Will Liquidnet Catch the Buy Side?

CEOs of nascent alternative trading systems (ATSs) know one thing above all others-the brief period of time immediately following launch is a dangerous one, filled with fear that an inauspicious start could have far-reaching consequences difficult to remedy. Seth Merrin, CEO of still-teething Liquidnet, which "opened its doors" on April 10, knows that as well as anyone.

"When you're starting out you have to have some traders entering orders, you don't have any time in this business to build up a critical mass of liquidity to keep traders interested in coming back," he says. "Traders will not spend the time to keep on entering orders into a system where they're not getting executions back out of it."

Merrin learned that lesson, along with the rest of the industry, watching Optimark launch its ill-fated trading network for buy-side institutions early in 1999, only to see the highly touted system fizzle and die (Optimark has altered its business model and is now helping Nasdaq build its new SuperMontage trading environment).

Industry analysts and insiders mused over the evaporation of Optimark's liquidity pool and came up with a few pointers for the next brave soul looking to create an oasis in cyberspace. The cardinal caveat, which Merrin seems to have heeded in devising Liquidnet's design, springs from the realization that buy-side traders, much like the rest of us, dislike anything that takes them out of their normal routine-let alone something that brazenly asks them to do what some have likened to composing algorithms.

So the Liquidnet plan would be different, a system which automatically gets first crack at every order on the asset managers blotter without asking the trader to make one mouse click-an industrial strength peer-to-peer network boasting advanced encryption, guaranteed message delivery and secure socket layer protection that maintains communication between the Windows-based Liquidnet applications at all participating buy-side institutions, notifying traders when a possible match to their order exists.

If that should happen, a trader would then decide if he or she wants to make the order active and begin negotiations for the price and amount of the security in question, with Liquidnet taking a two cents per share cut (ITG's well-established crossing network, Posit, charges the same). Sounds simple.

In fact, at the open, things haven't been too bad for the upstart ATS. As of the middle of May, Merrin says, the average Liquidnet trade was about 75,000 shares with 4.5 million shares moving through the system daily-not a bad start, he says, though 20 million shares a day is his goal. Also, nine out of ten executions were being done within the spread and 32 percent of those were traded at the midpoint.

Merrin, however, says more liquidity is what Liquidnet needs to really take off. By its very nature, the fact that all orders automatically go into the system, adding more firms virtually guarantees additional liquidity. At launch, Liquidnet had 38 firms on board and by the middle of May that number reach 43 with 30 more committed to incorporation and another 10 agreements close to bearing signatures.

Leo Smith, head of equity trading at Putnam Investments, a global money management firm with $345 billion in assets under management, who also sits on Liquidnet's advisory board, says one of the best features of the system is that its functionality is not contingent upon traders going out of their way. Additionally, the system seeks to solve the greatest issue traders face when trying to move large blocks of stock. "I think that at the end of the day the thing that's most appealing about Liquidnet is the fact that you can really and truly maintain your anonymity in the marketplace," he says.

And therein lies the attraction, the point, Liquidnet's raison d'etre-to provide a means for buy-side institutions to move large blocks of stock without moving the market, something brokers risk doing when they tip their hand while looking for a trade counterparty. Liquidnet logic says that cutting out the sell side and its agent, the broker, should help eliminate that risk of market impact-in Liquidnet, it's a buy-side affair.

But, according to Richard Rosenblatt, president of Richard Rosenblatt and Co., an equity-trading firm established in 1979, that line of thought assumes brokers bring nothing but bad manners to the table.

Liquidnet, he says, alters the fundamental way trading has been done for decades. "Once the two institutions find each other, if they're going to trade, they start negotiating price by themselves, whereas in the normal course of trading, the same institution would hire a broker to do that negotiating for them, so the presumption is that the broker has no special expertise in that area and is not value-added," he says.

