Factor Exposure Analysis with the Fama-French Six-Factor Model

This project applies the Fama-French five-factor model augmented with the Momentum factor to analyse the drivers of ETF returns. Both static and rolling regressions are performed to estimate factor exposures, evaluate alpha generation, and examine how tisk characteristics evolve over time.

Motivation

Understanding the sources of portfolio returns is a fundamental task in asset management.

Rather than evaluating performance only through total returns, factor models allow investors to identify whether returns are driven by systematic risk exposures or by manager skill (alpha).

This project implements a six-factor regression framework to decompose ETF returns into their underlying risk factors.

What it does

• Downloads monthly ETF returns via yfinance
• Fetches Fama-French 5-factor data and Momentum factor via getFamaFrenchFactors
• Runs a static OLS regression for each ETF, outputting factor loadings and annualized alpha
• Runs a rolling OLS regression (default 36-month window) and plots time-varying factor exposures

Assets analyzed

Ticker	Description
VTI	Vanguard Total Stock Market ETF
VLUE	iShares MSCI USA Value Factor ETF
AVUV	Avantis U.S. Small Cap Value ETF

The selected ETFs provide exposure to the broad US equity market, value investing and small-cap value strategies.

Factors

Factor	Description
Mkt-RF	Market excess return
SMB	Small Minus Big (size)
HML	High Minus Low (value)
RMW	Robust Minus Weak (profitability)
CMA	Conservative Minus Aggressive (investment)
Mom	Momentum

Key Findings

• Market risk (Mkt-RF) explains the largest share of returns across all three ETFs, with estimated market betas close to one.
• The factor model successfully distinguishes the investment styles of the selected ETFs through their factor exposures.
• AVUV exhibits strong positive exposure to the Size (SMB), Value (HML) and Profitability (RMW) factors, consistent with its small-cap value investment strategy.
• VLUE shows a pronounced Value (HML) exposure together with limited Size exposure, reflecting its focus on large-cap value stocks.
• VTI displays negligible exposure to style factors beyond the market, confirming its role as a broad market index ETF.
• Rolling regressions illustrate that factor sensitivities are not constant over time, highlighting the dynamic nature of portfolio risk exposures.

Investment Interpretation

The estimated factor loadings provide insight into the systematic sources of ETF returns.

Rather than evaluating performance solely through historical returns, the model identifies the underlying risk premia each portfolio is exposed to. This framework is commonly used in portfolio construction, performance attribution and risk management within quantitative asset management.

Rolling Factor Exposures

AVUV

VLUE

VTI

Regression Results

Static regression estimates for the selected ETFs. The estimates factor loadings are consistent with the investment styles of the underlying portfolios, while the rolling regressions illustrate how these exposures evolve over time.

Factor	AVUV	VLUE	VTI
Mkt-RF	1.06	1.04	0.99
SMB	0.90	-0.05	-0.01
HML	0.53	0.47	0.03
RMW	0.27	-0.36	0.02
CMA	-0.05	0.13	-0.01
MOM	-0.02	-0.06	0.00
Alpha	0.07%	0.06%	-0.03%
Annualized Alpha	0.84%	0.77%	-0.39%

Requirements

pip install yfinance statsmodels getFamaFrenchFactors lxml pandas numpy

Usage

Run the script directly:

python factor_regression.py

The script prints the static factor loadings table and displays rolling regression plots for each ETF.

Output example

Static regression returns a table with factor loadings, monthly alpha, and annualized alpha for each ETF. Rolling regression plots show how each factor exposure evolves over time using a 36-month sliding window.