Most metal brokers point to the years 1999-2001 as the dates when electronic trading of London Metal Exchange contracts began.
But Gavin Gross, who was then head of LME metals at voice-broker Spectron, has a more precise date in mind: June 16, 2000.
Gross should know, because he had spent the preceding six months signing up members of the London Metal Exchange to trade with each other across his company’s inter-dealer screen.
Click here if you missed Monday's introduction to our electronic trading series.
“We turned the system on June 15, by which time we had ten LME firms as members, and we know this only goes if there’s a trade. And the first trade was for five tonnes of nickel. People were curious, but the screen wasn’t full of a large amount of numbers," Gross told Metal Bulletin editor Alex Harrison.
“So day two, June 16, starts and a couple of traders looked at the screen and thought, ‘You know what, I’m a market-maker,’ and started to put in prices: good, tight markets for every metal. And then bing, bing, bing, buy that, sell that, boom, trade, ding, ding, ding."
“We were off and running. By the end of the year we had 30 firms as members.”
Gross was no stranger to metal markets. Before joining Spectron in January 2000, he had worked as an options trader for Billiton (which was then a ring-dealing member of the LME) and Koch Metals, having arrived in London in 1993 from New York, where he had been a physical metal trader for Philipp Brothers.
So when he started to sign LME member firms up to his new company’s platform, “I had an argument prepared. I knew enough about the market to know where the sensitivities were."
“The ring-dealers were getting commission from the category IIs. They were handling a lot of bank business, and they saw this thing as a threat. We were going to take those commissions therefore some of those meetings were fiery,” he said.
So Spectron laid out four main arguments to those it sought to bring on to its platform.
Four reasons to go electronic
First, other markets were going electronic for all kinds of good reasons and till Spectron there was no inter-dealer electronic platform in metals. Spectron’s platform was a good one, easy to use, and based on the platforms it leased from Trayport and had honed in the energy markets.
Second, inter-dealer phone markets, for all their adaptability and capacity to handle complex wholesale metal business, were also inefficient in some ways, particularly where price discovery was concerned.
In Gross’s words, “If someone is bidding at $9 on one phone and another person is offering at $6 on another line, we can bring all that business together at $7.5, everybody is better off, and we’re bringing transparency.”
Third: anonymity. “If you phone someone and say ‘I’m a buyer, I’m a seller’, they now know where you’re coming from. They know what you’re trying to do and everybody has their own positions trading as principals, whereas being on screen gave you anonymity,” Gross said.
Fourth, the commission to buy or sell on Spectron was below what category IIs were at that time paying to the ring-dealers. “The going commission rate for category IIs was 25 cents per tonne, and we were proposing to halve that to 12.5 cents per tonne. The cat IIs were thrilled: ‘Sign me up now.’"
“So we were bringing together transparency, liquidity, anonymity and a lower commission rate. And the prices on Spectron were real."
“At that time everybody had a Reuters screen where dealers would put up indicative prices, Refco four bid at eight on copper, say, then AMT is five bid at nine."
“But they were wide spreads and only indicative. They weren’t real markets," Gross said.
“I remember as a trader somebody would put up eight bid, and you’d call, and they’d say, ‘Oops, no, sorry, just sold there.’"
“Spectron was real markets: if Spectron said it was eight bid, it was eight bid.”
Though some of the ring-dealing members saw it as a threat to their lucrative trade with the banks, Gross had identified areas where the Spectron screen might potentially help them as well.
“A lot of the dealers were willing to quote [prices to] banks and make them markets. Banks had huge positions and sometimes they’d go to the ring-dealers, ask them for quotes and then hit them, leaving the ring-dealers with a lot of risk, all for a 25-cent commission.
“They’d sometimes get bashed up by these banks, or they’d take a lot of risk with their client business.
“And it wasn’t reciprocal, they couldn’t go to the banks and ask them to make a market. It was all one way.
“On the Spectron screen, though, all the bids and offers were open to them, whether for risk management or client business, giving them an opportunity to lay off positions on an equal footing.”
By November 9, 2000, over 65,000 lots of LME business had been transacted on Spectron, and Spectron counted ten LME ring-dealers among its 26 members.
Spectron’s trading inter-dealer platform was gaining traction, in other words.
“The ring-dealers that were upset and the firms that hadn’t signed up on day one were hearing about Spectron, ‘Look at all this Ali that’s traded, all this copper that’s traded.' There was a $1 bid-ask spread for 20 lots each side and these guys were shut out," Gross said.
