Empirical analysis of multi-factor models vs. CAPM for U.S. stock portfolio returns.

Post ID: 3299491

Overview

"Python Tips for Empirical Project 2" and the following is the link to the recorded video on TA Regression Demonstration 2.https://jhucarey.zoom.us/rec/share/MIvzUHPkqxCb0iqjLEgT95OwT8tJgZoDd7xekPC3lvsr3RN_mr3y_Yo2c9GwshKD.O51fxIb8EtyGW8Fo?startTime=1733443270000
need to submit one report up to 5 pages and also a html form file of python code for solving this
Primary Objective:Analyze the performance of multi-factor models compared to the single-factor CAPM model.
Tasks:
- For 25 stock portfolios (formed on Size and Book-to-Market):
  - Apply the three-factor Fama-French model: R(t)−RF(t)=a+b⋅[RM(t)−RF(t)]+s⋅SMB(t)+h⋅HML(t)+e(t)R(t) - RF(t) = a + b \cdot [RM(t) - RF(t)] + s \cdot SMB(t) + h \cdot HML(t) + e(t)
  - Report summary statistics for SMB(t) and HML(t), including correlations with [RM(t)−RF(t)][RM(t) - RF(t)].
  - Analyze the significance of the 25 portfolio loadings on the two new factors (SMB and HML).
  - Test the hypothesis that SMB(t) and HML(t) lack significant explanatory power beyond the market excess return.
- For 25 new portfolios (formed on Size and Operating Profitability) and another 25 (formed on Size and Investment):
  - Download the relevant return series from Ken French’s Data Library.
  - Summarize returns for these portfolios.
  - Test whether the three-factor model explains their variations effectively.
Deliverables:
- A project report containing:
  - Introduction: Summarize the research question, models, and findings.
  - Data Description: Highlight data sources, summary statistics, and differences between portfolio sets.
  - Model Analysis and Findings: Compare performance of CAPM vs. three-factor models across portfolios.
  - Conclusion: Present overall insights and comparative analysis.
- Python code for generating all results:
  - Ensure original implementation of regression procedures, as guided by TA sessions.
References:
- Use seminal articles like Fama and French (1993, 2015) for comparative benchmarks.