Quantitative Stock Selection

Campbell R. Harvey,
Fuqua School of Business, Duke University, Durham, NC USA
National Bureau of Economic Research, Cambridge, MA USA

Spring Term 4, 2005

Course Description

There has been a large growth in 'Alternative Investing' recently. Indeed, institutional investors routinely invest up 20% of their funds in alternative vehicles such as hedge funds. However, there is an important gap in the top MBA programs' curricula and the practice of investing. There is little or no mention of this important asset class in the standard textbooks.

The goal of the course is to bridge the theory of finance with the practice of finance in the area of stock selection. The course is very practically oriented, with students gaining hands-on experience in developing portfolios of stocks destined for a buy portfolio as well as stocks designated for short selling. While the course will examine the theoretical and empirical foundations of many different styles of alternative investing, most of the focus will be on the classic 'Long-Short'.

An important tool that the course will be employing is FACTSET's Alphatester. FACTSET is the premier tool for stock selection research. FACTSET is both a database of databases and a front-end program. In a few lines of script, students can sort through thousands of securities and forming portfolios based on attributes they consider important. Most top asset management firms have access to FACTSET. Through a special arrangement with FACTSET, and a generous grant from Duke's Global Capital Markets Center, FACTSET will be delivered to Duke students for the first time. Currently, Duke is the only top business school with access to FACTSET. We currently have three ports.

Another important ingredient is having the best data. The school already has access to databases such as Compustat and Worldscope, but for stock selection earnings forecasts are critical. Again, the Global Capital Markets Center has agreed to provide half the funding for I/B/E/S. This is a "special deal" in that students will receive the "real time" version of I/B/E/S - not the educational version.

While much of the course will be devoted to developing portfolio strategies, the key insight of the course is how to think about risk in the context of the theory of finance. Students will spend a considerable amount of time examining different risk models and whether we can explain the performance of the stock selection strategies in the context of risk, information advantage, or market inefficiency.

The Fuqua Honor Code is maintained in this course.

Finally, to do well in this course, you must be willing to set aside time. Assume you will spend 10 hours outside of the class time.


FINANCE 453, Global Asset Allocation and Stock Selection. There are no exceptions. Also, there are no Audits allowed in this course.


There is one major requirement in this course. The final project will count for 100% of your grade. The groups and project links are found:

Copy to student drive, and into the subdirectory created for your project materials. They will be posted to a space in my web structure for future reference. Importantly, I am not looking for a regurgitation of old ideas or the popular press. I am looking for some innovative ideas supplemented with empirical analysis.

All students are expected to attend the presentations in the last week. You must email me if you cannot attend the presentations.

Project publication

There is a possibility that some of the projects could be posted to the Financial Economics Network. When posting, my name will be appended as the final name on the case credits. [In finance, by the way, last name out of alphabetical order indicates the smallest contribution.] I will have more information on this during the term.

Outside Class Contact

This is my teaching term -- and I am generally available. Email is an efficient way to get me at odd hours, cam.harvey@duke.edu . If needed, call my assistant, Tara Bowens 660-7775 or email her at tbowens@duke.edu, to set up an office visit. I also don't mind taking cellular calls at 919-271-8156 (don't hesitate to call me at odd hours, if I am not available, I turn the telephone off).


As I mentioned above, 100% of the grade is based on the project. If you choose to work in groups, you will also need to complete a group self evaluation. I will email you near the end of the course to remind you about this.


I reserve the copyright for all parts of the course. Any commercial reproduction of any course materials including lecture notes taken by students during the class is not allowed unless explicit permission is given by me.

Data resources

We will be using the data available through FACTSET.

Outline and Recommended Reading Assignments

There is little written on Quantitative Stock Selection. Hence, there is not much reading.


Will be worked out during the first lecture. We will likely meet once per week for an extended session. At the beginning of each session, there will be a asset-management perspective-discussion of the recent economic data.

1. Wednesday March 23, 2005, 10:00 Seminar Room B

Course overview

Course overview.

Review of recent economic data. Review FINANCE 453 Assignment 5. Course goals and deliverables.

Guest Speaker: Robert S. Feldman, Portfolio Manager, FMR

Robert is a Portfolio Manager in Institutional Equity Group. He manages developed international (Europe + Asia) portfolios. Investment style is core approach utilizing proprietary fundamental and quantitative research. Focus on bottom-up stock selection. He will talk about low-tracking error selection. FMR manages $750 billion.



    2. Wednesday March 30, 2005, 11:00 Seminar Room B

    Univariate Screens

    Course overview.

    Review of recent economic data. Indepth analysis is leading univariate screens.

    Guest Speaker: Dr. Andrew L. Berkin, First Quadrant

    Andrew received his B.S. with honors in physics from the California Institute of Technology in 1983, and his Ph.D. in general relativity from the University of Texas at Austin in 1989. He subsequently was a postdoctoral research fellow in astrophysics at Waseda University in Japan for two years. In 1992 he started work at the Jet Propulsion Laboratory pursuing research into the visualization and analysis of large scientific data sets. The software that resulted, LinkWinds, was selected as winner of the 1996 NASA Software of the Year Award. Andrew joined the research group at First Quadrant in 1997, and is currently a Director, managing the firm's Small and Mid Cap equity products. His work includes modeling stock returns and asset allocation, estimating and analyzing transaction costs, developing analysis software, and investigating tax efficient investment strategies. He has published numerous articles in taxable investing, as well as scientific visualization and general relativity.



