), Merrill Lynch (MER
), and Goldman Sachs (GS
) stay in the game? Mainly by helping clients execute trades using sophisticated software that they once reserved for themselves. It's a move that's turning brokerage houses into software providers and traders into technical support staff.
Emily Portney, the chief operating officer of JP Morgan Equities Americas, believes that if it's done right, the change will bring success for JP Morgan even while the number of shares crossing over the old sales desk dwindles. Portney recently spoke with BusinessWeek Online reporter Burt Helm about which technologies are changing the market, how big firms are competing, and what the next "race" in investing technology will be. Following are edited excerpts of their conversation:
Q: So how is technology changing the trading floor these days?
A: Electronic marketplaces like LiquidNet and Archipelago (AX
) have had a massive effect on the equity marketplace. We have many, many more sources of liquidity than ever before, and clients are expecting us to provide tools directly so that they don't even need to go through us.
The traditional business in the marketplace was pretty much like this: A portfolio manager would come up with a trade idea, and then would ultimately communicate that to his or her own trading desk. That trading desk on the buy side would then generally contact a sales trader, who would pick up the phone and start making calls. But now the trading desk has a ton of alternatives they never had before, and there's a lot less stock moving across the "block" sales desk.
Q: One of the tools you and the other firms are now using is called algorithmic trading. How does it work?
A: Algorithmic strategies (see BW, 4/18/05, "Cracking the Street's New Math") are basically computer-based programs that allow a client to buy or sell a stock. Just like a good sales trader wouldn't dump an entire order into a marketplace, generally speaking, the program tries to minimize market impact. It does this by breaking a large order into tiny pieces, and then feeding the shares into the marketplace automatically. It allows clients to buy or sell stock pretty much on an anonymous basis, according to specific criteria he or she selects.
We are integrating our programs with whatever front-end terminal the client is already using, putting that power right at their fingertips and letting them execute electronically.
Q: With every brokerage house developing algorithmic software, what leads a client to select one broker over another? What's the competition like in developing this technology?
A: It's a race, really. At first the race was to try to get every client connected so that they could execute electronically. That has been accomplished, for the most part. The new race is to algorithmic trading: How many algorithms can you put out there, and how you show that your engine is better than their engine, and so on.
At this point they're relatively commoditized algorithms. Each of us has about 8 to 10 of them, designed for different trading objectives and stock types. A client has about 40 or so of them at his or her disposal, so we've spent a lot of time developing pretrade analytic [software] that helps clients decide which [tool] they want to use.
Q: What's the race after that?
A: We're trying to create algorithms for lists of stocks. Instead of evaluating each stock individually, it will take into account the relationship between all the names on the list. So if you want to trade IBM (IBM
) and another tech stock, it will realize that and trade accordingly to limit market impact, limit cost, and make sure you do it as quickly and as efficiently as possible.
With all these moves, sales traders aren't just handling a call about an order. They're answering questions like, "What does this algorithm do?" or "How do I use it?"
Q: Almost like tech-support staff.
A: Exactly. Sales traders are now much more tech-savvy. We've hired a lot of individuals as well who just build these tools. The real overarching theme here is that as technology becomes ever more important, a lot of the technology we used to use internally we are starting to push out to our clients. We're not only a broker, but we're becoming a software provider.