500 simulated revenue draws for each observed FCIP aggregated transaction, generated using the RMA M-13 framework with ADM parameters, ensuring all menu alternatives are evaluated under the same stochastic scenarios.

Official Pasture, Rangeland, Forage Pilot Insurance Program Data

📘 Overview

This release provides analysis-ready datasets supporting research on
Pasture, Rangeland, Forage (PRF) Insurance Program within the U.S.
Federal Crop Insurance Program (FCIP).

The data are generated and maintained by the rfcipPRF workflow and are
designed to support reproducible research using PRF rainfall index,
actuarial, and related spatial data products.

📁 Included Files

1. Historic CPC precipitation data

• cpc_historic_precipitation.rds
  Historic grid-level precipitation data used in PRF support
  workflows.

Columns

• grid_id = PRF grid identifier (numeric GRIDCODE from the
  official RMA RI grid shapefile)
• commodity_year = Commodity year (integer, e.g.,
  2011–present)
• month = Calendar month (integer 1–12)
• precipitation = Observed CPC CONUS 0.25° gauge-based daily
  precipitation summed to monthly totals for the grid cell, converted
  to inches
• precipitation_rma = RMA-adjusted monthly precipitation; scaled
  so that the interval-level sum matches the official RMA rainfall
  index (actual_index × long-run average precipitation)

2. County-level PRF actuarial data

• prf_adm_county_level_2007.rds through
  prf_adm_county_level_2026.rds
  Annual county-level PRF actuarial files containing rating and
  payment-factor information by county, interval, type, and coverage
  level.

Columns

• commodity_year = Commodity year (integer, 2007–present)
• state_code = FIPS state code (2-digit integer)
• county_code = FIPS county code (3-digit integer)
• interval_code = PRF insurance interval code (integer 1–11;
  each code represents a distinct 2-month precipitation window — see
  interval reference below)
• type_code = PRF type code (e.g., "RI" for Rainfall Index,
  "VI" for Vegetation Index)
• coverage_level_percent = Coverage level percentage (numeric;
  e.g., 70, 75, 80, 85, 90)
• base_rate = Effective base premium rate derived as
  total_premium_amount / liability_amount from the Summary of
  Business (SoB) TPU data
• payment_factor = Effective payment factor derived as
  indemnity_amount / liability_amount from the SoB TPU data;
  reflects realized loss experience at the county level

3. Grid-level PRF actuarial data

• prf_adm_grid_level_2011.rds through
  prf_adm_grid_level_2026.rds
  Annual grid-level PRF actuarial files combining grid identifiers,
  county-level values, actuarial rates, key program dates, and subsidy
  information.

Columns

• commodity_year = Commodity year (integer, 2011–present)
• state_code = FIPS state code (2-digit integer)
• county_code = FIPS county code (3-digit integer)
• grid_id = PRF grid identifier (numeric GRIDCODE from the
  official RMA RI grid shapefile)
• interval_code = PRF insurance interval code (integer 1–11; see
  interval reference below)
• type_code = PRF type code (e.g., "RI" for Rainfall Index,
  "VI" for Vegetation Index)
• coverage_level_percent = Coverage level percentage (numeric;
  e.g., 70, 75, 80, 85, 90)
• county_base_value = County base value (dollars per acre) used
  as the dollar amount of coverage per acre in PRF liability
  calculations
• base_rate = Base premium rate (mean of grid-level ADM rates
  within the county × interval × type × coverage-level cell)
• payment_factor = Payment factor used in PRF indemnity
  calculations (mean of grid-level ADM payment factors)
• contract_change_date = Contract change date
• sales_closing_date = Sales closing date
• modified_sales_closing_date = Modified sales closing date
• extended_sales_closing_date = Extended sales closing date
• earliest_planting_date = Earliest planting date
• final_planting_date = Final planting date
• extended_final_planting_date = Extended final planting date
• acreage_reporting_date = Acreage reporting date
• modified_acreage_reporting_date = Modified acreage reporting
  date
• end_of_insurance_date = End of insurance date
• cancellation_date = Cancellation date
• modified_cancellation_date = Modified cancellation date
• termination_date = Termination date
• premium_billing_date = Premium billing date
• last_released_date = Last released date
• released_date = Release date
• deleted_date = Deletion date
• filing_date = Filing date
• subsidy_percent = Premium subsidy percentage (fraction of
  total premium paid by the Federal government)

4. Official RMA PRF grid

• official_RMA_RI_grid_01.zip
  The official RMA Rainfall Index grid shapefile defining the spatial
  boundaries of all PRF grid cells across the conterminous United
  States.

