Meatpacking counties are economically concentrated, not demographically diverse — and that gap is exactly what makes an enforcement raid so disruptive

Summary

Nationally, meatpacking is one of the most foreign-born-dependent industries in the U.S. economy: about 45% of meatpacking workers were foreign-born in 2020 (American Immigration Council analysis of Census data), roughly 28 percentage points above the all-industry average, and individual plants have run far higher — an estimated 80–90% foreign-born at JBS's Greeley, Colorado beef plant, for example. But the county a plant sits in usually does not look demographically dependent on immigrants: our own analysis of the 15 U.S. counties with the most meatpacking (NAICS 3116) employment finds that the counties where meatpacking dominates the local job base (Scott Co., MS at 43.6% of private jobs; Grady Co., GA at 42.9%; Duplin Co., NC at 26.4%) actually have below-average county-wide foreign-born population shares (4.6%–11.8%), because the immigrant workforce there is concentrated specifically inside the plant, not spread across the county's other employers. Exposure to a sudden enforcement action is therefore best measured two ways at once, and they point in opposite directions: (1) how much of the local economy rides on the plant — which our county data can measure directly and which is highest in small rural counties in Mississippi, Georgia, North Carolina, and Alabama — and (2) how immigrant-dependent the plant's own workforce is, which no county-level statistic captures and which published labor-market research puts at roughly 20%–90% depending on the facility, with undocumented workers alone estimated at 30–50% of the industry. A single raid can remove the majority of a plant's line staff in a morning — ICE's July 2025 raid on Glenn Valley Foods in Omaha detained more than 100 of the plant's roughly 150 workers — even in a county (Douglas Co., NE) where meatpacking is a small enough share of the overall job base (under 2%) that a county-level economic-exposure score would have rated it low-risk.

1. What the national research says about foreign-born dependency

Meatpacking's reliance on immigrant labor is well documented in labor-market literature, independent of any single company or plant:

• 45.4% of meatpacking workers were foreign-born as of 2020 — 28 points above the all-industry average — per the American Immigration Council's analysis of Census microdata (Tending to America's Food Supply).
• Across the broader animal-agriculture/livestock workforce (a wider category than meatpacking alone), roughly 20% are foreign-born nationally, per the same report, versus 17.4% for all U.S. industries.
• Undocumented workers specifically are estimated at 30–50% of the meatpacking workforce in some accounts, and up to 60–90% in certain states for related animal-processing industries (FarmSTAND — an advocacy-affiliated source; treat as an upper-bound estimate rather than a government statistic).
• Plant-specific figures run even higher in the historical record: an estimated 80–90% of workers at the JBS Greeley, Colorado plant were foreign-born as of 2020 (High Country News).
• The industry has repeatedly turned to H-2B temporary visas to fill shortages — the Department of Labor issued more than 20,000 such visas to meat-industry employers in 2021 (Investigate Midwest), itself evidence of how hard domestic-only staffing is for this industry.

No table in this warehouse carries industry-specific nativity microdata (ACS Public Use Microdata, which is where these figures ultimately come from, is not loaded at the detail needed to isolate NAICS 3116 workers by nativity at county grain) — these national and plant-level figures come from the published literature, not a warehouse computation, and are reported as such rather than re-derived.

2. Where the plants are, and how concentrated they are locally (BLS QCEW, 2023)

Using BLS Quarterly Census of Employment and Wages (county x NAICS 3116, Animal Slaughtering and Processing, private ownership, 2023), the counties with the largest raw meatpacking/poultry-processing workforces are led by Hall Co., GA (Gainesville poultry cluster, 9,090 jobs), Cook Co., IL (8,226), Fresno Co., CA (7,320), Sussex Co., DE (7,096), and Marshall Co., AL (6,954). But raw headcount understates local dependency in a big metro and overstates it nowhere — what matters for exposure is what share of the county's private-sector jobs the plant workforce represents. On that measure the most exposed counties are:

Nationally, this employment is spread thin geographically — an HHI (Herfindahl-Hirschman Index) of the county-level distribution of BLS-disclosed NAICS 3116 employment computes to 192, in the “unconcentrated” DOJ/FTC band, with the single largest county (Hall Co., GA) holding only 4.9% of the disclosed national total. That reads as spread-out, low-concentration risk at the national level, but it hides an important data problem described next.

