Computers do not eat bread, wear jewellery or, for the most part, drive cars. Yet computer algorithms rather than humans are increasingly setting the prices of wheat, gold and fuel.
Automated trading systems account for large volumes of transactions in commodity futures markets. In grains and oilseeds they account for 49 per cent, precious metals 54 per cent and crude oil 63 per cent, according to the US Commodity Futures Trading Commission.
Such systems range from sophisticated black-box algorithms to simple programs that execute an order when prices reach a set level. At CME Group, the world’s biggest futures exchange operator, orders have to be tagged as automated or manual to ease identification.
Automation has fuelled a rise in volumes and fee revenue for commodity futures exchanges. But it has also sparked a debate over whether the mechanics of determining prices has changed for the better. Specifically, there are concerns that “fundamental” information such as weather or crop conditions is less useful in guiding short-term prices.
“The whole notion of fundamentals on any given day, for weeks at a time, months at a time, has completely gone out the window these days,” says Doug Duquette, an executive at Chicago-based Vertex Analytics, whose software analyses automated market orders.
“You get momentum of algos playing upon algos upon algos, and it will just drive markets to extremes that don’t seem to correlate or line up with fundamentals on any given day or time period,” he says.
John Saucer is vice-president of research and analysis at Mobius Risk Group in Houston, which helps oil and gas producers hedge their sales. He says that “trending” energy markets, in which prices lock into long paths, are a thing of the past. Volatility is instead measured in short bursts.
“It makes it more difficult to be a purely fundamental trader. In the past you had to take into account fundamental considerations, technical considerations and seasonal considerations — I do think you have to take into account transactional considerations and algos,” he says.
We don’t demonise and we are not as apprehensive about it as we used to be
In early 2016, a trade group for US cattle ranchers confronted CME Group with concerns that fast computer traders were causing costly price gyrations in the cattle market.
“We’re now in an electronic marketplace. The speed of computer advancement and technological advancement has absolutely exceeded our ability to adjust,” says Sarah Calhoun, government affairs manager of the National Cattlemen’s Beef Association.
The exchange group has since made amendments to its cattle futures contracts and changed how it regulates order messages, by limiting the number of messages a trader could enter per transaction.
Some beef producers use automated systems to manage their risks, Ms Calhoun acknowledges, but she says markets have become more volatile as automated trading grows. “We don’t demonise and we are not as apprehensive about automated traders as we used to be. But we are not saying that algorithms as a whole are not an issue.”
By adding to volumes, automated systems should make markets more liquid. Traders, though, say these volumes are concentrated in contracts with imminent delivery dates, such as natural gas for December 2017 rather than December 2020. “Long-dated” futures remain illiquid, traders say.
Some also complain that algorithms withdraw standing offers to buy or sell when another trader tries to transact with them, creating a mirage of market depth. “They’ve become a bit of a nuisance,” says Ernest Scalamandre of AC Investment Management. “They add volume but they don’t add liquidity.”
Misgivings aside, there is no nostalgia for the days when trading consisted of people shouting and waving hands on old exchange floors. There, information about the identity and size of orders was apparent to the privileged few who held exchange seats, often at the expense of others calling in orders by telephone.
“Electronic trading, from the point of view of a fund manager like me, is a hell of a lot better than pit trading because [automated trading systems] don’t screw you,” says John Bondurant, a fund manager in Memphis. “It used to be if I wanted to do 200 or 300 lots [contracts] of cotton it would take me all day. Now I can do it sometimes in 30 seconds. The liquidity is much deeper.”
He says it seems unfair that traders using fast algorithms can jump in front of his orders, obtaining a better price. But the larger size of markets has made it easier to trade.
FIA, the futures industry trade group, says automated trading has many benefits. Greg Wood, FIA senior vice-president of global industry operations, says: “It’s important to note that automation isn’t driving prices — fundamentals are still doing so, and automation makes price discovery more efficient.”