Originally published by OpenMarkets
In the trading world algos are used by investors via the platform’s open APIs. They are programs which can be classed as Artificial Intelligence (AI). While that’s a stretch for more basic algorithms, some of the more intricate algos draw on AI – and in the case of some sophisticated quant investment managers, the use of machine learning in the development of algos is applied. Science fiction brought to life yet again.
What is an algo?
For the uninitiated, I’ll start with the basics and borrow Investopedia’s definition:
“Algorithmic trading (automated trading, black-box trading or simply algo-trading) is the process of using computers programmed to follow a defined set of instructions (an algorithm) for placing a trade to generate trades at a speed and frequency that is impossible for a human trader.”
Long used by institutional investors, algos are primarily used to help with best execution. Rather than manual trading, algos provide an effective way to get into – or out of – market positions. Emotions are put to the side and investment decisions become rules based rather than driven by emotion.
Sentiment has put paid to many an investor’s dreams. Buying market darlings at their peak, panic selling after a spot of volatility, holding onto favourites long after they fail the pub test…these behaviours are surprisingly common and unsurprisingly detrimental.
Although US-based, Dalbar has undertaken an annual Quantitative Analysis of Investor Behaviour each year since 1994. Despite education about long-term investing, the power of compounding, dollar cost averaging and all those other investment maxims, US investors continue to make emotionally driven investment decisions, which, Dalbar consistently finds, costs them dearly.
Dalbar’s 2017 analysis reaffirmed behavioural biases that lead to poor investment decision-making are the single largest contributor to investors’ underperformance over time. While the research is US-based, the findings can be applied globally.
Back to algos…these little programs – which can be as simple as an excel formula or as complex as in-depth AI driven programming – level the playing field for non-institutional investors. Investment managers often rely on the emotional trading of others to take advantage of mis-pricing in the market – using algos evens the score somewhat for individuals.
Five reasons to use algos
I spoke to Phil Tauberman from Cannon Trading, a business built, in part, on providing algos to professional and retail investors. According to Phil, there are five reasons an investor would use algos:
- Algos allows for best execution.
- Instant and accurate order placement, which significantly increases the chance of execution at sought after levels.
- It’s set and forget – once the rules are in play and the trades are loaded, the algo directs the trades.
- There’s reduced risk of manual errors in placing the trades.
- Algos mean you don’t have to sit in front of a screen all day, so it’s a time saver, whether a regular trader or not.
Although traditionally the domain of investment banks and investment managers, algos are being increasingly adopted by individual investors. You know a trend is established when a website like Quantopian emerges. It’s designed for algorithmic traders who develop their own strategies; each month, the best algos win trading money from Quantopian. A select (and successful) few are invited to participate in its algo-driven hedge fund. And Quantopian is just one of many such sites.
How do algos work?
The days of chart watching are drawing to an end – the most common algos are trend following trading strategies that, as the name suggests, follow trends in technical indicators such as moving averages in price or volume, price breakouts or momentum. Trades are initiated based on the occurrence of the identified trend and traders benefit.
A common algorithmic strategy is TWAP, the time weighted average price algo. It essentially breaks up a large order and releases it to the market in smaller pieces, using evenly divided time slots across a given time. In the Australian market, it’s common to split the order across the six trading hours of a day.
The TWAP strategy aims to buy (or sell) at every possible interval; according to Phil Tauberman, investors will get a better price if they participate in every trade of the day rather than participating in just one or two trades. Consider it dollar cost averaging at a micro level. For investors rebalancing an SMSF portfolio, getting the best price with low transaction costs can make a substantial impact on the portfolio value over time.
This is a strategy adopted by many investment managers, including super funds – some of which may buy and sell large positions. Because they don’t want to influence a stock price – or the market – with large trades, the algo portions an order into a quantity of smaller pieces and executes the trades over a given time. The aim – for institutions and retail clients using TWAP algos – is to minimise transaction costs and obtain the best possible price.
The downside
As with anything to do with investment, there are risks to consider. Algos have been blamed for many a ‘flash crash’, whereby automated trading in a stock, an index or market is said to have caused a sudden drop, followed by a rebound. Such incidences are isolated and, thus far, have been short-lived.
However, there is a risk that when prices fall, algorithms can exacerbate that fall because the algo automatically responds to the falling price by selling more – an algo cannot apply common sense to a share price movement.
Consider the S&P/ASX 200, dominated by the big four banks, Rio Tinto (AX:RIO) and BHP (AX:BHP). If they’re down, the whole ASX200 is down, which could lead to unnecessary selling. Algos generally can’t read the news and can’t necessarily differentiate between a stock that’s down because of stock specific factors, or down because its larger brethren have dragged the market down.
Importantly, investors need to discuss transaction costs with their broker when using algos. How will they be charged? If investors don’t have a good arrangement with their broker, they could be charged for every line of execution…and that could become expensive.
Increasing demand from investors for algorithmic trading has led OpenMarkets to team up with Cannon Trading to offer algos to traders in the coming months.