ADVANCED SECURITY ANALYSIS
[BFF5040]
“THE FAMA-FRENCH CASE STUDY”
_____________________________________
GROUP ASSIGNMENT
GROUP 18
ALEX LEE [26268418]
JIANNAN ZHANG [25842528]
XUAN ANH NGO [26274736]
YIMING BAI [26413760]
ZHOUJING LI [25675087]
WORD COUNT: 2,918 WORDS
CONTENTS
EXECUTIVE SUMMARY 3
PART ONE. IN-SAMPLEAPPLICATION OF MODEL 3
1.1. FIRST-PASS REGRESSION OF 20 ASSETS 3
1.2 SECOND-PASS REGRESSION OF 20 ASSETS 4
PART TWO. OUT-OF-SAMPLE MODEL PERFORMANCE 5
2.1. CONSTRUCTION OF OUT-OF-SAMPLE PORTFOLIOS 5
2.2 EVALUATION OF OUT-OF-SAMPLE PORTFOLIOS 6
PART THREE. TESTS OF MOMENTUM-BASED PORTFOLIOS 8
3.1 CONSTRUCTION OF MOMENTUM-BASED PORTFOLIOS 8
3.2 EVALUATION OF MOMENTUM PORTFOLIOS 8
APPENDIX 12
A1 FIRST-PASS…show more content… The correlation between the market portfolio and HML and the correlation between intercept and HML is -0.335 and -0.070, which indicates a moderate negative relationship between market portfolio and HML, and weak negative relationship between intercept and HML. Also, the correlations between the market portfolio and SMB, and between the SMB and HML are 0.348 and 0.191 respectively, which means that there are some positive relationships between them.
1.2 SECOND-PASS REGRESSION OF 20 ASSETS
We are using the coefficients of excess returns of market, SMB and HML to run the second-pass (cross- section) regression. E[Ri-Rf] is from 20 observations of average excess return on assets, and the bi, si and hi are from 20 observations of beta on market risk premium, beta on SMB and beta on HML in first-pass regression estimation, respectively.
The estimated coefficients: γb, γs and γh correspond to risk factors in the first pass regression; hence, the second-pass regression is estimated by setting up the following hypothesis: γ0 = 0; γb = E[Mkt-rf]; γs = E[SMB]; γh = E[HML]
This hypothesis is from FF3, which indicates the excess return for assets i only depends on betas and as such market risk premium, SML portfolio premium and HML portfolio premium and any variables in the equation (except excess return for assets i) should be equal to mean zero. In additional, coefficients included from the second