CAPM vs. Fama–French Three-Factor Model

Comparing Asset Pricing Models for Equity Returns

Finance
Statistics
Econometrics
CAPM and Fama–French comparison for stock returns.
Published

June 8, 2025

Comparing asset pricing models to explain stock returns

This project evaluated two foundational financial models—the Capital Asset Pricing Model (CAPM) and the Fama–French Three-Factor Model—to determine which better explains equity returns. The analysis focused on estimating factor exposures, comparing model fit, and interpreting the economic significance of each factor.

View Source Code View Full Report ← Back to Projects

Role
Financial Data Analyst

Domain
Quantitative Finance

Stack
R · Linear Regression

Methods
CAPM · Fama–French · Factor Modeling

Project Snapshot

Question

Does the Fama–French Three-Factor Model explain stock returns better than CAPM?

Approach

Estimated both models using historical market data and compared explanatory power, residual behavior, and factor significance.

Finding

The Fama–French model captured additional variation in stock returns through size and value factors, providing a better overall fit than CAPM.

Overview

Asset pricing models attempt to explain why different stocks generate different expected returns. While CAPM attributes returns primarily to market risk, the Fama–French Three-Factor Model extends this framework by incorporating company size and value characteristics.

This project compared both models using historical financial data to evaluate whether the additional factors meaningfully improved model performance and interpretation.

The Challenge

Financial market variability

Stock returns are influenced by multiple sources of systematic risk that may not be captured by a single-factor model.

Model comparison

The project required comparing two competing statistical models using both quantitative metrics and economic interpretation.

Factor estimation

Regression coefficients were interpreted as measures of exposure to market, size, and value risk factors.

Model diagnostics

Residual analysis and goodness-of-fit metrics were used to evaluate model quality.

Analysis Pipeline

1
📥

Financial Data

Collected historical stock returns and factor data.

2
📊

Data Preparation

Computed excess returns and aligned market factor observations.

3
📈

CAPM

Estimated the single-factor market model using linear regression.

4
📉

Fama–French Model

Estimated the three-factor regression including market, size, and value factors.

5
⚖️

Model Comparison

Compared explanatory power, residuals, and statistical significance.

6

Interpretation

Evaluated which model better explained observed stock returns.

Regression workflow for comparing competing asset pricing models.

Technical Stack

  • R
  • Linear Regression
  • CAPM
  • Fama–French
  • Financial Modeling
  • Regression Diagnostics
  • Statistical Inference
  • Data Visualization
  • Quantitative Finance

What I Built

CAPM Regression

Estimated market beta and evaluated the relationship between market risk and excess stock returns.

Three-Factor Model

Extended the analysis by incorporating size (SMB) and value (HML) factors to explain additional return variation.

Model Evaluation

Compared model fit using regression diagnostics, residual analysis, and statistical significance tests.

Financial Interpretation

Connected regression coefficients to economic intuition and discussed how additional factors influence expected returns.

Key Results

📈
2

Asset pricing models compared

📊
3

Systematic risk factors analyzed

📉
OLS

Linear regression framework used for estimation

What I Learned

This project strengthened my understanding of factor-based regression models and how statistical modeling can be used to explain financial market behavior. Comparing CAPM and the Fama–French model demonstrated how incorporating additional explanatory variables can improve predictive performance and economic interpretation.

It also reinforced the importance of evaluating models using both statistical evidence and domain knowledge. A model with better fit should also produce results that align with established financial theory.

Research Deliverable

The complete report includes the regression methodology, coefficient estimates, diagnostic analysis, and model comparison.

View Source Code

View Full Report (PDF)

← Back to Projects