3. A structural data gap: the highest-exposure counties are the ones BLS won't disclose

Of the 1,774 county x NAICS-3116 employment cells in BLS QCEW for 2023, 1,503 (85%) are suppressed (disclosure_code = 'N') because employment in that county's meat-processing sector is dominated by one or two employers — disclosing the number would reveal that single company's headcount. This is not a random gap: it is precisely the small, single-employer-dominated rural counties — the ones where a plant closing or a raid would be most locally devastating — that BLS hides from public view. The 271 counties with disclosed employment sum to only 183,715 workers, well under the industry's true national total (BLS's national, non-suppressed NAICS 3116 employment count runs closer to 500,000). Any county-level ranking built from this table, including the one above, is therefore a lower-bound, disclosed-only picture that systematically undercounts the smallest, most single-employer, most exposed communities.

What every downstream figure in this report excludes, and why: every county-level statistic in Sections 2 and 4 (the top-employment rankings, the concentration percentages, the correlation coefficient, and the HHI) is computed only over the subset of county x NAICS-3116 cells with a real, unsuppressed annual_avg_emplvl value (271 of 1,774 cells, 15%) — the WHERE annual_avg_emplvl IS NOT NULL filter. Section 4's correlation additionally requires a matched, non-null ACS foreign-born figure for the same county (WHERE n.fb_share IS NOT NULL and county IS NOT NULL on the ACS side), narrowing the correlation sample from 271 disclosed counties to 253. The 1,503 excluded (suppressed) county-industry cells are not a random subset: they are disproportionately the smallest, most single-employer-dominated counties — exactly the highest-exposure profile this report is trying to identify. We cannot compute what the headline concentration or correlation figures would be with those cells included, because BLS never publishes the underlying employment number for them; the counties still appear in the ACS-side nativity data, they simply cannot be given a meatpacking-employment-share value here. This means the true population of most-exposed counties is almost certainly larger than the disclosed set reported in Sections 2 and 4, and may skew toward a distinct nativity composition than that set.

4. The core finding: county population diversity is not a good proxy for plant-workforce dependency — and it runs the wrong way

Across the 253 counties where BLS discloses meatpacking employment and Census reports a matched foreign-born share (2023), the correlation between a county's meatpacking-employment concentration (share of private jobs) and its overall foreign-born population share is weakly negative (r = -0.10, a plain Pearson correlation computed in-query, not a regression or causal design). Counties where meatpacking employment exceeds 5% of local private jobs (n=19) average a county-wide foreign-born share of just 6.4%, compared to 10.2% in counties where the industry is a smaller share of the local economy — largely because the highest-concentration counties are small, rural, and otherwise not very varied in their economic base (their other employers — retail, schools, local government — are staffed mostly by native-born residents), while large metros with generally high foreign-born population shares (Los Angeles, Cook Co./Chicago, Dallas) have meatpacking as a rounding error of their job base (0.1%–0.4% of jobs).

What this means for exposure: a county-wide nativity statistic is the wrong instrument for gauging a plant's foreign-born dependency. The industry's real foreign-born reliance lives inside the plant's own workforce and is invisible at the county-population level — the published 45% (national) to 80–90% (specific plants) figures from Section 1 have to be paired with a locally-specific exposure measure like the employment-concentration ratio in Section 2, not with the county's ACS nativity rate.

5. What a sudden enforcement action does in practice

ICE's July 2025 raid on Glenn Valley Foods in south Omaha (Douglas Co., NE) detained more than 100 of the plant's roughly 150 employees in a single morning — a manager described the plant as nearly impossible to operate that same day (Barchart/AP coverage). Notably, Douglas County's own meatpacking-industry employment share in our data is under 2% of local private jobs — a low score on the local-economic-exposure metric in Section 2 — illustrating the limit of any county-grain measure: exposure is ultimately an establishment-level fact (how concentrated is this one employer's own line staff) that a county aggregate, however carefully built, cannot fully see when the county itself is economically varied even though the plant is not. The 2018 Southeastern Provisions raid in rural Bean Station, TN removed roughly 100 workers, similarly gutting a single facility's line staff even though that plant sits in a small county (Wikipedia summary of coverage).

