With the growing appeal of algorithmic-trading tools to buy-side customers, brokerage houses are building up their algorithmic-trading teams. Often, that means raiding the competition. Over the past two years, a number of leading investment banking firms, including JPMorgan, Citigroup, Merrill Lynch and Nomura Securities, have been hiring quantitative analysts, developers and high-frequency traders with experience in statistical arbitrage or proprietary trading. It's all part of the next wave of automation sweeping trading desks, which is linked closely to profitable areas such as program trading and direct-market access.

But, because they are afraid that their new teams could be poached, none of these firms would discuss their new hires for this article.

Newer market entrants are focusing on wooing talent from Credit Suisse First Boston, Goldman Sachs, Morgan Stanley and Investment Technology Group, which are considered the market leaders in algorithmic trading, according to The Tabb Group's 2004 survey of buy-side traders. The relative newcomers - the Deutsche Banks, Citigroups, Dresdners and Barclays - have made a huge push to invest in experienced people in an effort to catch up.

"All of those houses are interested in the concept of algorithms so they don't lose revenue, and they will invest in [human] capital in building these businesses up," says David Korn, managing partner at Options Group, an executive recruiting firm based in London. The push has been to invest not only in people who can build the algorithms, Korn says, but also in the quantitative people who can talk about algorithms to both internal and external clients.

One of the more aggressive team builders has been JPMorgan Securities, which started ramping up its algorithmic trading capabilities in the middle of 2004 - at least two years behind the trailblazers. "Right off the bat, we knew we had to deliver product and get in flow," says Carl Carrie, vice president of JPMorgan Securities, who heads product development for algorithmic trading as part of the firm's electronic execution services (EES) initiative.

Carrie - a former fixed-income derivatives trader who founded The Beast.com, an Internet-based derivatives analytics firm, during the booming '90s - joined JPMorgan in June 2002 to run technology for a proprietary trading group in derivatives and then front-office applications for the entire equity trading floor. In June 2003, Carrie was tapped to build out a product group for the equity division's newly formed EES initiative.

"When we started hiring, we wanted to make sure this group worked well together as a team," says Carrie. "In the middle of the floor are program traders - they had to work well with the technology team, especially the auto-trading team, and the group needed to be self-sufficient and produce product," he says.

Recruiting the Competition

It wasn't until July 2004 that people started walking in the door, notes Carrie. But within six months, JPMorgan had hired seven people from other leading firms - a mix of statistical arbitrage traders, quantitative analysts, algorithmic developers and dedicated technologists.

Targeting its competitors, JPMorgan zeroed in on Citibank, from which it hired Robert Kissell, an expert in transaction-cost analysis and algorithmic trading. Previously, Kissell was a consultant to Citibank; prior to that, he worked at Instinet. Carrie declines to name other firms from which the new hires came from because he worries it could be construed as "rubbing it in."

But, given the scarcity of talent, competing firms obviously will compete for qualified people. "You can write all day about the shortage of qualified people," says Options Group's Korn. Because there is a shortage of quants in the market that have the right qualifications, candidates with PhDs who are just entering the job market will be of interest to firms, but many companies will look to their Wall Street competitors, says Korn. "It is inevitable that they will cherry-pick from these firm leaders," he asserts.

Nomura Securities built its algorithmic-trading technology from scratch about 20 months ago, hiring people with experience in first-generation algorithms, which date back to 1994. "What differentiates our team is the marriage of technology and quantitative skills," says John Comerford, managing director in charge of quantitative trading research at Nomura Securities. "We just didn't pull one or two people from the corporate IT department," adds Comerford, who joined Nomura from Schwab Capital Markets and previously worked at hedge funds.

Citigroup pulled people from three areas: the academic community; market practitioners who usually work in statistical arbitrage - at either hedge funds or in proprietary trading divisions at brokerage houses; and technologists, relates Will Geyer, managing director of the investment bank's Alternative Execution Group. Still, Citigroup prefers practitioners who have worked on trading floors. "They have to have the ability to apply theory into the real world," says Geyer. Though he declines to discuss specific individuals or headcount, he says, "We have the people now in place."

Merrill Lynch made a big splash in October 2004 with the hiring of Rohit D'Souza - then head of Morgan Stanley's North American equity businesses. In his new role, D'Souza is head of global equity trading, global markets and investment banking at Merrill. D'Souza's hiring has been called a coup for Merrill because he was the architect of Morgan Stanley's electronic-trading infrastructure. According to an industry source, D'Souza was responsible for building Morgan Stanley's equity trading lab, or ETL, a quantitative, technology-driven platform.

