Skip navigation


Download original document:
Brief thumbnail
This text is machine-read, and may contain errors. Check the original document to verify accuracy.
11/16/20 U. Chi. L. Rev. Online *4

Sharon Dolovich1

From the earliest days of the pandemic, it was clear that the novel
coronavirus posed an outsized danger to the more than two million
people locked inside America’s prisons and jails. Responding to the
risk, many public officials nationwide took modest steps to reduce the
incarcerated populations in their jurisdictions. But these efforts,
though welcome, were too minimal to make an appreciable difference.
By summer, infection rates in state and federal prisons dwarfed
national rates by a ratio of 5.5 to 1, and, accounting for age, people in
prison were dying at three times the rate of society as a whole.2 And by
October 2020, jail populations had begun to creep back up, prison
releases had largely ceased, and few signs remained of a more robust
decarceral strategy.
With the crisis still unfolding, we are only beginning to make sense
of the overall impact of COVID-19 on the people who live and work
inside American prisons and jails, and of what effect, if any, the
pandemic will have on the nation’s continued commitment to mass
incarceration under unduly harsh conditions. In this Essay, I take
stock of where things now stand. I also consider how we got to this
point, and how penal policy would need to change if we are to prevent
another round of needless suffering and death when the next pandemic
hits. For those who have followed the law, policy, and politics of the
American carceral system over the past 40 years, there are no
surprises here. Our COVID response reflects callous indifference to the
fate of people in custody, an attitude that has shaped the U.S. carceral
Professor of Law, UCLA School of Law; Director, UCLA Law COVID-19
Behind Bars Data Project. I thank Sasha Natapoff and Brendan Saloner for
helpful comments; John Boston, David Fathi, Aaron Littman, and Alan Mills
for their generous willingness to field my many questions; Liz DeWolf for her
editorial advice; and Kaitlyn Fryzek for excellent research assistance. I also
thank the entire UCLA Law COVID-19 Behind Bars Data Project team—
staff, leadership team, and volunteers—who have worked tirelessly since the
start of the pandemic to collect and analyze data of all kinds bearing on the
impact of COVID on people in custody. Your commitment to the enterprise
has been inspiring beyond measure. Thanks are due as well to the Vital
Projects Fund, Arnold Ventures, and the United States Centers for Disease
Control, for their generous support of our work. All views presented in this
Essay are solely my own.
2 See Brendan Saloner, Kalind Parish, Julie A. Ward, Grace DiLaura &
Sharon Dolovich, COVID-19 Cases and Deaths in Federal and State Prisons,
324 J. AM. MED. ASSOC. 602 (2020).

11/16/20 U. Chi. L. Rev. Online *5

experience since at least the Civil War and well into the “tough on
crime” era of the late twentieth century.3 This normative foundation is
impossible to disentangle from the structural racism that has driven
the glaring overincarceration of African Americans and other people of
color and helped shaped the brutality of the American carceral
experience. These ideological forces, long productive of a national
propensity to dehumanize the people we lock away, have collectively
generated a persistent regulatory refusal across all branches of
government to ensure meaningful protections for the incarcerated.4
Like any long-lasting system, mass incarceration under inhumane
conditions successfully entrenched itself in part by creating structural
obstacles to doing things differently. When COVID appeared, we saw
the power of this strategy manifest in real time, as even those
individual officials motivated to mitigate the threat found themselves

Although in the 1950s and 1960s, prison systems in some regions displayed
a brief commitment to humanism. See Sharon Dolovich, Creating the
PENALTY? (Charles Ogletree & Austin Sarat eds., 2012). But when viewed
over the broad sweep of American penal history, that period was anomalous,
with dehumanization the more typical orientation. There is, it bears noting,
an unmistakable connection between the callous indifference toward the
incarcerated persistently displayed by officials in all branches of government,
see Sharon Dolovich, The Regulation and Oversight of American Prisons,
ANNUAL R. CRIMINOL. (forthcoming 2021), and the role played by carceral
institutions in the aftermath of slavery and even during slavery itself. In the
Southern states after Emancipation, brutality against the enslaved readily
became brutality against prisoners, the vast majority of whom were Black.
THE ORDEAL OF JIM CROW JUSTICE (1997). And important new research is
excavating the sordid history of prisons and jails as institutions of American
chattel slavery, offering a new and chilling perspective on the endemic
inhumanity and thoroughly racialized character of present-day American
carceral practice. See Taja-Nia Henderson, Property, Penality and (Racial)
Profiling, 12 STAN. J. C.R. & C.L. 177 (2016) (describing the role local jails
played in supporting enslavers and the institution of chattel slavery in the
antebellum South); John Bardes, The Problem of Incarceration in the Age of
Slavery 5, 43–47 (draft copy on file with the author) (describing a network of
carceral institutions forming a “statewide penal system for enslaved convicts”
in Louisiana, Mississippi, Tennessee, and elsewhere in the antebellum South;
and describing the brutal methods of torture employed to humiliate and
“discipline” the enslaved people held in those facilities).
4 See Dolovich, supra note 3.

11/16/20 U. Chi. L. Rev. Online *6

hampered by political dynamics and legal regimes constructed over
decades to prioritize continued incarceration regardless of harm.
The picture, however, is not all bleak. Over the past few years,
despite a seeming bipartisan consensus on the need for prison reform,
the incarceration rate in most jurisdictions has remained stubbornly
immovable.5 Yet since the pandemic hit in March 2020, at least
120,000 people have been released from custody.6 True, these
reductions amount to only 5.6 percent of the roughly 2.15 million
people behind bars nationwide as of March 2020, and are insufficient
to allow corrections officials to get the virus under control, much less to
definitively shift the United States away from its decades-long practice
of mass incarceration. Still, the speed with which these releases were
undertaken—the overwhelming majority took place between March
and May—strongly suggests that, under the right conditions,
significant decarceration is indeed possible.
COVID also offers an opening to reframe the political conversation
around American carceral practice. Since at least the 1980s, the
dominant political narrative around crime and punishment has been
radically individualist,7 focused exclusively on the need to punish
people who commit crimes and to protect society against further
criminality. Not only has this orientation left us collectively unable to
reckon with the socioeconomic drivers of criminal activity, but we have
remained blind to the community costs of a default carceral response.
Prison sentences cause harm not only to the individuals serving time,
but also to their families and to the broader community. The same is
true of pretrial detention. This pandemic has forced a collective
recognition that what happens inside jails and prisons has serious
See Wendy Sawyer & Pete Wagner, Mass Incarceration: The Whole Pie 2020
(2020) (reporting that during the years 2017–2020, the number of people
incarcerated in state and federal prisons and jails held steady at between
2.140 and 2.157 million).
6 According to the UCLA Law COVID-19 Behind Bars Data Projects releases
tracker, at least 43,000 people have been released from prison and 77,000
from jail since the start of the pandemic. See COVID-19-Related Jail Releases
& COVID-19-Related Prison Releases, UCLA LAW COVID-19 BEHIND BARS
DATA PROJECT. These numbers are undercounts, since they reflect only
reported releases of classes of incarcerated people and do not include
individuals released through habeas corpus or compassionate release
petitions. This latter category includes 1,700 people granted compassionate
release from federal prison since March 2020. See Personal Communication
from Mary Price, General Counsel for Families Against Mandatory
Minimums, to Sharon Dolovich (Oct. 27, 2020) (on file with author).
7 See Sharon Dolovich, Exclusion and Control in the Carceral State, 16
BERKELEY J. CRIM. L. 259, 288–92 (2012).

11/16/20 U. Chi. L. Rev. Online *7

repercussions for the health of the broader community. In this way, it
has laid bare the folly of a penal model that measures public safety
benefits solely in terms of crime reduction. What is required instead is
a public health model that takes serious account, not only of
traditional penological objectives, but also of the harm imprisonment
inflicts on the incarcerated, on their families and communities, and on
the nation as a whole.
This Essay has two primary aims: to describe the impact of COVID
behind bars in the United States and the steps taken since March 2020
to mitigate the threat it poses, and to begin excavating the dynamics
driving the failure of the official response. Part I explains why the
incarcerated face an elevated risk of infection and potentially fatal
complications from COVID-19. Part II describes the measures various
corrections administrators took at the start of the pandemic to try to
limit viral spread inside jails and prisons, and why, however wellintentioned, these measures were insufficient to bring the virus under
control. Part III addresses the steps taken by public officials at all
levels to reduce the number of people in custody and offers initial
thoughts as to why, after a concerted push for releases on the part of
many public actors in the first months of the pandemic, these efforts
had already considerably slowed by the latter part of May 2020. (Here,
the focus is primarily, though not exclusively, on the federal courts’
nonresponse to urgent petitions from incarcerated plaintiffs.) Part IV
draws on the work of the UCLA Law COVID-19 Behind Bars Data
Project.8 It explores what the data shows regarding infection rates and
COVID deaths in custody, describes the limits of the available data,
and explains why the impact on people in jails and prisons is likely
even greater than the official numbers suggest. Part V zeroes in on the
culture of secrecy that American corrections administrators have long
been empowered to cultivate regarding what goes on behind bars. It
argues that this culture has exacerbated the threat COVID poses to
the incarcerated as well as to staff, that such secrecy is at odds with
the imperatives of a public institution, and that we need to replace the
reigning default posture of concealment with an ethos of transparency.

