By Anthony Currie
Last year it was equities. Last winter it was bonds. Now this summer foreign exchange, by far the largest market, finally embraces e-commerce. Single-dealer platforms will still be needed, but might only account for 25% of the volume. The seven-bank consortium behind FXall.com has grabbed all the headlines this month, but it is already a year or more behind two independent ventures, and years behind State Street’s FX Connect, which since April has allowed other dealers on to the system. Why has it taken so long for forex bankers to accept the multi-bank approach, and what are the consequences of leaving it this late?
Anthony Currie investigates
Josh Levy never intended to be part of an uprising. He left his job as a currency trader at Goldman Sachs early last year to join Valhalla Forex, based just a few minutes’ walk from his former employer, on the 15th floor of a building on downtown Broadway in New York. The grandeur of its Wagnerian name belies its size: it is a relatively small organization, specializing in proprietary trading in the forex spot markets, although it also dabbles in forwards.
Within weeks he had become exasperated with his dealers and market-makers. “We had good price discovery through our brokers so we knew within two pips where the price really was,” says Levy. “But the bid-ask quotes we were getting from the dealers were often quite far off the fair-market value. If the price was really 90-93 sometimes, especially in fast markets, we’d get prices which were four or five pips off the market, sometimes more.”
Having worked at a broker-dealer, he ought, one might think, to have been better prepared. For a start, Levy does not include Goldman in the group, which so frustrated him. “I’ve been on the other side, but I also didn’t think things should be as bad as they are. We saw widespread acute and chronic lack of service. It was eating into our P&L.”
Only equities investors tend to get much publicity with such complaints, although the execution costs of trading bonds is also coming under greater scrutiny. Somehow the foreign exchange market tends to be forgotten, which is surprising given that it is by far the largest of the three. On average $1.5 trillion is traded in foreign currency each day, which some estimates put at seven times the combined size of the debt and equity markets. Quoted prices can change 20 times in the space of a minute, the most active rates up to 18,000 times a day.
And yet the telephone and human intervention are the primary tools for doing business. Despite the introduction of systems such as Reuters and the electronic broker system (EBS) into the inter-dealer market in the early 1990s, the broker-client relationship saw virtually no technological advances and had to rely on the inefficient method of telephone trading in a fragmented, unregulated, rapidly-changing market until very recently.
Furthermore, despite having the word “exchange” in its title, the foreign exchange market bears none of the hallmarks of an exchange-trading environment. It is the least regulated and policed market, even by over-the-counter market standards, and lacks a coherent fee structure, which leaves customers at the whim of their liquidity providers. “Making money off the customer is the name of the game,” says Levy. “Reading and anticipating customer strategies before they deal, front running orders, using customer limit orders as free options for the market-maker, and running stop-loss orders prematurely for profit, all are commonplace to a dealer/market-maker in forex trading. The SEC and CFTC have put exchange-traded futures brokers and equity broker-dealers in jail for less.”
In their more candid moments the bankers will admit this, although only off the record. “We’d be more than happy to have formal fee structure put in place,” says a banker who works in e-forex at one of the leading banks. “We do have a fee structure of sorts, and that is to lie to our clients,” he continues with a wry smile.
And the foreign exchange desks can be big money spinners for the banks: in the mid 1990s, for example, Citibank was making 30% or more of its profits from that division.
Little wonder, then, that the banks have hardly been quick off the mark to extend the transparency and greater efficiency afforded by EBS to their clients. Not even JP Morgan, one of the more active of banks in the e-commerce arena, has forged ahead in the forex market. LabMorgan has been the incubator of several of the more innovative ideas, such as the multi-bank bond platform MarketAxess, but not in forex, even though the second-in-command (and the man who has run it since last year, before Nick Rohatyn was put in charge) is Thorkild Juncker, previously the JP Morgan’s head of foreign exchange.
The banks will provide reasons for forex lagging behind in adopting e-commerce solutions. One of the more cogent arguments is put forward by Paul Thrush, head of foreign exchange at Barclays Capital. “Banks have followed a huge range of different strategies with foreign exchange. Some regard it as an extension of research, some as a more efficient clearing and settlement process that sharpens price discovery. There’s never been a common view on how to use the product.”