Putting it together, a plant is most exposed to a sudden enforcement action when it combines (a) a documented or plausible high foreign-born/undocumented share of its own line workforce — the published range across the industry (20%-90%, with 30-50% undocumented in many facilities) — with (b) high concentration of local jobs in the plant itself, which determines how much of the surrounding community's economy and tax base would be disrupted if that workforce were suddenly removed. On the second axis, the most exposed communities identified here are Scott Co., MS; Grady Co., GA; Duplin Co., NC; Marshall Co., AL; Sioux Co., IA; and similar small rural counties in Nebraska, Iowa, Delaware, Virginia, and North Carolina — and the true list is almost certainly longer than what BLS discloses, given that the smallest, most single-employer counties are exactly the ones BLS suppresses.

What This Report Does Not Answer

• Broader food-processing workforce beyond meatpacking specifically: This corpus and the cited literature focus on meatpacking/animal-slaughtering (NAICS 3116) and the closely related livestock/dairy workforce; a comparably detailed foreign-born breakdown for broader food manufacturing (NAICS 311 overall, e.g. bakeries, dairy processing, produce packing) was not separately sourced. The AIC livestock figure (~20% foreign-born) is reported as the closest available proxy for the wider category.

Every query behind this report

7 warehouse calls ran in this session, in order. Each is reproducible against the same snapshot.

SELECT area_fips, annual_avg_emplvl, annual_avg_estabs, annual_avg_wkly_wage
FROM econ.county_qcew
WHERE industry_code = '3116' AND own_code = '5' AND "year" = '2023'
ORDER BY annual_avg_emplvl DESC NULLS LAST
LIMIT 25

WITH meat AS (
  SELECT area_fips, annual_avg_emplvl AS meat_emp, annual_avg_estabs AS meat_estabs
  FROM econ.county_qcew
  WHERE industry_code = '3116' AND own_code = '5' AND "year" = '2023'
),
total AS (
  SELECT area_fips, annual_avg_emplvl AS total_emp
  FROM econ.county_qcew
  WHERE industry_code = '10' AND own_code = '5' AND "year" = '2023'
),
nat AS (
  SELECT state, county, foreign_born, total_population,
         CAST(foreign_born AS DOUBLE)/NULLIF(total_population,0) AS fb_share
  FROM census.acs_nativity
  WHERE "year" = '2023' AND county IS NOT NULL
),
cty AS (
  SELECT county_fips, county_name, state_fips FROM geo.counties WHERE "year" = '2023'
)
SELECT m.area_fips, c.county_name, c.state_fips, m.meat_emp, m.meat_estabs,
       t.total_emp, ROUND(100.0*m.meat_emp/NULLIF(t.total_emp,0),1) AS pct_of_private_emp,
       n.total_population, ROUND(100.0*n.fb_share,1) AS county_fb_pct
FROM meat m
JOIN total t ON m.area_fips = t.area_fips
LEFT JOIN cty c ON c.county_fips = m.area_fips
LEFT JOIN nat n ON n.state = SUBSTR(m.area_fips,1,2) AND n.county = SUBSTR(m.area_fips,3,3)
ORDER BY m.meat_emp DESC
FETCH FIRST 20 ROWS ONLY

WITH meat AS (
  SELECT area_fips, annual_avg_emplvl AS meat_emp, annual_avg_estabs AS meat_estabs
  FROM econ.county_qcew
  WHERE industry_code = '3116' AND own_code = '5' AND "year" = '2023'
    AND annual_avg_emplvl IS NOT NULL
),
total AS (
  SELECT area_fips, annual_avg_emplvl AS total_emp
  FROM econ.county_qcew
  WHERE industry_code = '10' AND own_code = '5' AND "year" = '2023'
),
nat AS (
  SELECT state, county, foreign_born, total_population,
         CAST(foreign_born AS DOUBLE)/NULLIF(total_population,0) AS fb_share
  FROM census.acs_nativity
  WHERE "year" = '2023' AND county IS NOT NULL
),
cty AS (
  SELECT county_fips, county_name, state_fips FROM geo.counties WHERE "year" = '2023'
)
SELECT m.area_fips, c.county_name, c.state_fips, m.meat_emp, m.meat_estabs,
       t.total_emp, ROUND(100.0*m.meat_emp/NULLIF(t.total_emp,0),1) AS pct_of_private_emp,
       n.total_population, ROUND(100.0*n.fb_share,1) AS county_fb_pct
FROM meat m
JOIN total t ON m.area_fips = t.area_fips
LEFT JOIN cty c ON c.county_fips = m.area_fips
LEFT JOIN nat n ON n.state = SUBSTR(m.area_fips,1,2) AND n.county = SUBSTR(m.area_fips,3,3)
ORDER BY m.meat_emp DESC
FETCH FIRST 20 ROWS ONLY