Since the outbreak of the pandemic, the UCLA Law COVID-19 Behind Bars
Data Project has been tracking the impact of COVID in American prisons and
jails. It also tracks the impact of COVID in immigration detention and youth
facilities, along with jail and prison releases and grassroots organizing
campaigns on behalf of incarcerated people during COVID. In addition, the
Data Project, in collaboration with Columbia Law School, Bronx Defenders,
and Zealous, maintains a database of judicial filings and court orders relating
to COVID in custody.

11/16/20 U. Chi. L. Rev. Online *8

Part VI concludes with a call for a broad normative reorientation
toward assessing carceral policy through a public health lens.
From the early days of the pandemic, public health experts have
been clear about the measures necessary to guard against infection:
social distancing, interacting with others out of doors as much as
possible, washing hands frequently, disinfecting surfaces, wearing a
clean mask, and so on. Yet for the incarcerated, taking these measures
has proved close to impossible. For one thing, social distancing is
unavailable to people in prisons and jails, most of whom live in close
proximity to others, crowded into dorms or small cells where they eat,
sleep, and live within feet and often inches of one another. Even absent
overcrowding, there will be insufficient space to allow people to remain
six feet apart. By design, American penal institutions generally fail to
provide even the minimum living space corrections experts have long
regarded as necessary to avoid “physical, mental and emotional
deterioration.”9 Compounding the problem, many American carceral
facilities continue to operate well above their rated capacity,10 with
double celling—housing two people in cells designed for one person—a
standard practice. These rooms are often so small—they can be as little
as 55 square feet—that, once the bunk, commode, and storage space
are factored in, a cell’s inhabitants are generally unable to move about
at the same time without touching. As for dorm living, it is not unusual
for prison or jail dorms to hold 100 people or more. This population
density may require double or even triple bunks, often placed so close
together that a bed’s occupant can reach out and touch the adjacent
bunk. This setup means that, while people sleep, they may be within
six feet of as many as five or eight other people. During the day, with
people constantly moving around, it is generally impossible to keep
one’s distance.
People doubled up in small cells or crowded into dorms share sinks,
toilets, and showers, which provide another vector for transmission. In
many facilities, cleaning supplies for disinfecting purposes are in short
supply; in some places, people who want bleach or other cleansers
must bargain for them on the black market. As for hand hygiene, many
Rhodes v. Chapman, 452 U.S. 337, 371 (1981) (Marshall, J., dissenting).
According to the Bureau of Justice Statistics, between 12 to 25 state prison
systems and the BOP are currently overcrowded, accounting for between
26 percent and 49 percent of all people in prison. NAT’L ACADEMIES OF
(Emily A. Wang et al. eds., 2020).


11/16/20 U. Chi. L. Rev. Online *9

people in custody lack access to sufficient soap, warm water, and clean
towels. Hand sanitizer, which must be at least 92 percent alcohol to be
effective, has long been contraband in prisons and jails (although to
their credit, since the pandemic began, many state Departments of
Corrections (DOCs) have relaxed this prohibition). Ventilation, too, is
generally poor. As a result, people in custody are constantly breathing
recirculated air. And as Americans in general are being cautioned to
avoid gathering indoors and exhorted to do their socializing outside,
lockdowns are forcing people in custody to spend ever more time
Then there are the problems on the medical side. Prisons in
particular are full of people at disproportionately high risk of
complications should they contract the virus. Because the incarcerated
population is on average younger than the overall national
population,12 one might expect the risk of complications from COVID to
be lower than in the broader society. But this is not the case. Prison
takes a physical toll, and prison health care is often grossly
inadequate, with preventative care in particular minimal at best. As a
result, people age faster when incarcerated.13 People in custody are
also disproportionately likely “to have experienced profound stress
and/or trauma over their lifetime, to have a history of substance use
disorder and/or homelessness, and to have had limited access to
quality health-care and education.”14 For this combination of reasons,
people in prison are generally physiologically older than their
Assuming they have any outdoor access at all: depending on their facility
and classification, some people in prison are only rarely allowed yard time,
and in jails, there is often no outdoor space available to detainees at the best
of times.
12 See Jeffery T. Ulmer & Darrell Steffensmeier, The Age and Crime
Relationship: Social Variation, Social Explanations, in THE NURTURE
BEHAVIOR AND CRIMINALITY 377, 377–78 (2014); Matt Vogel & Lauren C.
Porter, Toward a Demographic Understanding of Incarceration Disparities:
Race, Ethnicity, and Age Structure, 32 J. QUANT. CRIMINOL. 516, 517 (2016).
13 See Brie A. Williams et al., Addressing the Aging Crisis in U.S. Criminal
Justice Health Care, 60 J. AM. GERIATRIC SOC’Y 1150, 1151 (2012) (“The age
that a prisoner is considered to have reached the ‘older’ or ‘geriatric’
threshold varies by jurisdiction. In general the age cutoff is lower than for
non-prisoners because of the common perception that many incarcerated
persons experience ‘accelerated aging’. . . [A]t least 20 state departments of
correction and the National Commission on Correctional Health Care now set
the age cutoff for ‘older’ prisoners at 50 or 55.”).
14 Rachael Bedard, et al., Ageing Prisoners: An Introduction to Geriatric
Health-Care Challenges in Correctional Facilities, INT’L REV. RED CROSS 917,
919 (2016).

11/16/20 U. Chi. L. Rev. Online *10

chronological age would suggest, and are consequently more likely
than other members of society to have a host of medical issues known
to exacerbate complications from COVID, including heart disease,
asthma, hypertension, and diabetes.15
At the start of the pandemic, corrections administrators attempted
to avoid viral spread in their facilities by canceling visits, closing their
facilities to volunteers, and instituting lockdowns that required people
to remain in their cells or dorms almost all the time. But there is no
getting around the fact that, multiple times a day, someone still has to
pass through every corner of a facility to deliver meals and to check on
the people stuck inside their cell or dorm. Being human, people in
custody have basic needs that must be met daily if they are to survive:
food, medication, medical and mental health care, and so on. And the
imperative to attend regularly to the needs of the incarcerated is
especially acute during COVID. Some people may have contracted the
virus and urgently need medical attention. Others may be experiencing
serious anxiety, suicidal ideations, or other mental health
complications due to the stress of being locked down for extended
periods during a time of unprecedented uncertainty16—especially in
facilities that have experienced COVID-related fatalities.17 Still others
may be suffering medical complications unrelated to COVID yet find
themselves unable to get to the infirmary due to restrictions on
movement. Some daily circulation of custody officers and medical and
mental health staff is therefore unavoidable. And whatever their
purpose, every person who circulates through a facility may spread the
virus or be at risk of infection themselves.
See Peter Wagner & Emily Widra, No Need to Wait for Pandemics: The
Public Health Case for Criminal Justice Reform (2020) (reporting
disproportionate rates of asthma, high blood pressure/hypertension, diabetes,
heart problems, tuberculosis, and HIV in American jails and prisons as
compared with the American population as a whole); Brie A. Williams et al.,
Addressing the Aging Crisis in U.S. Criminal Justice Health Care, 60 J. AM.
GERIATRIC SOC’Y 1150, 1151 (2012) (“On average, older prisoners nationwide
have three chronic medical conditions and a substantially higher burden of
chronic conditions like hypertension, diabetes and pulmonary disease than
both younger prisoners and older non-prisoners.”).
16 See, e.g., Hannah Riley, SCHR Calls on U.S. Department of Justice to
Intervene as Georgia Prisons Descend into COVID-19-Related Chaos,
SOUTHERN CTR. FOR HUMAN RIGHTS (Sept. 15, 2020) (reporting that the
suicide rate in Georgia prisons in 2020 is double that of previous years).
17 See also this viral video, recorded on a contraband cellphone by a resident
of FCI Elkton on April 5, 2020, which powerfully conveys the desperation
experienced by people locked inside prisons that have had multiple COVID

11/16/20 U. Chi. L. Rev. Online *11

This last point bears emphasizing: staff too face an elevated risk of
contracting the virus. They also spend hours every day inside carceral
institutions, where they cannot socially distance and must breathe
recirculated air for many hours together. And every day, at the end of
their shifts, staff leave their facilities, potentially bringing the virus
with them into their homes and communities.
The danger COVID poses to people in congregate settings was
evident from the earliest days of the pandemic. In March 2020, the
U.S. Centers for Disease Control (CDC) issued a guidance document
enumerating best practices for “correctional and detention facilities.”
Among other issues, the recommendations covered
“cleaning/disinfecting and hygiene practices”; “social distancing
strategies to increase space between individuals in the facilities”;
screening protocols for staff, visitors, and “incoming
incarcerated/detained individuals”; and the quarantining of people
with confirmed infections or exposure to someone with COVID.18
For their part, corrections administrators around the country
began implementing measures to address the threat. Family visits
were canceled, programs were suspended, and lockdowns were
instituted for all residents not performing essential labor. A flurry of
additional policies were also adopted, including those establishing
enhanced cleaning protocols; providing for the distribution of masks,
gloves, and cleaning supplies; requiring isolation of the infected;
limiting movement and transfers between facilities; and ordering
residents to socially distance as much as possible.
As policies, these responses largely reflected the best early
understanding of how to reduce COVID transmission. But policies are
only effective if they are followed.19 With over 1,800 prisons and almost
3,200 jails across the United States, it is impossible to offer a fully
comprehensive account of how scrupulously the new policies were
implemented in all facilities. There is, however, already considerable
evidence that in countless institutions, a yawning gap quickly emerged

AND DETENTION FACILITIES (Mar. 23, 2020) (on file with author). In July
2020, the CDC issued a revised guidance document, which included
recommended protocols for testing and contact tracing.
19 See Ahlman v. Barnes, 445 F. Supp. 671, 660 (C.D. Cal. 2020) (“[A]lthough
defendants may have a policy to comply with CDC guidelines, actual
compliance has been piecemeal and inadequate.”).

