Division of the Humanities and Social Sciences California Institute of Technology Pasadena, California 91125 Using Information from Trading in Trading and Portfolio Management: Ten Years Later

The author is a visiting faculty member in economics at Caltech. At the time the original version of this paper was written, he was Managing Director at First Quadrant 2 , with responsibility for global active equity portfolios totaling $6 billon. Dr. Leinweber has a long background in electronic trading. He is the inventor of MarketMind, the expert system now incorporated in ITG's Quantex product, and a founder of Codexa, which provided internet information filters for institutional investors and traders. He holds a Ph.D. in applied mathematics from Harvard University and undergraduate degrees from the Massachusetts Institute of Technology. The centerpiece of this paper is a comprehensive analysis of all of the trading by a large US pension fund in 1991. The first version was published in the Summer 1995 issue of the Journal of Investing and later incorporated in the AIMR's CFA reading. This updated version, requested by the AIMR, reflects the technological and market structure changes of the last ten years, and adds a new empirical analysis of all the 2001 trading by a $7 billion US equity manager. Trading technology has changed in many ways, yet many of the same characteristics of institutional trading are seen just as strongly as they were before. The attention given to large difficult orders still shows in lower than expected trading costs. Small " no-brainer " orders still represent the largest component of overall trading costs. These are precisely the type of costs that modern electronic trading systems are designed to reduce. The idea that trading costs were a substantial drag on performance was a relatively novel in 1991. Today, it is a central concern for many managers, including those profiled here.