Incarceration and the COVID-19 Crisis

The Issue:

The novel coronavirus outbreak has focused attention on the potential for disease transmission in prisons and jails. Prisons and jails present significant challenges to containing disease outbreak. Many called for immediate reductions in prison and jail populations in order to protect the health of inmates, correctional staff and visitors, and the communities in which detention facilities are located. Others expressed apprehension about the potential impacts of widespread inmate releases on public safety. The net effect on jail and prison populations -- and spread of the virus within those facilities -- has varied across places. Facilities are doing the best they can with limited information, but more guidance on optimal policy is crucial.

Those released from jails and prisons to limit the spread of COVID-19 may face difficult living conditions, scarce jobs and limited reentry services.

The Facts:

• The United States has the highest incarceration rate in the world, with an estimated incarcerated population of 2.3 million inmates. Slightly over half of the incarcerated population, or approximately 1.3 million people, is being held in state prisons. Another third of that population, or approximately 750,000 people, is being detained in county jails. The remainder are being held in federal detention facilities. The static measure of the county jail population understates the scope of county jail detention; county jails process approximately 11 million new admissions every year. Of those detained in county jails, approximately 75% are being held pretrial (that is, before a conviction). 
• Jails and prisons present significant challenges to containing disease transmission. Maintaining social distance in correctional facilities is generally difficult, and incarcerated people often lack in-room access to hygiene products and water, factors that make it more difficult to contain the spread of the highly contagious novel coronavirus. Moreover, approximately 10% of the incarcerated population is 55 or older, and at higher risk for adverse outcomes. However, jails and prisons differ substantially in ways that may affect their ability to combat the spread of the virus. Jails tend to keep inmates in close quarters, and are not built for long stays; individuals often cycle from jails to communities and back over relatively short time horizons. Prisons are intended to house inmates for a year or more; social distancing may be easier in such spaces.
• Jail populations dropped dramatically in mid-March after the White House issued guidelines recommending social distancing. In a sample of 508 jails that post daily jail rosters online, representing a total population on March 16 of 130,508 detainees, the median jail population dropped by 30% between March 16 and May 17 (the figure above reports the daily total jail population and a 7-day moving average of county-level COVID cases for the subset of 193 counties for which continuous daily data are available). Prison populations saw much smaller reductions; the typical prison reduced its population by only 5%. These changes appear to be exacerbating racial disparities in incarceration rates. Between March 16 and April 12, the white jail population in a sample of 341 jails reporting detainee data by race dropped by 25.5%, while the Hispanic jail population dropped by 20.3%, and the black jail population dropped by only 16.8%. It is not yet clear what is driving these disparate effects: underlying disparities in charges/characteristics that determine eligibility for release, differential treatment based on race when discretion is applied, or some combination of the two. The Public Safety Lab at New York University is currently conducting research on detainee-level data to determine the proportions of jail population reductions due to decreases in admissions and to increases in releases; the correlates of increased release probabilities; and the causal impacts of releases on public safety outcomes. 
• Correctional facilities around the country are presenting some of the largest clusters of positive coronavirus cases. However, testing is neither widespread among incarcerated populations nor consistent across the country, so the true extent of the spread of COVID-19 within facilities is not known. As of May 18th, nearly 25,000 people incarcerated in state and federal prisons had tested positive for COVID-19. This represents 1.5% of the total prison population. Other data on the spread of COVID-19 in prisons that would be useful to have is more limited. Based on official reports, over 68,000 people incarcerated in state and federal prisons have been tested; this is only 5.3% of the reporting population, suggesting that prisons are primarily testing people who are symptomatic. In the Marion Correctional Institution in Ohio, for example, mass testing of every inmate found that nearly 2,000 people – or approximately 80% of those incarcerated - had COVID-19; this suggests that the virus is much more widespread than official data reveal. (By May 20th, the number of positive cases among those incarcerated in the Marion Correctional Institution had increased to 2,439.) States such as Louisiana and Delaware that report information on asymptomatic positives provide evidence to suggest that a large share of people with the virus are asymptomatic. At least 340 people incarcerated (0.02% of the total prison population) have died due to COVID-19, though this number is likely an underestimate. States vary not only in which data they release but how often they release it. Some states report only on weekdays (e.g. Alabama, Maine), while others report even less frequently (e.g. Maryland reports weekly). These delays make it difficult to understand the spread of COVID-19 within prisons and for policy makers to quickly respond to changing conditions. Data on COVID-19 in jails are even more limited than data from prisons. Though no unified source yet exists for data from all of the over 3,000 jails in the United States, data collected for a small number of jails suggests that conditions are similar to those in prisons. These data - drawn primarily from large, urban jails - show that large numbers of both staff and people incarcerated have reported positive for COVID-19. 
• Outbreaks in prisons impact facility staffing as well as the communities beyond their walls. As of May 18th, approximately 6,600 staff members in state and federal prisons were reported to have COVID-19. Information on cases from staff, who may carry COVID-19 between prisons and their local communities, generally rely on self-reports by staff members instead of positive test results, and not every state has made these data available. While this approach avoids the issue of limited tests, it also reduces the accuracy of these data and the comparability to incarcerated people who must test positive. Though many states do report the number of staff who have COVID-19, few say how many staff members are on leave due to being exposed to COVID-19. Data from Colorado suggest that the effect of COVID-19 on prison staffing may be far higher than positive cases alone would indicate. As of May 19th, the most recent date available, 11 staff members reported an active COVID-19 case while 42 staff members in total were on leave due to either an active case or exposure. Data from previous days show similar disparities between the number of staff reported with an active case and the number of staff on leave. This suggests that prison systems are operating with several times fewer staff than may be apparent from their positive report data.
• Those who are released from jails and prisons to limit the spread of COVID-19 may face difficult conditions in their communities. They may not have a safe place to live where social distancing is possible; it is therefore unclear how much risk of infection falls upon release. In addition, jobs are scarce, and many community-based reentry services may be unavailable. Previous research shows that being released at a time when the local labor market is weak increases recidivism rates (see here). On the other hand, if community supervision requirements are weakened -- for instance, requiring phone calls with probation officers instead of in-person meetings -- this could reduce the number of technical violations recorded that might typically lead to (re-)incarceration. Existing research shows that this is unlikely to have any detrimental effect on public safety.
• Financial assistance could help reduce recidivism among the recently released. The federal government is providing financial assistance to most low-income Americans. Those who are currently incarcerated are not eligible, but those who have recently been released from jail or prison are eligible. However, takeup may be low. Individuals typically need to have filed federal taxes in recent years in order to receive a payment; this will likely exclude many people who have been released from jail or prison in recent months. On average, this group is less attached to the formal labor market; their incomes may be too low to necessitate filing taxes. Research shows that financial assistance to those with recent criminal justice involvement can reduce recidivism (see here). Making it more difficult for this group to obtain the federal financial assistance to which they are entitled could therefore have the opposite effect, leading to more criminal behavior going forward.