11/16/20 U. Chi. L. Rev. Online *12

between announced policies and the daily experience of incarcerated
people on the ground.
Some examples: In Valentine v. Collier, the district court heard
“undisputed testimony” that in the Wallace Pack Unit in Texas,
“despite the prison’s claim of enhanced cleaning measures, its cleaning
protocol and practice remained virtually [unchanged],” with prison
officials neither increasing the number of janitors nor providing them
with the means to ensure the facility was kept properly clean. One
janitor “received only one pair of gloves to share with his co-janitor, an
arrangement medical experts described as tantamount to no gloves at
all.” In Ahlman v. Barnes, officials at the Santa Ana Jail in Orange
County “claimed that [the jail] had already achieved proper social
distancing, provided [residents] enough soap for frequent
handwashing, and isolated and tested all symptomatic individuals.”
Yet the detainees told a different story, filing sworn affidavits that
described “being transported back and forth to the jail in crammed
buses, socializing in dayrooms with no space to distance physically,
lining up next to each other to wait for the phone, sleeping in bunk
beds two to three feet apart, and even being ordered to stand closer
than six feet apart when [they] tried to socially distance.” In addition,
the district court in Ahlman found that detainees did “not receive
sufficient cleaning supplies to keep their living areas clean and
disinfected,” and heard testimony that “the cloth masks provided” were
“not replaced for weeks” or were “made from blood- and feces-stained
In Marlow v. LeBlanc, the district court found “credible
testimony” out of Rayburn Correctional Center that the Louisiana
Department of Corrections was not following its own COVID response
guidelines. As a consequence, “no procedures have been implemented
to avoid chokepoints in the walkways” of the dorms, so that “foot traffic
often results in [residents] and staff ‘almost touching’ each other.”20 At
mealtimes, residents were standing “in a heel to toe fashion” while
awaiting their trays and eating while sitting “directly next to one
another.” Kitchen workers “only occasionally wear masks . . . while
serving food,” and those waiting to be seen at the medical clinic stand
“shoulder-to-shoulder.” In Ohio’s Pickaway Prison—an early COVID
hotspot—residents living in crowded dorms reported resorting to
“hanging bed sheets from the top rack of their bunks to protect
themselves from others’ coughing, sneezing, and breathing.” In federal
Marlowe v. LeBlanc, No. 18-63-BAJ-EWD, 2020 U.S. Dist. LEXIS 72146, at
*12 (M.D. La. Apr. 23, 2020), but see Marlowe v. LeBlanc, 810 F. App'x 302,
303-04 (5th Cir. 2020) (granting defendant’s request for a stay of injunctive
relief ordered by district court).

11/16/20 U. Chi. L. Rev. Online *13

Bureau of Prisons (BOP) facilities, correctional officers (COs) and other
staff reported feeling “pressure to work even after being exposed to sick
prisoners.” A resident of FCI Elkton in Ohio posted a YouTube video
made on a contraband cell phone showing “men packed together in
their cubicles, sleeping and wheezing.”
Some of this institutional failure may be a product of bad faith
on the part of corrections officials. But even prison administrators and
staff motivated to try to mitigate the COVID risk face impossible odds.
Carceral institutions are simply not conducive to limiting the spread of
a highly contagious airborne virus. At the best of times, typical staffing
ratios generally allow correctional officers to do only the bare
minimum in terms of providing food, medications, access to showers,
visits, and other basics. The physical design and layout of the
institutions themselves are not conducive to healthy living; sanitation
and hygiene are perennial challenges, and people have little personal
space or access to fresh air. Nor is there generally much trust between
staff and residents, which makes it difficult to get the buy-in necessary
to effectively implement even well-conceived policies. To be sure, this
absence of trust is well earned; to take just one example, after years of
COs responding to internal disorder by imposing solitary
confinement—an experience condemned by the UN Special Rapporteur
on Torture as “cruel, inhuman, or degrading”21—many people in
custody have reported being hesitant to admit to being COVIDsymptomatic out of fear of being sent to “the hole.” During a pandemic,
when the margin of error is small, all these dynamics only exacerbate
matters, making it even less likely that institutions will operate in
ways necessary to ensure the health and safety of everyone involved.
In short, it did not take long to see that, however wellintentioned the staff and however well-conceived the policies, prisons
and jails would be unable to effectively contain the spread of COVID
without considerable reductions in population density. Some
corrections officials sought to achieve this end without releases, but
these efforts were largely ill-conceived. For example, in some
jurisdictions, administrators tried to achieve lower density by
reopening and transferring people into individual units or entire
facilities that had previously been decommissioned. But, unless
preceded by a strict fourteen-day quarantine, transfers create their
own risk of transmission, as the residents and staff of California’s San
Quentin prison found when transfers from the California Institute for
Men in Chino, effected in late May 2020 with insufficient precautions
against COVID spread, sparked a viral explosion in San Quentin.
The phrasing used is taken directly from the UN Convention Against
Torture and Other Cruel, Inhuman or Degrading Treatment or Punishment.

11/16/20 U. Chi. L. Rev. Online *14

Since then, over 2,200 residents of San Quentin have tested positive
for COVID and at least 28 people have died.
Furthermore, if carceral facilities are standing empty, it is
generally for good reason. This became clear in the Cook County jail,
when in April 2020 the sheriff sought to spread out the jail’s
population by reopening tiers “that had been closed for some time.” 22 In
the ensuing lawsuit, plaintiffs reported finding conditions in these
units to be “filthy, [to] lack heat, and [to] lack running or potable
water.” Similar events played out in Alabama, where the Department
of Corrections, seeking space for COVID quarantine, reopened Draper
Correctional Center, which had been closed in 2018 after federal
investigators “found the facility plagued by rats, maggots, open sewage
and toxic fumes.” Those transferred into Draper for quarantine
reported finding the conditions “unsanitary and inhumane,” lacking
even working toilets.23
These experiences in Cook County and Alabama are reminders
that the state’s obligation to ensure decent living conditions for the
incarcerated is not suspended during a pandemic.24 Strategies to
reduce viral spread must instead be undertaken with an ongoing
recognition of the humanity of the incarcerated and the constitutional
imperative that they be treated humanely. Among those taking this
imperative seriously, it was obvious from the first that, if there was to
be a reduction in the population density of prisons and jails, there was
only one fit method: people needed to be released from custody in
numbers sufficient to enable social distancing by those who remained

Reply in Support of Plaintiff’s Renewed Motion for Preliminary Injunction
at 6, Mays v. Dart, No. 1:20-CV- 02134 (N.D. Ill. Apr. 19, 2020), ECF No. 64.
23 Given the conditions that led the Alabama DOC to close the facility, this
situation should not be surprising. Had the facility been habitable, it would
not have been closed. On the state of the Alabama prisons pre-COVID, see
24 On this obligation more generally, see Sharon Dolovich, Canons of Evasion
in Constitutional Criminal Law, in THE NEW CRIMINAL JUSTICE THINKING
111, 130–32 (Sharon Dolovich & Alexandra Natapoff eds., 2017); Sharon
Dolovich, Evading the Eighth Amendment: Prison Conditions and the Courts,
133, 137–40 (Meghan J. Ryan & William W. Berry III eds., 2020).

11/16/20 U. Chi. L. Rev. Online *15

The need for significant releases to reduce the danger of
widespread infection and COVID-related fatalities in custody was clear
to many observers and stakeholders from the earliest days of the
pandemic. Nationwide, numerous public officials with the authority to
order releases quickly mobilized. On the jail side, the effect was
considerable, with the median jail population dropping 31 percent
between early March and mid-May. In the prisons, the drop was less
pronounced, though still substantial—roughly 8 percent between
March and June. These efforts, however, did not last. By the end of
May, the rate of releases had fallen considerably. Jail populations
began to creep back up, and have continued a steady upward climb. As
for prisons, temporary population reduction efforts—several of which
consisted of suspending transfers from local jails to state prisons—
have been halted, driving a subsequent increase in population. Some
minimal prison releases continue to occur, but eight months on, with a
few notable exceptions, they have slowed to a trickle. In what follows, I
provide a brief snapshot of the ways public officials have sought to
combat COVID through releases. I then speculate as to why, after an
initial push from March to May 2020, these efforts largely ceased.
The range of available population reduction strategies varied by
context, with jail officials having more levers to drive down numbers
than did prison officials. This disparity is a consequence of the
different populations these institutions house. Prisons are places
where people go to stay: they hold people serving sentences for felonies,
i.e., crimes that carry penalties of more than one year. Jails, by
contrast, are mostly temporary waystations25: in the main, they house
people who are awaiting trial, awaiting sentencing, on probation
violations or immigration holds, or serving short sentences for
misdemeanors. The fact that jails hold people who are pretrial or who
have committed relatively minor transgressions, coupled with the
sheer volume of people moving in and out of jail every year—more than
10 million people churn annually through the nation’s roughly 3,200
jails—means that, apart from a few high-profile detainees, the public
pays little attention to who is in jail. Those officials who take steps to
shrink jail populations consequently face relatively little political risk

The average jail term is less than 30 days. But some people may spend
years fighting their cases. And in California, as a consequence of
Realignment (AB 109), people convicted of non-serious, non-violent, non-sexrelated offenses will serve their full sentences in county jails, which can mean
years in custody in facilities designed to hold people only temporarily.

