1	"Presentation by:" Presentation by: Bryan Durand Josh Amoss Suri Thummala Steve Beuchaw Matthew Malouin
2	Agenda Introduction Methodology Factors Analyzed Summary Scoring Model Selected Factors
3	Establishing Long-Short Trading Strategy Objective Limit universe of stocks to firms with middle capitalization ($500M to $2B) market values We feel this is a less efficient universe We feel these stocks will be liquid enough (low market impact due to trading) Establish long / short portfolios based on quantitative stock screens Rebalance portfolios monthly Quantitative stock screen Eleven factors Find predictive powers on positive and negative returns Select factors with strong predictive powers Go long stocks in top quintile Go short stocks in bottom quintile
4	Stock Selection Process Screen Parameters US equities listed on both NYSE and NASDAQ Market capitalization above $500 million to $2,000 million Monthly data In-sample time frame: 1989 – 2001 Out-of-sample time frame: 2001 – 2004 Selected eleven fundamental, expectational, and momentum factors to predict future stock returns
5	Description of Factors Fundamental Book to Price: book value per share / price per share Dividend Yield: dividends per share / price per share FCF Yield: (cash flow from operations – capex) / price per share Return on Assets: annual net earnings / total assets Return on Equity: annual net earnings / total shareholder equity Expectational Percent Change in FY1 Estimates over 3 Months: percent of analysts changing their FY1 estimates over the last three months Estimate FY1 EPS Yield: consensus estimate of FY1 EPS / price per share SUE Score: standard unexpected earnings Momentum Momentum 3 Months: one month – one year 3 month price return 1-Year EPS Growth: historical one year earnings per share growth rate 3-Year EPS Growth: historical three year earnings per share growth rate
6	Two Factors had rock-hard performance
7	Our Scoring System Factor 1: FCF Yield Portfolio does well in both up and down markets Catastrophic loss in 1999 (-46%) High turnover in portfolio – high cost to implement We scored the long portfolio a 3 and the short portfolio -3 Factor 2: Percent Change in FY1 Estimates over 3 Months Historical returns and consistency are good Recent (in sample) returns not as strong Factor works well especially during market anomalies such as 1999 We scored the long portfolio a 2 and the short portfolio -2 Overall System (Factor 1)(3/5) + (Factor 2)(2/5)
8	Value Weighted Portfolio shows intriguing results Scoring Strategy has good performance both In Sample and Out of Sample There is a step down in returns between Quintiles 1 and 5 Only 1 year with negative return (1999) Moderate turnover compared to FCF Yield only – lower cost to implement
9	Heat map demonstrates strong consistency
10	Distribution of returns for the scoring model are positively skewed
11	Value Weighted Portfolio shows intriguing results Quintile 1 and Quintile 2 have a solid Alphas spread over Quintile 4 and Quintile 5
12	Conclusions We have found two factors that give us very attractive returns both in sample and out of sample The value weighted portfolio (long portfolio 1 and short portfolio 5) Sharpe ratio is 1.71 versus an S&P 500 sharp ratio of 1.21 The average market caps is stable at approximately $1B across all portfolios Portfolio 1 beat the benchmark 65.5% of the time while portfolio 5 beat the benchmark 43.4% of the time (similar performance in up and down markets) Portfolio 1 has a Beta of 0.949 while Portfolio 5 has a Beta of 0.977 The Alpha of Portfolio 1 is 13.421 versus Portfolio 5 being -7.671 The T-stats are over 2 This appears to be a very attractive screening method by any measure
13	FCF Yield is a suitable factor for a long-short strategy FCF Yield has good performance both In Sample and Out of Sample There is a step down in returns between Quintiles 1 and 5 Quintile 5 does better than Quintile 4 Quintile 5 has an average FCF yield of -9% and contains several growth companies Further research might consider limiting to positive FCF yield companies (a possible knock out screen for growth companies)
14	FCF Yield’s heat map demonstrates strong consistency
15	FCF Yield provides attractive Alphas Quintile 1 and Quintile 2 have positive Alphas while Quintile 4 and Quintile 5 have negative Alphas for a good Alpha spread
16	% Change in FY1 EPS Estimates was selected for its hedging ability We have chosen to sacrifice some return in order to attempt to prevent catastrophic (and career ending) portfolio losses
17	Heat map demonstrates some consistency and is able to limit catastrophic losses (example: positive 1999 return)
18	The in sample Alphas are suitable and we are willing to accept marginal Alphas out of sample due to the factor’s hedging ability
19	"Back-up" Back-up
20