We are introducing an updated version of the Hudson Labs Forensic Risk Score, a model that measures risk signals in corporate disclosures.
Traditional forensic models rely on financial ratios and accounting measures. However, many warning signs appear in the narrative disclosures in company filings. Analysts and short sellers often find these signals by reading disclosures closely and noting unusual explanations, complex arrangements, or heavy reliance on accounting assumptions and estimates.
The Hudson Labs process involves analyzing disclosures in 10-K and 8-K filings and identifies patterns linked to accounting choices, financing structures, and other issues that may increase the likelihood of future problems. Because we focuses on narrative disclosures rather than financial statement data, we captures risks that traditional ratio-based forensic models often miss.
Hudson Labs' process extract red flags from filings and use them to produce a single risk score. The score is machine-learned rather than additive. It reflects how multiple risk signals interact rather than simply counting individual flags.
The Risk Score
Risk scores range from 1 to 100, from low risk to high risk. Scores above 70 are considered high risk.
Several features of the score are important for interpretation:
- High scores are uncommon. Fewer than six percent of mid-cap or larger companies score above 70.
- High scores are predictive. High-scoring companies are more likely to face regulatory scrutiny, securities class action lawsuits, and significant stock drawdowns.
- The model emphasizes precision. When a company receives a high score, the probability of future problems is elevated. However, a low score does not guarantee safety.
- Signals often appear early. In many cases, elevated scores appear months before negative events become widely recognized by the market.
Hudson Labs trained the model on a proprietary dataset that includes SEC enforcement actions, regulatory investigations, and securities class actions related to financial reporting.
The score is designed to help investors identify companies that may warrant closer review and to surface risks that traditional quantitative screens often miss.
See our research and case studies showing how the Hudson Labs forensic risk rcore predicts securities class action lawsuits and stock underperformance.
Hudson Labs Forensic Risk Categories
As part of this update, users can now see the inputs behind the score more easily. In addition to the overall risk score, the platform displays a risk rating and a summary of the issues identified across eight forensic risk categories. This allows users to see the signals that contribute to the score and the areas of disclosure that drive the risk assessment.
Hudson Labs groups disclosure signals into eight forensic risk categories:
Earnings and Accounting – Evaluates indicators and opportunities for possible earnings manipulation or smoothing.
Liquidity, Credit, and Financing Structure – Evaluates whether liquidity is supported by operating performance or by refinancing, structured financing, or asset monetization.
Related-Party Relationships – Evaluates whether insiders or affiliated counterparties influence the company’s reported results.
Resignations & Dismissals – Evaluates turnover in CEO, finance leadership, and oversight roles responsible for financial reporting and governance.
Internal Controls – Evaluates whether the company has effective processes and oversight to produce reliable financial statements.
Legal and Regulatory Exposure – Evaluates the severity of investigations, enforcement actions, or regulatory conflicts.
Governance – Evaluates whether leadership behavior reflects strong stewardship or weak internal discipline.
Strategic Complexity – Evaluates whether acquisitions, restructurings, or repeated strategic repositioning make performance difficult to interpret.
See additional details on what each risk category captures, why it matters, and examples of cases where elevated risk preceded broader problems at the company here.
How Clients Use the Score
Fundamental Equity Investors
Fundamental investors use the Forensic Risk Score in several ways.
Short sellers use the score for idea generation. Elevated or rising scores highlight companies whose reported performance depends more on accounting choices, financing structures, or other mechanisms than on underlying business execution.
Long-short investors use the score for both idea generation and risk control. High scores identify potential short candidates, while rising scores on existing holdings trigger deeper review and position adjustments.
Long-only investors use the score as a risk management tool. Elevated scores prompt analysts to examine disclosures more closely and inform position sizing and portfolio exposure decisions.
The category breakdown shows analysts where the risk signals appear in the filings. Instead of reviewing the entire document, they can focus on the specific disclosures that drive the score.
Quantitative and Systematic Investors
Quantitative investors use the Forensic Risk Score and its underlying category signals as model features.
Because the signals come from disclosure structure and narrative risk rather than prices or accounting aggregates, they add information that is largely orthogonal to traditional financial factors.
Common uses include:
- Incorporating the score or category signals
- Using the score as a risk control
- Building long–short strategies
- Combining the signals with quality, momentum, or valuation factors
D&O Insurance and Litigation Risk Teams
Directors and Officers (D&O) insurance teams and litigation risk analysts use the Forensic Risk Score as an early-warning indicator of governance and disclosure risk.
Elevated or rising scores predict a higher likelihood of securities class action lawsuits and stock underperformance. This relationship appears consistently in Hudson Labs' own research and has been independently confirmed by actuarial teams at our insurance clients.
Typical uses include:
- Supporting underwriting decisions and premium differentiation
- Identifying companies that warrant deeper governance review
- Monitoring insured portfolios for rising disclosure risk
- Supplementing traditional governance checklists with systematic narrative signals
Risk and Portfolio Oversight Teams
Risk teams use the score to monitor how much of a portfolio is exposed to companies with structurally fragile disclosures or business models.
Because the score updates as disclosures evolve, it can provide early signals of deterioration before problems appear in reported results or guidance.
Common uses include:
- tracking exposure to high-risk percentiles over time
- flagging companies whose scores are rising rapidly
- analyzing which risk categories are driving emerging vulnerabilities
The Hudson Labs Forensic Risk Score provides a systematic way to evaluate structural risk embedded in corporate disclosures. By combining narrative signal extraction, category-level analysis, and machine learning, the score translates complex disclosure patterns into a clear and comparable measure of disclosure-based risk.
Interested in exploring the Hudson Labs forensic risk score?
👉 Book a demo to learn about the institutional tier features.