11/16/20 U. Chi. L. Rev. Online *16

of the sort that has largely thwarted meaningful decarceration efforts
since the “tough on crime” era began.26
When COVID-19 hit, there were thus several available mechanisms
for reducing jail populations that were unlikely to draw much political
blowback. Some involved diverting people from jail altogether.27 In
East Baton Rouge, the local sheriff ordered his officers to stop
arresting people for most misdemeanors. In Fort Worth, law
enforcement officials were directed to issue citations for low-level
offenses instead of arresting suspects and booking them into the jail.
In Racine County, Wisconsin, the Sheriff’s Office restricted jail
admission solely to those individuals suspected of violent crimes. In
Maine, the chiefs of the state’s superior and district courts issued an
order vacating over 12,000 outstanding warrants for failure to appear
or for any unpaid fines or fees. In South Carolina, the state Supreme
Court directed courts not to issue bench warrants for failure to appear,
and to release on their own recognizance, without bond, anyone
charged with a non-capital crime, unless found to pose “an
unreasonable danger to the community” or “an extreme flight risk.” 28
In other instances, steps were taken to release from the jails people
already in custody. In San Francisco, the district attorney ordered
prosecutors not to oppose motions to release people facing
misdemeanor charges or felony drug charges unless they were found to
pose a public safety threat. In Los Angeles, the sheriff ordered the
release of 1,700 people who had been sentenced to jail time for nonviolent offenses and had less than 30 days left to serve. In New York
City, the mayor ordered the release of three hundred elderly, medically
compromised individuals from Rikers Island. In several cases, releases
from county jails were the product of collaborative efforts among
various stakeholders. In Washington County, Arkansas, local jail
officials worked with local prosecutors and circuit judges to release
approximately 150 people on home monitoring, and sought (and
received) state approval to release 33 people serving 90-day sentences
See Dolovich, supra note 3 (examining the political dynamics that have
driven the forces of mass incarceration over the past five decades).
27 During this same period, some state prison systems also halted or limited
intake. But because the people affected were bound for state prison, these
efforts only created a backlog in county jails. It is therefore only diversionary
efforts at the local level, which prevent the intake of new arrivals into the
jail, that would reduce population density in carceral facilities as a whole.
28 All examples in this and the following paragraph are taken from the UCLA
Law COVID-19 Behind Bars Data Project jail releases tracker. See COVID19-Related Jail Releases, UCLA LAW COVID-19 BEHIND BARS DATA PROJECT.
I thank Maddy DeLone and her team for their commitment to this enterprise.

11/16/20 U. Chi. L. Rev. Online *17

for parole violations. In New Jersey, following mediation involving the
Attorney General, the County Prosecutors’ Association, the Office of
the Public Defender, and the ACLU of New Jersey, the state’s Supreme
Court ordered the release of anyone serving time in jail as a condition
of probation, on a probation violation, pursuant to a municipal court
conviction,29 or for a misdemeanor.
These combined efforts had a notable effect. By mid-May 2020, the
median national jail population had dropped by 31 percent from the
start of the pandemic. Unfortunately, rather than continuing on this
path, officials across the board began to pull back on their efforts to
reduce jail populations. Starting in mid-May, the jail numbers began to
creep back up, and as of October 1, at least 50 percent of the reductions
had been erased by new admissions, with all signs indicating that this
upward trend will continue.
Early in the pandemic, some decarcerative efforts were also seen on
the prison side, although without the low-hanging fruit available in the
jails, their impact was more modest. Some examples: In California, the
governor accelerated by up to 60 days the releases of 3,500 people who
had already been found suitable for parole but still awaited expiration
of the statutory waiting period. In Kentucky, the governor commuted
the sentences of over 900 people serving prison sentences for “nonviolent, non-sexual” crimes. Pennsylvania’s governor used his reprieve
power to accelerate the releases of over four hundred people with
medical conditions that put them at high risk of complications from the
virus. In North Dakota, the state’s parole board held a special session
and granted early parole release to 120 individuals. The Nebraska
parole board, likewise in a special session, approved early release for
56 people.30
These efforts continued as 2020 progressed. In early June, the
Arkansas Board of Corrections certified more than 1,200 people as
eligible for parole consideration. As of early July, 730 people had been
released, leaving the Arkansas prison system roughly at capacity for
the first time since 2007. More recently, federal judges in Connecticut
and California ordered the federal Bureau of Prisons (BOP) to compile

See Alexandra Natapoff, Criminal Municipal Courts, 134 HARV. L. REV.
(forthcoming 2020).
30 All examples in this paragraph are taken from the UCLA Law COVID-19
Behind Bars Data Project prison releases tracker. See COVID-19-Related
thank Maddy DeLone and her team for their important work.

11/16/20 U. Chi. L. Rev. Online *18

a list of all medically vulnerable residents at FCI Danbury31 and FCI
Lompoc,32 and to review each person on that list for their suitability for
home confinement. Following completion of that process, 119 people
were released from Danbury, and to date, 165 people have been
released from Lompoc (44 to home confinement and 121 to halfway
houses), with an additional 81 granted compassionate release after the
process was kickstarted by the federal litigation.33 In addition, federal
courts have been exercising their authority under the First Step Act to
consider compassionate release petitions from people in BOP custody.
Although the vast majority of these petitions are denied, to date,
roughly 1,700 people34 have been granted compassionate release from
federal custody via the courts. In addition, since the pandemic began,
at least one legislature has pursued a statutory response. In late
September 2020, the New Jersey legislature passed a law that, with
some stipulated exclusions, awards “public health emergency credits”
of six months’ remission from prison sentences “for each month or
portion of each month, served during the declared [COVID]
emergency.” The impact of this bold legislative move serves as a
reminder that legislatures hold broad decarcerative power; in early
November 2020, pursuant to the new law, 2,258 people who were
within a year of completing their sentences were released from New
Jersey prisons, with another 1,167 to be freed in the coming months.35
Given the threat of viral spread in crowded facilities, all these steps
were welcome ones. They were not, however, enough to contain the
risk. What is required is a reduction in population density sufficient to
allow those who remain to socially distance. As we have seen, even
facilities that are not officially overcrowded but are simply operating at
capacity are too crowded to adequately address the COVID threat.
Although the precise target is unclear, a June 2020 study of a large
urban jail found that an increase from 26 percent to 54 percent in the
number of people in single cells, along with a suite of other protective
measures (screening, testing, paid sick leave for staff, etc.), produced
See Martinez-Brooks v. Easter, No. 3:20-cv-00569, 2020 U.S. Dist. LEXIS
83300, at *102 (D. Conn. May 12, 2020).
32 See Torres v. Milusnic, No. CV 20-4450-CBM-PVC, 2020 U.S. Dist. LEXIS
131446, at *67 (C.D. Cal. July 14, 2020).
33 Personal Communication from Naeun Rim, Principal Attorney at Bird,
Marella, Boxer, Wolpert, Nessim, Drooks, Lincenberg & Rhow, P.C., to
Sharon Dolovich (Oct. 19, 2020) (on file with author).
34 Personal Communication from Mary Price, General Counsel for Families
Against Mandatory Minimums, to Sharon Dolovich (Oct. 27, 2020) (on file
with author).
35 See Tracey Tully, 2,258 New Jersey Prisoners to be Released in a Single
Day, N.Y. Times (Nov. 6, 2020).

11/16/20 U. Chi. L. Rev. Online *19

an estimated “83% reduction in predicted symptomatic cases and
hospitalizations and an 89% reduction in predicted deaths.” Double
celling has long been standard practice in most state prison systems,
and the 8-percent drop in the national prison population is far from
enough to enable prisons to single-cell half their occupants. Even in
California, which has reduced its prison population by 30 percent since
early March 2020, these efforts have only been sufficient to reduce
prison overcrowding to 109 percent of capacity.
Prior to COVID, such a rapid reduction in the number of people
incarcerated would have been nothing short of remarkable. Over the
past decade, although public condemnation of mass incarceration had
become commonplace and policymakers across the political spectrum
joined the call for change, the number of people in custody has
remained stubbornly high. The ability of public officials to so rapidly
shrink the American carceral footprint in just two months (mid-March
to mid-May) has demonstrated that, with sufficient political will,
meaningful reductions are possible. And yet, despite the ongoing
danger COVID-19 poses to the health and safety of the incarcerated,
most of the efforts by public officials to reduce the number of people in
their facilities slowed almost to a halt two months into the pandemic.
The question is why in the main these early efforts were so
precipitously abandoned. It is too soon for definitive assessments, but
having closely watched these events unfold, I see at least four possible
explanations.36 First, there was simple COVID fatigue. After a strong
initial national response to the virus, including the extraordinary step
of self-quarantine adopted by millions of people across the country, the
problem did not go away, but continued to grow and spread. As the
public, feeling exhausted and overwhelmed, watched this progression,
people seemed to stop expecting a robust and effective governmental
response. Among the ill effects of this fatigue and the consequent
resetting of expectations appears to have been an easing of pressure on
state and local officials to address the looming crisis in the prisons and
Second, after the murder of George Floyd on May 25, 2020 and the
explosion of political protests that quickly followed, many of the
journalists who had been covering the impact of COVID on
incarcerated people shifted their attentions to the Black Lives Matter
movement and the urgent issue of police violence against Black
Americans and other people of color. For those committed to the cause
of racial justice and to a radical rethinking of the American criminal
system, this media focus was vital. It also, however, seems that,

This list is not intended to be exhaustive.

