the orders extends to 20 pages ( yes twenty pages ).

But, there are ways to minimise and work with it some of the time.

I encourage all to regularly read;

www.nanex.net

One particularly amusing story on that site ( that caused a change in Federal Reserve data release policy !!!!) is called something like '… Einstein and the great Fed robbery …'

Look it up- it will change your world.

Surely a candidate to be made into a film one day.


Comments

Name

Email

Website

Speak your mind

3 Comments so far

  1. G man on March 30, 2015 7:47 pm

    Consider reading Virtu’s IPO paperwork to get an idea of profits, its not peanuts!

  2. Spekulatn on March 31, 2015 12:21 am

    MONEY AND SPEED: INSIDE THE BLACK BOX

    https://www.youtube.com/watch?v=H4BzsevJthw

  3. Bic on March 31, 2015 6:14 pm

    Well, I guess Nanex is at least a step up from Zerohedge…

Archives

Resources & Links

Search