Most banks had in any case been concentrating on developing their own platforms, whether closed-system or web-based. Again, though, this is on the whole more beneficial to the banks, allowing them to cut costs by transferring on-line what they have done over the phone for years. Few try to address fee structures or market manipulation.
Another reason is that the overall service structure of foreign exchange is so different to other markets. “Fixed income has been driven heavily by competition over research, says Lars Olesen, e-commerce manager for Citigroup's global foreign exchange group. “The idea of co-mingling the research and the pricing on a multi-bank site fits well in that market where pricing is more transparent. It’s just not been a theme in forex.”
Then the arguments get more flimsy – it was a Y2K issue, say some, despite the fact that several initiatives in debt and equity were announced last year, and a slew of debt initiatives early this year. Or they claim that clients don’t want it. Internet delivery only ranks eighth in order of priority for customers in Euromoneys forex survey last May. True, but clients consistently demand good pricing, execution and processing from their brokers, something which the internet and other electronically-provided systems can, if done well, improve upon. What is true, of course, is that for many investors and traders the internet is still not fast enough, and security worries still persist.
Gary Craft, the e-finance analyst for Deutsche Banc Alex Brown in California, puts it another way. He produces a piece of research called E-finance Weekly, which offers an overview of the major e-finance developments. In his May 12th edition he looked at foreign exchange, calling the section ‘Mr Money center, do you really want to give up those valuable FX economics?’
The news that week was that a major money center bank had struck an agreement with Intellisys to offer foreign exchange content. Craft’s response: “Where is the competition for FX trade execution? Spreads in the FX market are incredibly rich and they will not fall without a true marketplace that facilitates FX trade execution competition. Content itself is wonderful...but the electronic commerce community needs a marketplace where the price can he improved.”
That, he says, is unlikely to come from the incumbents. “Money center banks are not likely to build FX marketplaces – they are likely to he forced into them. Look for innovation in FX trading, transacting and settling to come from outside the dealer cartel. No dealer today wants to price improve and arguably must be forced into any market efficiency developments.”
It’s the pricing argument again: why overturn that which has brought you such good fortune in the past? The answer is that if you don’t someone else will, and as the equities market in the US has proved so effectively, that someone is more than likely to be a new entrant with much less or nothing to lose.
And, as in the equities market, the banks have brought it on themselves. Stuck with the old phone system, which brought in such good revenues while also knowing that technology was bound to change the model, the vast majority spent their efforts on single-dealer platforms. This was by no means an illogical thing to do, first because that’s how business was traditionally done, each bank competing against all the others. But also, says Peter Gerhard, managing director, foreign exchange at Goldman Sachs, “we all need to have a proprietary platform which will form the basis for the multi-bank offerings. The multi-bank site is just the common portal.
But the banks’ over-concentration on the single-dealer platforms for so long could have set them back against the insurgents. These multi-bank sites are not simply about showing several dealers’ prices at once to a client. Just as important, and a huge way of saving time and money and avoiding confusion is to merge as much of the middle and back office as possible. “Clients want to have ubiquitous, inexpensive access to the markets they work in,” says Lars Loesen, e-commerce manager for Citigroup’s global foreign exchange group. “And what they have been telling all of us for the last two years is that they don’t want to have 15 different systems and have to in-put 15 different log-in passwords.”
As a result, much of the investment in the single-dealer platforms has, thus far, gone without reward. “In general, clients have been reluctant to sign on to many single-dealer platforms because they want to avoid the hassle. They want multi-bank sites, and seem prepared to wait for them and continue using the phones for now.”
Longer-term of course it is not a waste. As Gerhard points out, each bank needs its own system to route the customer to and from the common platform, and there will always be customers who want or need to stay on the proprietary platform along. “I estimate that perhaps 75% of electronic volume will go through multi—bank sites, and 25% directly to the individual bank’s site,” says Gerhard. “That’s probably the midpoint of what my peers at the other firms think.”
by ignoring the multi-bank offering until very recently the large dealers may
have lost valuable time to start-ups, and to the only bank-run multi-bank
platform in operation,
Valhalla Forex became one of the start-ups. “We were frustrated, but rather than banging the phones on the desk and breaking headsets out of frustration we decided to do something about it,” says Levy.
