Let a computer make up your mind about when it's best to buy and sell
Take your pick: algorithm trading works because the trader buys more when the share price is low and less when high. photograph: toru hanai/reuters
Algorithmic trading suppresses two potentially costly states of mind – ‘optimism’ and ‘panic’
Buy when everyone is selling and sell when everyone is buying. This sensible advice from Warren Buffett is often hard to put into practice. Our emotions can interfere, making us buy at inflated prices along with everyone else in a boom and sell at distressed prices when a panic emerges.
Algorithm trading can avoid this. Here, a computer or some other automated process makes the decision to buy or sell. Computers aren’t emotional and therefore can often outperform experts.
Simple to complex
Automated or algorithmic trading can range from the simple to the complex. At a basic level, a trader might put aside €1,000 each month to invest in shares in company X.
The €1,000 is spent regardless of the share price. In January the share price is €8, in February it is €10 and in March it is €12. Table A (right) shows that the average share price over the period is €10 but the shareholder manages to buy 308 at an average of €9.73.
Table B (below right) shows that when the share price is more volatile, the trader buys more shares for the same amount of money even though the average for the period stays at €10.
This explains why some algorithm traders can make impressive profits when the markets are volatile.
This form of trading works because the trader buys more when the share price is low and less when high. It suppresses two states of mind – “optimism” where we buy overpriced shares and “panic”, when we sell at distressed prices.
Simple algorithmic trading produces only small profits which are often absorbed by transaction charges. It works better when trading currencies than shares because transaction charges are a lot lower and there is no stamp duty.
Another variation on algorithm trading which encourages buying low and selling high is “mean reversion”.
Here, traders calculate the average share price over, say, one year, and take the view that if the current price is well above the average, the likelihood is that the share will fall back to the long-term average and vice versa.
For instance, a trader might calculate the average share price for company X at €10 and either commit himself or programme a computer into buying 1,000 shares if the share price falls below €9 or sell if the share price goes above €11.
Algorithm trading can become more sophisticated and lucrative when derivatives are involved.
Susan Hayes, a mentor with Optionstradingireland.com, says there is growing interest among Irish investors and adds that “they tend to use American stockbrokers because of the wider choice and lower transaction charges”.
Call and put
Some traders, for instance, sell “call” and “put” options on shares which compel them to sell when the share price rises and buy when the share price falls.
Call and put options are similar to insurance contracts. The seller receives a premium but guarantees that he will buy the share if it falls below a certain price (put option) or sell the share if it goes above a certain price (call option).
The benefit for the trader selling the options is that he receives the premium from selling calls and puts. More importantly, it forces him to buy when the share price falls and sell when the share price rises.
Irish stockbroker Goodbody offers a limited range of derivatives to its wealthier clients but only after a vigorous vetting procedure to make sure that investors understand the risks.
Spotting anomalies is a more sophisticated form of algorithm trading. It works best when carried out across different asset classes.
For instance, in 2007, some investors noted that bank share prices were rising at a time when credit traders were charging a much higher premium to insure the same banks against default. Algorithms saw this anomaly and profited from it. Traders (or more precisely computers) sold bank shares and repurchased them at much lower prices. Even traders who knew nothing at all about the looming banking crisis benefited.
Simple algorithm trades won’t make you millions but they are worth understanding. Algorithm trading over different asset classes is more for the dedicated investor and may involve investing in trader screens to get data on shares, currencies, commodities and interest rate volatility.
Many traders make the mistake of mixing leverage with algorithm trading. In essence, leverage is where you borrow money to buy shares. If you own shares on borrowed money, you are forced to sell when the share price falls and therefore you do the opposite of what Buffett suggests.
Derivatives can make you rich but, as Seán Quinn discovered, they’re risky and can change your life very quickly. As financial derivatives business IG Index enters the Irish market, the best advice is to start off with small amounts and set strict loss limits. Otherwise emotions are bound to overrule any algorithm you create.
Cormac Butler is the author of Accounting for Financial Instruments and has led training seminars for bank regulators and investors on financial risk. He has traded equities and options