    3. Friday April 8, 2005, 9:00am Seminar Room B

    Multivariate Screens

    Combining Factors.

    Review of recent economic data. Multivariate screening techniques.

    Guest Speaker: Peter Swank, Goldman Sachs Asset Management

    Peter is a senior member of the Global Quantitative Equity Research team conducting research on fundamental-based quantitative investment strategies. Peter joined GSAM in September 2004. Previously he was a senior vice president and director of equity risk management at ING Investment Management, and director of quantitative equity research at its predecessor firm Aeltus Investment Management. He has also held senior quantitative research positions at First Quadrant and Asset Strategy. Peter began his investment career as a faculty member at the Marshall School of Business at the University of Southern California.

    Peter received a A.B. in mathematics and economics from the University of Illinois, where he was elected to Phi Beta Kappa, and M.A. and Ph.D. degrees in economics from the University of California, Los Angeles. He has previously served as President of the Los Angeles Quantitative Investment Association.



    4. Wednesday April 13, 2005, 11:00am. Seminar Room B

    Dynamic Factor Weights

    Time varying factor loadings which depend on economic conditions.

    Review of recent economic data. Methods for time-varying factor weights.

    Guest Speaker: 1. Gregory M. McMurran, Chief Investment Officer, Portfolio Manager, Analytic Investors

    Greg McMurran is responsible for management and oversight of the implementation of the firmís investment strategies. He is a major contributor to the firmís ongoing research efforts as well as new product development and strategy applications. Greg is also a member of the firmís management committee as well as the Compliance Committee and the firmís Best Execution Task-Force. With more than 25 years of quantitative research, portfolio management and trading experience, Greg has an extensive background in managing quantitative investment portfolios. Greg is also a recognized authority on options valuation and hedging strategies and has authored several articles on these subjects. Greg received a Bachelor of Arts degree in Economics from the University of California, Irvine. He also received an M.A. in Economics at California State University, Fullerton where he wrote his masters thesis on estimating the probability of default for third world debt.

    Guest Speaker: Steven Sapra, CFA, Portfolio Manager, Analytic Investors

    Steve Sapra is a Portfolio Manager responsible for the ongoing research effort for the firm's U.S. equity strategies as well as the day-to-day portfolio management and trading of such accounts. Prior to joining Analytic Investors, Steve was employed as a Senior Consultant at BARRA, Inc. in Berkley, CA. He consulted with investment managers in the area of risk control and strategy implementation. Steve received an MA in Economics from the University of Southern California and a BA in Economics from California State Polytechnic University, Pomona. Steve, a Chartered Financial Analyst, is currently a lecturer at California State Polytechnic University, Pomona, teaching Economics part-time.



    5. Tuesday April 19, 2005, 9:00am Seminar Room B

    Scoring Screens and Resampling

    Using mean-variance optimization along with other types of optimization to form a scoring screen..

    Review of recent economic data.

    Guest Speaker: TBA,

    Speaker description.



    6. Tueday April 26, 2005, 8:30am Seminar Room B

    Migration Tracking

    Following the performance of constituents and using the memory.

    Review of recent economic data.

    Carolie Burroughs, Portfolio Manager, ING Aeltus Investment

    Carolie is a Portfolio Manager for ING Aeltus Investment Management, Inc. in Hartford, Connecticut. She manages the ING Small Company Fund, and the real estate securities portfolio within the ING Strategic Allocation Funds. She was previously an Assistant Portfolio Manager for ING Aeltus' Institutional Small Cap strategies, and has been a member of the Small Cap team since joining the firm in 1998 as a Quantitative Equity Analyst. Prior to joining the firm, she worked in the fixed income department at Loomis Sayles and before that at Colonial Management Associates. She has conducted research on real estate securities for the Center for Real Estate and Urban Economic Studies at the University of Connecticut. Ms. Burroughs received her B.S. in Mathematics from the University of Connecticut and M.B.A. from the Fuqua School of Business at Duke University. She holds the Chartered Financial Analyst designation.

    William Weng, Director, Credit Swisse Asset Management

    William is Director and a senior quantitative analyst in the quantitative resources group where he focuses on quantitative equity research and products. He joined in 2004 from Banc of America Capital Management, where he was a Vice President and senior quantitative analyst, a role in which he conducted quantitative research, and managed top performing quantitative portfolios. William holds a B.E. in management information systems and an M.E. in systems engineering from Tsinghua University in Beijing, and a Ph.D. in economics from Boston University. He has co-authored articles in academic journals including Journal of Derivatives.

    Alex Shabshis, Vice President and Senior Quantitative Analyst, Alliance Capital Management

    Alex joined Sanford C. Bernstein, Inc. (now part of Alliance Capital) in 1988, shortly after graduating from The Cooper Union with the degree in Electrical Engineering. During the past decade, he either led or played a senior role in numerous projects related to stock selection strategies, quantitative aspects of global portfolio management and product design. Key responsibilities included design of numerous global and international services (including hedge funds), application of optimization to portfolio construction and development of the Bernstein's multi-factor global quantitative stock selection model (in use since Q1 of 2001). Since March 2002, Alex is also a co-manager of the Bernstein Global Diversified Hedge Fund.



    7. Wednesday April 27, 2005, 11:00am Seminar Room B

    Student projects

    Presentation of final projects