Contents of zip

The zip archive contains a point/polygon shapefile with one record per
PRF grid cell. Key attribute:

• GRIDCODE = PRF grid identifier (integer); used as grid_id
  throughout all other release files

Source: Grazing Management Systems. official_RMA_RI_grid, Edition
1.0, published 2009-08-22.
Online: http://prfri-rma-map.tamu.edu/default.aspx

5. RMA Rainfall Index data

• rmaRainfallIndices.rds
  Grid- and interval-level RMA Rainfall Index values derived from the
  latest official RMA PRF rainfall index archive. One row per grid ×
  insurance plan × commodity × year × interval combination.

Columns

• grid_id = PRF grid identifier (numeric GRIDCODE)
• insurance_plan_code = RMA insurance plan code (integer; 13 =
  PRF Rainfall Index, 14 = PRF Vegetation Index)
• Commodity0088 = Commodity 88 indicator (1 if the grid
  participates in the PRF program under commodity code 88)
• Commodity1191 = Commodity 1191 indicator (analogous flag for
  commodity code 1191)
• commodity_year = Commodity year (integer)
• interval_code = PRF insurance interval code (integer 1–11; see
  interval reference below)
• actual_index = Mean realized rainfall index value for the grid
  × interval × year (numeric; expressed as a fraction of the long-run
  average, where values < 1 indicate below-average precipitation)

6. PRF Summary of Business (SoB TPU)

• prf_sobtpu.rds
  County-level PRF insurance Summary of Business (TPU) records
  aggregated from USDA RMA for commodity years 2007 to present.
  Filtered to insurance plan code 13 (PRF) and commodity code 88.

Columns

• commodity_year = Commodity year (integer, 2007–present)
• state_code = FIPS state code (2-digit integer)
• county_code = FIPS county code (3-digit integer)
• type_code = PRF type code
• interval_code = PRF insurance interval code (integer 1–11;
  renamed from practice_code)
• interval_name = PRF interval name (character; renamed from
  practice_name)
• coverage_level_percent = Coverage level percentage (numeric;
  e.g., 70, 75, 80, 85, 90)
• insured_acres = Total net reporting level amount (insured
  acres; sum of net_reporting_level_amount)
• liability_amount = Total liability amount (dollars)
• total_premium_amount = Total premium amount (dollars)
• subsidy_amount = Total subsidy amount (dollars; Federal
  portion of the premium)
• indemnity_amount = Total indemnity amount (dollars; payments
  made to policyholders)

7. Potential PRF area by grid and county

• potential_prf_grids.rds
  Grid-level estimate of land eligible for PRF coverage within each
  PRF grid cell and county, measured in acres. Derived by intersecting
  USDA Forest Service Rangelands V1, USDA NASS Cropland Data
  Layer (CDL) pasture/forage classes, and USGS 2010 population density
  (to exclude developed areas). One row per state × county × grid
  combination.

Columns

• state_code = FIPS state code (2-digit integer)
• county_code = FIPS county code (3-digit integer)
• grid_id = PRF grid identifier (numeric GRIDCODE)
• potential_pasture = Estimated acres of CDL-classified pasture,
  grassland, shrubland, or hay within the grid cell and county
  (excluding populated areas)
• potential_rangeland = Estimated acres of USDA Forest Service
  Rangelands V1 rangeland or transitional rangeland within the grid
  cell and county (excluding populated areas)
• potential_range_pasture = Estimated acres of co...

Various items aggregated from FCIP Data

📘 Overview

This release provides analysis-ready datasets supporting research on
supplemental crop insurance products within the U.S. Federal Crop
Insurance Program (FCIP).

The data are generated and maintained by the fcipSupplementalLab
workflow and are designed to support:

• Measurement of supplemental insurance availability across crops and
  regions
• Analysis of adoption patterns over time
• Agent-level participation analysis for upplemental crop insurance
  products
• Policy analysis related to FCIP subsidy reforms and program design

All files are serialized as .rds objects and intended for direct use
in R. Most users do not need to regenerate these data.

📁 Included Files

1) Reproducible study environment

A serialized study-environment object created by setup_environment()
to standardize the data build across machines and runs. The object
records:

• Analysis window (year_beg, year_end)
• Fixed random seed
• Project name
• Required local directory structure

This file is reloaded by subsequent build stages to ensure
reproducibility.

File:
study_environment.rds

2) Cleaned RMA Summary of Business data (supplemental plans)

An analysis-ready SOB-TPU dataset restricted to add-on /
supplemental insurance products, harmonized across years and coding
schemes.

Key features: - Acres-only records - Supplemental insurance plans
(including SCO and ECO) - Harmonized insurance plan codes - Supplemental
plan share calculations

This dataset is the backbone for adoption, availability, and agent-level
analysis.

File:
cleaned_rma_sobtpu.rds

3) Annual ADM tables for SCO and ECO

A collection of year-specific Actuarial Data Master (ADM) extracts
capturing supplemental insurance availability, including expanded
coverage variants (e.g., SCO-88 and SCO-90).

Each file corresponds to a single commodity year.

