Program Allocation Disparity Index (PADI)

GMI-AGR-052Agriculture and Farm OperationsIntegration-RequiredPattern O-16: Algorithmic Equity Bias

Definition

The residual disparity in program eligibility, credit decisions, carbon enrollment and credit issuance, and claim outcomes under AI models across farm size, tenure (owner versus tenant), production system, geography, and operator population, after controlling for the risk and eligibility factors the program legitimately uses.

The Measurement Gap

There is no metric that tells you how program eligibility, credit, and claim outcomes under AI models differ across farm size, tenure, system, and operator population after controlling for risk, which means allocation disparity is produced by the model and invisible in the approval rate.

Formula

PADI = max over populations p of |rate_p,adjusted − rate_reference,adjusted| / rate_reference,adjusted, where adjusted rates come from a model controlling for the program's documented eligibility and risk factors; the population with the maximum is named

Proposed formula, pending validation pilots.

At a Glance

Unit
Difference
Trend Direction
Lower; approaching 0
Category
Economic Alignment and Fairness

Decision Supported

Whether a model's allocation outcomes are produced by risk or by the model, and which populations need a manual review lane or a model audit before the next allocation cycle.

← Detection-Oriented Practice Distortion (DOPD)Recommendation Crowding Amplification Index (RCAI) →

Assess Your Organization

The AI-GPS Assessment evaluates your AI governance maturity and recommends KPIs like this one for your implementation roadmap.

Take the Free AI-GPS Assessment