Algorithms Making Their Way Into Stock Buyback Arena

By Anjali Cordeiro
Of DOW JONES NEWSWIRES
28 November 2005
Dow Jones News Service
(c) 2005 Dow Jones & Company, Inc.

NEW YORK (Dow Jones)--Trading-algorithm programs are slowly making a foray into the realm of corporate stock buybacks.

The idea of using software programs that take stock repurchase orders and execute them has begun to pick up steam amid a surge in buyback activity. A number of securities firms are capitalizing on this opportunity by using technology to expand their business and create efficiencies. Stock repurchases in S&P 500 companies soared to $81 billion in the second quarter of 2005 from $42 billion a year earlier, according to Standard & Poor's Corp.

Corporate buybacks can be complex due to regulatory guidelines companies need to follow to avoid legal liability. For instance, under the Securities and Exchange Commission's Rule 10b-18, companies have certain restrictions on buying back their securities at the opening and during the last half hour of trading. There are also constraints on the price and volume of the shares being bought back.

These issues can make the buyback process a tough one for a human trader to juggle, and here is where algorithms may carry an edge. The software executes stock trades electronically and can be programmed to guard against slip-ups on the regulatory front.

Tradetrek Securities, a Newark, N.J., broker dealer, says it has created an algorithm to help companies execute a stock buyback more effectively. The product is designed "to help corporate treasurers achieve their buyback with minimal market impact and best price, while ensuring there isn't a violation of SEC rules," said Managing Director George Rodriguez.

"Typically, once we know what a company wants, we do a historical analysis on that company's (stock) and marry that with real-time analysis. We put in 10b-18 requirements and blend it all together," said Rodriguez. The algorithm then decides the most appropriate time to send the order and the size of the order.

The Tradetrek algorithm doesn't require human intervention once it has been programmed and is connected electronically to stock exchanges and to electronic communication networks. "Our traders act more as pilots monitoring a plane," said Rodriguez. "There are always traders and engineers to ensure it is functioning as designed but for the most part the algorithm works on autopilot." The corporate treasurer can, through an electronic connection, see every transaction as it takes place in real time.

So, does this kind of new technology strike a death knell for traders at buyback desks at large Wall Street firms? Not quite. A few large firms have created their own buyback algorithms, but say that human traders can't be completely replaced by technology. These firms are looking at buyback algorithms as tools to boost the productivity of their traders, rather than as replacements for the members of their buyback desks.

JPMorgan Chase & Co. (JPM) says it has buyback algorithms in place, but it isn't using the technology to replace traders and the firm's buyback desk hasn't been downsized.

Instead, the buyback algorithms help traders become more efficient and handle more order flows, said Carl Carrie, head of product development in JPMorgan's global executions services group.

This algorithm can operate without human intervention once it has been programmed. Still, the buyback process continues to be supervised by a special equity trader, who looks for ways to add value, Carrie added. For instance, the trader may decide that it would be most effective to carry out the buyback through a mix of manual block trades and algorithmic trades. Also, human traders may play a more hands-on role in a buyback involving a stock that is less liquid.

"Algorithms are not only for the buyside, but are also useful for making the sellside more efficient and effective with our client base," added Carrie.

Before the firm began using algorithms for its buyback desk, a trader handled each deal manually. Assistant traders had to constantly watch to ensure that all the trades carried out on any particular day met regulatory requirements. Algorithms remove the need for this constant scrutiny.

"The buyback algorithm removes the pain of having to focus on non-value added activity," said Carrie. "We are doing more business, enhanced by having algorithms to increase productivity," he added. "The amount of buybacks our firm has done has increased." Specific data on the size and growth of the firm's buyback business wasn't available.

Another firm, Jefferies & Co. (JEF), has a buyback algorithm called the Corporate Buyback Autoserver. While Jefferies traders can choose to do trades manually or through the algorithm, most covering corporations find it beneficial to use the buyback technology, said John Shaw, head of portfolio and algorithmic trading at Jefferies.

"The algorithm manages the minutiae of trading, while the trader watches to see that the overall goal of the customer is achieved," Shaw said. "Because markets trade in such small orders, it is difficult for people to keep trading themselves and concentrate on best execution all the time."

Firms that offer buyback algorithms list a slew of benefits of the technology. For one, trading algorithms typically break up large orders into tiny parts, and can help maintain anonymity during buybacks. They can also help companies buy back their stock at a more effective price.

Still, Adam Sussman, a senior consultant at the Tabb Group, a Westborough, Mass., research firm, said the concept of the buyback algorithm is still in its early stages. While a few firms might offer the technology, most are still in the process of developing the software, he said.

And although algorithms have their uses, they also have their pitfalls. "Their weakness is in trading illiquid stocks," said Sussman. So, for instance, buyback algorithms may not perform very well in certain small-capitalization stocks.