Summary
- Forensic value portfolios using Beneish, Altman, Ohlson, and Montier algorithms have outperformed the S&P 500 since 2017, with negative portfolios showing higher returns but greater volatility.
- Over nearly 9 years of live forward testing, Forensic Negative portfolios returned +791.6% and Forensic Positive portfolios +466.1%, versus the S&P 500’s +192.2%.
- Applying market timing tools like Momentum Gauges can further enhance returns and mitigate downturn risks, as recent portfolio performance demonstrates.
- Positive forensic stocks offer stable returns and low delisting rates, while negative forensic stocks deliver outsized gains with higher risk and frequent delistings.
Introduction
Every financial statement tells two stories: the one management wants you to read, and the one the numbers expose.” ~ fraud examiner
I have spent the last 35 years trading, researching, and constructing algorithms to identify and leverage the value across fundamental, technical, and behavioral finance models. Of the ten portfolio models designed for optimal portfolio mixes for members to beat the market at Value & Momentum Breakouts, eight come from enhancing well-tested anomaly research in published financial journals.
Summary
- Forensic value portfolios using four top algorithms (Beneish, Altman, Ohlson, Montier) have consistently outperformed the S&P 500 since 2017, with negative forensic portfolios showing higher returns but greater volatility.
- Over nearly 9 years of live forward testing on SA, the Forensic Negative portfolios have returned +791.6% and the Forensic Positive portfolios have returned +466.1%, not including dividends, compared to the S&P 500 +192.2% including dividends over the same period.
- Applying market timing tools like the Momentum Gauges may further enhance returns and help avoid major downturns, as evidenced by recent portfolio performance.
- This multi-year study demonstrates the value of forensic analysis using a purely fundamental approach for detecting financial irregularities to improve risk-adjusted returns.
The Four Forensic Algorithms
Each selection is evaluated against four independent forensic models from peer-reviewed financial literature, together applying 28 different fundamental variables:
1. Beneish M-Score: Eight variables detecting earnings manipulation. An M-Score above -2.22 signals likely manipulation. Correctly identified 71% of major accounting fraud cases in advance of public disclosure.
2. Altman Z-Score: Predicts bankruptcy risk within 2 years. Values below 1.81 indicate high distress.
3. Ohlson O-Score: Multi-factor bankruptcy probability model. Values above 50% indicate elevated 2-year bankruptcy risk.
4. Montier C-Score: Simple scoring system for “cooking the books” with values above 4 flagging firms at risk of financial irregularities.
This algorithmic combination is a purely fundamental financial study with new selections for midyear 2026 to continue testing the top four forensic algorithms applied to detect bankruptcy risk, earnings manipulation, and financial irregularities. As a former certified fraud examiner and anti-money laundering specialist, I was concerned by some of the findings in published reports that could adversely impact my investments. This led me to apply forensic models to stock portfolios to see if forensic models alone can produce discernible differences in returns.
Over the past 9 years from 2017 on Seeking Alpha, I have published 54 different forensic portfolios in one-year and two-year test periods. As a result of this long-term analysis, we are seeing strong differentiation in results between negative and positive forensic portfolios. Most notably, the delisting of stock symbols (merger, acquisition, leaving the exchange) is approximately 10x higher among negative forensic stock selections than for positive forensic stocks.
