Methodology

Scoring Methodology

The State Payment Integrity Index provides an evidence-based view of financial control risk across U.S. jurisdictions. Each state receives a composite score on a 0–100 scale derived entirely from publicly available data.

Data Sources

State Auditor Reports

Published findings from state auditors of public accounts (NC, CT, VA, TX, FL and expanding). Reports are scraped, parsed via OCR and LLM extraction, and classified by finding type and severity.

Open Expenditure Portals

State transparency portals via Socrata Open Data API and direct data downloads. Vendor payment records are analyzed for anomalies including duplicate payments, unusual vendor concentration, and Benford's Law deviations.

Legislative & Budget Records

Public procurement thresholds, legislative audit mandates, and budget documents that contextualize spending patterns and compliance requirements.

Composite Score

Each jurisdiction's composite risk score blends three dimensions:

45%

Severity

Average severity of audit findings. Material Weakness carries 3x weight, Significant Deficiency 2x, Other 1x.

35%

Frequency

Total distinct findings per jurisdiction, bounded to prevent outlier dominance. Higher counts indicate systemic rather than isolated issues.

20%

Trend

Year-over-year comparison. Increasing finding counts raise the risk weighting; declining counts lower it.

Trend Classification

WorseningRecent 12-month findings exceed prior period by >20%
ImprovingRecent 12-month findings below 80% of prior period
StableWithin 80%–120% band of prior period

Check Taxonomy

Findings are classified against a taxonomy of 104 compliance checks organized across eight categories. Each check maps to a specific contractual or regulatory control that can be verified against source documents.

A

Rate & Pricing

Discrepancies between invoiced rates and current contract rate schedules, including escalation clause misapplication.

B

Scope & Authorization

Work billed outside the authorized scope, period of performance, or without required task order authorization.

C

Billing Mechanics

Mathematical errors, duplicate line items, unit-of-measure mismatches, and fiscal year boundary violations.

D

Subcontractor & Personnel

Unauthorized subcontractors, key personnel substitutions, and labor category misclassification.

E

Acceptance & Deliverables

Missing acceptance documentation, billing for undelivered milestones, and conditional acceptance violations.

F

Compliance & Reimbursables

Unapproved reimbursable expenses, insurance and bonding lapses, and regulatory non-compliance.

K

Financial Controls

Duplicate payments, tax computation errors, retainage withholding failures, and payment timing violations.

P

Pricing & Credits

Discount non-application, credit memo omissions, and cumulative spend threshold violations.

Spend Anomaly Detection

In addition to audit report findings, the Index incorporates statistical anomaly signals derived from public expenditure data. These are patterns that warrant further review, not conclusions of wrongdoing.

Duplicate Payment Detection

Identifies payments to the same vendor for the same amount within a narrow time window, flagging potential processing errors.

Vendor Concentration Analysis

Flags agencies where a single vendor captures a disproportionate share of category spend, indicating potential competitive procurement gaps.

Year-End Spending Patterns

Quantifies fiscal year-end spending acceleration relative to monthly averages, highlighting periods of elevated payment risk.

Benford's Law Distribution

Tests the first-digit distribution of payment amounts against the expected statistical distribution, identifying unusual clustering.

Limitations & Disclaimers

Scores are relative indicators, not absolute quality measures. States with more transparent reporting may show higher risk scores due to greater data availability. We continuously refine the scoring model and expand coverage as additional data sources become available.

Anomaly signals represent statistical patterns that may warrant further review. They are not allegations of fraud, waste, or impropriety. Many flagged patterns have legitimate explanations, such as recurring service contracts that produce identical monthly amounts. All data is sourced from publicly available government records.