Content area
Full Text
News flash: Barr Rosenberg does not use Barra models. In fact, when Rosenberg left Barra in 1984, he immediately built another set of quantitative models for his new firm, Rosenberg Institutional Equity Management, which he runs today.
Rosenberg, 55, first became intrigued with actively managing money through quantitative principles at Barra. But the system proved so comprehensive that he couldn't sell it to even Why? For a fund to sign him on meant assigning to a subadviser too central a strategic role. He also believed that running the unit within Barra would compete with customers, a view on which he and his friend and Barra CEO Andrew Rudd came to differ. "The choice was to give that up or found [my] own firm," says Rosenberg.
So he left, raising funds from Harold Arbit, who was then head of Concord Capital Management and who had helped Rosenberg in his first efforts a decade earlier. (Arbit remains Rosenberg's limited partner.) Ten days after the Orinda, California-based firm opened in May 1985, Aluminum Co. of America's pension plan signed on as its first client. For years Rosenberg delivered exactly what he had predicted - 4 percent consistent alpha in core large- and small-cap portfolios. Funds flowed in, including generous amounts from Japanese clients of Nomura Securities Co., and assets peaked in 1990 at more than $9 billion.
Today Rosenberg is an uncommonly modest legend. "If you believe things you can't deliver, then you make a fool of yourself," he says. In 1990 and 1991 he underperformed his U.S. benchmarks by about 4 percent. After the recession ended in 1991, he installed new earnings and investor-sentiment models, reversing his performance, although assets still skidded to a low of $3.1 billion in 1996. In the past year, however, assets more than doubled, to $6.7 billion, including $800 million in four U.S. mutual funds, which the firm distributes to institutions and retail clients, both directly and through fund supermarkets.
Rosenberg - ever the calm intellectual with Buddhist leanings - still professes his belief that the models count, albeit new, improved ones. its own. "The ongoing frustration was that the competition was lousy," says another former insider. "If we could have done half a good job, we could have killed the other...