Trustnet Magazine Issue 44 October 2018 | Page 26

Your portfolio blamed to some extent for corrections of every shape and size in recent years. However, while the market tends to quickly rebound from these temporary setbacks, is there a danger that algorithms could eventually pose a long-term threat to investments left on “autopilot” in tracker funds? Christopher Gannatti, head of research at WisdomTree, thinks this is unlikely. High-frequency trading “Most of the time when the question comes up about algorithmic trading, it is under the umbrella of high- frequency trading,” he says. “A lot of work has been done on this and it found the biggest issue for these algorithms is the location of the servers. You have these high-frequency trading hedge funds for which money is no object which spent vast amounts of money on fibre optic cables in a perfectly straight line between Chicago and New York to ensure they had the fastest overall trading experience.” Gannatti says the reason these funds are so obsessive about speed is they use the algorithms for front running. This means that through analysing patterns they can determine whether a large order is about to be placed for a company’s shares, allowing them to buy in first to benefit from the subsequent uplift in value. FE TRUSTNET [ ALGORITHMS ] 24 / 25 “Having done it before, you would have a trading record as long as your arm and high-frequency traders would rip your legs off” “The analogy you frequently see is picking up pennies in front of a steam roller,” Gannatti continues. “The reasoning being that if you find a way of making a tiny amount of money, but you do this many times and have a bit of leverage, ultimately you have an attractive investment strategy.” But with these hedge funds using algorithms to profit from the actions of professional fund managers, surely it is only a matter of time before a “pump and dump” version is developed that could take advantage of the vast amounts of retail investors’ cash flowing into passives? For example, with passive funds buying every stock on an index, an algorithm could take advantage of this by pumping money into one stock, forcing the tracker funds to buy it at the close of trading, pushing up its price. The algorithm could then short it and sell out, so forcing the index trackers to sell out as well, deflating its price – before repeating the process ad infinitum. However, when this question was put to the experts, they said there was so much wrong with it, they didn’t know where to start. Alex Tarver, analyst at QuotedData, says the first problem is that the market would “see you coming from a mile off”. “Extraordinarily large pockets” “The securities regulator and the FCA would be all over it, because it is market manipulation,” he explains. “And you would need extraordinarily large pockets to lift the share price and keep it there. Having done it before, you would have a trading record as long as your arm and high frequency traders would rip your legs off.” However, while pushing up the price of individual stocks like this would be incredibly difficult – not to mention illegal – in the likes of the S&P 500 and FTSE 100, what about in more obscure markets where the stocks are a fraction of the size and the regulators take more of a “hands off” approach? Gannatti points out the problem with this hypothesis is that although trustnet.com