📊 Accounting Integrity as Alpha: Testing the Beneish M-Score in Portfolio Strategy

Author: Louis Sellier
Course: Earnings Quality and Fundamental Analysis (EQFA)
Institution: Columbia Business School
Date: Spring 2025

---

🧠 Overview

This project investigates whether forensic accounting insights—specifically the Beneish M-Score—can be used not only to detect potential earnings manipulation, but also to construct equity portfolios that outperform the market.

Using financial statement data from the CFA Institute Level II curriculum dataset and return data from Alpha Vantage, we:

• Compute M-Scores for a representative sample of U.S. firms (2001–2022)
• Run firm-level regressions to test whether M-Scores explain returns beyond market beta
• Build equally weighted portfolios based on M-Score rankings and compare them to the S&P 500

---

🧮 Methodology Summary

• Model: Beneish M-Score
• Sample Firms: Apple, UnitedHealth, NVIDIA, ExxonMobil, General Electric, Costco, Johnson Controls, Intercontinental Exchange
• Data Period:
  • Financials: Quarterly from 2001–2022
  • Returns: Annual from 2010–2022
• Data Sources:
  • CFA Institute Training Dataset
  • Alpha Vantage API (price data)
• Analysis:
  • Annual rebalancing of Best/Worst M-Score portfolios
  • Return comparison against SPY
  • HC1-robust firm-level regressions