Send to a friend
06
Sep

Algorithmic Trading is Anti-Investing

No doubt, recent turbulence in global markets has many investors asking why and how such volatility can arise.

Answers to these questions are often replete with trite references to sentiment and behaviour. Indeed, these answers may offer partial explanations. However, as technology and artificial intelligence infuses more and more aspects of our lives, it reveals itself in a potentially insidious way in the markets.

*Darron West represents Foord Asset Management and is also a Senior Lecturer at the University of Cape Town.

No doubt, recent turbulence in global markets has many investors asking why and how such volatility can arise.

Answers to these questions are often replete with trite references to sentiment and behaviour.  Indeed, these answers may offer partial explanations.  However, as technology and artificial intelligence infuses more and more aspects of our lives, it reveals itself in a potentially insidious way in the markets.

It is no secret that much of the activity inherent in market movements results from actions taken by computers that have been programmed to trade according to certain rules and patterns.  This type of programme trading is also known by other monikers such as high frequency trading or algorithmic (“algo”) trading.

In essence, trading of this nature is predicated on the notion that a computer (if programmed appropriately) can receive information at near light speed and react to that information almost as quickly (and certainly more quickly) than the average human trader.

It should be plainly apparent that one of the intrinsic characteristics of such trading is its patently short term nature.  Equally, it must be clear that an armada of computers executing trades on an exchange can, should and does increase trading volumes.

This begs the question, “Who benefits?”  Stock exchanges themselves endorse an algo approach to trading, since higher trading volumes mean higher revenues for the exchanges.  Other proponents of algo trading argue that it aids price discovery and makes markets more efficient.

These views may be short-sighted and disputable.  In the first instance, algos are designed and built by human beings and as such they are not foolproof.  One has merely to look at history for examples: the market crash of 1987 was somewhat attributable to the programmed execution of sell orders to liquidate positions that had fallen through stop loss levels.  More recently, US markets witnessed what appeared to be an algo trade gone wrong when natural gas prices dropped precipitously in seconds.

Secondly, there may be an inherent conflict of interest in algo trading.  For a computer to execute a programmed trade, it requires information from an exchange.  The information sought is detail relating to the price and quantities of buy and sell orders (which may well have been placed by human investors).  Whilst it is a given that the information is paid for legitimately, it does lead one to ask whether the acquisition of and reaction to this information isn’t tantamount to front running, since an instantaneous reaction to purchased information to which no cognitive participant has reacted is hardly a reaction to information in the public domain.

Thirdly, exchanges might need to be more attentive to their oversight role as a consequence of their endorsement and facilitation of algo trading.  It is common knowledge that exchanges are often shut down during times of extreme panic or euphoria so as to allow those respective sentiments to calm.  Since algos might very well cause price trends inducing extreme fear or greed, exchanges should be especially vigilant; the corollary to this is that exchanges might need to reconsider their endorsement of algo trading too.

What cannot be in dispute is that algo trading is a speculative activity.  It is not investing.  The decision inputs for algos are based purely on price movements and have absolutely nothing to do with the fundamental value of the securities bought or sold.  Price discovery is about the cognitive activity of matching a willing buyer and seller – it has nothing to do with an artificial and seemingly mindless response to a price only.

If anything, the foil to the purported advantages of algos as a mechanism of market efficiency might very well be their contribution to increased market volatility as their trading rules amplify trends, making price spikes self-fulfilling.  Whilst volatility is an expected consequence of investing in the markets, excessive volatility is destructive.  It distorts a fair representation of value and risk, both of which are key to successful investing.

Ends