Market sell-offs used to be caused by people phoning their stockbrokers and telling them to sell.
These days, a widespread fall in share prices is much more likely to be the result of robots responding automatically to predetermined commands.
Last week, fears about the spread of coronavirus helped stock markets record the quickest correction – a greater than 10% fall from a peak – since the Great Depression of 1929.
Dickie Hodges, of investment company Nomura Asset Management, said coronavirus created the perfect scenario for algorithms to bring down markets.
“Algorithms make automatic trades based on positive or negative headlines."
"With the coronavirus outbreak we have seen bad headlines spike, which has led to stock selling,”
“With the outbreak now hitting America, this will only accelerate.”
Mr Hodges expected the selling to get worse before it gets better because any vaccine, if found, will be slow to pass clinical tests and be distributed globally.
It is also suggested that the rise of exchange-traded funds (ETFs) enable whole portfolios to be be bought and sold like a stock, is exacerbating the crash. When people withdraw money from ETFs they trigger selling in all stocks in the benchmark that ETF tracks, which contributes to a broader sell-off. Forced sellers without buyers create more extreme losses.
There is some £340bn invested in funds which trade when algorithms detect signals in UK markets, and the desire of fund managers to design and use an ‘automatic trading’ system has involved the belief that there are hidden patterns in financial markets to watch for, rather than spending time studying individual companies or meeting business leaders as most fund managers do.
Hidden patterns form the basis of algorithm stock trading strategies.
For example, weather data might hold clues about commodity price movements leading to the algorithm to buy or sell mining stocks. Satellite images of shopping centre car parks might reveal secrets about consumer spending. If you can prove there is a link between two data points – and no one else knows about it – then you can make money by moving first.
Swiss investment group Gam has had success in using computers to read financial reports using a technique called natural language processing.
Computers crunch huge amounts of written data to assess whether companies are in or out of favour among stock analysts. Its algorithms then make automatic trading decisions without human interference.
They can buy businesses tipped on the expectation of share price bumps and sell those likely to come under pressure, long before most other investors can.
This is much more efficient than employing humans to read financial reports.
Incredibly, despite a team of PhD scientists working on the development of algos, they normally fail to find patterns that work and hold up consistently.
Even when they find something that makes it into a trading strategy, it only makes money just over 50% of the time.
However, this is enough – similar to the casino winning in the long-run on a roulette table.
Performance has been less than exciting, though...
Gam has two funds available to British investors.
The Gam Systematic Alternative Risk Premia Fund is available through fund supermarket AJ Bell. It has £2bn in assets and charges 0.71%.
The £1.2bn Gam Systematic Core Macro is available through rival shop Hargreaves Lansdown and charges 0.70%.
These funds invest in a wide range of markets but all the trading decisions are made by algorithms.
However, as the chart shows, returns have been poor when compared to global stock markets.
Investors must make a judgment about whether the virus will cause a short-term shock fall followed by a recovery, or whether it poses a longer-term threat which will take longer to overcome.
If retail investors join the sell-off then this could lead to a more serious market crash, but the evidence suggests this is not happening.
Most managers are likely to be measuring their response at this stage.
Holding tight seems to be the order of the day when markets are this volatile, and as soon as there is any apparent ‘bottom’ buyers will return quickly.
For any further, personal advice about this subject, please do not hesitate to contact me
You can view and download a PDF version of this article here
The full library of Blog article PDF files can be found here