“And as a trading company you can’t be shut out, so eventually, and it could have been through clenched teeth, people said, ‘Sign me up. Give me the screen.’”
For the better part of a year then, Spectron had the field to itself, as the LME worked on its own electronic platform, Select, which it launched in February 2001.
“Eventually we knew the LME would have its own system and members would pay for it and surely the LME would want that system to carry its own volume and not have the embarrassment of this outside broker doing the trade,” Gross said, a recognition that encouraged Spectron to cut its own commission by 2.5 cents per tonne and instead adopt a model under which the aggressors on individual trades paid that sum to the initiator of the trade.
Some who were active in the market at the time said the LME was developing Select as a tool which it would deploy as a last resort, against Spectron or another rival platform Emetra.
A source at one ring-dealing firm who was close to the debate at the time, but declined to be named in this article, said: “People were pretty miffed that the LME was going to allow a platform under which clients would have some form of direct access to the exchange, though that would not happen till 2003. Select, to my mind, was first developed to give the LME a defensive posture against the other trading platforms.”
To the ring-dealers, the traditional members of the exchange, the LME was evolving rapidly and in ways that looked as though it might damage not only them but also the marketplace itself.
“The concern at the time was that everything was changing. With the demutualisation of the LME and other trading floors closing, a new breed of members developing from the world of finance and the traditional industry-based ownership of seats diminishing, and the category IIs eating the ring-dealers’ lunches, people were understandably fearful that there was a move to close the ring," the unnamed source said.
“And more than that, they were also fearful that the executive and new members wanted the LME to be the same as all the other markets: global, standardised, financial,” he explained.
Simon Heale, who was the ceo of the exchange from mid-2001, does not recall it in the same terms, however.
“The LME, of which I was and am a huge fan, was at the time quite a conservative institution and, if you look back, our attempts to innovate and introduce new products had not been enormously successful. Partly that was a reflection of the ownership structure and how it operated,” he told Metal Bulletin recently.
The LME, for example, launched silver futures in 1999 and was to scrap them in 2002, and in 2000 launched its index, tracking six base metals, which was not a conspicuous success either.
“There was no doubt whatsoever that it would have been madness not to offer an electronic platform, though,” Heale said.
Glen Chalkley, who still works at the LME today as head of electronic business development, argues that the move towards electronic trading was inevitable, driven by a desire to cut costs to trade in the market by the brokers’ clients.
“Some members were concerned about how the market would evolve. But what’s clear is that it did evolve: members changed their business models to allow more electronic trading or direct order-routing for their clients," he said.
“And their clients meanwhile were looking for straight-through processing. They wanted to execute a trade without human intervention to cut their costs and reduce the possibility of errors,” he told Metal Bulletin.
No overnight success
But, as Heale acknowledges, LME Select (unlike Spectron) took time to pick up pace.
“During some of those early times when we were only trading tens of lots, there’s no doubt I was thinking, ‘Oh, this is hard work, I hope it’s going to come true.’"
“But then, bit by bit, volumes started to rise with different players coming in, trading a bit here and a bit there, and at that point I thought, ‘It’s going to take time, but this is going to work. We’re going to make it happen but we just have to be patient.’"
“Being effectively a not-for-profit institution, we were able to take a long-term view of it. It allowed us a degree of patience and time which was very helpful,” he said.
Others identify the specific factors that drew business on to LME Select.
“The move to Cinnober in December 2003, which was a more user-friendly system, definitely helped improve usage and liquidity,” Chalkley said.
The exchange also created a system, akin to that which Spectron had used, under which those that hit orders paid, and those that received the orders got paid.
Enter the D-Team
While it is clear that many LME members had grave reservations about the arrival of electronic trading, category II firm Dresdner Bank was moving quickly to position itself in the new electronic marketplace.
The team there included John Browning, Steve Bingley, Fabian Somerville-Cotton and Robert Rees, all of whom are still active in the market, and was headed by Alex Wilkinson.
In particular, Dresdner had experience in developing and offering electronic platforms to its clients in forex markets, through a tool called Piranha, and was in general a seat of technological innovation under Ramji Al-Noor, the head of IT.
They identified a chance to replicate the success they had had with Piranha in the metal markets, and are widely acknowledged to have been at the forefront of the moves to push the LME into the electronic age.
The team there saw the market as one in which clients were unable to participate directly in the creation of prices on the LME since while members, and particularly ring dealers, made markets to each other and to clients there was no compulsion to show clients’ bids or offers to the market.