In late spring the firm took the decision to change its business completely, moving out of proprietary trading as its core concern and into an online trading venue. The result was MatchbookFX.com, an electronic system for institutional and qualified individual investors to trade foreign exchange. It’s not a revolutionary model, at least not by broader market standards, but Levy, who is an executive vice-president at the new firm, states that it is a big departure from the norm for foreign currency trading. "Broker-dealers generally don’t charge commission for forex trades, so the only way to make money is to trade against your clients. What we’ve done is to shift the balance of power back to the end users, making them price makers rather than price takers.”
To do so MatchbookFX has taken a leaf out of the US equities market. One of the main backers, and software providers, is NexTrade, one of the second-tier electronic commission networks for trading in US equities. And as in the ECN world, and as the name of the company suggests, the business is based on matching buy and sell side orders. Price discovery is provided by feeding in data from the major providers, be it Reuters, Globex, EBS or others. Buyers and sellers then agree on a price with no other middleman in MatchbookFX’s online marketplace.
The time from frustration to pro-action to going live was very short, in part because MatchbookFX adapted NexTrade’s existing technology rather than building a new platform. The first deal went through in September last year, and now the company handles roughly $100 million in trades each day. “They can be as little as $200,000 in size,” says Levy. “But the standard is $1 million, and we have had trades executed of $5 million in one clip, and more”.
So it’s no behemoth, and on those terms not an immediate threat to the incumbent dealers – the big institutions only do such small tickets as a favor. But the forex bankers at the majority of these firms are more than aware that even a small operation such as this is a potential threat; they’ve had three years to watch and learn from the huge shifts in tile market structure of the equities market, especially in the US.
Nor is it the only model out there, which has developed independently of the broker-dealers. On the other side of America in California, Lori Mirek is running another forex-trading insurgent, Currenex, whose trading product is called FXtrades.
Her path to challenging the established forex hierarchy is even stranger than Levy’s. She had never worked in foreign exchange before – her last job was as head of America Online’s business-to-business e-commerce division, following a career which has taken her through some of the big tech companies in the US, such as Ameritech, Sun Microsystems and Oracle. “Part of my job at AOL, was to seek out enormous markets with huge inefficiencies and therefore huge opportunities for a B2B exchange,” says Mirek. “The forex markets fit the hill perfectly.”
And the vehicle chosen was a small company which for four years had been examining forex processes and cash flows, and had a small trading operation and a small customer base. Currenex was formed as a holding company last year after buying the company in the fourth quarter (strangely, Currenex won’t release the name of the company), and started to build an infrastructure and develop a more comprehensive web-based service. Staffing the firm are a number of forex traders from some of the major buy and sell side firms, such as Citigroup, JP Morgan, CSFB and Barclays Gobal Investors.
Mirek officially took over as president and CEO at the end of April, at the same time as the firm announced strategic investment from firms including TH Lee Putnam Internet Partners, DLJ’s venture-capital arm and WR Hambrecht and Company, which has been developing a web-based Dutch auction model for bringing IPOs to the market.
In the course of her work there, Mirek has come across conflicts similar to Levy’s. She has, she says, seen forex basis-point mark-up sheets for different currencies at different times of the day at one firm. In another incident she was making a pitch to a major forex user, and one of the people at the meeting was a woman who sat there smiling throughout as Mirek described how her platform ought to help cut costs by giving the user more power.
“After a while I started wondering whether I had a piece of spinach in my tooth or something,” says Mirek, recalling the ever-present smile on the other side of tile table. But at the end of her pitch, the corporate’s CEO turned to the woman, who as it turned out happened to be the former senior FX trader for a bank, and asked her opinion. “She told him that everything we had said was true,” says Mirek. “That her job as a dealer was to maximize the bank’s profits at the expense of the corporates and funds and that she had been hired by the corporate to prevent that happening there.”
