Institutional investors are important both because of the magnitude of their holdings and trading activity and because their decisions to buy or sell can signal a good takeover candidate as well as in some instances facilitate the takeover. However, little is known as to why institutions prefer one stock to another. Through discriminant analysis seven variables (six accounting ratios and size) are identified that reveal the differences in financial characteristics between neglected and strongly held firms. Regression analysis is then used to show that four variables from the set of seven discriminant variables are consistent predictors of the actual fraction of outstanding shares held by financial institutions. The analysis provides some support for the claims of corporate chief financial officers that institutions focus only on short-term performance. However long-term performance is also shown to be important.

No abstract

simulation AUGUSTO A. LEGASTO, Jr., is an associate professor of the Department of Management at the Baruch College of the City University of New York. He received his MA in administrative sciences from Yale University and his PhD in production systems in 1975 from Columbia University. He is currently active in the art of problem definition and modeling in the realm of higher management decision making, particularly in the public sector. He has published several articles in books and in journals including Management Science. He has co-edited the book, System Dynamics, with Jay W. Forrester of M.I.T. in the TIMS He has extensive industrial experience in the design and simulation of computer systems and is currently active in applying pattern recognition and discriminant analysis to problems in economics and finance. He and co-workers are now engaged in applying fixed-point analysis to well-known simulation models, such as those developed by Forrester, to determine if the emerging field of attractors and repellors can offer insights into the behavior of those models. ABSTRACTA crisis triggered procedure is applied to analysis of the criminal justice system of the United States; in particular, the development of a "crisis triggered" computer simulation model of the criminal justice system is reported. The model is then used to identify potential major crises that have occurred or are expected to occur in the criminal justice system, to simulate them, and to study alternative regimes or patterns of recovery that the system may undergo.