11/16/20 U. Chi. L. Rev. Online *20

although some BLM leaders centered in their advocacy both mass
incarceration and the treatment of the incarcerated, the danger
COVID continued to pose to people in custody largely fell off the radar.
Again, the waning of attention seemed to have eased the pressure
public officials felt to continue their efforts to decarcerate.
Third, there is the problem of inadequate data. I return to this
issue in more detail below, but for now, the key point is that many of
the prisons that have been reporting COVID infections and death rates
lack adequate testing, which has kept official numbers artificially low.
The effect has been to propagate an unduly rosy picture that allows
officials to claim—and perhaps even to believe—that whatever
mitigation efforts they have undertaken must have staved off the
worst, making further releases unnecessary. On the jail side, with
most facilities nationwide failing even to track infections, officials have
been able to remain oblivious to the real possibility of wide viral spread
in their jails. Whether these actors are knowingly dissembling or
whether, by failing to investigate, they have convinced themselves that
no problem exists, officials in many jurisdictions have, for lack of
contradictory evidence, felt able to publicly state that COVID no longer
poses a threat in their facilities and thus that no further releases are
Fourth, and likely most consequentially, by mid-May, it was clear
that the federal courts were not going to be an effective channel for
release, thus negating any pressure correctional officials may have felt
to take ameliorative steps to avoid possible legal liability. In the early
days of the pandemic, advocates around the country began filing
petitions in the federal courts, seeking orders requiring the adoption of
measures to mitigate viral spread and the release of people whose age
or medical condition made them especially vulnerable. These efforts
had some initial success, with several courts granting preliminary
injunctions or temporary restraining orders directing correctional
officials to improve conditions inside or to identify facility residents at
highest risk from COVID in preparation for their release.37 But

See, e.g., Wilson v. Williams, No. 4:20-cv-00794, 2020 U.S. Dist. LEXIS
70674, at *25 (N.D. Ohio Apr. 22, 2020), see also Wilson v. Williams, No. 4:20cv-00794, 2020 U.S. Dist. LEXIS 87607, at *6 (N.D. Ohio May 19, 2020) (FCI
Elkton); Valentine v. Collier, No. 4:20-CV-1115, 2020 U.S. Dist. LEXIS 68644,
at *33 (S.D. Tex. Apr. 20, 2020) (Texas Wallace Pack Unit Prison); Mays v.
Dart, No. 20 C 2134, 2020 U.S. Dist. LEXIS 62326, at *51 (N.D. Ill. Apr. 9,
2020) (Cook County jail); Swain v. Junior, No. 1:20-cv-21457-KMW, 2020
U.S. Dist. LEXIS 60878, at *3 (S.D. Fla. Apr. 7, 2020) (Miami-Dade jail).

11/16/20 U. Chi. L. Rev. Online *21

plaintiffs did not prevail at trial everywhere.38 And even when they
did, in most instances, it did not take long for the appellate courts to
step in on the side of defendants.39 In case after case, appeals courts
granted stays of district court orders on grounds strongly suggesting a
general lack of sympathy with plaintiffs’ arguments.
Not every appeals court issued a stay. Wilson v. Williams was a
case out of FCI Elkton, a low-security federal facility in Ohio that
emerged as an early COVID hotspot, with six people dead by April 22.
In that case, the Sixth Circuit twice declined to stay the preliminary
injunction issued by the district court. But this victory proved shortlived. Having lost in the Sixth Circuit, BOP officials took their case to
the Supreme Court, and in the process elicited an unmistakable sign
that corrections officials need not fear legal liability should they fail to
protect people in custody from the threat of COVID.
It is worth spending a moment on the details of Wilson. Perhaps
more than any other litigation, this case supports the conclusion that,
even on the most compelling facts and even during a state of
emergency, the federal courts, guided by the Supreme Court, will be
putting no pressure on corrections administrators to do anything more
than the bare minimum to protect the incarcerated from COVID—
despite the constitutional duty of care that jail and prison officials
plainly bear towards the people in their custody.
The case was brought in mid-April on behalf of all current and
future residents of FCI Elkton, requesting that prison administrators
create a list of all medically vulnerable residents and find some way to
remove them from the facility. Judge James Gwin of the Northern
District of Ohio certified the class and issued a preliminary injunction
directing BOP officials to create the list and, within two weeks, to find
alternative housing arrangements for the people on it, whether
through release to home confinement or transfers to other facilities. 40
In reaching this result, the district court acknowledged that BOP
See, e.g., Money v. Pritzker, No. 20-cv-2093, 2020 U.S. Dist. LEXIS 63599
(N.D. Ill. Apr. 10, 2020) (Illinois state prisons).
39 See, e.g., Swain v. Junior, 958 F.3d 1081, 1092 (11th Cir. 2020) (staying
preliminary injunction); Valentine v. Collier, 956 F.3d 797 (5th Cir. 2020)
40 Wilson v. Williams, No. 4:20-cv-00794, 2020 U.S. Dist. LEXIS 70674, at *2
(N.D. Ohio Apr. 22, 2020). Given what is now known about the risk of COVID
spread via transfers, it is clear that releases to home confinement represent
the safest option. At a minimum, any inter-prison transfers would need to be
preceded by a strict fourteen-day quarantine (as would release to home
confinement, to prevent people from spreading the virus to their families and
communities once they have been released from custody).

11/16/20 U. Chi. L. Rev. Online *22

officials had taken some steps “to lessen the Covid-19 threat.” They
had been segregating new arrivals for fourteen days. They had
developed a protocol for evaluating and responding to symptomatic
individuals. They had established a system for regular temperature
checks. But Judge Gwin found that officials were “fight[ing] a losing
battle.” People in Elkton lived in dorms. There was no testing protocol
and seemingly minimal testing. Six Elkton residents had already died
of COVID, yet the prison at the time was reporting only 59 positive
test results among its residents. As the district court pointed out, by
that time, officials at Marion Correctional Institution, an Ohio state
prison with a population comparable to Elkton’s, had already
conducted mass testing, with at least 1,950 Marion residents testing
positive. This finding at Marion, coupled with the number of fatalities
already recorded at Elkton by mid-April, strongly suggested a more
widespread outbreak in Elkton than official data acknowledged.
Although Elkton officials had restricted movement through the prison,
residents remained in units of 150 people, which, as Judge Gwin noted,
were nothing like “family units,” as Elkton officials sought to
characterize them. Some incarcerated workers continued to circulate
through the facility, as did the staff, who were also daily exposed to the
risk of viral spread. On the basis of these facts, Judge Gwin issued the
preliminary injunction plaintiffs sought. In turn, BOP officials sought
a stay pending appeal, which the Sixth Circuit denied.41
Meanwhile, at Elkton, the virus continued to rage. Although, as of
early May, testing in Elkton still lagged, of those people who were
tested, 24 percent were found to be infected (a number the district
court later labeled “unacceptable”).42 Of 2,417 people then housed in
the facility, Elkton officials had identified 837 individuals as being
over 65 “or as having significant pre[-existing] health conditions
making them especially vulnerable to COVID-19.” Yet by May 8, they
had judged only five people on that list as appropriate for home
confinement and a further six as “maybe qualifying.” At this point,
plaintiffs went back to the district court with an “Emergency Motion to
Enforce the Preliminary Injunction.” In his subsequent order, Judge
Gwin found the BOP response to that point to have been beyond
inadequate, emphasizing that the BOP had statutory authority to
release to home confinement people who were elderly or terminally
ill—authority that had been expanded by the CARES Act passed by
Congress in late March in response to COVID. What’s more, U.S.
Wilson v. Williams, No. 20-3447, 2020 U.S. App. LEXIS 14291, at *6 (6th
Cir. May 4, 2020).
42 Wilson v. Williams, No. 4:20-cv-00794, 2020 U.S. Dist. LEXIS 87607, at *4
(N.D. Ohio May 19, 2020).

