docs/us_economy/economy_poverty_county_year

economy_poverty_county_year

Census SAIPE official county poverty, 2019-2023. child_poverty_rate (ages 0-17) is the CHR "Children in Poverty" Select Measure, surfaced as a plain variable; overall poverty rate and median household income come from the same model. Tier B (model-based estimate).

tier bTier B: SAIPE is a model-based small-area estimate combining federal income tax records, SNAP benefit data, the ACS, and the Census population base. It is the Census Bureau's official county poverty series but is modeled, not a direct count.povertychild-povertyincomesdohsaipe
grain
county-year
years
2019 – 2023
cadence
Annual
overview

overview

~3,140 counties x 5 years (2019-2023).

current vintage — 2023

history — Yes — 1989–2023 (this table loads 2019–2023)

provenance

source & licensing

authority
U.S. Census Bureau
dataset
Small Area Income and Poverty Estimates (SAIPE)
license
Public domain (U.S. Federal Government work)
citation
U.S. Census Bureau, Small Area Income and Poverty Estimates (SAIPE), 2019-2023.
schema

fields

nametypedefinition
_ingested_attimestampTimestamp when this record was written to BigQuery.Pipeline metadata field.
_ingestor_versionstringVersion of the ingestor script that produced this record.Pipeline metadata field.
_reliability_tierstringData quality tier (A/B/C) assigned at ingestion.Pipeline metadata field.
_source_dataset_idstringInternal dataset identifier linking to meta.datasets.Pipeline metadata field.
child_poverty_countint64 · countSAIPE-modeled number of related children ages 0–17 living in poverty in the county-year (SAEPOV0_17_PT). Higher values indicate more children in poverty.
child_poverty_coveragestringCoverage flag for child_poverty_rate: SUPPORTED when a source value is present, UNSUPPORTED when the source value is missing. Values are never imputed.
child_poverty_ratefloat64 · percent (0–100)SAIPE-modeled percent of related children ages 0–17 living in poverty (SAEPOVRT0_17_PT); this is the County Health Rankings "Children in Poverty" measure surfaced as a plain county-year variable. Higher values indicate greater child poverty.
country_idstringISO alpha-2 country code (always 'US' for domestic tables).
county_idkeystring5-character FIPS code identifying the county.Part of primary key. Joins dim.counties on county_id.
county_namestringHuman-readable county name corresponding to county_id.
geo_basisstringGeography provenance flag: 'source' for rows at SAIPE's native canonical geography; 'crosswalk_estimate' for Connecticut (2019-2021) and Alaska Valdez-Cordova (2019) legacy-geography rows reallocated to modern successors via the population-weighted legacy-county crosswalk (us_dimension.crosswalk_legacy_county).
median_household_incomeint64 · dollars (nominal)SAIPE-modeled median household income for the county-year (SAEMHI_PT). Higher values indicate greater household income.
poverty_rate_all_agesfloat64 · percent (0–100)SAIPE-modeled percent of people of all ages living below the federal poverty line in the county-year (SAEPOVRTALL_PT). Higher values indicate greater poverty.
state_idstring2-character FIPS code identifying the state.Joins dim.states on state_id.
yearkeyint64Reference year of the observation.Part of primary key.
relationships

joins

primary key
county_id, year
common joins
not provided
usage

how to use this table

method

SAIPE child poverty rate = modeled percent of related children ages 0-17 in poverty. Carried through as published (percent). Median household income is the modeled SAIPE estimate. No aggregation or weighting applied.

do not use for

Sub-county geographies, single-year change detection in very small counties (model smoothing), or as a direct count (these are modeled estimates).

known issues

Connecticut and the Alaska Valdez-Cordova split span a county-equivalent geography change. SAIPE's 2019-2021 CT rows use the legacy 8 counties (09001-09015) and its 2019 AK row uses Valdez-Cordova (02261); none map 1:1 into the current planning-region / successor-borough geography. Rather than drop them, these rows are reallocated to the modern successors via us_dimension.crosswalk_legacy_county (exact 2020 town-population weights — rates population-weighted, counts split by population share) and flagged geo_basis='crosswalk_estimate'. All other rows are geo_basis='source'. 2022+ CT and 2020+ AK are reported natively by SAIPE and pass through unchanged.

last updated · Jul 7, 2026