Scope/Coverage of Data
The 2025 edition of the USDA, Economic Research Service (ERS) County Typology Codes classifies all counties in the 50 U.S. States and Washington, D.C. for which data are available. County-equivalents in U.S. territories are not classified due to limited data availability. Data availability for a given code also varies by data source and by data year. The 2018–22 5-year American Community Survey (ACS) enumerated data for Connecticut’s nine planning regions, which replaced counties as a Census geographic unit in 2023. Classifications that use the 2018–22 ACS—low employment, low post-secondary education, and housing stress—are determined for 3,144 counties and county-equivalent geographies.
The 2017–21 ACS used for the persistent-poverty codes, the net migration data used for the retirement-destination codes, and the data from the U.S. Department of Commerce, Bureau of Economic Analysis used for the Economic Typology Codes enumerated data for Connecticut’s eight counties, so these classifications are determined for 3,143 counties and county-equivalents. The 2023 delineation of core based statistical areas (CBSAs) does not classify Connecticut’s county geographies. For the purposes of this analysis, Litchfield County, CT and Windam County, CT were treated as nonmetropolitan (nonmetro) and all other counties in Connecticut were treated as metropolitan (metro).
The population-loss code uses Decennial Census data from 2000, 2010, and 2020. Counties that did not exist in 2000 or 2010 are not assigned population-loss codes. Bedford city, VA was added to Bedford County, VA in 2013 and Clifton Forge city, VA was added to Alleghany County, VA in 2001. The population data for these county/city combinations were summed in the Decennial Census years prior to their mergers to create a consistent geography for the population-loss county code. This results in 3,135 counties and county-equivalents for which population-loss status was determined. See the U.S. Department of Commerce, Census Bureau page on Substantial Changes to Counties and County Equivalent Entities: 1970-Present for more information on changes to county geographies.
U.S. Department of Commerce, Bureau of Economic Analysis (BEA) combines Virginia's independent cities with nearby counties and reports data for the combined area only. The BEA also combines Maui, HI and Kalawao, HI into a single geographic area. The Economic Typology Codes are based on the earnings and job shares of these county/city combinations, but codes are reported separately for these counties in the data file. The full list of county/city combinations used by BEA is in the documentation tab of the ERS County Typology Codes, 2025 Edition Excel download file.
The Economic Typology Codes, 2025 Edition
The 2025 edition of the USDA, ERS County Economic Typology Codes is based on the industrial composition of nonmetro counties. The ‘High Industry Concentration’ Economic Typology Codes indicate if a county has a high share of earnings or jobs in farming, mining, manufacturing, Federal and State Government, or recreation relative to other counties. These codes are not mutually exclusive, meaning that a county may be classified as having a high concentration in multiple industries. The mutually exclusive ‘Industry Dependence’ Economic Typology Code indicates if any of those five industries have a high combined share of earnings and jobs and which industry with a high share is the most prevalent in a county relative to the other industries.
Selection of the five industries examined is guided by export base theory. Farming, mining, manufacturing, Federal and State Government, and recreation are industries that many nonmetro counties have comparative advantages in (due primarily to the availability of natural resources and low production costs) and that allow them to produce goods and services for non-local markets. Agriculture, mining, and manufacturing are industries that export goods to outside markets and are traditionally thought of as ‘export’ industries. Federal and State Government represents an opportunity for workers in a local market (county) to provide government services to residents of a larger market (State or national). The recreation industry allows local producers to provide goods and services to visitors who otherwise would not be consumers in the local market. Exporting goods and services brings additional income and earnings into the local market that would not be available if selling in the local market alone. However, a high concentration in or dependence on a single or few ‘export’ industries can also leave a county’s economy more susceptible to market volatility in the industry.
Economic Typology Codes: Methods
Labor and proprietors’ earnings and jobs by place of work are the basis for the economic typology codes. Each industry’s earnings and jobs are calculated separately as a percentage of total earnings or total jobs in the county in 2019, 2021, and 2022. These percentages are summed and divided by three to obtain three-year-average percentages. This averaging is done to minimize the effects of any one-year anomaly on an industry’s share of earnings or jobs. Data for 2020 are omitted to avoid any outsized effects from the widespread shutdown of many businesses due to the COVID-19 pandemic.
