The rise of algorithmic trading over the past decade is the primary reason for record-low volatility, but it has also created hidden risks which will wreak havoc on markets.

Algorithmic trading, or computerized trading, is a broad term referring to computer algorithms or programs with built-in trading rules to determine when to buy or sell specific asset classes as well as how much to buy or sell. Therefore, it removes human input from trading. The most common form of algorithmic trading is known as high-frequency trading whereby computer programs look to profit by buying and selling assets in milliseconds. Fractions of a cent can be made on such rapid trading, multiplied by millions of shares, for example, and profits start adding up quickly.

Other forms of computerized trading include exchange-traded funds, as they are required, per their rules, to track a particular index without regard to fundamentals or human decision-making. There are other forms for algorithmic trading, but these two are the most common. Combined, they comprise a multitrillion-dollar industry that was virtually non-existent prior to 1998.

The true size of this industry varies according to different sources. However, a reliable source used by The Economist and Financial Times called Aite Group shows that as of 2014 (see Chart 1), a majority of equities and futures trading in the US is currently in the hands of algorithms, while other asset classes such as options, foreign currencies and fixed income securities are quickly catching up. Asia and Europe are also moving in line with this trend in the US market, which means that in a few years’ time if this trend continues, nearly all trading across all asset classes globally will be managed by computer trading programs.
In July 2017, JPMorgan issued a report on this trend stating that only 10% of equity trading today has any human input.

This algorithmic trading is reinforcing what is known on Wall Street as the ‘herd mentality’, everyone (ie all the computer programs) is following the same trends and buying the same assets. Thus, it is creating a steady rise in values of these assets while at the same time witnessing very low volatility.

This low volatility is reflected in the VIX Index, which measures volatility, but is more commonly known as the investor ‘fear index’. The index has been reaching new lows over the past couple of years as you can see in Chart 2. The interpretation of this by some market experts is that investors are either very confident in the future outlook of the market and economy or very complacent, meaning they are not seeing future potential risks.

Quorum Center believes that the latter is the most probable explanation of the record low VIX Index.

What market experts have failed to recognize so far have been the effects of algorithmic trading on market volatility. The experts have been puzzled as to why negative economic news and recent geopolitical events have failed to raise volatility. The answer is right in front of them: a majority of trading today is done without human interaction, thus removing the human knee-jerk reactions from market trading.
Many investors have cheered the takeover of financial markets by computers because it has reduced volatility, increased liquidity and created one of the longest-running bull markets in US equities in history.

Algorithmic trading, however, is not perfect and has its flaws. It’s these flaws that are of concern. Algorithms have been known to go wrong. Just as they have been known to generate small profits over a high volume of trading, they can also generate losses which can become magnified if they are not stopped in time. Here are three examples:

1. 2010 Flash Crash
On the 6th May 2010, a US$4.1 billion trade on the New York Stock Exchange caused the Dow Jones Industrial Average Index to fall more than 1,000 points only to rise to approximately the previous level all in a span of 15 minutes. The reason for this dramatic rise and fall is still disputed but the US Department of Justice filed a criminal case, including fraud and market manipulation, against Navinder Singh Sarao, a high-frequency trader. He was accused of using spoofing algorithms designed to trick the exchanges into believing that they were receiving quotes for securities.

2. Knight Capital algorithm malfunction
On the 1st August 2012, Knight Capital launched an untested software program on its trading platform. Programmers had forgotten to add specific code related to Knight’s automated routing system for equity orders. As a result, the firm lost over US$440 million in less than 45 minutes before it could be stopped. The firm had to be bailed out by Blackstone Group, TD Ameritrade and Jefferies Group, among others.
3. 2017 Ethereum Flash Crash
On the 22nd June 2017, the price of Ethereum, the second-largest digital currency after Bitcoin, dropped from US$317 to 10 US cents in minutes on the GDAX Exchange before recovering to its previous price level. The crash was triggered by a multimillion-dollar sell order, which brought the price down, from US$317.81 to US$224.48, and caused a flood of 800 stop-loss and margin funding liquidation orders triggered by computer programs’ response to the price drop.

The aforementioned examples were one-off events and did not cause an entire asset class or market to fall. However, that’s not to say that it will not happen. US Federal Reserve Chair Janet Yellen spoke of this threat in a speech she gave at the annual central bankers’ meeting in Jackson Hole, Wyoming on the 25th August 2017. In it she said: “… in addition, algorithmic traders and institutional investors are a larger presence in various markets than previously, and the willingness of these institutions to support liquidity in stressful conditions is uncertain.” This was the first time the Federal Reserve acknowledged the dominant role algorithmic trading is currently playing in markets and highlighted the potential risk. The risk is in the uncertainty over how these algorithms will react during a market correction.

The truth is that markets are in uncharted territory. We have never been in a position where computers dominate market trading. In the aforementioned examples, you can see how quickly markets can change. A small drop in prices could trigger all the algorithms to issue sell orders at the same time causing widespread market panic and magnifying losses. We are in the eighth year of a bull market run in US and European equities. What is unique about this bull run is that it has risen without a major correction of 10-20% or more along the way as has been the case with past bull runs. Much of this can be attributed to the herd mentality generated by algorithmic trading.

Nearly every big bank on Wall Street today is calling for a correction, which would be viewed as a healthy sign that the bull market is taking a break while still on an upward trajectory. However, a modest correction of 10% or more has not been tested on algorithmic trading. We believe that such a correction would be amplified by the fact that the herd mentality will be intact just as it has been in buying assets on the way up. The herd will be rushing to sell together on the way down causing larger losses than anyone is expecting. It could quite easily end this already aging bull market.

The bottom line
The record-low volatility in equity markets can be attributed to the takeover by computers of market trading from humans. This has been good for investors in rising markets, but it will wreak havoc in a down market as all the computer programs rush for the exits at the same time.