Getting statistics for persons incarcerated in the U.S. by state is more difficult than one might expect. Counts of persons in jails, which do not include short-term police lock-ups, are not regularly available by state in the U.S. Bureau of Justice Statistics (BJS) series for prison and jail inmates at midyear. The most recent year for which jail inmates are available by state is 2005.
Alaska, Connecticut, Delaware, Hawaii, Rhode Island, and Vermont have integrated local-state prison systems. Hence separate jail inmate counts for these states do not exist. Jail inmates counts for these states can be estimated by assuming that their jail inmates totals have the same proportion in the national jail total as their combined jail-prison totals have in the national combined jail-prison total.
Prisoners held in state and federal prisons are reported annually by state in BJS prisoner publications. The U.S. Bureau of Justice Statistics’ Prisoners Series (Prisoners in (year)) and its Prison and Jail Inmates at Midyear Series (Prison Inmates at Midyear (year)) both include tables of prisoners by state. These tables are entitled “Prisoners under the jurisdiction of state or federal correctional authorities, by jurisdiction.” The state totals in such tables do not include federal-jurisdiction prisoners held within that state. The U.S. Federal Bureau of Prisons provides annual counts of federal prisoners by prison in its annual State of the Bureau report, but it does not aggregate these counts by state. The state distribution of federal prisoners differs considerably from the state distribution of state prisoners. In particular, Kentucky and West Virginia have large shares of federal prisoners relative to their state prisoner shares.
Another complication is the difference between custodial and jurisdictional prisoner counts. Jails hold in addition to local prisoners some state and federal prisoners because of bed rental agreements, overcrowding constraints, and temporary holding responsibilities. Hence counts of prisoners under the jurisdiction of local officials (local prisoners) are not the same as counts of persons held in local jails.
Jail and prison inmate counts on a specific day also differ from average daily populations across a year. Seasonality in jail populations is larger than seasonality for prisons, because the average jail stay is less than thirty days, while the average prison stay is over a year.
The Census of Prisons for 2005 and the Census of Jails for 2005 allow these technical complications to be disentangled and analyzed. Based on data from those sources, here’s prison and jail populations, subtotaled by states.