County-level estimates of earnings and jobs by place of work come from U.S. Department of Commerce, Bureau of Economic Analysis, Regional Economic Accounts, released in November 2023. Non-publicly available, unsuppressed earnings and jobs estimates are used for counties in 40 States and Washington, D.C. Estimates for counties in Florida, Hawaii, Kentucky, Louisiana, Massachusetts, Michigan, New Hampshire, North Carolina, Rhode Island, and Vermont come from the publicly available BEA files where possible. Earnings and jobs shares are calculated using alternative data if values for 2019, 2021, or 2022 are suppressed. If complete BEA data for 2019, 2021, and 2022 are not available, the following alternatives are used:
- If one or two of the three years (2019, 2021, or 2022) used to calculate the three-year average are suppressed, then a single year or a two-year average of the available years of data is used.
- If data for 2019, 2021, and 2022 are suppressed, the most recent single year of BEA data going back to and including 2013 is used.
- If no BEA data are available from 2013 onward, the following alternatives are used:
- Available subindustry BEA earnings data from 2019, 2021, and 2022 are summed and the shares are averaged to create an underestimate of earnings for the high mining, manufacturing, and recreation concentration codes and industry dependence code.
- Employment data from the S2403 table of the 2018–22 ACS are used for the high mining, manufacturing, and recreation concentration codes and industry dependence code.
- For the High Government Concentration code, BEA Federal, State, and local Government earnings and jobs from 2019, 2021, and 2022 are used.
- Separate thresholds are calculated for counties that use ACS employment data and for counties that use Federal, State, and local Government data for the High Government Concentration code.
- For the Industry Dependence code, an alternative method is used to determine if a county is government dependent. In cases where the State government (alone) earnings and jobs are not available in the BEA data, a county's Federal and State Government earnings and jobs shares are estimated by subtracting the nonmetro mean share of earnings and jobs in local government from the county’s share of earnings and jobs in Federal, State, and local Government. State and local government earnings and employment data are available as a combined industry in the BEA data. Neither Federal Government nor the combined State and local industry data are suppressed for any counties.
| Counties using BEA data from 2013 to 2018 | Counties using BEA subindustry earnings, and/or ACS employment data | |||||
|---|---|---|---|---|---|---|
| Total | Classified as high concentration | Classified as dependent | Total | Classified as high concentration | Classified as dependent | |
| Farming | 0 | 0 | 0 | 0 | 0 | 0 |
| Mining | 66 | 4 | 1 | 104 | 2 | 0 |
| Manufacturing | 10 | 1 | 1 | 15 | 3 | 2 |
| Government | 4 | 0 | 0 | 16 | 6 | 3 |
| Recreation | 22 | 1 | 0 | 60 | 7 | 2 |
| Source: USDA, Economic Research Service using data from the U.S. Department of Commerce, Bureau of Economic Analysis, Regional Economic Accounts and the U.S. Department of Commerce, Bureau of the Census, 2018–22 5-year American Community Survey and 2020 Decennial Census. | ||||||
The industries reported by the BEA and used to create the economic typologies are based on the 2017 North American Industry Classification System (NAICS). Farming includes the earnings and jobs of sole proprietors, partners, and hired laborers involved in the current production of agricultural commodities. Mining includes mining, quarrying, and oil and gas extraction. Manufacturing includes the manufacturing of both durable and non-durable goods. Government includes earnings and jobs from Federal and State Government but excludes local government. Recreation combines earnings and jobs from arts, entertainment, and recreation, accommodation and food services, and real estate and rental and leasing services. The percentage of vacant housing for seasonal, recreational, or occasional use—based on data from the 2020 Decennial Census—is also used to determine if a county has a high concentration in recreation or is recreation dependent.
High Industry Concentration Code Thresholds
The thresholds used to determine if a county had a high concentration of earnings or jobs in an industry are based on the distribution of earnings and jobs for that industry in nonmetro counties. High farming, mining, manufacturing, and government concentration thresholds were set at the nonmetro mean-plus-one standard deviation level and rounded to the nearest whole number. If an industry’s share of earnings or jobs in a county were greater than or equal to the threshold level, the county was classified as having a high concentration in that industry. As such, these measures indicate those counties where employment or earnings in the industry are especially high compared to other nonmetro counties.
High recreation concentration counties are classified using a weighted z-score that incorporates shares of recreation earnings, jobs, and vacant housing for seasonal, recreational, or occasional use into a composite measure. Separate z-scores are calculated using the nonmetro mean and standard deviation values for the annual average percentage of total earnings in recreation, the annual average percentage of total jobs in recreation, and the percentage of vacant housing used for recreation purposes. The earnings and jobs z-scores are each assigned a weight of 0.3 and the housing z-score is assigned a weight of 0.4, and the three weighted z-scores are summed to produce the composite weighted z-score. A county is classified as ‘high recreation concentration’ if its composite weighted z-score is at least two-thirds of a standard deviation above the nonmetro mean (is greater than or equal to 0.67). This method is a variation of the original developed by Ken Johnson and Calvin Beale in their nonmetro recreation classification completed in 2002. See the USDA, Economic Research Service report, Rural America, Vol. 17, Issue 4 for more information.
