Skip to main content
ABC News
Jails Are Coronavirus Hotbeds. How Many People Should Be Released To Slow The Spread?

This article was published in partnership with The Marshall Project, a nonprofit news organization covering the U.S. criminal justice system.

At the start of the COVID-19 pandemic, epidemiologists warned that jails and prisons would be breeding grounds for infectious disease, thanks to their densely packed populations and uncertain access to hygiene products and medical care.

And, sure enough, the four biggest known clusters of outbreaks in the U.S. on June 3 were all linked to correctional facilities, according to a tracking project by The New York Times.

But while experts are unsurprised by the outbreaks in correctional facilities, it’s less obvious what should be done about it. Computer simulations suggest that reducing the jailed population could help control the spread of the virus inside and outside the jails. And officials have been doing that — but they’re wrestling with where to draw the line. Over the last week, police across the country have arrested at least 9,300 people protesting the police killing of George Floyd, prompting more urgent questions about law enforcement’s handling of the crisis.

To try to understand just how big a role jails could play in spreading COVID-19 in the U.S., a team of researchers built a statistical model of how the virus moves in and around jails, simulating the spread from early April onward. Several epidemiologists we spoke to said the model, developed by researchers at Washington State University, the University of Pennsylvania, the American Civil Liberties Union, Tusculum University and the University of Tennessee, is the first they’ve seen that considers the effects of jails on the disease’s spread, which leading models of the virus have ignored. The study suggests common estimates are vastly undercounting the toll of COVID-19 because they fail to account for the flow of people into and out of jails; thousands could die inside facilities, and many more could die in the communities surrounding jails.

Perhaps counterintuitively, the model found that if social distancing measures outside jails succeed in controlling the virus, jails would become a primary vector of spread, leading to large numbers of preventable infections, hospitalizations and deaths. An otherwise effective virus response would thus be undone by ignoring jails.

On the other hand, the worse the pandemic is outside jails, the smaller jails’ role in spreading the disease becomes. Say businesses, schools, parks and other social hubs open without restriction all at once. As the virus spreads like wildfire in the outside world, jails are no longer a primary source of infection, simply because the infection is rampant everywhere.

Of course, the model’s predictions are just estimates — and they depend on several assumptions that do not mirror reality. (Anyone who wants to build a COVID-19 model has to make some educated guesses, not all of which will turn out to be right, as FiveThirtyEight has explained in the past.) The model, for example, only factored in larger jails, and it uses estimates for important characteristics of the virus, like exactly how the virus spreads and what percentage of the population is getting it, since there is still so much we don’t know about it. The researchers also assumed that the country would be living under a nationwide shelter-in-place order for the full time period covered by the model, which extends for several more months. We already know that’s not the case.

But while the model’s creators acknowledge that the model’s estimates won’t align precisely with reality, it can still tell us something about the dynamics of disease spread.

“There’s no doubt the impact will be very, very large,” said Lucia Tian, chief analytics officer for the ACLU, who took part in the research.

In the team’s simulations, the more jail populations are reduced, the more it flattens the curve inside jails and out. Dramatic cuts in jail populations — for instance by stopping 95 percent of arrests and doubling release rates, which the authors calculate would reduce the detainee population by more than 85 percent — are forecasted to be most effective in preventing deaths.

By demonstrating the relative effectiveness (or ineffectiveness) of several policies, Tian hopes the model can aid leaders and policymakers as they attempt to minimize the damage. She explained with a classic adage of statistics: “All models are wrong, but some are useful.”

Amid the pandemic, many agencies have made fast and dramatic changes in an effort to empty out jails. The number of people in state prisons — where people, generally those sentenced to a year or more, serve their time after they’ve been convicted — has barely budged. But the typical jail, which holds mainly people who have not yet been tried or those serving short misdemeanor sentences, has lowered its population by over 30 percent, with a few reducing their populations by around 60 percent, according to a report by the criminal justice think tank Prison Policy Initiative. Cuts like these would have been hard to imagine before the virus — but are they enough?

“There’s no golden or perfect amount by which a population should be reduced,” said Wanda Bertram, who contributed to the report and works as the Prison Policy Initiative’s communications strategist. “There’s simply the fact that compared to any other measure that a jail or prison can take, nothing is nearly as effective at slowing down the spread of the coronavirus as letting people go.”