11/16/20 U. Chi. L. Rev. Online *23

Attorney General William Barr had in an April 3 memo directed BOP
officials to “prioritize the use of [their] various statutory authorities to
grant home confinement” to people in BOP custody “seeking transfer in
connection with the ongoing COVID-19 pandemic.” By “thumbing their
nose at their authority to authorize home confinement,” Judge Gwin
found, Elkton’s administrators “threaten[ed] staff and . . . low security
[residents].” On May 19, therefore, the district court ordered
defendants “to make full use of the[ir] home confinement authority
beyond the paltry grants . . . it has already issued”; to respond in short
order, on terms consistent with statutory directives, to all the petitions
for compassionate release filed by Elkton residents; and to transfer as
many Elkton residents as possible to other facilities “where social
distancing is possible.”
What happened next set the stage, not only for the trajectory of
Wilson v. Williams, but for virtually all cases brought on behalf of
incarcerated plaintiffs seeking relief in the wake of COVID. On May
21, the Wilson defendants appealed to the U.S. Supreme Court for a
stay of the preliminary injunction, an effort that received vehement
support from the Solicitor General, who strongly asserted the
inappropriateness of the federal courts directing the BOP’s COVID
response. Although the Supreme Court denied the request, it
emphasized in its brief order that the issue was procedural: defendants
had not yet appealed to the Sixth Circuit for a stay of the May 19
order. And despite this procedural infelicity, the Court’s order noted
that three Justices—Clarence Thomas, Samuel Alito, and Neil
Gorsuch—were already prepared even then to grant the BOP relief.
The defendants duly returned to the Sixth Circuit, which ordered
expedited briefing but denied their request for a stay. Procedural
infelicity resolved, the BOP went back to the Supreme Court, which on
June 4, to no one’s surprise, gave the government the relief it sought.
Days later, the Sixth Circuit issued an opinion on the merits vacating
the preliminary injunction.43
I leave for another day an account of the legal grounds on which the
federal courts have rejected claims for relief from the COVID threat
brought by incarcerated plaintiffs.44 For present purposes, what
See Wilson v. Williams, 961 F.3d 829, 845 (6th Cir. 2020) (vacating April
22, 2020 preliminary injunction); Wilson v. Williams, No. 20-3547, 2020 U.S.
App. LEXIS 29862, at *2 (6th Cir. Sep. 17, 2020) (vacating May 19, 2020
44 The main issues have been a mix of procedure (e.g., whether habeas corpus
is the appropriate vehicle for the relief sought, or whether the requirements
of 18 U.S.C. § 3626 have been satisfied) and substance (e.g., whether

11/16/20 U. Chi. L. Rev. Online *24

matters is this: In Wilson, the district court found that plaintiffs—
residents of a low-security facility, many close to their release dates—
faced a substantial risk of serious harm and even premature death,
and that, despite having over 800 medically vulnerable people in their
custody and the legal authority to provide for their protection, prison
officials had failed to take readily available steps to keep them safe. If
ever there was a strong imperative for judicial action in the face of
regulatory failure on the part of the political branches, it was here.
And yet, in literal short order, with no substantive analysis, the
Supreme Court strongly signaled its opposition to judicial intervention,
a move raising the obvious question of which cases would merit
judicial enforcement of constitutional protections for the incarcerated.
Supreme Court orders in later cases,45 along with a slew of
subsequent circuit court opinions, have confirmed that the federal
courts have largely maintained a hands-off posture in the face of the
COVID threat.46 But even by the third week of May, by which time a
number of appellate decisions had already been entered and the first
Supreme Court order issued in Wilson, corrections officials would have
seen enough to know which way the wind was blowing. Although it is
hard to say by how much, it is certain that, to some extent, the fear of
future liability drove prison officials early in the pandemic to try to
plaintiffs can demonstrate the deliberate indifference required to succeed on
their Eighth Amendment claims), with several opinions generally invoking
the familiar theme of the need for judicial deference to corrections officials.
On the centrality of judicial deference in prison law doctrine more generally,
see Sharon Dolovich, Forms of Deference in Prison Law, 24 FED. SENTENCING
REP. 245 (2012).
45 See Barnes v. Ahlman, 140 S. Ct. 2620, 2621-22 (2020) (Sotomayor, J.,
dissenting); Valentine v. Collier, 140 S. Ct. 1598, 1599 (2020) (Sotomayor, J.,
46 Since the start of the pandemic, an as-yet unknown number of people have
been released from custody on individual habeas petitions. Yet to date, of the
innumerable class actions that have been brought by incarcerated plaintiffs
since March, only two—Martinez-Brooks v. Easter and Torres v. Milusnic—
have yielded releases. Each involved a single federal facility—FCI Danbury
in Connecticut (Martinez-Brooks) and FCI Lompoc in California (Torres). And
in each case, the number of people released has thus far been relatively
small, 119 from Danbury and 165 from Lompoc. The grounds for relief in
those cases also bear noting: in each, as in Wilson, the judge found that BOP
officials had not exercised the authority to expand the use of home
confinement granted them in the April 3 Barr memo, and ordered them to do
so. The ability of the district courts in this suite of cases to rely on such
executive direction suggests that, given the current state of the governing
legal doctrine, the possibility of judicial action may depend on meaningful
engagement in the decarcerative enterprise by the political branches.

11/16/20 U. Chi. L. Rev. Online *25

reduce the size of their incarcerated populations. The evident
unwillingness of the federal courts to intervene on behalf of plaintiffs
even at the height of the emergency and even under the most
compelling circumstances almost certainly contributed to the waning of
these efforts.
By the end of March 2020, most state Departments of Corrections
and some of the larger jails had begun to track the virus in their
facilities and to post their findings on online dashboards. The metrics
reported were not uniform and included a range of variables, such as
COVID tests conducted, positive infections, active cases, recovered
cases, and deaths from COVID. Some DOCs reported this data only for
prison residents, while others included data concerning staff. Despite
the variation, by persistently tracking the posted data, it was possible
to get a rough picture of the officially reported numbers of COVID
infections and deaths in state and federal prisons. And by comparing
this data with the number of reported COVID infections and COVIDrelated deaths in the country as a whole, one could determine the
relative reported impact of COVID-19 in American prisons.
The first such analysis (on which I was a co-author) was published
in mid-July 2020 and confirmed the dire predictions as to the spread
and impact of COVID behind bars.47 Relying on data gathered through
June 6, 2020 from BOP and state DOC dashboards by the UCLA Law
COVID-19 Behind Bars Data Project48 (which I direct), we found that
the COVID-19 infection rate was 5.5 times higher in prisons than in
the U.S. population. Even given our expectation of some disparity, this
number was eye-popping. As for COVID deaths, making sense of the
relative rate was not as straightforward. In terms of the broad
numbers, we found an elevated risk among people in custody, with a
crude death rate in prisons of 39 deaths per 100,000 prison residents,
as opposed to 29 deaths per 100,000 people in the United States more
broadly. On these numbers, people in prison had a 34 percent higher
risk of dying from COVID than did people in society at large. But
See Brendan Saloner, Kalind Parish, Julie A. Ward, Grace DiLaura &
Sharon Dolovich, COVID-19 Cases and Deaths in Federal and State Prisons,
324 J. AM. MED. ASSOC. 602 (2020).
48 From the end of March until mid-June, the daily work of gathering this
data was conducted by Grace DiLaura and Kalind Parish, who served as CoLeads of the UCLA Law COVID-19 Behind Bars Data Project’s data team.
From late June to the present, the work has been a collective effort by
Michael Everett, Chase Hommeyer, Hope Johnson, Neal Marquez, and
Kalind Parish. I thank them all for their dedication and their consistently
excellent work.

11/16/20 U. Chi. L. Rev. Online *26

compared with the size of the disparity on infections, with prison
residents facing a risk of infection 550 percent greater than the general
public, this death rate at first seemed puzzlingly low, and perhaps to
suggest that predictions of disproportionate risk of death from COVID
in prison were off the mark.
However, the puzzle clarified once age disparities were taken into
account. In American society as a whole, people over age 65 have died
from COVID at much higher rates than people in younger age
brackets. And because most people who commit serious crimes do so
when they are young, people in prison are on average much younger
than the general public. Given this age disparity, the fact that, on the
raw numbers alone, the death rate among people in prison was
34 percent greater than in society as a whole is striking in itself. But to
make the comparison meaningful, it is necessary to calculate relative
death rates assuming a comparable age distribution between the two
populations. And once we adjusted for age and sex —meaning we
calculated what the COVID death rate in prison would be if the
incarcerated population had the same age and sex distribution as the
U.S. population—we found that people were dying of COVID-19 in
prison at 3.0 times the rate seen in the country as a whole.
Even taken alone, these findings are enough to demonstrate a
public health disaster. But if anything, they very likely understate the
disparities. Take infection rates. Our findings were based on the
number of officially reported positive tests. But the data indicates that
many prison systems are not systematically testing, if they are testing
at all. Even now, almost nine months into the pandemic, the most
commonly reported number of cumulative confirmed cases among the
1,406 unique facilities being tracked daily by the UCLA Law COVID19 Behind Bars Data Project—the cumulative number of COVID
infections reported by 363 separate facilities—is zero. An additional
265 prisons are reporting fewer than 10 infections since March, with
another 114 facilities reporting fewer than twenty.49 There may well be
prisons into which COVID-19 has not yet penetrated, as well as those
in which, despite some confirmed cases, the virus has not widely
spread. But these numbers strongly suggest that in many facilities,
prison officials are testing only the most highly symptomatic
individuals, if they are testing at all, and that they are consequently
underreporting the extent of viral proliferation.
To be sure, testing is also highly variable in American society in
general, which means that the number of reported positive COVID
tests in the nation as a whole that we used for comparison purposes

I thank Kalind Parish and Hope Johnson for running these numbers.