Counties where none of the five industries met or surpassed their respective thresholds are classified as ‘nonspecialized’.
| ERS typology industry | U.S. Department of Commerce, Bureau of Economic Analysis component industries/descriptions | Earnings share threshold, percent | Jobs share threshold, percent | |
|---|---|---|---|---|
| Farming | Sole proprietors, partners, and hired laborers involved in the current production of agricultural commodities | 20 | 17 | |
| Mining | Mining, quarrying, and oil and gas extraction | 11 | 7 | |
| Manufacturing | Manufacturing | 25 | 17 | |
| Government | Federal civilian; State government | 13 | 8 | |
| Recreation | Arts, entertainment, and recreation; accommodation and food services; real estate and rental and leasing | 11 | 17 | |
| Note: Earnings share and jobs share thresholds are calculated as the nonmetro mean-plus-one standard deviation value for each industry, and are rounded to the nearest whole percentage. High recreation-concentration counties are not classified using the mean-plus-one standard deviation thresholds and include a housing component, but earnings and jobs values are included in the table for comparison. Source: USDA, Economic Research Service using data from the U.S. Department of Commerce, Bureau of Economic Analysis. |
||||
Industry Dependence Code Thresholds
The Industry Dependence Code differs from the High Industry Concentration Codes in three ways:
- A common threshold is used for all industries to determine if an industry’s share of earnings and jobs is high.
- An industry’s share of earnings and its share of jobs in a county are combined in a composite weighted z-score, so both earnings and jobs must be jointly high enough for the industry to meet or surpass the threshold.
- The industry with the highest combined share of earnings and jobs (highest weighted z-score) that meets or surpasses the threshold is selected as the industry on which the county ‘depends’. This means that the code is mutually exclusive, and counties can only be ‘dependent’ upon one industry.
The common threshold is determined by taking the mean and standard deviation values of annual average earnings shares and of annual average jobs shares for all five industries, combined in a single distribution of all nonmetro counties (5 industries x 1,958 nonmetro counties = 9,790 observations). For earnings, this results in a mean value of about 7 percent and a standard deviation of about 10 percent. For jobs, this results in a mean value of about 7 percent and a standard deviation value of about 7 percent. These mean and standard deviation values are used to calculate z-scores for earnings and jobs for each of the five industries. The earnings and jobs z-scores are both assigned weights of 0.5 and are summed to create composite weighted z-scores for farming, mining, manufacturing, and Federal and State Government. The recreation earnings and jobs z-scores are each assigned weights of 0.3, the recreation housing z-score is assigned a weight of 0.4, and those three weighted z-scores are summed to create the composite weighted z-score for recreation.
A z-score value of 1 is equal to the mean-plus-one standard deviation value of the data distribution. The value was selected as the industry dependence threshold to reflect the mean-plus-one standard deviation thresholds used in the High Concentration Codes. If one industry has a composite weighted z-score of at least 1, then the county is determined to be dependent on that industry. If multiple industries have composite weighted z-scores of at least 1, the county is determined to be dependent on the industry with the largest composite weighted z-score. If no industry has a composite weighted z-score of at least 1, then the county is classified as ‘not dependent’.
The Demographic Typology Codes
The low post-secondary education (2018–22), low employment (2018–22), persistent poverty (1990, 2000, 2007–11, 2017–21), housing stress (2018–22), and population loss (2000, 2010, and 2020) classifications are based on the U.S. Census Bureau’s Decennial Census and 5-year American Community Survey (ACS) data from the years in parentheses after their names. The thresholds used to determine low post-secondary education, housing stress, and retirement-destination counties are based on the nonmetro mean-plus-one standard deviation values of the applicable variables. For more information on the persistent-poverty code, please see the USDA, Economic Research Service, Poverty Area Measures data product and The Poverty Area Measures Data Product (TB-1967, July 2024).
