In this edition of Frank Thoughts, Frank discusses the rise of high-frequency algorithmic computer trading and its inherent disadvantages to long-term investing. 

THE 800 POUND GORILLA

      Everybody assumes “professional” investors are sophisticated and knowledgeable about stocks but are they necessarily knowledgeable or even “investors?”  The overwhelming amount of daily volume on the New York Stock Exchange -upwards of over 60%- is simply directional betting … otherwise known as Quantitative Trading.  Algorithmic computer trading programs are written to instantly buy and sell stocks based on current order flow and immediate news.  This can, and often does, result in a large percentage of stocks becoming correlated … that is to say moving up or down together at the same time.  The historical correlation of stocks moving in the same direction is 39%.  At the beginning of 2010 the percentage started to increase peaking out in the third quarter at a ten year high of 80%. 

      There is little to no distinction made in quantitative trading between individual stocks.  All are essentially treated the same … as a trading commodity.  Consequently, almost all stocks will trade in the moment ...  expressed by traders as “risk on, risk off.”  But this is not investing in any traditional sense.  Moreover what results when a majority attempts to do the same thing?  Answer:  It no longer works.  It is analogous to a 200 pound man and an 800 pound gorilla attempting to ride a seesaw. 

     According to Financial Times columnist, Michael Mac Kenzie, high frequency trading now accounts for upwards of 70% of daily U.S. share trading, up from 30% percent a few years ago.  This growth was driven by the explosive formation of new quantitative hedge funds.  Growth peaked in 2007 at $1.2 trillion, according to research firm eVestment Alliance, and is now at $467B ...  a 61% decline!  This year the average hedge fund is up just +1.5% creating further pressure on their performance.  Many, if not most, put similar investment criteria into their algorithms.  Hence they will often trade together … both up and down.  This has resulted in more market correlation and correspondingly lower profit margins and poorer performance.  As more quantitative shops were established, trading increasingly became a negative feedback loop.  In effect, high frequency trading was making the market more efficient when what the quant funds needed -to make money- was market inefficiency.

      Quantitative trading needs to be better regulated but it is now an integral part of price discovery.    Correlation is slowly, but inexorably, starting to wind down.  Hedge funds are now often canceling out each other’s algorithms … which is why, I believe, a third of the volume on the N.Y.S.E. has disappeared.

      It is not what an investment is worth at the moment, however, that makes real investors money.  Rather it is … what will the business be worth in the future?  And that judgment call is a uniquely human experience ... one that cannot be expressed in an algorithmic formula.  Advantage us.
 

FRANCIS PATRICK BOLAND

10-15-2010