Algo traders want to do high-frequency trading in markets they can agitate," says Darryl Hooker, Emerging Market Manager at ICAP Electronic Broking, in the course of a lengthy discussion of the impact of specifically algo trading on commodities markets (in ICAP's experience, gold, silver, platinum, palladium) and vice-versa. "Algo trading drives average ticket size down, but it encourages a far greater number of trades," says Hooker. Interestingly, Hooker adds that there's also a time-zone factor: however deep and liquid a product might look, there's much less take-up via technology in Asia and Japan than in Europe and the US.
That last observation might seem counter-intuitive, given the commodities-market significance of China and India (and more generally, Asia's history of IT innovation), but - leaving aside the variety of other ways in which Hooker's observations are thought-provoking - they serve here to illustrate a wider point. Commodities markets may be opening up to electronic trading, as Robert Brady has already argued, but any illusion of uniformity between different commodity markets and different traders is just that: an illusion. Indeed, while the arguments for trading the mainstream commodities might have a kind of family resemblance to the arguments for trading any other asset class, there are other commodities where the attraction is obscurity, or rarity, or nobody else has thought of it, or the fact that they are traded out of caves in wine-growing regions of France.
David Nahmanovici argues (page 20) for drilling down to the level of individual commodities; perhaps there's also a case for calling a halt to the drilling once you know the trading venues and the main participants in that individual commodity. Find out that information, and you might already have a potentially significant determinant of short-term future price behaviour.
Why? For the (perhaps obvious) reason that a screen-traded market behaves differently from a market dominated by, let's just say for old times' sake, youngsters in colourful jackets shouting at each other, and for the (perhaps less obvious) reason that, for the adventurous investor, there are individual commodity markets at various stages of the transition from jacket-wearers to screens. We may be past the big, headline-grabbing transitions, but they don't account for everything. One suggestion is that those electronic traders looking to expand their trading range into "new" commodities might usefully look out for anomalies thrown up not by the characteristics of the commodity itself, but by the expanding population of the market for that commodity. Commodities are, after all, an asset class in which the supply and innovation of new instruments (wind, wine, fish, et cetera) are matched only by traders' appetite for something new to trade.
And prevailing trading practices can be part of the case to trade. The "algo effect" on a commodity's price behaviour will almost certainly prove significant in one common situation, and almost certainly not apply at all in another. The first is the individual-commodity market in a state of transition, where a critical mass of participants, or the market itself, is "going electronic" or has just gone (the algos now have access), or where newcomers are bringing in their own trading practices. Such a market will often be characterised by the "age" of the commodity itself; people have been trading this stuff since it was brought in on camels from Samarkand. There will tend to be a cash market here, and more importantly, the range of contracts/opportunities offered will be limited by the functionality shortfall. But there will be tension between the old and the new.
The second situation, where an "algo effect" will almost certainly not throw out anomalies, is where a "leapfrog commodity" (the term "mushroom commodity" went quickly out of use for obvious reasons) has come into being in, effectively, contract form only. These are markets without a history, which have "leapfrogged" any evolutionary (legacy) development process in trading practice. You can trade various forms of pollution, for example, and yes, you could argue that there is something like delivery, but nobody's going to be delivering the physical product to your office. You hope.
So what happens to a commodity market when the algos start trading, and what could you do about it? Citing a conspicuous example, Shaun Downey chief technical analyst at CQG says: "When the CBOT went electronic, that made a huge difference. There are a lot of people trading the relationships in the bean complex: soya beans, the spreads; but then also people doing bean oil against soya meal, corn against soya, wheat against corn, all those relationships. Because they're now trading electronically, there's a lot more take-up in that area. You only have to look at the increase in the number of bid:asks. That's the sort of thing that tells you computers are getting involved." Notice that take-up and notice that creativity. Obviously, new liquidity arrives with the algos, but so does a change in the pattern of activity. As Hooker says above, ticket sizes tend to drop, but trading activity increases.
And that has an obvious knock-on effect. Downey says: "When computers get involved, you get an increase in short-term volatility. They all jump in on top of each other, then come back very quickly." But this is short term. The window for that volatility lasts until the market "matures" as an electronic market. Over time, thereafter, pricing becomes more consistent, and we begin to see spread compression. There's less gapping in the market. You've got yourself an established commodity market. Carry on trading, but don't miss the change.