Sang Lee, an analyst covering institutional e-trading with Celent Communications, a consulting and research company which helps financial institutions create new business strategies, says that sentiment reflects a larger suspicion in the marketplace that brokers may, at times, be a supernumerary link in the trading chain. "You will see increasing pressure on the broker-dealers. They are getting pressure from every side, really in every asset type, everyone is trying to figure out if they are really needed."

In Liquidnet's system, additionally, it's not only the broker who finds he or she has become superfluous but, in a sense, the trader as well. Rosenblatt says that one of a traders' main functions is to optimize market timing by knowing when and through which avenue to expose their orders, again a task automated by Liquidnet. "Usually the trader for the institution, who is also value-added, makes the decision on which stocks he wishes to expose to the marketplace," he says.

Merrin says that is still possible-traders still have the ability to shield certain orders from exposure to Liquidnet's eye. "By default every order comes into Liquidnet, now the members can filter out certain orders they don't want to come in," says Merrin.

That facet of Liquidnet's design has become a major point of contention, creating conflict between Liquidnet and one of the industry's leading trade order management (TOM) systems. Think of it this way: Liquidnet's success depends on liquidity that can only be attained by getting as many firms on the network as possible. In order to get a firm on board, an interface to its TOM needs to be built which is an involved and costly endeavor. When approaching a firm with Liquidnet's application in-hand, making that sale is much easier if it has already been integrated with the firm's TOM, leaving little integration pain to the firm itself.

For the most part, TOM vendors have facilitated their clients integration process by constructing interfaces to the system, working with Liquidnet independent of their clients' involvement. Some firms, however, like leading TOM vendors The LongView Group (over 60 clients) and Charles River Development (also over 60 clients) have not built that interface, leaving the onus of doing so on their clients, at least for the time being. Until those two firms are brought into the Liquidnet fold, Merrin and his team stare out at a vast liquidity pool which sits relatively untapped.

"They force their clients basically to write the interface themselves which they've done and in our ideal world we would be working closely with the vendors, which we're doing with MacGregor, Advent, Eze Castle and Decalog," says Merrin, "seems to me, perhaps, that those firms are more responsive to their clients needs for whatever the reason is."

Merrin says, though, that Liquidnet can still get clients of those firms onto the system rather easily as LongView and Charles River clients who have built a Liquidnet interface are allowing others to use it. The logic for such generosity being that for any client attending the Liquidnet party, the more people, the better.

"There's no cost in signing on to Liquidnet, no service fee, no minimum commitment," says Merrin. "All they have to do is sign a two-page agreement and we'll come in and do the install." He adds that installation process only takes about a half-an-hour.

A spokesperson for Charles River Development says that at the time the company was approached by Liquidnet, the commitment that an interface would require was not supported by a high degree of interest among Charles River clients. "We're deal driven here so we really needed to see demand before we went and invested the development resources," he says. Over the past few months, says the spokesperson, client interest in the network has grown, leaving Charles River ready and willing to help its clients tap in.

Annie Mariani, vice president global client services with The LongView Group, echoes the sentiment that client interest must drive any resource-consuming projects. She says that when LongView was approached by Liquidnet, her company also had more pressing initiatives on its to-do list.

"We have some clients that are interested in using it and want to commit to using it, there is a very small handful of those and a couple of those clients are using it today," she says. "There are another few clients who say we will never use Liquidnet, they don't believe in the way they do things and then I would say the majority of our clients are either not that interested or they are on the fence and waiting to see if it takes off."

Mariani says the reason interfacing with Liquidnet must be carefully weighed against the cost is because integration with the ATS is no mean feat. Problems arise because Liquidnet is not Financial Information Exchange (FIX) compliant, using a proprietary stored-procedure API with a daemon that pulls the database. A daemon is a UNIX program that executes in the background ready to perform an operation when required-usually an unattended process that is initiated at startup.

She says that when first meeting with Liquidnet last summer, she learned that the ATS needed a brand new custom interface which would require two things: writing a stored procedure to pull out the data from the blotters and then, for the return of trade-execution information back into the TOM, using Structured Query Language (SQL)-a language used to interrogate and process data in a relational database.

