Pacific and Southwest/Southeast drought & storm belts run the worst crop-insurance loss ratios; concentration is flat, not trending

Summary

Over 2016-2025, the worst federal crop-insurance loss ratios (indemnities paid divided by premium collected) were run by a cluster of drought- and storm-exposed states rather than any single region: Washington (1.45), Florida (1.36), Texas (1.32), Arkansas (1.32), Utah (1.28), California (1.28), New Mexico (1.24), Arizona (1.22), Georgia (1.15), and Louisiana (1.14) all paid out more in indemnities than farmers and taxpayers paid in premium. These fall into two distinct exposure groups: the arid Southwest/Mountain West (UT, NM, AZ, plus CA and parts of TX) driven by chronic and episodic drought, and the Gulf/Southeast/Pacific tier (FL, GA, LA, WA) driven by hurricanes, excess moisture, and orchard-specific perils. By contrast, the Corn Belt (Indiana 0.31, Illinois 0.31, Iowa 0.44, Wisconsin 0.45, Ohio 0.50, Missouri 0.55, Michigan 0.60) ran comfortably profitable for the program the whole decade.

On concentration: state-level indemnity shares show no statistically meaningful trend toward either more or less concentration over the decade. A Herfindahl-Hirschman Index (HHI), computed with the hhi_concentration tool on each year's state shares of national crop-insurance indemnities, swings between 496 (2016) and 969 (2022) with no consistent direction (regression of HHI on year: slope +7.6/year, r²=0.02, correlation r=0.15 - not distinguishable from noise). Every single year falls in the DOJ/FTC 'unconcentrated' band (HHI<1000). The swings track which specific region got hit by a major drought or storm in a given year (e.g., the 2022 spike is the Texas/Kansas/Southern Plains drought concentrating over $4.6B of that year's indemnities into Texas alone, 23.9% of the national total that year), not a structural drift toward fewer or more states dominating losses. 2024-2025 concentration (548, 520) is back near where the decade started (496, 549 in 2016-2017).

Which regions ran the worst loss ratios

Loss ratio = SUM(indemnity_amount)/SUM(total_premium_amount), summed across all crops, plans, and counties within a state for 2016-2025 (the RMA Summary of Business table's most recent full decade), restricted to states with over $50M in decade premium to avoid noise from tiny programs.

Texas and California carry the largest absolute dollar volumes of any loss-making state ($18.8B and $7.2B indemnity respectively over the decade) - so in dollar terms, the 'worst region' story is really a Texas story reinforced by a Southwest-drought and Gulf-hurricane story. By contrast, the Corn Belt states (IN, IL, IA, WI, OH, MO, MI) all ran loss ratios well under 0.6, meaning the program's Midwest grain book has been subsidizing the payouts concentrated in the drought/hurricane-exposed states.

External corroboration: USDA's own reporting on Pasture, Rangeland and Forage (PRF) insurance - a drought-specific crop-insurance line - found the Southwest region received about 40% of national PRF indemnities from 2017-2023, with Arizona alone receiving 12.3% and Nevada 8.7% of all PRF payments nationwide, consistent with the Southwest states' elevated loss ratios found here.

Is the pattern getting more concentrated or more spread out

To test concentration, we computed the Herfindahl-Hirschman Index (HHI, 0-10,000 scale) via the hhi_concentration tool on each year's state shares of total national crop-insurance indemnities (not loss ratio itself, since loss ratio isn't a summable quantity - indemnity dollars are the right basis for a concentration measure). Each year was run as a separate tool call against that year's state-level indemnity totals:

Every year falls in the DOJ/FTC 'unconcentrated' band (HHI<1000) throughout the decade - crop-insurance losses have never been dominated by a handful of states the way, say, a concentrated industry would be. A regression of HHI on year across these 10 points gives a slope of +7.6 HHI points/year with r²=0.024 and correlation r=0.15 - far too weak to call a trend. The apparent 2018-2022 climb and 2023-2025 decline is better explained by which specific disaster hit which state in a given year (2022's spike to 969 is almost entirely the Southern Plains/Texas-Kansas drought, which alone drove Texas to 23.9% of that year's national indemnities) than by any structural shift toward payouts concentrating in fewer places or spreading across more. By 2024-2025 concentration is back down near 2016-2017 levels. The honest answer is: neither - concentration is episodic and disaster-driven, not trending in either direction over the decade.

