PolicyEngine
Research
APIPython
Donate
us
PolicyEngineModel
Overview
PipelineCalibration targetsValidation
Behavioral responses

Data

Microdata pipeline

PolicyEngine constructs its representative household dataset through a 14-step pipeline. Public survey data is merged, stratified, and cloned to 10 geographic variants per household. Each clone is simulated through PolicyEngine US with stochastic take-up, then calibrated via L0-regularized optimization against administrative targets at the national, state, and congressional district levels simultaneously, producing 488 geographically representative datasets.

National
Geography-specific

Census CPS ASEC

Base microdata
Input
~150,000 persons
→
Output
~150,000 persons
Census Bureau CPS ASEC (March supplement)

The Current Population Survey Annual Social and Economic Supplement (CPS ASEC) provides the baseline microdata. The March 2025 supplement covers tax year 2024 with ~200,000 individuals across ~60,000 households.

  • 93+ person-level columns: demographics, income, employment, disability
  • 10 tax unit columns: AGI, federal/state taxes, credits (EITC, CTC, ACTC)
  • 35 SPM unit columns: income, benefits, taxes, poverty thresholds
  • Geographic identifiers: state FIPS, county FIPS, NYC flag
PolicyEngine

PolicyEngine 2026

Compare
PolicyEngine US with…
Pick one or more peers to switch this model page into side-by-side view.