Although Mariani says LongView never reached the point of detailed analysis that would have put a dollar figure on building the interface, she says the company is prepared to do whatever its clients deem most important. In the event that Liquidnet should reach such a favored status, she says LongView has recently asked for Liquidnet's final specifications so developers can have a better idea of how to tackle the problem if an interface gets the green light.

If LongView does decide to go ahead with the interface, however, Mariani says the company will do so in a manner corresponding with its clients wishes. She describes the requirement that all TOM orders flow into the Liquidnet system as a "functional discrepancy" because almost all LongView clients find that to be a major negative.

"Liquidnet is requesting the entire book or the entire blotter from the traders and all of the traders have made the statement to us that, 'We don't want to show our entire blotter, we want to have control over our blotter, and we want to be able to designate what Liquidnet sees,'" she says.

While Mariani says the two companies are not at an impasse and hopes to work through the point of contention, she says, "We'd much rather develop an interface that meets our traders desires rather than just build a standard interface to the specification from an ATS like Liquidnet."

If the time should come when traders working with LongView gain access to Liquidnet, they will still find themselves confronted with the same situation Ryen Munro, head trader with Essex Investment Management, an investment advisory firm founded in 1976 and specializing in growth equity and fixed income, faces when he sits down at his trading desk-where to look for liquidity.

While Liquidnet's paradigm does not require an active search on the system, the decision must still be made, when a possible order match is found, whether or not to do the deal over the network. "We have great relationships on the street (sell-side firms) and it pays off for us. Do I want to jeopardize that by doing too much business away from them?" asks Munro, who also sits on the Liquidnet advisory board. "So it's hard, it's difficult."

A senior industry player, who insists that his name not be used, says punishment from slighted sell-side firms could take on a number of forms. "They might be less willing to provide capital to make bids and offers in less liquid stocks. I don't think there's any question that firms that commit capital for the benefit of their clients commit more capital and at tighter prices for their better clients, so that is one area that certainly could be affected," says the source.

He also mentions the fact that sell-side firms, which sometimes shop orders around to their best customers might be less inclined to dial the numbers of an institution that no longer regards the involvement of brokers as beneficial.

Munro says that although he is on the Liquidnet advisory board and although relationships on the Street may be affected, he is always committed to getting the best prices for his clients and will thus go wherever he needs to go.

Rosenblatt questions whether a trading environment like Liquidnet, which he likens to an OTC model and calls "trading in a closet," is one where best price is likely to be found. By design, the model sees a bid and offer matched, followed by negotiation between the two parties during which the details of the trade are finalized. He says that atmosphere precludes other firms from injecting better bids and offers which often results in an improved price for the customer.

Merrin stands by his model as an easy way to move large blocks of stock without moving the market. He says feedback from participating members has been positive with most just looking for the pipelined firms to come on-line ASAP, delivering an instant infusion of liquidity. Specifically, he says, new customer-requested features such as instant chat, which could function as a means for clients to entice each other into activating dormant orders, will be included in the next major version release which is due out near the end of the third quarter.

Merrin says he will consider the venture a success if liquidity on the network reaches and maintains his goal of 20 million shares a day but, as LongView's Mariani says, since liquidity for Liquidnet is all about participation, the wait and see attitude of many firms can mean simply watching a slow decline. "For a company like Liquidnet that's tough because it's the wait and see people that they need the liquidity from," she says.

Liquidnet's best hope for success, however, may lie in a sentiment expressed by Essex Investment's Munro, "I need to have resources such as Liquidnet available, I need access to it and I think it's good for my clients to have the potential to have orders or trades in there. I wouldn't want to be left out."

---

Some of Liquidnet Participants Include:

-AIM Management Group

-Essex Investment Management

-Putnam Investment Management, Inc.

-SSB Citi Asset Management Group

-Scudder Kemper Investments

-T. Rowe Price

-Pimco Equity Advisors

-Northern Trust Global Advisors

-American Century Investment Management