March 3, 2000

Role as Big Nasdaq Market Maker Helps Knight/Trimark's Portfolio

By GREG IP
Staff Reporter of THE WALL STREET JOURNAL

When egghead.com Inc. released good news early one morning in December, before the Nasdaq Stock Market opened, many online investors thought its stock would open higher. Kenneth Pasternak knew it would.

Mr. Pasternak sat before a screen at Knight/Trimark Group Inc., a market maker whose job it is to execute trades. His screen showed that orders to buy Egghead exceeded orders to sell by 100,000 shares. Because it would be Knight's job to fill those orders, Mr. Pasternak quickly went to work, buying up 50,000 shares in informal trading before the market opened. When it did open -- with Egghead sharply higher -- he sold them to online buyers, nailing a quick $15,000 profit.

Most online investors know little about Knight, but Knight knows a lot about them. In five years since its founding, Knight has ridden the online-trading explosion to become the country's largest market maker, executing a huge 21% of Nasdaq dealers' trading volume (and a third of the smallest issues), as well as 7% of volume in shares listed on the New York Stock Exchange.

Seeing all those orders gives Knight what Mr. Pasternak, its chief executive, calls an "informational advantage": exclusive intelligence on which it can trade for its own profit. "We're smarter than the market in aggregate and we're able therefore to make a determination whether the stock will go up or down," he says.

So even as Knight executes the trades of thousands of amateur day-traders, the firm is, in effect, a day-trader itself. Most of its nearly 400 traders are paid solely on the basis of profits they earn for the firm. As a day-trader, Knight is surely one of the most successful: It hasn't had a single losing day this year.

Moreover, unlike most of the online brokerage firms it serves, Knight is very profitable. Its net income more than doubled to $168 million last year, and its stock is up sevenfold in the 20 months it has been public.

All this success is a magnet for both admiration and criticism. "They've built a great firm in terms of automating processes and serving people," says Bill Burnham of venture-capital firm Softbank Capital Partners. But, he complains, the people at Knight are "taking information about retail customers' intentions to trade and using that information to improve their own proprietary trading profit, at the expense of their customers and of other participants in the market." Similar criticism is leveled at other market makers who fill orders from discount brokers' customers, such as Schwab Capital Markets, a unit of Charles Schwab Corp.

Knight is hardly alone in trying to profit from the prices at which it trades with its customers. But Knight's end-customers aren't big institutions but mostly small investors, often not aware of the mechanics of order execution. Nor do they have any choice in the matter, because many major online brokers send all their orders in particular stocks to Knight or another such firm for execution, in return for payment.

Brokers defend this practice by praising Knight's service. The firm has "very robust liquidity, their service and support have always been good, their speed of execution is right up there," says John Chapel, head of U.S. brokerage operations at broker TD Waterhouse Group Inc., which sends about half of its Nasdaq orders to Knight.

As for Knight, it deserves no criticism, Mr. Pasternak says. He argues that Knight, by promising to buy when investors want to sell and to sell when investors want to buy, is giving them free and valuable access to its capital, plus instant execution of most orders at the best price posted in the country.

Mr. Pasternak wasn't among the Wall Street chiefs debating the stock market's structure before the Senate Banking Committee earlier this week. But he had a lot to do with why they were there. His firm, more than any other, has thrived on the fragmentation of stock trading that most of the chiefs bemoaned. It has done so by, in effect, becoming the biggest fragment.

It happened almost by accident. Mr. Pasternak studied to be a teacher but taught just one semester before quitting to join a market-maker firm. In 1995, he and a colleague, Walter Raquet, set up their own market maker, which evolved into Knight. They made more than a dozen discount brokers co-owners. The pitch was twofold: If the brokers sent the firm their orders for execution, they would benefit both indirectly by their ownership, and directly by the payments the firm made for their "order flow."

The subsequent explosion in online trading would have been enough to make Knight a success. But something else crucial happened. Nasdaq market makers historically profited from the spread between the bid price at which they bought stocks from customers and the offer price at which they sold. But in 1997, a federal probe of dealers' practices resulted in new rules that squeezed these spreads. The squeeze was made tighter still when stocks began trading in sixteenths instead of eighths.

With profitability collapsing, many dealers stopped making markets and sent their Nasdaq orders to "wholesalers" such as Knight or Schwab Capital Markets, which make markets in thousands of stocks for other brokers. Such firms, Mr. Pasternak told investors in the 1998 prospectus to Knight/Trimark's initial public offering, would no longer profit primarily on spreads. Rather, they would "take advantage of the profit opportunities represented by each trade."

