H-2A Worker Housing

Limitaciones

Federal fiscal years (October 1 - September 30), not calendar years: `year` is the fiscal year in which DOL's Office of Foreign Labor Certification issued its final determination on the H-2A application (the decision date), the basis DOL reports on. FY2026 is PARTIAL: it holds decisions through June 30, 2026 only (DOL's cumulative Q3 file) and will be replaced when DOL posts the full year; do not compare it with complete years. No decision is dated before October 31, 2025 nationally or November 3, 2025 on a job order with Georgia housing, so October 2025 is all but empty - a gap that falls within the October 1 - November 12, 2025 federal government shutdown. The universe is every housing site physically in Georgia listed on an H-2A job order DOL certified in full or in part in the year (certifications DOL marks '(Expired)' included): the primary housing site in Section D of the job order and each additional site in its Addendum B. Housing on withdrawn and denied job orders is excluded (FY2020 35 sites / 1,264 beds, FY2021 105 sites / 2,514 beds, FY2022 59 sites / 1,813 beds, FY2023 55 sites / 2,630 beds, FY2024 62 sites / 2,146 beds, FY2025 52 sites / 1,702 beds, FY2026 33 sites / 1,606 beds). Each fiscal year counts the determinations issued in it, so a job order re-determined in a later year counts again, as in DOL's own figures. 624 FY2020 applications nationally (18 with a Georgia first worksite, 13 of them certified) left every housing field blank (an older version of the form) and list no housing addendum; their housing is unknown and not counted. Beds are the total occupancy capacity the employer listed for each housing site (TOTAL_OCCUPANCY) - capacity listed on certified job orders, not unique beds, beds built, or workers actually housed. A camp listed on several job orders in a year (an employer's successive seasons, or several employers sharing it) counts once per job order, and the listed capacity is a maximum, so Georgia's listed beds exceed the H-2A positions certified at Georgia worksites (dataset h2a_certifications): FY2020 40,036 beds vs 27,614 positions; FY2021 50,206 beds vs 35,205 positions; FY2022 51,931 beds vs 34,974 positions; FY2023 54,493 beds vs 37,536 positions; FY2024 66,485 beds vs 43,436 positions; FY2025 71,648 beds vs 42,723 positions; FY2026 62,105 beds vs 39,586 positions. Units (TOTAL_UNITS) are the buildings, trailers or apartments as the employer counts them. Sites are counted as DOL lists them. 11 sites list 0 beds and 1 site lists 0 units with beds (e.g. one barracks); they are kept as filed. The largest single listing is 420 beds (listed in FY2024, FY2025, FY2026); large camps are real and kept. Some listings repeat another listing of the same job order exactly (same street address, city, ZIP, units and beds) - FY2020 21 (520 beds); FY2021 10 (326 beds); FY2022 10 (674 beds); FY2023 11 (775 beds); FY2024 28 (1,480 beds); FY2025 31 (1,384 beds); FY2026 15 (724 beds) - and are counted as listed, because DOL's record cannot tell a second identical building at one address from a double entry; they bound the possible double counting. DOL's own count of a job order's housing records (TOTAL_HOUSING_RECORDS) is not used: the record layout defines it as the Addendum B rows, but on 4,343 of the 4,623 certified job orders with Georgia housing it equals the addendum rows plus the primary site, on 186 the addendum rows alone, and on the rest neither. A site is in Georgia by its own address (the housing state), whatever the job order's first worksite: over all years 173 Georgia sites (6,769 beds, 102 job orders) belong to job orders whose first worksite is in another state, and are counted; 22 certified job orders with a Georgia first worksite list no Georgia housing (13 of them left the housing fields blank; the rest list their primary housing in FL, LA, SC) and are not counted. So this table's job orders are not the same set as h2a_certifications' applications. The state field decides: 1 site (3 beds) whose ZIP code and county are Georgia's but whose state field names another state (FY2021 Preston, stated LA) is not counted. County is DOL's own housing county (HOUSING_COUNTY on the main file, PHYSICAL_LOCATION_COUNTY on the addendum), corrected only by the certain-only ZIP rule the worksite topics also use; housing street addresses are not geocoded, unlike those topics' worksite counties. Where DOL's housing county is impossible for the housing ZIP code - none of the ZIP's 2020 population lives in it - it is replaced only when the right one is certain: at least 99% of the ZIP's population lives in one other county, and the housing city is the ZIP's postal name or a Census place in that county, and not also a place or postal name in the stated county. 68 sites (2,515 beds) were moved this way (FY2020 5, FY2021 18, FY2022 9, FY2023 10, FY2024 13, FY2025 9, FY2026 4). Another 105 sites (4,835 beds) keep a stated county that is impossible for their ZIP, because the evidence falls short of certain: in 79 the ZIP straddles a county line, in 24 the city does not match the ZIP (the ZIP itself may be mistyped) and in 2 the city lies on both sides of the county line. Their county may be wrong. DOL left the housing county blank on 360 sites (12,455 beds): FY2025 321 (317 of them on job orders decided from July 1, the last-quarter gap that also blanks FY2025 worksite counties; the rest decided from April 3, 2025), and a few in other years (FY2020 4, FY2021 18, FY2022 8, FY2023 5, FY2024 4). DOL's FY2026 Addendum B file gives one county per job order, copied onto every additional site (no job order in it lists two counties; in earlier years 81%-85% of job orders with sites in three or more ZIP codes list more than one county), so for those 862 sites (25,561 beds) the county is not used. One FY2021 site in Waycross names Hampden County, Massachusetts, and its county is not used. Those sites take the county holding most of their housing ZIP code's 2020 population: 1,203 sites (37,483 beds), 738 of them in ZIP codes that straddle a county line (dominant share as low as 44%); 20 sites with no usable ZIP count in the state row only (never hand-mapped). Where DOL does name the county (as corrected above), the ZIP's dominant county disagrees with it on 622 of the 8,883 Georgia sites whose ZIP is in the crosswalk (7.0%), so expect roughly that share of the ZIP-placed sites to sit in a neighboring county. County rows sum to the state row for beds, units and sites except for the unplaced sites (FY2022 1 site, 15 beds, 1 unit; FY2025 8 sites, 304 beds, 39 units; FY2026 11 sites, 216 beds, 11 units). num_applications is not additive across counties: a job order with housing in two counties counts in each, so the county rows sum to more job orders than the state row. DOL's FY2020/FY2021 main files repeat 6 case numbers (none with Georgia housing); the latest determination is kept. No suppression: rows are housing sites on job orders, not people, so the domain's county small-cell rule does not apply, small county counts are published, and every zero is a real zero. Housing type, inspection fields, addresses and employer identity are not published here.