“Access to Spectron was only available to category I and category II LME members and its prices were unavailable to the data vendors such as Reuters, so for clients, Spectron remained a broker-to-broker black box into which they could not see,” Browning told Metal Bulletin.
The LME launched Select in February 2001, but it was not till the second half of 2003 that Dresdner began enabling its clients to trade directly on the LME through a platform called Gator, short for global access to order-routing.
Once Dresdner began routing clients directly into LME Select, volumes and liquidity on the screen rose again, Browning recalled.
Inevitable as this may seem to have been in hindsight it was not an easy step, as those brokers that opposed allowing clients access to prices on LME Select fought a strong rearguard action.
Having been unable to prevent an electronic platform being developed by the LME, the pressure increased when it became clear that Dresdner was going to allow clients to plug in directly.
Part of the challenge to the broker that wanted to plug its clients into LME Select came in the form of questions about whether or not LME rules allowed it to make such a move, those active at the time said.
Hamish Purdey was operations manager of Ffastfill from September 2000-October 2002, a company where he later became ceo. In that role, he faced challenging questions about the service that Dresdner and Ffastill would be offering.
“When Ffastfill first provided the software to Dresdner that enabled them to route their clients’ orders into the system, it was quite controversial. We had other LME members call us up and say, ‘Why is it you can provide these prices electronically?’” he told Metal Bulletin.
“I was personally accused of destroying the market with my iceberg order. In markets with such thin liquidity you didn’t want to show your whole book, just as you wouldn’t in the ring. So I provided an order type, an iceberg, where if you wanted to sell 100 lots you’d only show two lots at a time, for example, with the volumes and the time between orders both randomised.”
And if Purdey felt the pressure, working for the software provider, which eventually had metal clients all over the world as a result of the new users that entered the market through electronic trading, Browning did so too.
“There was a great deal of anger, all of it, it seemed at the time, directed at me. Remember we had seen the closure of the LIFFE and IPE would close soon. People expected, including myself, that the LME floor would close within a year,” he said.
Dresdner was well-placed to deal with the protests, questions and complaints though, because of the experience of its staff, and the positions they had within the LME’s own member structure.
As Wilkinson recalled, “We built the system, and there was some question from others about whether or not we would be breaching any rules by deploying the front-end screens to people.
"But John Browning’s access to the LME and position on the board enabled us to assure ourselves that we weren’t.”
Browning himself, who joined the LME board in 2002 and was chair of the newly created e-commerce committee, is as unequivocal now as he was then.
In his view while brokers might have been entitled to discuss the advisability of giving clients electronic access, no LME rules disallowed it.
“Let us be clear: there was never a rule prohibiting electronic access. The market consensus was that it could not be done. There was an understanding that client orders could not be placed directly onto LME Select,” he said.
Dresdner’s team argued that in a principal-to-principal market whether the order was made on the floor, on the phone or on LME Select the bid or offer would always ultimately be Dresdner’s, on behalf of its client.
“Some felt [where LME Select was concerned] this was a semantic argument, hair-splitting in fact, but Dresdner was only replicating the floor methodology,” Browning said.
Ultimately the question of whether or not an LME member could allow clients direct electronic access to LME Select was debated at a board meeting, soon after the arrival of Donald Brydon as LME chairman in September 2003.
“The ring brokers dominated the policy of the LME, and when it became clear Dresdner had given access they expected to be able to close that tap off,” Browning said.
“On the day, the LME board took a different view, it was a very highly charged and bitterly debated moment as everyone present understood that ratifying client access to LME Select irrevocably broke the LME business model and the ring members’ grip on the LME board of the previous 125 years.”
The spark for Select
But while the improved interface once Cinnober took over LME Select, the incentives to trade electronically and clients’ moves to embrace lower-cost, straight-through processing all helped encourage the development of an electronic market, the flames were lit by users themselves, as Browning explained.
“Following some early press comment [about Dresdner’s order-routing system], I received a call from a USA-based hedge fund, which already traded the LME in a small way but were frustrated by the inter-dealer broker/ring methodology.
“They wanted to trade in larger volumes and would do so if they could attach their ‘black box’ to our infrastructure. Of course I said ‘yes’. But really at that moment I didn't know if it could be done," Browning said.
“Within about six weeks we were in testing and when we went live the volumes the fund sent through were phenomenal. The spark of liquidity was suddenly generated because there were two venues of price discovery, which now frequently ran in opposite directions."