11/16/20 U. Chi. L. Rev. Online *27

may also represent a considerable undercount. There is, however,
strong evidence that the lack of mass testing is cloaking far higher
numbers in the prisons than in society as a whole. 50 The first signs
came in April, from two Ohio prisons, Marion CI and Pickaway CI,
when the state’s governor, Mike DeWine, ordered mass testing in those
facilities. At Marion, population 2,500, 2,300 tests were administered,
and 2,028 residents tested positive.51 At Pickaway, at least 77 percent
of the prison population was found to be infected. The same pattern
emerged at other prisons around the country, with expanded testing
protocols in select institutions revealing extremely high infection rates
inside. In North Carolina’s Neuse CI, 60 percent of prison residents
have tested positive. In Arkansas’ Cummins Unit Prison, capacity
1,876, 956 prison residents were found to have the virus. At Butner
Low FCI, there were 638 infections out of a population of 1,197. At FCI
Elkton—the facility at issue in Wilson—933 out of 2,004 prison
residents have now tested positive. In Michigan’s Lakeland prison, out
of 1,440 tests administered, 813 came back positive. And in California,
mass outbreaks have been recorded at prisons across the state,
including CSP Chuckawalla (1,397 positive tests of 2,822 residents),
ASP Avenal (2,938 of 3,887), and San Quentin (2,239 of 3,989). Given
the relative uniformity of conditions across prisons, the same forces
that produced these infection rates at the prisons listed are likely
producing similar infection rates at other facilities, which remain
unreported only because testing has not been done. There is unlikely to
be any county in the United States in which mass testing in the
population as a whole would reveal infection rates this stratospheric. 52
As COVID infections and deaths have spiked nationally, the
disparity between the impact on the incarcerated and on society as a
whole has narrowed somewhat, but it remains considerable. As of
September 22, 2020, the disparity in infection rates in federal and
“Mass testing” occurs when a critical mass of a prison’s population is
tested. Universal testing would mean that every person in a facility, staff as
well as residents, is regularly tested on an ongoing basis. To my knowledge,
there is no prison in the country that has yet instituted universal testing.
51 Unless otherwise specified, all data on infection rates listed in this
paragraph reflect cumulative cases as of October 27, 2020 and are drawn
from the UCLA Law COVID-19 Behind Bars Data Project. All population
data cited in this paragraph are taken from the most recent data reported in
the HIFLD. I thank Michael Everett for pulling these data together.
52 In mid-November 2020, 19 percent of people tested in Newark, New Jersey,
were found to have the virus. This finding, regarded by local officials as the
alarming figure it is, prompted sweeping containment measures across the
city aimed at minimizing social interaction and enabling effective social

11/16/20 U. Chi. L. Rev. Online *28

state prisons as compared with the broader population had dropped
from 5.5 to 1 to 4.8 to 1.53 Likewise, as of October 10, 2020, the
adjusted death rate had dropped from 3.0 to 2.7. 54 We should not let
the fact of these reductions obscure the bottom line: these disparities
remain extremely high and confirm that COVID-19 continues to have a
disproportionate effect on people in custody, with many suffering from
preventable illness and premature death. Moreover, given what we
now know about the possible long-term effects of the virus, some
number of those in custody who were symptomatic but did not die will
likely be left with serious chronic health problems and even cognitive
Yet there remains much we do not know. As noted, testing in many
facilities continues to be minimal or nonexistent. There is still no
uniformity in what data are being reported. Anecdotal evidence
suggests that not even all COVID deaths are reported as such: in some
jurisdictions, a death is not officially attributed to COVID unless the
decedent had received a positive COVID test before they died. Such
tests may be administered postmortem,55 a practice that would allow a
more complete accounting of the COVID death rate behind bars. But at
present, this step does not appear to be standard practice among
medical examiners. And although the biggest jail systems now publish
their data on dashboards of the sort found on DOC websites, most jails
around the country post no data at all. Indeed, it appears that many
jails do not even collect such data, leaving not only the public but jail
officials themselves in the dark as to the impact of the virus in their
It has become commonplace to observe that, as the novel
coronavirus moves through society, it exposes deep pathologies that
have long been hiding in plain sight. This effect is certainly true here.
We have already seen that a host of longstanding conditions, from
overcrowding to grossly inadequate medical care, has provided COVID19 an environment perfectly adapted for viral spread. But as in society
in general, the pathologies the virus exposes in the carceral
environment are not only the physically tangible. In addition, several
normative dynamics that have long shaped the American penal system
Personal Communication from Kalind Parish to Sharon Dolovich (Sept. 28,
2020) (on file with author).
54 Personal Communication from Kalind Parish to Sharon Dolovich (Oct. 19,
2020) (on file with author).
55 See Michelle Andrews, With Postmortem Testing, ‘Last Responders’ Shed
Light on Pandemic's Spread, NPR (May 19, 2020).

11/16/20 U. Chi. L. Rev. Online *29

have impaired the carceral response to COVID. These include the “us
versus them” attitude that often pits prison administrators against
those in custody; the callous indifference to the health and safety of the
incarcerated, which has led some officials to choose not to order the
mass testing that would provide an accurate picture of viral
proliferation in their facilities; and the longstanding ideological
commitment to being “tough on crime,” which, having been vociferously
endorsed by lawmakers for more than four decades, now poses a
political obstacle to releasing from custody anyone originally convicted
of a serious crime, no matter how medically compromised they may be,
how much time they have already served, or how great their risk of
But perhaps the most damaging carceral norm in this moment is
the official culture of secrecy that has long kept the public from having
a full and accurate picture of what goes on inside prisons and jails. The
power of prison officials to control the outflow of information from their
facilities has been a standard feature of American carceral practice
from the earliest days of the Northeastern penitentiaries and the
Southern plantation prisons.56 With state legislatures historically
performing virtually no oversight and the federal courts—the only
plausible venue for enforcement of prisoners’ constitutional rights—
committing to a “hands-off” posture with respect to prisoner suits,
prison officials long enjoyed virtually unimpeded authority over what
took place inside the walls.57 With this authority came the power to
exclude. Although there is no valid reason why carceral institutions
must be hidden from public scrutiny, strict limits on access became a
standard feature of carceral life.
This broad regime reigned virtually unchecked through much of the
twentieth century, until the 1960s and 1970s, when for a brief period
the federal courts displayed a new willingness to enforce constitutional
protections for people in custody. During this period, the Supreme
Court repeatedly affirmed the status of the incarcerated as rightsbearing subjects and, in case after case, expanded the scope of the

See Sharon Dolovich, The Regulation and Oversight of American Prisons,
ANNUAL R. CRIMINOL. (forthcoming 2021).
57 See, e.g., Malcolm M. Feeley & Van Swearingen, The Prison
Conditions Cases and the Bureaucratization of American Corrections:
Influences, Impacts and Implications, 24 PACE L. REV. 433, 438 (2004)
(describing the “feudal-like” character of the plantation prisons that, with the
demise of convict leasing, emerged across the American South in the early
twentieth century).

11/16/20 U. Chi. L. Rev. Online *30

rights they could claim.58 Yet at the same time, the Court still made
sure to enshrine in modern constitutional law the power of corrections
officials to restrict access to their facilities. In two cases from 1974, Pell
v. Procunier and Saxbe v. Washington Post, the Court held that
journalists have no First Amendment right of access to prisons and
jails beyond that enjoyed by members of the general public. And in
1978, in Houchins v. KQED, the Court established that corrections
administrators have total discretion to set the terms of public access to
their facilities. Thanks to this trio of cases, jail and prison
administrators may without fear of constitutional liability deny access
to any parts of their institutions they wish to keep hidden—even from
members of the media, through whose work citizens “receive that free
flow of information and ideas essential to intelligent self-government.”
Lost in this legal regime—and in the culture of secrecy it has
engendered—is the fact that prisons and jails are public institutions,
operated on behalf of society as a whole. Corrections officials are not
sovereign over the people in their custody. They are public servants
whose sole job is to administer carceral facilities in ways consistent
with the public interest. Yet instead of operating as if they are
accountable to the public for what happens to the people in their
custody, corrections officials, with the imprimatur of the courts,
generally treat information as to goings-on in the prisons and jails as
proprietary, theirs to withhold or share as they see fit.
This attitude, and the overall culture of secrecy it engenders,
impedes efforts to assess the conditions in which the incarcerated are
held and the treatment they receive. It also requires that advocates
expend their limited resources on efforts to secure discovery and on
public records requests, exercises that often seem to inspire further
evasive maneuvers by corrections officials, eating up even more of
advocates’ resources. As the parties tussle over information, the harms
suffered by people in custody, whether from affirmative abuse or
failures of care, remain unremedied.
If there is to be meaningful oversight over prisons and jails, there is
a strong need for transparency, and for free access to information
During this period, incarcerated plaintiffs prevailed in the Supreme Court
on a host of claims. See, e.g., Procunier v. Martinez, 416 U.S. 396 (1974) (First
Amendment expression); Wolff v. McDonnell, 418 U.S. 539 (1974) (Fourteenth
Amendment procedural due process); Johnson v. Avery, 393 U.S. 483 (1969)
(Fourteenth Amendment due process right of access to the courts); Bounds v.
Smith, 430 U.S. 817 (1977) (same). In 1976, the Court also made clear for the
first time that the Eighth Amendment prohibition on “cruel and unusual
punishment” applies to prison conditions. Estelle v. Gamble, 429 U.S. 97, 104