As Mr. Burnham puts it, "Nasdaq has been transformed from a market where people make money off spreads to one where people make money off information. When the market was so much more fragmented, it was hard to be right. But when you have 30% of the order flow, you can make some damn good guesses."

Mr. Pasternak concurs. Computers automatically fill the vast majority of orders Knight is charged with executing, leaving most of its 393 traders free to try to take advantage of the information these orders reveal about the market.

The trader ethic begins with Mr. Pasternak, who mixes the jargon of financial theory with the expletives of a trader. Seated in a vast trading room in a Jersey City, N.J., tower overlooking lower Manhattan, the 46-year-old CEO explains that Knight profits by combining many bits of information about market trends with calculated risks.

Oracles and Eggheads

One morning last summer, Oracle was trading at $34. What would happen if it fell a point? Mr. Pasternak opened his file of limit orders, those that can be executed only if the stock hits a price the customer specifies. There were 13,000 shares' worth of buy orders between $33 and $34, but orders to sell 2 1/2 times that many shares between $34 and $35. Knowing of this selling pressure, Mr. Pasternak would hesitate to buy Oracle, and he might even sell it short, betting on a decline.

Or consider that Egghead morning, late in December, when the company put out the news that it ranked in the top 10 e-commerce sites. Faced with a 100,000-share imbalance of buy orders over sell orders, all of which Knight had to fill, Mr. Pasternak bought half that many Egghead shares in unofficial pre-opening trading, which takes place mostly among brokers and other institutions.

That would take care of half of the buy orders, but now he still was obligated to sell 50,000 more shares to online investors. He didn't own them. So he decided to go short-selling the investors 50,000 Egghead shares that Knight had borrowed, to be replaced later after a hoped-for fall in the price.

The move looked smart as the stock weakened slightly just after the opening. "I was informationally advantaged," Mr. Pasternak says.

But then it turned dicey. A second wave of buyers sent Egghead shares climbing. Knight at one point had a paper loss of $250,000. But the stock slid sharply by the end of the day. Knight made a profit of $100,000.

But what about small investors? Softbank's Mr. Burnham says that while buyers got Egghead stock at the opening price, as promised, perhaps that opening price would have been lower if not for Mr. Pasternak's heavy pre-opening buying. "Kenny wouldn't have bought those 50,000 shares if he didn't know they wanted to trade at the open," Mr. Burnham says. "He used their own information against them."

Crossing the Market

Knight promises to execute, at the day's opening price, the first 250,000 shares' worth of buy orders sent to it before the opening bell. But other traders say that before the opening, wholesalers -- and Knight in particular -- regularly use aggressive trading tactics to push a stock up or down to favor the positions they will have to take when they execute the orders. All the wholesaler "cares about is getting the stock up to a level where he can fill all his orders profitably," says Matthew Johnson, head of Nasdaq trading at Lehman Brothers. But "where it opens is not necessarily in the best interest of their customers."

In normal markets, the highest bid (to buy) is just below the lowest offer (to sell). Yet it's not uncommon, traders say, for Knight to bid more for a stock than the lowest offer to sell it, and to offer to sell a stock for less than the highest bid to buy it -- an anomalous situation known as "locking" or "crossing" a market. This anomaly leaves the best-priced order unfilled. But it forces the market in the direction the firm wants it go. Nasdaq restricts crossing during the day but permits it before the opening.

"It's not unusual to see the large wholesale firms leading the pack on some of these locked and crossed markets on most openings, and clearly Knight is the name that's pre-eminent," says Patrick Ryan, president of Ryan, Lee & Co., a small brokerage firm in Washington, D.C. Still, he says the problem results more from the behavior of Knight's end-customers than from Knight itself. If Knight is "sitting there with unsolicited orders from a group of gamblers -- who figure 'P.T. Barnum was right, if I pay $90, someone will pay $92' -- clearly it's buyer beware."

Circling theglobe

Consider the day theglobe.com went public, in November 1998. The new issue was priced at $9 a share. Small investors swamped dealers with orders to buy at the start of trading. Shortly before the opening, Nasdaq records show, underwriter Bear Stearns & Co. was offering to sell shares at $70. Yet shortly afterward, Knight bid $75 for them. Then Schwab Capital Markets bid $80. Bear lifted the price at which it offered to sell shares several times, finally to $90, but Knight and Schwab again bid even more than the offer price.