Currenex’s business model is similar to MatchbookFX’s in just one respect, and that is that it is not particularly radical. It matches the neutral third-party role in trades you would associate with equities or derivatives exchanges with a multi-bank platform such as seen in the bond markets with either EuroMTS or TradeWeb.
The big difference between Currenex and TradeWeb – the main multi-bank electronic platform success story in any product thus far, now accounting for roughly 30% of all trades in US government bonds – is that the banks do not own a stake in it, nor do they sit on the hoard.
This, says Mirek, solves two big problems. “First it gives tile end users peace of mind that a group of banks wont he in a position to use the data to trade against them. We are the dragon guarding the castle containing the data. Second, a lot of smaller banks don’t want the big banks controlling the development of and access to this market.”
So the banks - there are at least 20 signed up now although no names are being released – act merely as liquidity providers in a reverse auction process. They are alerted to provide a price when a client on the system wants to make a trade, and have 25 seconds to do so, depending on the complexity of the transaction. The client then has five seconds to make a decision on which price, if any, to accept. “The added transparency this brings to the client can bring considerable cost savings,” claims Mirek. “One client saved anywhere between 3.4 basis points and 6.2hp on major currencies over six months depending on the size of the trade. Another monitored the pricing relative to Reuters, and FXtrades beat it 80% of the time.”
One of Currenex's most obvious advantages is that it is already up and running. There is only one other multi-batik site in operation, and that’s FX Connect, formerly the proprietary system developed by Boston-based State Street. But the two, at least as yet, are hardly competitors. Currenex’s clients are mainly corporates – Intel and Mastercard are fully signed-up users, the latter, according to Mirek, transacting 97% of its volume over FXtrades. Stare Street’s system, on the other hand, focuses exclusively on the buy-side institutional market, and only opened its doors to other dealers in April, when Deutsche Bank, and then Société Génénale and one other bank that prefers to remain nameless, were invited to come on board.
Mark Snyder, State Street’s head of foreign exchange, says that they are also in discussion with 15 other financial institutions about using the system, and that FX Connect was always envisioned as a multi-bank. “This was always the plan, and in 1998 and 1999 there were clients who signed up mainly because they knew we were going to open it up to more counterparties.” But the real decision-makers were the clients. “Our mutual clients wanted more banks on it, and some even said that they might not do business with the banks if they didn’t jump on board,” continues Snyder. “One of the major lessons of the B2B world is that the buyers win.”
For State Street it has been very successful -witness their rise up the Euromoney rankings this year to 44th place up from 44th position in 1999. “We had 90 clients trading on it last year, and now over 250 buy-side firms are signed top,” says Snyder. “We have roughly40% of the top 500 institutional investors using the system.” At present, FX Connect is only available through a private network, “It is internet ready, but clients still prefer a high-speed, reliable closed system.”
The real competition to Currenex ought in theory at least to be provided as a consequence of an announcement at the start of June. Seven of the larger forex dealers are banding together to form Fxall.com, namely Bank of America, Credit Suisse First Boston, Goldman Sachs, HSBC, JP Morgan, Morgan Stanley Dean Witter and UBS Warburg. “We’d all been having bilateral conversations for a year or more on the impact of e-commerce on our industry,” says Paul Kimball, co-head of foreign exchange at Morgan Stanley Dean Witter and chairman of Fxall.com. “Our clients had all been telling us the same thing: that they wanted to have a multi-bank electronic platform. It’s a response to client complaints and desires. Finally, about six weeks before the announcement, we decided to get together as a group to discuss options.”
That would put the start of the group to discussions at about the time that State Street had opened its platform up to other banks and that Currenex was launching. It does sound as if it is a defensive move. The head of foreign exchange at one of the member banks admits that to a degree the banks’ hands had been forced. “There was a general reluctance to avoid taking this kind of step for as long as possible, in part because we were all concentrating on developing our own internal web sites. But that allowed some non-bank platforms to develop in our absence.” And there are those who still question the need for it, although a declining number. “Clients have been demanding this for some time, and in some cases have been reluctant to sign on to single-dealer platforms to avoid the hassle of having to do 15 different installations of straight through processing,” says Philip Vasan, head of e-commerce for CSFB. “Dealers meanwhile thought the market was so vast that there was plenty of room for stand-alone systems. That view is changing rapidly.”
