Prism Disclosure Alpha  |  Organizational Psychometrics

Early-warning signals
from public filings

Prism identifies corporate misconduct and deterioration before it surfaces — from the behavioral patterns organizations leave in their own public filings.

Not sentiment. Not keywords. Not ESG.

Request Evaluation View Products
9,156
Companies
508K+
Documents
25 yrs
History
Daily
Updates
Signal Suite

What Prism delivers

Prism operationalizes "tone from the top" into a quantitative, cross-sectional measure. All inputs are public documents. No web scraping, no PII, no MNPI.

Conduct Risk Index R3K
+22.5% alpha
Sharpe 1.30 · t = 5.69
Flagship long/short alpha signal. Walk-forward validated on the Russell 3000 using a 7-factor model (Carhart + QMJ). Negative market beta — the signal is defensive. Alpha survives every standard correction.
16 of 17 formation years positive · Max drawdown −24.1% · p < 0.001 · QMJ loading negative (not repackaged quality)
Prosecution Risk Score S&P 500
AUC 0.944
Nested CV · 95% CI [0.91, 0.97]
DOJ criminal prosecution probability. Identifies firms whose public communications show declining prosocial and moral tone over time — the behavioral signature of enforcement-prone organizations.
Temporal drift signal · 94.7% variance unexplained by sector + size · Bootstrap N = 1,000
Enforcement Risk Index R3K
AUC 0.839
13× separation · 9,156 firms
SEC and DOJ enforcement probability across the full Russell 3000. Designed for compliance screening, D&O insurance risk assessment, and ESG risk integration.
Daily delivery · Russell 3000 universe · Compatible with compliance workflows
Global Coverage
International equity markets under validation. Publication follows the same walk-forward diligence standard as the US indices.
Coming Soon
Tier 2 — Raw Behavioral Panel
Full 57-dimension quarterly panel. 9,156 companies, 2001–2026. Build your own signals from the underlying behavioral data.
Case Study

Boeing vs. NVIDIA, 2016–2023

Boeing — Conduct Risk Percentile

The 737 MAX signal was visible one year before the crash.

Boeing's Conduct Risk percentile peaked at the 94th in 2017 — one full year before the 737 MAX crashes in 2018 and 2019. The signal captured the behavioral deterioration embedded in Boeing's public communications before it materialized in physical outcomes.

NVIDIA shows the contrast. A firm with consistently low conduct risk across the same window, with no directional deterioration. Same instrument. Different organizations.

Boeing peaked at the 94th percentile in 2017. The 737 MAX MCAS failure was disclosed October 2018. The signal preceded the event by four quarters.

Performance

Conduct Risk Index — verified statistics

Walk-forward 7-factor (Carhart+QMJ) regression on the Russell 3000. Formation years 2005–2024, 194 months. All numbers are out-of-sample.

Alpha & Risk
Carhart alpha (annualized) +22.5%
7-factor alpha (Carhart + QMJ) +22.5%
Sharpe ratio 1.30
t-statistic +5.69 p < 0.001
Max drawdown −24.1%
Formation years positive 16 / 17
Worst single year −7.6% (2014)
Factor Independence
Market beta −0.333 defensive
SMB (size) beta −0.130 not size-driven
HML (value) beta −0.022 not value
UMD (momentum) beta −0.062 not momentum
QMJ (quality) beta Negative not repackaged quality
Universe Russell 3000+
Out-of-sample window 194 months
Conduct Risk — Annual Returns (L/S)
'06+25%
'07+3%
'08+13%
'09+4%
'10+14%
'11+37%
'12−10%
'13+23%
'14+9%
'15+40%
'16+26%
'17+6%
'18+9%
'19+11%
'20+48%
'21+16%
'22+10%
'23+27%
'24+19%
'25+22%

Full backtest range shown. Validated walk-forward window (formation years 2005–2024, 194 months) delivers +22.5% net of 7 factors. Past performance is not indicative of future results.

How It Works

Computational behavioral science

Prism applies established psychometric instrument design to the mandatory disclosures organizations make to regulators. Public filings are mandatory cultural performances — they reflect the ethos of the organization that wrote them.

01
Collect
Every SEC filing and earnings call transcript for 9,156 companies. Sources include 10-K, 10-Q, DEF 14A, and earnings transcripts.
02
Score
Proprietary psychometric scoring across 57 behavioral dimensions. Computational behavioral science applied to organizational language. Psychometric instrument design is proprietary.
03
Signal
Behavioral dimensions aggregated into composite risk and alpha signals. Temporal behavioral measurement captures organizational change over time. Delivered via CSV, API, or S3.
Signal Behavioral scores from regulatory filings and earnings transcripts
Scoring method Proprietary psychometric methodology. Instrument design is proprietary.
Scoring Proprietary scoring across filing types and transcripts
Universe US Russell 3000 (9,156 companies)
Dimensions 57 behavioral constructs per company
Sources SEC EDGAR (10-K, 10-Q, DEF 14A) + earnings call transcripts
Rebalance Annual (portfolio); daily scoring updates
Built by PhD in psychology and Columbia Business School postdoctoral researcher
Access

Evaluation and licensing

Evaluation access available on request. All engagements begin with an evaluation package: sample data, dictionary, catalogue, and methodology overview.

Evaluation
Evaluation Package
Sample data, dimension dictionary, signal catalogue, case studies, and methodology overview. Delivered as a structured data package for technical diligence.
Tier 1 — Alpha
Conduct Risk Index
Walk-forward validated alpha signal. Delivered daily via CSV, API, or S3. Prosecution Risk and Enforcement Risk available as compliance overlays.
Tier 2 — Panel
Raw Behavioral Panel
Full 57-dimension quarterly panel for 9,156 companies from 2001 to present. Build proprietary signals from the underlying behavioral data.
Get Started

Request Evaluation

Sample data, dimension dictionary, methodology overview, and case studies. All engagements begin with a structured evaluation package.

prism@stronglight.io