But Merrill is not new to algorithmic trading. The global broker began investing in low-touch, engine-based trading three years ago, when, in March 2002, it hired Mike Stewart from Schwab Capital Markets to focus on program trading and algorithmic trading. "We definitely traditionally had somewhat of an under emphasis on program and low-touch trading, and by definition we closed that gap with talent and product," says Stewart.

Over the course of 2003 and 2004, Merrill dedicated resources to portfolio trading and cash (equity) trading specifically to build out a set of execution algorithms (known as MLX-ACT) for its customers and internal sales traders, continues Stewart. "There's no question there was a dedicated recruitment and that we focused on making sure that we had an equivalent distributed set of products to our client base," he says. Stewart adds that Merrill has distributed that quantitative skill set and focus on shared infrastructure across the globe.

Without discussing specific hires or plans to expand or tweak the existing team, Stewart says Merrill has hired people who had direct and relevant trading and quantitative experience. He notes that they came "from frankly everywhere over the course of three years."

While some firms went outside their own firewalls to hire staff for algorithmic trading, Instinet recognized that it had someone talented on the inside. In December, the institutional agency-only broker promoted Steve Brain to head of algorithmic trading in New York. Brain transferred from Instinet's London office, where he ran development for Newport, its global portfolio trading system. Brain, whose title is senior vice president and global head of systematic trading, brought people with him from the London team. His role is to build up sales, support, infrastructure and trading rules as well as marketing for what is to be a separate offering from Instinet's other electronic-trading services.

Industry observers expect the hunt for talent to continue as sell-side firms continue to push to improve their algorithms in response to changing market conditions, regulation and client requests. In the meantime, JPMorgan is reaping the fruits of its hiring spree, rolling out pre-trade analysis tools and focusing on developing new algorithms in its lab. "When we started we had a shortfall of algorithms," the firm's Carrie says. "At this point, we have essentially caught up to the top-tier players."

---

Niche Players Hire Brokerage Talent

A number of algorithmic-trading start-ups have emerged over the past year, fueling the hiring frenzy that's pulling talent away from leading brokerage houses.

In January, EdgeTrade, an institutional broker that offers direct-market access technology, announced that it had hired Frank Brown, a former professor of theoretical physics at Columbia University who has 10 years of experience in quantitative and algorithmic trading. Brown most recently ran the quantitative sub-group at Bear Stearns, where he was also a member of the portfolio-trading team. Prior to that, he worked on the portfolio-trading desk for Credit Suisse First Boston's Advanced Execution Services and at Estep Trading Partners, a quantitative hedge fund that he cofounded.

"The pedigree that he brings to the table was key to us," says Joseph Wald, chief executive officer at EdgeTrade. "We wanted to have a leader in the space."

Last summer, Frederick Graboyes bought TradeTrek Securities, a Newark, N.J.-based institutional broker-dealer, with the explicit purpose of starting Algorithm Trading Solutions as a pure play on algorithmic trading. Graboyes led a quantitative equity group and the program trading desk at Bank of New York's program trading desk. Previously, he was a portfolio manager and head trader at INTECH, a money-management division of Prudential Financial, where he built algorithms. "I've been doing this for 10 years," says Graboyes, who notes that the firm's director of research is a scientist who previously worked at Morgan Stanley.

Miletus Trading was formed about a year ago to focus on algorithmic trading. Two of its cofounders - Richard Johnson and Mike Capelli - left Investment Technology Group, while the rest have worked for a number of well-known firms, including Reuters, Instinet and Cantor Fitzgerald. "We're constantly hiring," says Capelli. "We're always looking for the right people." - Ivy Schmerken

---

Citigroup Immerses Teams In International Markets

Rather than design its algorithms centrally in New York and ship the intellectual property to overseas offices, Citigroup has taken a global approach to building its algorithmic trading teams. The teams developing its smart servers are located in three regions: The North American team is based out of New York and Boston; the team in London covers the European markets; and the team in Australia covers Asia. "The team is global - and it is local, because we think that the high-frequency research that is needed in the coming generations of these servers is best understood by being immersed in that market center," explains Will Geyer, managing director, Alternative Execution Group, Citigroup.

For example, to understand the particular nuances of the European market, the European team builds the engines and tailors the factors that underlie those engines based on their local knowledge. "We don't attempt for every market to go off the exact same model," says Geyer. "We think that's a differentiator."