Because New York City has been the epicenter of the virus in the United States, it provides a useful case study of the debate about the threat of COVID-19 to incarcerated populations. Jail populations in the city have been dropping for years, down by more than half in the last decade. Since the epidemic began, they have dropped another 25 percent — to their lowest levels since the 1940s — which city officials say is even more meaningful given the reductions of the last several years. “We’re diverting, we’re issuing probation, we’re sending less people into the system. I think that’s all happening, even before this pandemic hit,” said New York City Police Commissioner Dermot Shea.

Shea pushes back against calls for further cuts, arguing that a significant portion of the people still in jail are charged with serious or violent crimes, or have previously been convicted of multiple serious offenses. “Each of these releases has a potential impact on public safety, and you try to weigh that against the humanity issue of having someone contract the disease in jail,” Shea said. “We’re trying to strike that balance.”

But advocates still worry about the roughly 4,000 people still being held and point to hundreds of arrests made during the demonstrations of the last several days, which included clashes that led the Legal Aid Society to say that police were “brutalizing” protestors. Many are calling for further reductions at Rikers Island and other city jails, which the president of the city’s corrections officers union referred to as “the epicenter of the epicenter.”

Agencies don’t always agree about who should be considered for release. The Brooklyn district attorney’s office, which has consented to releasing hundreds of people awaiting trial during the COVID-19 outbreak, has clashed in the past with NYPD leadership over jail reduction policies.

But even for an office willing to consider alternatives to incarceration in more serious or gun-related cases, the pandemic has prompted some difficult decisions. Jill Harris, chief of policy and strategy for the Brooklyn district attorney, is particularly worried about releasing those charged with domestic violence, which may be rising even as other types of crime decline in shuttered cities. Harris, who worked at the ACLU and the Drug Policy Alliance before coming to the district attorney’s office, says the pandemic adds to the existing complexities that make domestic violence especially difficult for victims to report and for law enforcement to disrupt. Victims may feel even more conflicted about reporting abuse if it means potentially exposing their partner or family member to the virus in jail. And as families shelter at home, there may be fewer opportunities to report an abuser who is physically there all the time.

“Reducing incarceration and increasing public safety go hand in hand,” she said. “But there does get to be a point where the people who are there need to be there. And I feel like we are approaching that point in Brooklyn.”

The virus has led to fast and unprecedented reductions in jail populations, and reformers are witnessing the nation’s jail system make the types of changes they’ve been told for years are impossible. After the last three months, legal experts and advocates have the same question: If jails can do it, can other parts of the country’s vast justice system change, too?

“It’s kind of an exciting time to work in criminal justice, because in a way, I think that [the] emergency could be the spark that leads to innovation,” said Lucy Lang, a former assistant district attorney in Manhattan who now heads the John Jay College of Criminal Justice’s Institute for Innovation in Prosecution. “If we mean that we’re going to really stick with the decreased population, then what are we going to do instead?”

Lang said maintaining lower levels of incarceration is made possible by an all-hands-on-deck approach, where many agencies intervene to relieve some of the conditions that put people in danger of arrest. Reformers also say more extensive use of alternative crime responses like special courts, violence interrupters (people who try to prevent violent crimes by intervening with those likely to commit them before they happen), community policing, diversion programs (sending offenders to a rehabilitation program rather than prison), reduced sentencing and bail reform can pave the way for less incarceration.

Some of the ideas Lang described question long-held beliefs about crime and punishment. She recounted a recent response to a homicide that included a restorative justice approach as an alternative to incarceration or trial. In that instance, those harmed by the crime met with those responsible to ask the perpetrators to confront the consequences of their actions and reach consensus on how amends might be made and further harms might be prevented. “It was clear [the victims] experienced a much greater sense of resolution and justice than if there was a trial.”

Eventually, though, the pandemic will ebb and the criminal justice system will have to figure out the new normal. Christine Tartaro, professor of criminal justice at Stockton University, hopes advocates are right that this moment could be a turning point. “I am curious … as we reduce the census throughout all of these jails and prisons, if after this it makes us think about whether we need to go back to filling them again.”

But that’s far in the future. For the moment, people are still thinking about how full jails should be right now.

Additional reporting by Laura Bronner.

Sign up for The Marshall Project’s newsletter, or follow them on Facebook or Twitter.

Most Americans still trust scientists l FiveThirtyEight

Anna Flagg is The Marshall Project’s senior data reporter, covering criminal justice topics including immigration, crime, race, policing and incarceration.