Every query behind this report

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

SELECT state_abbr, SUM(indemnity_amount) AS indemnity, SUM(total_premium_amount) AS premium,
SUM(indemnity_amount)/NULLIF(SUM(total_premium_amount),0) AS loss_ratio
FROM ag.rma_crop_insurance
WHERE year BETWEEN 2016 AND 2025
GROUP BY state_abbr
HAVING SUM(total_premium_amount) > 50000000
ORDER BY loss_ratio DESC
LIMIT 15

SELECT state_abbr, SUM(indemnity_amount) AS indemnity, SUM(total_premium_amount) AS premium,
SUM(indemnity_amount)/NULLIF(SUM(total_premium_amount),0) AS loss_ratio
FROM ag.rma_crop_insurance
WHERE year BETWEEN 2016 AND 2025
GROUP BY state_abbr
HAVING SUM(total_premium_amount) > 50000000
ORDER BY loss_ratio ASC
LIMIT 10

SELECT year, state_abbr, SUM(indemnity_amount) AS indemnity
FROM ag.rma_crop_insurance
WHERE year BETWEEN 2016 AND 2025 AND indemnity_amount > 0
GROUP BY year, state_abbr

WITH state_year AS (
  SELECT year, state_abbr, SUM(indemnity_amount) AS indemnity
  FROM ag.rma_crop_insurance
  WHERE year BETWEEN 2016 AND 2025 AND indemnity_amount > 0
  GROUP BY year, state_abbr
),
tot AS (
  SELECT year, SUM(indemnity) AS total_indemnity FROM state_year GROUP BY year
)
SELECT sy.year, SUM(POWER(100.0*sy.indemnity/t.total_indemnity,2)) AS hhi,
  SUM(sy.indemnity) AS total_indemnity
FROM state_year sy JOIN tot t ON sy.year = t.year
GROUP BY sy.year
ORDER BY sy.year

WITH state_year AS (
  SELECT year, state_abbr, SUM(indemnity_amount) AS indemnity
  FROM ag.rma_crop_insurance
  WHERE year BETWEEN 2016 AND 2025 AND indemnity_amount > 0
  GROUP BY year, state_abbr
),
tot AS (
  SELECT year, SUM(indemnity) AS total_indemnity FROM state_year GROUP BY year
),
hhi_tbl AS (
SELECT sy.year AS yr, SUM(POWER(100.0*sy.indemnity/t.total_indemnity,2)) AS hhi
FROM state_year sy JOIN tot t ON sy.year = t.year
GROUP BY sy.year
)
SELECT corr(yr, hhi) AS corr_year_hhi, regr_slope(hhi, yr) AS slope, regr_r2(hhi, yr) AS r2
FROM hhi_tbl

SELECT state_abbr, SUM(indemnity_amount) AS indemnity FROM ag.rma_crop_insurance WHERE year=2016 AND indemnity_amount>0 GROUP BY state_abbr

SELECT state_abbr, SUM(indemnity_amount) AS indemnity FROM ag.rma_crop_insurance WHERE year=2017 AND indemnity_amount>0 GROUP BY state_abbr

SELECT state_abbr, SUM(indemnity_amount) AS indemnity FROM ag.rma_crop_insurance WHERE year=2018 AND indemnity_amount>0 GROUP BY state_abbr

SELECT state_abbr, SUM(indemnity_amount) AS indemnity FROM ag.rma_crop_insurance WHERE year=2019 AND indemnity_amount>0 GROUP BY state_abbr

SELECT state_abbr, SUM(indemnity_amount) AS indemnity FROM ag.rma_crop_insurance WHERE year=2020 AND indemnity_amount>0 GROUP BY state_abbr

SELECT state_abbr, SUM(indemnity_amount) AS indemnity FROM ag.rma_crop_insurance WHERE year=2021 AND indemnity_amount>0 GROUP BY state_abbr

SELECT state_abbr, SUM(indemnity_amount) AS indemnity FROM ag.rma_crop_insurance WHERE year=2022 AND indemnity_amount>0 GROUP BY state_abbr

SELECT state_abbr, SUM(indemnity_amount) AS indemnity FROM ag.rma_crop_insurance WHERE year=2023 AND indemnity_amount>0 GROUP BY state_abbr

SELECT state_abbr, SUM(indemnity_amount) AS indemnity FROM ag.rma_crop_insurance WHERE year=2024 AND indemnity_amount>0 GROUP BY state_abbr

SELECT state_abbr, SUM(indemnity_amount) AS indemnity FROM ag.rma_crop_insurance WHERE year=2025 AND indemnity_amount>0 GROUP BY state_abbr