Data and Methods


Vera identified the county jail systems that serve the 50 largest U.S. cities and the largest cities in the states that are not home to one of the 50 largest cities. The six states that do not have county jails—Alaska, Connecticut, Delaware, Hawaii, Rhode Island, and Vermont—were excluded because the state governments operate unified prison-jail correctional systems. Similarly, the City of Baltimore was excluded because the Baltimore City Detention Center is a division of the Maryland Department of Public Safety & Correctional Services. These criteria limited the sample to 62 counties.

Vera researchers, with the support of volunteers from BNY Mellon and Goldman Sachs, obtained data for 48 counties by accessing each county’s online budget publications (43 counties) or by submitting a request to the county budget office or sheriff’s department (five counties: Ada County, ID; Cass County, ND; Hennepin County, MN; Polk County, IA; and Wayne County, MI).

Data is unavailable in 14 counties because the budget was not available online and the county did not respond to Vera’s request (six counties: Harris County, TX; Hillsborough County, NH; San Francisco County, CA; Essex County, NJ; Pulaski County, AR; and Hinds County, MS) or the sheriff’s department budget does not disaggregate jail costs from other department functions, such as law enforcement (eight counties: El Paso County, CO; Fresno County, CA; Jefferson County, AL; Marion County, IN; Oklahoma City, OK; Sacramento County, CA; Suffolk County, MA; and Tulsa, OK).

There were ten counties for which Vera obtained jail budget data but jail personnel data was not available (Ada County, ID; Clark County, NV; Dallas County, TX; Douglas County, NE; Hennepin County, MN; Hillsborough County, FL; Kanawha County, WV; Fulton County, GA; Minnehaha County, SD; and Travis County, TX).


Vera researchers collected actual spending and employee headcount data for fiscal year 2011 and the adopted budget and employee headcount data for fiscal year 2021. In 18 counties (Bexar, TX; Cook, IL; Cumberland, ME; Davidson, TN; Washington DC; Franklin, OH; Fulton County, GA; Hillsborough, FL: Jackson, MO; Los Angeles, CA; Milwaukee, WI; Orleans, LA; Salt Lake, UT; Sedgwick, KS; Tarrant, TX; Wayne, MI; and Yellowstone, MT), Vera collected proposed or recommended fiscal year 2021 budget data because the fiscal year 2021 adopted budget was not yet available online. King County, WA represents fiscal year 2021-2022 data as the county has a biennial budget.

Jail spending comprises all the expenses related to the operation of the local jail, including expenses related to the jail that are paid by other county agencies. Vera did not collect data related to juvenile justice, community corrections, or any private jails that operate in these counties (usually under federal contract). The fiscal year 2011 data is adjusted for inflation using the Consumer Price Index (CPI-U).

The county jail population in 2011 is the average daily population reported in the Bureau of Justice Statistics Annual Survey of Jails, accessed through Vera’s Incarceration Trends jurisdiction-level datafile. The county jail population in 2021 is the jail population on the most recent day data was available in Vera’s jail population data.

Cost savings calculation

Vera’s estimation of cost savings computes savings for three types of jail costs: (1) variable costs, such as for food and uniforms, which are directly related to the care and custody of the jail population and change immediately as the jail population changes; (2) step-fixed costs for corrections officer salaries that remain constant until the jail population crosses a threshold at which a housing unit can be closed; and (3) step-fixed costs for administrative staff and other operating expenses.

Vera estimated the variable costs to be 20 percent of the agency’s non-personnel costs. This value was benchmarked to the short-run marginal cost for Washington State jails that was computed by the Washington State Institute for Public Policy—one of the most careful analyses of jail marginal costs that is available. The size of housing units was conservatively estimated to be 100 beds, since a 2015 survey found that housing units average 60 beds. Vera assumed that savings for administrative staff can be taken in stages when the jail population falls below 75 percent, 50 percent, and 25 percent of the jail’s current population. The justification for these thresholds is that rightsizing the jail’s administration and maintenance staff is necessary and appropriate as the jail population declines by large margins.

Vera did not make any reduction to the portion of the jail personnel budget (assumed to be 10 percent) that supports programming, social services, and medical care. The most recent data available indicates that 6 percent of jail staff are educational or professional staff, which includes psychiatrists, psychologists, social workers, counselors, medical doctors, nurses, paramedics, chaplains, and legal specialists.