The shares opened at $90, and within minutes, Knight executed purchase orders by selling more than 450,000 shares at $90, Nasdaq records show. The stock got as high as $97 that morning, but closed the day at $63.50. Many investors were shocked by how much they ended up paying.

Schwab Capital Markets President Lon Gorman says there was an "irrationality" in the market that day, and he has since led an investor-education campaign "to make sure that never happens again." Speaking more generally of the criticism of wholesalers, he says: "The notion that there's something going on in the back room, that you get an execution that's inferior, is totally bogus."

As for Knight, Mr. Pasternak says it loses money almost every morning because of its guarantee to fill orders at the opening price. That crazy morning, he notes, the buy orders that market makers had to fill at the opening bell exceeded all the shares in the IPO. Mr. Pasternak says he will occasionally place a bid higher than the offer "if I don't like how the market is pricing."

Knight promises online brokerage firms that when it gets an order, it will automatically execute it at the best price anywhere, even if it's not Knight's quote. To execute buy orders, for example, the firm buys shares and keeps them briefly in inventory, risking a price decline before it gets rid of them.

But Knight takes steps to limit its risk. For example, it chooses whom to trade with. Mr. Pasternak welcomes the "uninformed" orders of thousands of individual investors, because he is confident that, on average, Knight will be smarter than them. And just as a casino bars gamblers who consistently beat the house, Knight's systems watch for investors who consistently make money trading against the firm. For such a customer, Knight may restrict or suspend the promise to automatically execute all trades at the best price posted anywhere.

Knight also occasionally suspends this promise during "fast markets." Suppose a mention on CNBC triggers a surge of buying or selling in a stock; Knight can suspend automatic execution after it has accumulated a long or short position of, say, 25,000 shares. Then it switches to manual execution and fills orders only against another customer or another dealer -- a slow process during which the stock may move a lot.

Knight tells online brokers when it has restricted automatic execution, but the brokers typically don't notify investors. In a volatile market, an investor may not get his order executed for several minutes.

Christopher Vu of Houston entered a market-price order for 10,000 shares of Books-A-Million with broker Brown & Co. one day in November 1998, with the stock at just over $29. Brown sent the order to Knight, but Knight didn't fill it -- that is, sell Mr. Vu the shares -- until six minutes later, some at $37 and the rest at $38.

With the stock just below $40, Mr. Vu, who thought he had bought the 10,000 shares but wasn't sure, sent an order to sell them. Five minutes after receiving that order, Knight executed it -- at $33.50. Mr. Vu, who had expected a $9,500 profit, lost $4,500. "I was totally taken aback," he says, adding that he wasn't aware Brown didn't execute orders itself.

While his order was awaiting execution, trades took place elsewhere at better prices, notes Mr. Vu's lawyer, Philip Aidikoff, in an arbitration against Brown. He argues that a broker that doesn't go to other venues -- having agreed to send orders to a single market maker -- may be breaching a fiduciary duty.

Brown says Mr. Vu has no case, because Brown didn't do anything to slow the execution, and Knight's service wasn't in any way defective. A spokeswoman for Brown's parent company, Chase Manhatten Corp., adds that Knight is "consistently one of the best in terms of their execution speed." Mr. Pasternak says that he isn't familiar with this case, but that such incidents typically arise when automatic execution is suspended; he says he would like to find a way to notify investors when it is.

Regulators and some Wall Street firms are increasingly wondering if dedicated order-flow arrangements are hurting the quality of the markets. Such deals enable a firm like Knight to "cordon off" a portion of the market that only it can see and trade with, says Ed Nicoll, CEO of Datek Online Holdings Corp. That means Knight's customers don't benefit from competitive bidding by other investors that might improve their price or force Knight to improve its price, he says.

Mr. Nicoll's firm is a competitor to Knight through its Island ECN. Electronic communications networks, unlike market makers, merely display customer orders, against which other customers can trade directly. They don't pay brokers for dedicated order flow, and they don't make trading bets.

Mr. Pasternak says ECNs don't provide what individual investors want and what Knight provides: certainty of execution, at the best price anyone is posting, and the rock-bottom commissions that payment for order flow helps make possible. "Put a button on everyone's computer," he suggests, "an ECN choice and market-maker choice -- and let the customers choose."

Write to Greg Ip at gregory.ip@wsj.com