High frequency trading (HFT), sometimes known as ‘algo trading’ is computer controlled trading that happens at high speed, depending on the algorithms that drive the trading software.
These trades happen at a speed much faster than trades done by humans. As the trades happen at such a great speed, multiple trades can occur simultaneously. As with all share trading, the profit or loss is based on the difference between the buy and sell prices.
However, unlike regular trading, an investment made using HFT may be held for only seconds, or even fractions of a second. The computerised trading platform trades shares in and out, buying and selling tens of thousands of times a day.
The aim of HFT is to capture just a fraction of a cent per share on every trade made. This is why there is so much trading, so rapidly. These fractions of a cent accumulate quickly to produce a significant positive return at the end of each day.
Companies active in HFT tend to liquidate the entire portfolio at the end of every day. Unlike traditional investors, who buy and hold their shares and await dividends. Data from U.S. government economists states that up to 70 per cent of equity market trading in 2010 came from high frequency trading.
How does it work?
The rapid trading, and short periods of holding stock, is based on decisions made by the trading software. So the success of the trading hangs on algorithm. And yes, you guessed it: the algorithms are a closely guarded secret by HFT providers.
Most HFT strategies fall within one of four trading strategies:
This is when a company quotes both a buy/bid price and a sell/offer price. As HFT software allows rapid fire buying and selling, the intention is to buy at a low price and then sell at the high price. So, the range between what is offered, what the stock is bought at, and then what it is immediately sold for, is where the profit lies.
This is said to provide market liquidity. For some HFT firms, this is their primary strategy.
Ticker tape trading
Quote and volume data is embedded in raw market data. HFT software is capable of extracting this information that has not yet crossed the ticker-tape screens on trading room floors.
Even though the data is not yet visible to other traders, it is public data. Thus, having computer software access and assess the data and use this for HFT is legal.
The HFT software will easily identify stocks that have unusual price changes or trade volumes. The system will then generate buy or sell orders depending on the assessment of data.
There are certain recurring events that result in predictable, short-term outcomes with specific securities. So HFT systems are programmed for these and take advantage of the event when they occur.
Statistical arbitrage at high frequencies is used in the trading of bonds, futures and foreign currencies. A predictable temporary deviation from a known stable statistical relationship, between multiple securities, is called a statistical arbitrage.
As an example, the software may be programmed to respond when there is interest rate parity in the foreign exchange market. This will mean the relationship between the prices of a domestic bond, a denominated foreign currency bond, the spot price of currency, and the price of the forward contract of the currency will all come into play for trading.
And the examples only get more complex from there, with multiple statistical arbitrages flagged and more than four securities being traded.
What are the benefits?
There is much debate about the possible benefits of HFT. Some industry experts say that high frequency trading improves market liquidity, narrows the difference between bid/offer prices, lowers the market volatility, and makes trading cheap and easy for investors.
However, this is not the view of the whole industry, nor the executives of many of the companies whose shares are being traded in this way.
Advocates of HFT tend to be, not surprisingly, the traders that operate the platforms and algorithms that underpin this style of trading.
Does it hurt the market?
There concerns that HFT actually creates market volatility. So great is the concern, the Australian government and the Australian Securities and Investments Commission (ASIC) are considering new trading regulations.
The U.S. Securities Exchange Commission found that HFT contributed to the market volatility of the 2010 Flash Crash.
On May 6, 2010 the Dow Jones Industrial Average plunged to its largest intraday point loss of around 1,000 points, only to recover within minutes. What set off the Flash Crash was the single sale of $4.1 billion in futures contracts by a mutual fund. Immediately after that, high frequency traders magnified the sell. The large volume of the trade would have been detected in the raw data (See ticker tape trading).
That single massive sale of future contracts wiped out available buyers in the market. And the HFT firms began to quickly buy and then resell the contracts to each other. Think of it as a game of hot potato with futures contracts. All this rapid activity drove the price down on the contracts by 3 per cent in just four minutes.
In Australia last year, Oz Minerals believes it became the subject of volatility created by HFT. The company saw big volumes of shares traded for relatively small values. The price fluctuated as a result. There were no clear motives or market announcements that drove the trading volumes. The company is not alone in experiencing such rapid, unmotivated fluctuations in trading and pricing.
In a 2012 survey by Colonial First State Global Asset Management, 35 brokers stated they were concerned that HFT was having a negative effect on the integrity and transparency of the ASX.
Does it hurt the retail trader?
The belief that the traditional trader cannot win in this market is beginning to dominate discussion.
Traditional retail traders are not using HFT platforms. They are interested in investing long term in companies, taking dividends and building a wealth portfolio, often with superannuation funds.
Meanwhile, U.S. government economists have found that HFT firms are taking significant profits away from retail investors. The same research found that in 2010, HFT firms averaged a daily profit of around $45,000 at the expense of other, traditional traders.
What is ASIC considering?
With billions being invested by trading firms into computing technology, high-speed networks, trading software and algorithms, the speed of trading is only set to increase.
ASIC is planning to implement new rules on automated trading including a requirement for a ‘kill switch’. That is, there will need to be an immediate way to cease trading if an algorithm is faulty or significantly damaging market stability.
Additional controls are also planned to prevent trades from occurring where unusual or extreme volatility occurs. The new rules will be compulsory for all companies undertaking HFT.
Relevant terms you should know
Algos – slang term for the algorithms that drive the high frequency trading platforms.
Dark pools – a system that enables securities to be traded away from public exchanges like the ASX. They were originally created to allow traders to anonymously execute large share transactions that would otherwise cause a massive price movement on the market. So, the trades were ‘done in the dark’, and details released once the trade was done, minimising any effect on the market and company share price.
The drawback is that the without public transparency, prices may be lower than if they were publically traded. Plus, traders can jump the queue for orders that are on the public market. They can also be done far more quickly and can leverage HFT platforms to do so. Overall, dark pools can lead to a lack of transparency in the market.
Flash trading – a concerning area where certain market participants buy the right to see incoming buy/sell orders around 30 milliseconds ahead of other participants. That is more than enough time for an algorithm to assess and make a judgment on how to trade.
Ultra low latency direct market access – often known by its acronym ULLDMA, this is a separate class of trading that relies on speed to gain miniscule advantages in particular security trading, simultaneously on disparate markets. The boast is that some ULLDMA systems can transmit, order and receive an acknowledgement in less than 10 milliseconds.
For regular updates, Like BigPond Money on Facebook.