| ERS demographic type | Data source | Definition |
|---|---|---|
| Low post-secondary education | 2018–22 5-year ACS | At least 57 percent of residents ages 25 to 64 did not have any post-secondary education. |
| Low employment | 2018–22 5-year ACS | Less than 63 percent of residents ages 25 to 54 were employed. |
| Population loss | 2000, 2010, and 2020 Decennial Censuses | County lost population from 2000 to 2010 and from 2010 to 2020. |
| Housing stress | 2018–22 5-year ACS | At least 29 percent of owner- and renter-occupied housing units had at least one of the following conditions: 1) lacking complete plumbing facilities, 2) lacking complete kitchen facilities, 3) with 1.01 or more occupants per room, 4) selected monthly owner costs as a percentage of household income greater than 30 percent, and 5) gross rent as a percentage of household income greater than 30 percent. |
| Retirement destination | Applied Population Laboratory, University of Wisconsin-Madison | Number of residents ages 55 to 74 increased by at least 15 percent from 2010 to 2020 because of migration. |
| Persistent poverty | 1990 and 2000 Decennial Censuses and 2007–11 and 2017–21 5-year ACS | Poverty rate was 20 percent or higher in 1990, 2000, 2007–11, and 2017–21. |
The retirement-destination classification was created with data from the Applied Population Laboratory at the University of Wisconsin-Madison. See the Applied Population Laboratory website for more information. Net migration for the population ages 55 to 74 for the period 2010 to 2020 was calculated using the residual method. Using this method, net migration equals the change in population between 2010 and 2020 after removing the change in population due to deaths and population aging. If the net migration of people ages 55 to 74 from 2010 to 2020 added 15 percent or more to the number of people that were in that age group in 2010, the county was classified as a retirement destination. The age group 55 to 74 is used to determine retirement destinations because research shows that people in this age group are more mobile and have different migration patterns than those in the 75 and over age group.
Strengths and Limitations
The ERS County Typology Codes data product is designed to be an easy-to-use tool for capturing the economic and demographic heterogeneity of nonmetro counties in the United States. The Codes capture a subset of relevant phenomena that are often of interest to rural researchers and policymakers but are not an exhaustive accounting of relevant issues. The Codes are descriptive in nature, so users should carefully consider whether the Codes appropriately identify counties for their intended use.
The ERS County Typology Codes are primarily intended for the evaluation of conditions in nonmetropolitan counties. The thresholds used to determine if a county has a high industry concentration or industry dependence are based on the earnings and jobs distributions of nonmetropolitan counties only. Economic typology codes are also assigned to metro counties, but researchers should carefully test whether these classifications are meaningful in the metro county context.
The High Industry Concentration and Industry Dependence Codes offer slightly different perspectives on industrial composition in nonmetro counties. The non-mutually exclusive High Industry Concentration Codes identify the counties that have a high share of earnings or jobs in an industry relative to that industry in other nonmetro counties. These codes will likely be of greater interest to users who want to examine the prevalence of a particular industry independent of other industries. The mutually exclusive Industry Dependence Code identifies which of the five industries is the most prevalent based on each industry’s share of total earnings and jobs. This code is more reflective of nonmetro counties’ industrial structures and identifies relatively more manufacturing dependent counties and relatively fewer farming, mining, government, and recreation dependent counties than the High Industry Concentration Codes. This code will likely be of greater interest to users who want to identify which of the five industries is the most dominant in a county.
Suppression of BEA earnings and jobs figures necessitate the use of alternative data years, industry codes, and data sources for some counties in Florida, Hawaii, Kentucky, Louisiana, Massachusetts, Michigan, New Hampshire, North Carolina, Rhode Island, and Vermont. Use of these alternative data means that some of the Economic Typology Codes are determined using data that may represent a different time in the economic cycle, may exclude the earnings of some subindustry categories (for example, food manufacturing may be excluded from the total earnings from manufacturing), and may capture county employment by place of residence (ACS) rather than jobs by place of work (BEA). Alternative thresholds are used where appropriate to increase the comparability of codes that use these alternative data sources to those that use the unsuppressed BEA data.
Data Sources
Egan-Robertson, D., Curtis, K.J., Winkler, R.L., Johnson, K.M., & Bourbeau, C. (2024). Age-Specific Net Migration Estimates for U.S. Counties, 1950–2020. Applied Population Laboratory, University of Wisconsin-Madison, 2024. Web. 8/14/2024.
U.S. Office of Management and Budget (OMB), Metropolitan and Micropolitan Area Delineation Files, July 2023
U.S. Department of Agriculture, Economic Research Service, Poverty Area Measures, 2023 edition
U.S. Department of Commerce, Bureau of the Census, 1990, 2000, 2010, and 2020 Decennial Censuses of Population and Housing
U.S. Department of Commerce, Bureau of the Census, 2007–11, 2017–21, and 2018–22 American Community Survey 5-year Estimates
U.S. Department of Commerce, Bureau of Economic Analysis, Regional Economic Accounts, November 2023
Recommended Citation
U.S. Department of Agriculture, Economic Research Service. (2025, April). County Typology Codes, 2025 Edition.