It’s not just defensive, however. While all in the group are loath either to grant or to claim the honours for starting the discussions, it’s more than coincidence that Vasan was only appointed head of e-commerce towards the end of April (the public confirmation was 25th April); before that he had been head of foreign exchange. Several of the initial meetings took place at CSFB’s headquarters on New York’s Madison Avenue.
The inspiration for Fxall.com came in part, says Gerhard at Goldman Sachs, from the success of TradeWeb. “It really grew out of discussions we were having at board meetings at TradeWeb. Obviously that’s a very US-centric model which wouldn’t suit the foreign exchange market. We needed a more geographically diverse consortium.” In part, he says, that need also played a role in determining the size of the group. “For a product as global as foreign exchange we had to have a geographically broad array of banks to he able to represent the major markets effectively.”
Given that tile foreign exchange activities of the seven are so much broader and diverse than the US Treasuries strategies of those in ‘TradeWeb, how did they all manage to agree on what to do? “We tried to focus on where we had common ground,” says CSFB’s Vasan, “So we’re concentrating on bringing down the cost of dealing, of straight through processing and increasing transparency for clients in spot markets, forwards, options and research.” Which seems to cover a fair amount, rather than concentrating on a specific area. “We’re avoiding the more contentious areas, such as retail, which don’t affect us uniformly,” says Vasan. “We’re staying purely institutionally focused.”
For the next three months or so the group will work under what Kimball calls a swat team, under himself and interim CEO Philip Weisberg, from LabMorgan. Then longer-term management structures will be put into place. Those outside the group wonder whether seven hoard members might he too unwieldy for effective governance, although Gerhard counters that “it’s no surprise that there’s an odd number in the group. For most decisions we’ll need a simple majority, and reserve the 5-2 majority for those critical issues where a supermajority becomes necessary. We all expect this to succeed, we are investing a lot of resources in it, and have demonstrated over the last few months that we can work together.”
According to Kimball, seven was a good mean figure. “Having just two or three was too small, 12 or more too unwieldy. We think this is just right to begin with.” But from the start all participants are stressing that Fxall.com is so named because it is open to all. So what happens if they are called on to prove that if and when others join? We’re thinking about non-voting equity stakes as one option that seems to have worked elsewhere, and possibly introduce a revolving board seat for new members,” says Kimball. “But we do want as many as possible to participate in using the system. That’s the great irony of the new economy, that the m ore who work together, the greater the rewards.”
Other than that, not much else is being revealed, more a sign that not much else has been decided. “We have no rigid set of rules, and intend to be very open-minded going forward,” says Gerhard. “Our client advisory board will be very influential in the implementation and subsequent development of this site.”
It’s no reason for Currenex to quake in its boots just yet. As with most start-ups in the e-finance space, all that’s been announced is that the entity exists, that there are seven banks on the board, and that they are open to all players to join in some fashion or other.
But this is the honeymoon period, when the polygamists still revel in the knowledge that they all share the same basic idea. They have yet to announce a technology provider, and decisions have to be made by seven banks still busy dealing with telephone broking and perfecting their proprietary systems. Going live won’t be an issue until this time next year, if the example of Brokertec is anything to go by. This inter-dealer electronic broker was set up last June to offer cash and futures trading in government bonds, run initially by seven banks, increased to 12, and still not out yet. “It’s a wonder they ever managed to get together for the meetings there’s so many of them,” says a senior forex banker not in the syndicate. Kimball isn’t shrinking from the challenge. “Working together with competitors will be as hard a task as it was to get the technology to a point where we could consider doing this. It needs resolve, focus and single-mindedness from all of us.”
The battleground preparations for the multi-bank forex platform are not yet complete, however. The announcement of Fxall.com is as important for who has been left out as much for who has been included. Conspicuous by their absence are the three largest forex banks, Deutsche Bank, Chase Manhattan, and Citigroup, all of whom were invited to the discussions.