SELECT COUNT(*) AS n_counties, SUM(annual_avg_emplvl) AS total_emp,
       SUM(CASE WHEN disclosure_code='N' THEN 1 ELSE 0 END) AS n_suppressed
FROM econ.county_qcew
WHERE industry_code = '3116' AND own_code = '5' AND "year" = '2023'

SELECT area_fips, annual_avg_emplvl AS meat_emp
FROM econ.county_qcew
WHERE industry_code = '3116' AND own_code = '5' AND "year" = '2023' AND annual_avg_emplvl IS NOT NULL

WITH meat AS (
  SELECT area_fips, annual_avg_emplvl AS meat_emp
  FROM econ.county_qcew
  WHERE industry_code = '3116' AND own_code = '5' AND "year" = '2023' AND annual_avg_emplvl IS NOT NULL
),
total AS (
  SELECT area_fips, annual_avg_emplvl AS total_emp
  FROM econ.county_qcew
  WHERE industry_code = '10' AND own_code = '5' AND "year" = '2023'
),
nat AS (
  SELECT state, county, CAST(foreign_born AS DOUBLE)/NULLIF(total_population,0) AS fb_share, total_population
  FROM census.acs_nativity
  WHERE "year" = '2023' AND county IS NOT NULL
)
SELECT m.area_fips, 100.0*m.meat_emp/NULLIF(t.total_emp,0) AS pct_meat_emp, 100.0*n.fb_share AS fb_pct, n.total_population
FROM meat m JOIN total t ON m.area_fips=t.area_fips
LEFT JOIN nat n ON n.state=SUBSTR(m.area_fips,1,2) AND n.county=SUBSTR(m.area_fips,3,3)
WHERE n.fb_share IS NOT NULL

WITH meat AS (
  SELECT area_fips, annual_avg_emplvl AS meat_emp
  FROM econ.county_qcew
  WHERE industry_code = '3116' AND own_code = '5' AND "year" = '2023' AND annual_avg_emplvl IS NOT NULL
),
total AS (
  SELECT area_fips, annual_avg_emplvl AS total_emp
  FROM econ.county_qcew
  WHERE industry_code = '10' AND own_code = '5' AND "year" = '2023'
),
nat AS (
  SELECT state, county, CAST(foreign_born AS DOUBLE)/NULLIF(total_population,0) AS fb_share, total_population
  FROM census.acs_nativity WHERE "year" = '2023' AND county IS NOT NULL
),
j AS (
  SELECT m.area_fips, 100.0*m.meat_emp/NULLIF(t.total_emp,0) AS pct_meat_emp, 100.0*n.fb_share AS fb_pct, n.total_population
  FROM meat m JOIN total t ON m.area_fips=t.area_fips
  LEFT JOIN nat n ON n.state=SUBSTR(m.area_fips,1,2) AND n.county=SUBSTR(m.area_fips,3,3)
  WHERE n.fb_share IS NOT NULL
)
SELECT corr(pct_meat_emp, fb_pct) AS r_all, corr(pct_meat_emp, total_population) AS r_pop,
       AVG(CASE WHEN pct_meat_emp>5 THEN fb_pct END) AS avg_fb_high_conc,
       AVG(CASE WHEN pct_meat_emp<=5 THEN fb_pct END) AS avg_fb_low_conc,
       COUNT(CASE WHEN pct_meat_emp>5 THEN 1 END) AS n_high_conc
FROM j