Files:
cleaned_rma_adm_supplemental_<year>.rds
(e.g., cleaned_rma_adm_supplemental_2023.rds)

4) Agent-level supplemental insurance datasets

Annual agent-level datasets linking licensed FCIP agents to:

• SCO availability
• ECO 90 percent participation
• ECO 95 percent participation

Agent records are constructed by merging cleaned ADM data with SOB
adoption information for the corresponding year.

Files:
agentdata_<year>.rds
(e.g., agentdata_2022.rds)

5) Panel of supplemental insurance availability and adoption

A longitudinal panel dataset capturing:

• Whether SCO and ECO products are offered
• Whether they are adopted
• How availability and uptake evolve over time

This dataset is designed for descriptive analysis, econometric modeling,
and policy evaluation.

File:
supplemental_offering_and_adoption.rds

6) Calibrated ADM for The One Big Beautiful Bill Act (OBBBA) Policy Analysis

This dataset provides recalibrated ADM tables used to evaluate the
implications of coverage changes introduced under the One Big
Beautiful Bill Act (OBBBA).

These tables approximate premium parameters under modified supplemental
coverage bands, enabling forward-looking policy analysis prior to the
release of official ADM updates.

Files

📦 Data Access and Versioning

All datasets are distributed via a GitHub Release tagged data and
uploaded using the piggyback package. This ensures:

• Stable download URLs
• Transparent versioning
• Reproducibility for collaborators and reviewers

📌 Intended Use

These data are intended for:

• Research on supplemental crop insurance design and adoption
• Policy analysis under current and proposed FCIP reforms
• Internal ARPC and academic workflows
• Replication and extension by external researchers

They are not official USDA or RMA products and should not be
interpreted as such.

⭐ If you find this data release useful, please consider starring the
repository.

Prepared FCIP data for in FCIP-related calibrations.

Calibrated yield estimates for each observed FCIP aggregated transaction, based on the methods described in Tsiboe et al. (2025).

Tsiboe, Francis, Dylan Turner, and Jisang Yu. 2025. Utilizing large-scale insurance data sets to calibrate sub-county level crop yields. Journal of Risk and Insurance, 92(1), 139–165. https://onlinelibrary.wiley.com/doi/10.1111/jori.12494

Summary of Business data breaks out FCIP participation at variaous levels:

sobtpu_all aggregates loss experience for groups of producers who are similarly defined by their contract choice (i), the insurance pool they selected (j), and the crop year (t). Contract choices combine insurance plan (e.g., APH, RP), coverage level, and unit structure (e.g., Optional [OU], Enterprise [EU]). Pools are the most granular rate‐setting level and are distinguished by county, commodity, crop type, and practice (e.g., irrigated, organic).

sobcov_all aggregates loss experience for groups of producers who are similarly defined by their coverage level, county, commodity, and commodity year.

sobscc_all aggregates loss experience for groups of producers who are similarly defined by their county, commodity, and commodity year.

Cause of Loss breaks out FCIP participation by peril

All data and models were processed on Beocat, the high-performance computing cluster at Kansas State University (https://beocat.ksu.edu/).

Weather variables are derived from the Parameter-elevation Regressions on Independent Slopes Model (PRISM) dataset, which provides daily gridded weather data at approximately 2.5 × 2.5 mile resolution. Degree-day measures follow the interpolation procedures described in Schlenker and Roberts (2009) and Tack et al. (2015).

Daily PRISM variables were aggregated to the county level as weighted means of all PRISM grid cells intersecting each county. Weights were based on the proportion of each grid cell associated with the relevant crop’s planted area, as derived from NASS Cropland Data Layers (CropScape), 2008–2024 (USDA NASS, 2019). County-level daily degree days and precipitation were then summed or averaged over each year’s growing season.

The growing season was defined using USDA crop progress reports. For each crop and state, the season begins when at least 10 percent of acreage has been planted and ends when at least 90 percent has been harvested. Weekly progress reports were converted to calendar dates to determine season length. Crops planted in one year and harvested in the next (e.g., winter wheat) were handled accordingly. When crop-state-year observations were missing, long-run average planting and harvest dates were used for imputation.

References

Schlenker, Wolfram, and Michael J. Roberts. 2009. “Nonlinear Temperature Effects Indicate Severe Damages to U.S. Crop Yields under Climate Change.” Proceedings of the National Academy of Sciences of the United States of America 106 (37): 15594–15598. https://doi.org/10.1073/pnas.0906865106.

Tack, Jesse, Andrew Barkley, and Lawton Lanier Nalley. 2015. “Effect of Warming Temperatures on US Wheat Yields.” Proceedings of the National Academy of Sciences 112 (22): 6931–6936. https://doi.org/10.1073/pnas.1415181112.

USDA National Agricultural Statistics Service. 2019. “CropScape – Cropland Data Layer (2019 Release).” https://nassgeodata.gmu.edu/CropScape/.

A collection of frequently used spatial features