DYNAMIC WEIGHTED SCORING MODEL 2272006 Huimei Li Xing

DYNAMIC WEIGHTED SCORING MODEL 2/27/2006 Huimei Li Xing Shen Ben Xie Xuan Zheng

Agenda u Model Summary u Factor Analysis- a Test u Subjective Scoring u Dynamic Weighted Scoring u Summary 9/16/2020 2

Model Summary -- Basic u Objective: u u u Universe: 500 largest companies in US Rebalance period: 1 month Model period: u u Explore various factors Explore different scoring methods generate consistent positive return with minimum market exposure In Sample 1995 -2002 Out of Sample 2003 -2005 Sorting fractile: Quintile Scoring method u u Subjective Objective: dynamic weighted 9/16/2020 3

Model Summary – Factors Choice u Fundamental factors u u u Expectational factors u u u FCF Yield Dividend Yield Price to Book Ratio CFO Growth Rate Expected Earnings Yield (FY 1) Expected Earnings Yield (FY 2) Change in Consensus Unexpected Earning Momentum u u Short Position % Price Momentum 9/16/2020 4

Factor Analysis -- Summary In US large cap equity market, the relative performance of various factors are: FCF Yield, FY 1, FY 2, Price Momentum Y ie e. M ld om en tu m FY 2 ie ld Y FY Pr ic 1 st Quintile 9/16/2020 1 ie ld FC F sit io Po rt Y n s su en ns Co ei n Sh o ie l Y nd ng Ch a D iv id e e/ Bo ok e Pr ic ur ct l. S tru ita Ca p d 1. 6 1. 4 1. 2 1 0. 8 0. 6 0. 4 0. 2 0 Average return (%) u 5 th Quintile 5

Factor Analysis -- FCF Yield Formula: CQ_CF_FREE_PS(0 L 45 D)/CM_P(0) Results: Best-performing quintile: 1 Worst-performing quintile: 5 Exception: bubble year, fractile 1 underperformed other quintiles. 9/16/2020 6

Factor Analysis -- (FY 1/FY 2) Formula: IH_Meadian_FY 1 R(0)/CM_P(0), IH_Meadian_FY 2 R(0)/CM_P(0) Results: Best-performing quintile: 1 Worst-performing quintile: 5 Exception: 1998, 1999, quintile 5 outperform 1 9/16/2020 7

Factor Analysis -- Price Momentum Formula: CM_RET(-1 -13) Results: Best-performing quintile: 1; Worst-performing quintile 5 ; Y 2000 performance: Past stock price movement 9/16/2020 8

Subjective Scoring Choose four best performing factors based on the outcome of individual factor analysis Four factors combine fundamental, expectational, and momentum variables Scores are assigned based on subjective judgment u u u Price Momentum Fractile 1 Fractile 5 9/16/2020 FCF Yield Expected Earning Yield (FY 1) Expected Earning Yield (FY 2) 3 4 2 2 -3 -4 -2 -2 9

Subjective Scoring u u u Scoring screen provides a consistent result Only incur a 9. 4% loss in 1999 at in-sample test Incur a 18. 9% loss in 2003 at the out-of-sample test This portfolio has a 13. 7% annualized average return with an 15% standard deviation from 1995 to 2005 The sharp ratio is 0. 92 and the alpha is 19. 78% The outcome of the subjective scoring screen depends on the experience of the portfolio manager 9/16/2020 10

Subjective Scoring 9/16/2020 11

Subjective Scoring – High/Low Score Difference 9/16/2020 12

Dynamic Scoring -- Methodology u Dummy Variable u 3 month T Bills- Long term Treasury Bond u Lagged 30 days u Quintiles u FCF_1, FCF_5, FY 2_1, FY 2_5 u FY 1_5, MM_1, MM_5 u Synthetic Assets u Run Optimizer u Run scoring with dynamic weighting 9/16/2020 13

Optimization 9/16/2020 14

Dynamic Scoring -- Result Value Weighted Result Bubble period and postbubble period 9/16/2020 15

Summary 9/16/2020 16