While foreign exchange moved to an electronic broking system in 1992 and equities became gradually more automated – in Europe at least – the credit market in the main sat back and watched.
Much of the reason is to do with the structure of the market. Three major factors differentiate the bond markets from other products. First, there is no central exchange, for cash products at least. As a result there has been neither a central market-place for price discovery nor outdated structures nor malleable regulations to exploit in the way the ECNs have in the equities market.
Second, they are the most complex of the financial products, not homogenous as equities and foreign exchange largely are. The complexity of subproducts in the credit markets – plain vanilla, varying maturities, zero coupon, inflation linked, credit derivatives, to name but a few – has increased in the past decade, as has the sophistication of the major players.
Third, investment banks chose to integrate these changes, adapting their existing structures to do so. This has helped to keep the credit markets more opaque, so allowing investment banks to charge for their services of interpretation.
But that is now beginning to change. CSFB, for example, regards the credit market as a special case to the extent that it has created a separate e-commerce division for it with its own P&L account, run by Ben Cohen.
Why now? “The trading of the major benchmarks has become so commoditized that it is no longer possible to cover the fully loaded cost of sales via the bid-offer spread,” says Michael Paull, managing director, European fixed-income markets at Instinet. “It can be a half-million-dollar decision to add a guy to the floor, taking into account salary, cost of floor space, IT charge-off costs, in-house charge to support research, and so on. To hit your RoE targets you need to generate at least three times this in revenues. If the salesman is concentrating his focus on commoditized bond product, he is never going to realize the full value of his seat on the floor. This trading needs to go fully electronic, in order to thus free up traders’ and salespeoples’ time so they can concentrate on the value-added transactions.”
Already several systems are now either up and running or in development to bring the benefits of technology to the bond markets. At least, it is a benefit to the market-place, if not to all the players.
When Tullett & Tokyo and Liberty Brokerage announced their merger in July, the two firms said a priority would be to develop an electronic brokerage system. The earlier merger of Garban and Intercapital is promising the same. They could well be too late.
Cantor Fitzgerald, the largest of the inter-dealer brokers, launched Cantor e-speed in July. EuroMTS has been trading benchmark European government bonds from three countries since April, based on the bond dealing system introduced for domestic Italian bonds in 1988, and by the end of September will have extended the system to cover four more countries’ benchmark debt. And in June seven banks, all involved in EuroMTS, announced the formation of Brokertec, which aims to become a global electronic inter-dealer broking network for a variety of fixed-income products, beginning with US and European government bonds.
In the broader world of broker end-investor dealing, several systems are already in operation. One that is to be launched later this year is Instinet’s fixed-income brokerage. It has the benefit of being able to tap into an existing network, parent Reuters’ high performance network, and it is playing that as one of its strengths. “A lot may jump in with new systems,” says managing director Rustam Lam. “But who is able to reassure a highly critical market-place with high standards that they can deliver and maintain and support it 24 hours a day, 365 days a year?”
Once it’s accepted, says Paull, it will take off rapidly. “The markets are ruthless and aggressive pursuers of liquidity wherever they can find it. It’s platforms like ours that will add the high-octane fuel to this. Markets will always migrate to technology for this reason.”
Of those already in operation, two are run as a consortium, Tradeweb and LimiTrade. Tradeweb began in September 1997 and deals exclusively in US treasury products. Although web-based, it needs additional software installed on the dealer’s computer to make it usable. LimiTrade was formed in May last year and trades corporate bonds. In most respects, Tradeweb simply automates how the market operated beforehand; LimiTrade allows for direct trades between investors, so calling into question what the middlemen do.
To gauge what was needed for its system, LimiTrade canvassed both the buy and sell sides and used for its stress test the model of the illiquidity crisis in US high-yield bonds after the Milken débâcle in the early 1990s. “Even today these markets lack mature mechanisms that provide centrality, reference prices and the control of market fade [the inability to trade bonds after a period of time],” reads its mission statement. “LimiTrade provides all three of these developed market supports.”
But the use of such systems has hardly been astounding to date. “The Bond Markets Association conducted a survey in 1997 detailing 11 electronic bond-trading systems in North America,” says Chris Carroll, a managing director at Deutsche Bank and the man responsible for developing its proprietary bond trading system, Autobahn. “They did the same again in 1998 and this time there were 26 systems. Only a very small number of them are viable currently, and the percentage of flow is still moderate.”
But a quick look at the futures market shows just what effect electronic trading can have on business. “It’s the most advanced of the credit markets in terms of electronic trading,” says an investment banker at one of the major houses, “except for one dinosaur in Chicago, of course.” But even the CBOT is now attempting to keep up with developments, voting yes to a tie-up with Eurex and even exploring business options with IBM.
Barriers lowered
Two other exchanges provide better examples of adaptation, though. The first is Liffe, which at the start of 1998 was still intending to build a new, larger, trading floor for open outcry. That idea was scrapped but meantime Liffe corralled its 200 or so members to get the backing for a new electronic trading system, Liffe Connect, which officials claim is more than 10 times as fast as the Eurex system. It has been built to be not just fast but also flexible and open-ended. Users of the system can use whichever front end they choose and simply plug in Liffe Connect. The system has the potential to be used to trade anything for which there is a liquid market. Not only that, but it is being delivered on time.
The other example comes from the French derivatives exchange, Matif. When it decided to extend trading in its benchmark Notionnel 10-year bond contract to an electronic market, the new NSC system, alongside open outcry, it estimated that it would take at least 18 months for liquidity to migrate. It took just six weeks and the volume of trading increased.
There are three reasons why this happens, explains Instinet’s Paull. “Electronic trading lowers a number of barriers to entry: first, it makes the market more transparent and so shows a truer depth to the market; second, the efficiencies of electronics push down the cost of each execution; and third it makes transaction executions easier for traders.”
If liquidity is defined as the amount of penalty one pays to exit a trade, then electronic broking severely reduces that penalty, so making trading more attractive.
Additionally, in the fixed-income markets, says Paull, “ticket error costs can be much higher than they are in other areas of trading. Electronics can virtually, if not entirely, eliminate “ticker errors” thus greatly increasing the overall efficiency of trading and vastly reducing one area of infrastructure cost associated with supporting a fixed-income trading operation”
Brokertec is analyzed in more detail below, as is the CreditTrade system developed by a small wholesale finance e-commerce firm in London, Mutant Technologies. First, though, a look at a web-based research tool which DLJ has been offering its clients over the last year.