"Let’s remember at that time in the pre-market there were lots of different prices for copper as each broker would make his own. But in the ring the copper price and the entire market place became unified and there was one price: the price in the ring," he added.
“Many of our clients withdrew their online orders during the rings. But our black-box fund client did not."
“Select now started to move independently of the rings. Let’s say the market was trading at $60 on the opening of the ring, on closing orders in the last 15 seconds the ring is offered down to trade at $55. Select on the other hand might start $60 bid, then trade at $60 then suddenly more orders would appear at $59, $58, $57, $56 and $55, and floor traders would race to buy at $55 in the ring and sell Select," Browning explained.
“Someone was mis-pricing but it was difficult to say who. My client? The floor? Who could say who was right?"
“Within days I was in the USA and signed up another black-box client. When the second black box connected, its methodology was that of a market-maker. It would bid and offer three months all day, and from that point the market place was never the same, it moved relentlessly on-line.”
While Browning focused on ensuring that the electronic interface worked for clients and could see the ways in which the market might change as a result, the Dresdner team also understood how winning metal clients through their platform could power their business.
More than brokerage
When other LME brokers questioned what Dresdner stood to gain from giving clients access, “to me it was patently obvious, the metal derivatives business was a way to collect margin and collateral. If I can put screens in front of people, not only do I get order flow but I also collect margins and collateral”, Wilkinson said.
“Sitting above LME Select was the Dresdner crossing engine. Clients thought they had direct market access but they actually got better liquidity matching with our other clients within Dresdner."
“It wasn’t about the brokerage, it was the liquidity: we wanted to net off the margins and take the collateral. Everything that we could net off we then put into other assets, T-bills and so on, and remember this is a time when US treasuries were paying 6 or 7% on an annualised basis,” he told Metal Bulletin.
Indeed, Wilkinson’s view was that putting screens in front of so many clients meant Dresdner could be “incredibly aggressive on brokerage rates because that’s not where we wanted to make money, we were interested in flow, in margins, in collateral”.
The move on to LME Select generated large profits for Dresdner, both from the brokerage and the use of gross margins in other assets, but the market in general benefited too as transparency and liquidity increased.
Volumes from clients grew, as they became increasingly empowered.
“It transferred enormous power from the brokers to the clients as it reinvented previous trading methodology," Browning said.
“At Dresdner we did not have to sell the product because senior clients, household names, were contacting Dresdner on a daily basis,” he added.
And as time went by, and increasing volumes of trade were transacted on LME Select, the dealers who had so opposed its launch and its extension on to clients’ own screens, began to reap the benefits themselves.
“As the liquidity rose on LME Select the broking community found they could use it to augment their market-making, and particular things such as the execution of large stops in volatile markets became considerably easier. Floor brokers moved to adopt the new methodology and soon enjoyed the risk-free commissionable income stream," Browning said.
“After several months the atmosphere changed, clearly I was not forgiven, but I was no longer hated.”
And just as Spectron’s Gross identified transparency for LME members as a key selling point on Spectron, the early supporters of LME Select and its usage by clients highlight a similar aspect of its importance.
“Electronic trading has produced both more liquidity and more transparency. Remember that LME Select was publishing prices outside the times that the ring was trading and that brought the market to the attention of new users, who were attracted to the increased transparency and visibility around the prices, which they wanted to see on a screen,” Chalkley said.
And as for the future?
Inevitably comparisons will be drawn between the dawn of electronic trading, and the LME’s plans today to attract more volume into the monthly dates, aiming at a comparable constituency as was drawn to the market from 2000 onwards by the prospect of electronic access.
“In some ways the situation now is similar to what it was 15 years ago, though I don’t think what we are proposing with the monthly prompts is such a fundamental change because it does not change the way the LME operates," Chalkley said.
“But we do have members already publishing monthly prices, though it is done less aggressively than in the days of Enron, and there are definitely clients that want transparent, liquid LME monthly prices, just as back then there were clients that wanted electronic prices and access,” Chalkley said.
Perhaps the last word should go to Browning, who moved to Hong Kong several years ago.
"Today the rationale is the same, but the venue is different. We have the internal Chinese Shanghai Futures Exchange price and the international LME price. Because of their regional dynamics frequently their pricing moves in opposite directions, someone is mis-pricing but its difficult to say whom. East or west, who is right? But it is in that intersection the good broker will always make money."
Most metal brokers point to the years 1999-2001 as the dates when electronic trading of London Metal Exchange contracts began.