Extreme stock returns
- PDF / 371,293 Bytes
- 21 Pages / 602.009 x 771.004 pts Page_size
- 64 Downloads / 254 Views
Daniel Glickman* is a Principal of State Street Global Advisors. Currently, he is the project leader for several projects on Stock Selection in the US, Japan, and Emerging Markets. Dan holds an MBA in Analytic Finance and Econometrics from the University of Chicago, an MS in Operations Research from Columbia University and a SB in Mathematics from Massachusetts Institute of Technology.
A. Gregory DiRienzo is an Assistant Professor at Harvard University in the Department of Biostatistics; there he conducts applied and theoretical research in clinical trials, survival analysis and bioinformatics. Greg received his PhD in Mathematics and Statistics from the University at Albany — SUNY.
Richard Ochman is an Associate of State Street Global Advisors. His current research projects in US Stock Selection and Currency Overlay Management. Richard is a recent graduate from the Boston University School of Management, where he received a Bachelor of Science in Finance and Management Information Systems. *State Street Global Advisors, 2 International Place, 21st Floor, Boston, MA, USA. Tel:⫹1 617 664 2451
Keywords: volatility; market efficiency; behavioural finance; stock returns; fundamental analysis
Abstract We identify characteristics of stocks in the Russell 1000 and 2000 that exhibit extreme (in the top and bottom 2.5 per cent) total returns over the next quarter. Using these characteristics, we develop a model to identify 50 (100) stocks as expected extreme performers in the Russell 1000 (2000). Over 22 per cent (16 per cent) of firms that we identify as expected extreme performers actually are extreme performers in the next quarter. Moreover, we contrast the characteristics of rockets (in the top 2.5 per cent) with torpedoes (in the bottom 2.5 per cent). Using these characteristics, we develop a model to identify expected rockets and torpedoes. We develop a strategy that forms portfolios long on expected rockets and short on expected torpedoes. This strategy returns an average of 8.7 per cent (7.3 per cent) and standard deviation of 19.4 per cent (10.5 per cent) per quarter using the Russell 1000 (2000) universe. Defining risk as the probability of exhibiting extreme total returns over the next quarter, we find that growth stocks are more risky than value stocks in the Russell 1000 and 2000 in the period from 1992 to 2000.
Introduction When Willie Sutton was asked why he robbed banks, he replied, ‘because that’s where the money is’. When our colleagues ask us why we target the
䉷 Henry Stewart Publications 1470-8272 (2001)
characteristics of US stocks that exhibit extreme returns (total returns in the top or bottom 2.5 per cent of the given universe over the next quarter), we too reply, ‘because that’s where the money
Vol. 2, 2, 107-127
Journal of Asset Management
107
Glickman et al.
is’. For example, in the period from 1992 to 2000, stocks in the top 2.5 per cent of the Russell 1000 average over 58 per cent return per quarter. In contrast, stocks in the bottom 2.5 per cent of that universe average under ⫺37 p
Data Loading...