11/16/20 U. Chi. L. Rev. Online *31

concerning what happens inside and what those happenings mean for
the health and safety of the incarcerated. This has always been true,
but it is especially so during COVID. We have already seen the
insufficiency of the available data concerning the impact of the virus on
the incarcerated. Testing data is incomplete, and even those
institutions reporting deaths among residents generally include only
cases where COVID was the confirmed cause, although it is very likely
that many people are dying inside both from COVID and because of
it.59 As a matter of course, corrections administrators should be
conducting mass testing in their facilities. They should also, at a
minimum and as a matter of course, be publishing:

cumulative infection counts;
data on all deaths in custody during this period, including
official cause of death;60
the number of people in their custody who are at high risk of
complications from the virus;
the locations where those at high risk are being housed; and

In addition to people dying due to complications from COVID, it is certain
that many people will also die in custody during this period, not from the
virus, but from conditions that, were it not for constraints on prison medical
care, could have been treated and death prevented. Although COVID may not
have been the direct cause of death, these fatalities should still be classified
as “because of Covid.” There are also sure to be people who die from
untreated mental health issues during this period. There has been little
attention paid to suicides in carceral facilities since the start of the pandemic,
but my strong hunch is that, when the dust settles and a more complete
picture can be assembled, we will find that the suicide rate during this period
will have exceeded that of previous years, perhaps considerably. See, e.g.,
Maxine Bernstein, Federal Inmates in Sheridan Are Finding Coronavirus
Lockdown ‘Overwhelming,’ Public Defender Says, OREGONIAN (Apr. 15, 2020)
(“Oregon’s federal public defender said Wednesday she fears an inmate’s
suicide and other reports she’s received of inmates harming themselves stem
in part from the continued coronavirus lockdown at the federal prison in
60 Prison officials will often cite HIPAA restrictions, which require that state
actors preserve the confidentiality of medical information, as grounds for
withholding this information. But this is a red herring. Under HIPAA, next of
kin have the power to waive confidentiality. And when family members lose a
loved one who dies while incarcerated, they are typically willing and even
eager to waive HIPAA protections to the extent necessary to publicize the
circumstances of their loved one’s death. Prison officials committed to
transparency around the impact of COVID in prison would assist advocates
or journalists in securing family members’ consent to disclosure—or would
seek to do so themselves.

11/16/20 U. Chi. L. Rev. Online *32


the steps that are being taken to prevent the spread of
COVID throughout their facilities.

Nor should publishing this data be the extent of the transparency.
Innumerable media reports during this period feature first-person
testimonials from incarcerated individuals directly contradicting
official claims regarding conditions inside their facilities. It is not
enough that corrections officials describe what is happening on the
ground and what steps they are taking to keep people safe. In addition,
every facility should be open to ensuring that neutral observers with a
commitment to wide dissemination of information concerning the
functioning of public institutions are able to see with their own eyes,
and to broadcast, what is going on inside carceral institutions.
Reasonable time, place, and manner restrictions would of course be
warranted: facilities could not operate effectively if every reporter who
wanted to get inside could enter at will, especially when the need to
limit viral spread is paramount. At the same time, steps should be
taken to ensure that some neutral parties get an accurate picture of
what is happening throughout an institution, even—and perhaps
especially—when staff think they are not being observed. To restrict
outside access to the carceral equivalent of a Potemkin village would
defeat the purpose.
Senator Elizabeth Warren has introduced legislation mandating
standardized reporting on COVID-19 “in federal, state and local
correctional facilities.” This is an important step. The challenge is to
establish sufficiently compelling mechanisms to secure compliance. In
its early iteration, Senator Warren’s bill set as the penalty for
noncompliance a 10-percent reduction in Byrne Grants, which are
federal funds available to state and local jurisdictions to support law
enforcement and other criminal legal policies. Unfortunately, this
move seems unlikely to provide an adequate incentive for compliance.
Byrne grants vary in size according to jurisdiction, but they are
typically quite modest. According to the U.S. Department of Justice
Bureau of Justice Assistance, in 2015, the total amount distributed
under the Byrne program was $255.7 million, with 1,143 jurisdictions
eligible for receipt. The size of grants generally tracks the size of
counties. But even assuming a roughly equal distribution, Byrne
grants would average $223,710, which means a 10-percent penalty for
failing to comply with federal COVID reporting requirements would
cost a jurisdiction just $22,371. Especially given the size of corrections
budgets, corrections officials who perceive information relating to their
facilities as proprietary, and who might therefore resent federal efforts

11/16/20 U. Chi. L. Rev. Online *33

to force transparency, are unlikely to be moved to release requested
data by the threatened loss of such a small amount.61
The challenge of identifying a sufficiently effective compliance
mechanism makes vivid just how deeply entrenched is the culture of
secrecy in American prisons and jails. COVID forcefully reminds us of
the high health and safety costs of such a regime. In the short term, we
need effective legal mechanisms for ensuring broad public access to the
relevant data. More broadly, we need to normatively recast our
collective understanding of the professional obligations of corrections
officials. The authority they enjoy is not theirs by right. It has been
granted to them only in order that they may fulfill their delegated
responsibility, which is to run the prisons and jails in a way that
ensures the health and safety of the people we incarcerate while they
serve their time. When, as now, the fulfillment of those responsibilities
requires information as to what is happening inside to be broadly
disseminated, prison officials should not only stop trying to keep the
relevant data from getting out, but they should be actively publicizing
The carceral conditions that have made the pandemic dangerous
for people in custody have been in place for decades. Long before
anyone had heard of the novel coronavirus, American prisons and jails
were overcrowded, unhygienic, and full of aging people with chronic
illnesses poorly managed by medical and mental health staff, if they
were managed at all. A thick veil of secrecy has consistently made it
hard for advocates, lawmakers, journalists, or private citizens to get an
accurate sense of what goes on inside. And in this sphere, deep
regulatory failure has been normalized, as the various institutional
actors with the legal authority, and thus a duty, to ensure basic
protections for the incarcerated have routinely failed to fulfill their

It can be hard to know what measures would more effectively prompt
compliance, especially given the limited levers available to Congress to
require action by state and local officials. One obvious move would be to
threaten forfeiture of 100 percent of Byrne funds for noncompliance. Or
perhaps Congress might establish a tripling of attorneys’ fees for any case
where litigation is required to shake loose information that would have been
openly reported had the legislative requirements been followed, a possibility
proposed by Aaron Littman. In any case, it seems clear that a 10-percent
reduction in Byrne funds is unlikely to elicit the desired compliance.

11/16/20 U. Chi. L. Rev. Online *34

charge, instead displaying a callous indifference to the fate of the
human beings we have locked away.62
If COVID has taught us anything, it is that our collective failure
to recognize the embeddedness of carceral institutions in the broader
community and to ensure humane conditions for those individuals
living inside is not only cruel but also self-defeating. When people are
incarcerated, they do not disappear. Despite high walls, the facilities to
which they are removed are still very much within society itself. Every
day, residents cycle through Receiving and Release,63 family and
friends of the incarcerated come inside to visit their loved ones, and
staff and volunteers leave at the end of their shifts to return to their
families and communities.64 Through these and other channels, what
happens in prisons and jails cannot help but have a direct and often
immediate impact on society as a whole. My own view is that, even
were this not the case, we would still have a strong moral and
constitutional obligation to ensure the safety and well-being of the
people we choose to incarcerate. But one need not share this view to
recognize the imperative of humane treatment. COVID makes clear
what advocates for the incarcerated have long emphasized: mass
incarceration under inhumane conditions harms everyone. Until we
collectively recognize that carceral policy is a central determinant of
public health, the American carceral system—and thus American
society as a whole—will be as unprepared to address the next
pandemic as it was to effectively manage this one.
Sharon Dolovich, The Regulation and Oversight of American Prisons,
ANNUAL R. CRIMINOL. (forthcoming 2021).
63 See Danielle Kaeble, Time Served in State Prison, 2016, Bureau of Justice
Statistics, at 1 (2018) (reporting that the average stay in state prison among
those released in 2016 was 2.6 years). Every year, more than 10 million
people churn through local jails. ZHEN ZENG, BUREAU OF JUSTICE STATISTICS,
JAIL INMATES IN 2018 1 (2020) (“In 2018, jails reported 10.7 million
admissions”). On the prison side, close to 600,000 people are admitted to
prison annually and around 600,000 are released. See E. Ann Carson,
Prisoners in 2019, Bureau of Justice Statistics, 13 at Table 8 (2020) (noting a
U.S. total of 596,407 prison admissions in 2018 and 576,956 admissions in
2019; a total of 614,860 prisoners were released in 2018 and 608,026 released
in 2019).
64 The most recent data indicate that 658,055 people work in American
prisons and jails. See James J. Stephan, Census Of State And Federal
Correctional Facilities, 2005, Bureau of Justice Statistics at Appendix Tables
12, 13 (2008) (as of 2005, federal and state prisons employed 445,055 people);
see also Zhen Zeng, Jail Inmates in 2016, Bureau of Justice Statistics, at 6,
Table 8 (2018) (as of 2015, there were 213,000 total employees in city and
county jails).