ECON 675 Capstone
Comprehensive guide for the ECON 675 capstone: Fama-French factor analysis of AI stocks vs. the dotcom bubble, with scatter plots, factor loading comparisons, and full team speaker scripts.
Overview
This is the comprehensive reference guide for our ECON 675 capstone project: applying the Fama-French three- and five-factor models to the Magnificent Seven AI portfolio and comparing factor loadings to the 1990s dotcom bubble.
The site covers all three team members' presentation portions, revised scatter plots per professor feedback, and empirical factor loading estimates sourced from the academic literature.
Research Question
Do standard risk factors (CAPM, FF3, FF5) explain excess returns in the Magnificent Seven AI portfolio? How do factor loadings compare across the dotcom era (1997–2001) and the AI era (2020–2025)?
Key Finding
The RMW (profitability) factor loading flips sign between eras. Dotcom stocks loaded negatively on RMW — roughly 86% of internet companies were unprofitable at the March 2000 peak. The Magnificent Seven load positively — NVIDIA's net margin exceeds 53%, and Mag-7 average net margin is 25.8%. The FF5 model detects this structural difference where FF3 cannot.
What's Inside
- Mustafa's introduction script (Slide 1)
- Arnav's factor analysis scripts (Slides 6–8)
- Rinad's team roles script (Slide 9)
- Factor loading comparison table (dotcom vs. AI, FF5)
- Four scatter plots with corrected axes (returns Y, factor X)
- Dotcom vs. AI bubble valuation comparison (CAPE, P/E, profitability)
- Sep 13 revision checklist based on professor feedback
- Full literature section with 8 key papers
Stack
| Layer | Technology |
|---|---|
| Frontend | HTML, CSS, JavaScript |
| Charts | Canvas API (scatter plots with seeded PRNG) |
| Hosting | Portfolio static assets on Vercel |
| Data | Ken French library, JRFM, JF, AER literature |