Kimball sounds genuinely disappointed, but realistic, that none of the big three joined up. “We have to start somewhere. Clients don’t want us to wait, and in internet time each week we wait is a month, each month a quarter. And what we have is a diverse, broad group with a lot of liquidity behind them.”
It’s not that they are against multi-bank sites. All three, to varying degrees, are involved already. Deutsche is a founding member of Brokertec, has signed up for State Street’s system and is even going on joint calls to clients. One of the first acts Simon Lack took after becoming head of e-commerce for the investment-banking businesses of Chase was to start negotiations with Bear Stearns and instigator JP Morgan to set up MarketAxes, a joint fixed-income platform. Citi has been less open, if we excluded Salomon Smith Barney’s strategy, although Citi is one of the founder members of Volbroker.com, an on-line inter-dealer currency options trading system announced in March. And, says Citi’s e-commerce manager for global for global foreign exchange Lars Olesen, “we decided back in February that we needed to do something in this space. Citi is prepared for the night opportunity and will participate fully in the multi-bank market as it develops.”
As for Chase, says Lack, “we didn’t think that it quite met our needs. We believe that smaller is better from a governance point of view, say three or four participants, at least to begin with. And we want to be part of something more substantive rather than just announcing intentions. Attention spans can only be measured in weeks, so you can’t afford to have a slow start.”
Joe Norena, head of e-forex at Deutsche Bank, echoes the view: “We want to take a leading role in developing this market, whether in dealer-client offerings or the inter-dealer space such as Volbroker.com. We want to be involved in initiatives that have concrete plans and milestones to head for, not just good intentions.”
Which means we can expect at least one other multi-bank offering involving the top three. That has to be more of a worry for Mirek at Currenex. “We see these announcements as a direct competitive response to us, so it totally validates our approach: she says. “At the end of last year there was hardly a noise out of the e-forex corner, then come January they all rushed out their single-dealer systems, version 1.01. These multi-bank offerings won’t be live for at least a year, and customers get better pricing and operational efficiencies today with Currenex.”
It looks as if there will be at least two medium-term consequences of this e-forex arms race. First, that as with equities exchanges before, the path to greater efficiency must first pass through a village called fragmentation. Single-dealer platforms will still be with us, and there may be a s many as six multi-dealer plat forms serving clients, whether independent or bank-run.
Clients may demand choice, but having to log on to half-a-dozen different sites hardly solves their stated desires over the last two years to reduce the number of systems and processes they have to deal with. But it appears that this course is inevitable. “I guess we have to accept the process of other markets,” says Kimball. “Where they first went through fragmentation and then a mixture of consolidation and withering.”
Second, it will create a battle to win liquidity, and this could be a very bloody, drawn-out affair (drawn out in internet time, of course). And for the first time the banks will not have as much control. Certainly forcex services are often linked to other. And it would make sense for banks to promote the use of their own investments. “Clearly if we’re going to have to cannibalize our own business we’d prefer to own a piece of whatever platform is going to carry it out, says an e-forex banker at one of the big dealers. “We’ve said no to a number of offerings, including Currenex, as a result of not being able to take a stake.” Mirek seems to be fixed on remaining third-party neutral, although a clever membership or ownership structure which avoids the messy conflicts on some of the existing US equities and derivatives exchanges might be found, especially if it’s a matter of continuing the business or shutting up shop.
But Currenex has a year or more to develop a customer base on a system which empowers the customer more than the banks have done in the past, and system which, says Mirek, a majority of the FXall.com investors have signed up to participate in. How banks will deal with watching a platform they invested in – and it will need significant investment, Brokertec has spent over $120 million on technology alone – fail because clients have the power to force them to trade on one they have no stake in, will be one of the defining stories of the e-finance revolution.
But if the battle goes on too long, it might be irrelevant. “Who knows, maybe five years from now deals could be getting done which ate so hug that no bank will be willing to accent the risk-reward,” says State Street’s Snyder. “Institution might find a way of pooling their own liquidity.”