HELP US END MASS INCARCERATIONThe Prison Policy Initiative uses research, advocacy, and organizing to dismantle mass incarceration. We’ve been in this movement for 23 years, thanks to individual donors like you.
We outline five things to keep in mind about crime data trends during the pandemic, including a few tips for where to look for information about your local area.
Crime rates have fallen in recent weeks, with most of the country under “stay at home” orders. As TIME and Bloomberg News have recently reported, cities across the country have seen changes in offense patterns as well as in the number of total crimes reported, with most places showing a significant decrease in overall crime. But crime data analysis isn’t cut-and-dry, so here we outline five things to keep in mind about crime data, including a few tips for where to look for information about your local area.
Local crime trends may vary a lot and appear dramatic: here’s why ⤵
Where to find local crime data, and what questions to ask about it ⤵
1. The types of crimes, and targets of crimes, are changing during the pandemic
Among the crime trends observed during this pandemic are changes in the types of crimes and sites of crimes. Notably, violent crime has dropped, and burglaries have shifted away from home break-ins to target closed businesses instead. These changes make sense in a world where more people are staying home during work hours and on weekends, and most brick-and-mortar businesses are either closed or operating fewer hours and with a fraction of their usual staff. For violent crimes like assault or robbery to occur, people have to come into close contact. And why would someone try to steal from an occupied house when there’s an empty shop downtown? Similarly, car thefts have gone up in a few cities, which is unsurprising when many drivers have little reason to move the car from where it’s been parked for the past month. Experts have weighed in to explain some of the recent crime trends for TIME and Bloomberg, offering similar common-sense explanations; the idea that crime changes as people’s “routine activities” change is also the subject of decades of criminological study.
There are more disturbing changes, likely fueled by the emotional strain of social isolation and collective grief, not to mention the economic strain of a sudden loss of income for millions of people. Some cities have reported an increase in calls related to domestic violence, for example, and with schools closed, children are more vulnerable to abuse and neglect at home. With commerce moved largely online, we may see more internet-based “white collar” crime, such as the new crop of scams reported by the FBI.
For millions of people, the strain caused by this virus will not end when shops open up and people can get together again, and evidence suggests that some crimes may increase because of the looming economic crisis. It will continue to be important to monitor crime trends to see how this pandemic affects Americans, particularly those on the margins, and to recognize that most crimes signal unmet needs that require help, not punishment.
2. Changes in policing will mean fewer arrests and lower crime rates
It’s important to remember that official crime data comes from law enforcement agencies, so it is a record of crimes reported to, or by, police. Therefore, as police practices change, so do crime statistics. This pandemic has impacted police departments in a number of ways: police officers have gotten sick, and officials have directed police to avoid unnecessary contact with the public and even to respond to some offenses differently, such as issuing citations in lieu of arrest. Most of these changes will result in fewer reported crimes, as there are fewer police officers available to patrol the streets, fewer face-to-face traffic and street stops, and fewer arrests overall. So some of the “decrease” in crime is due to changes in policing rather than changes in criminal behavior.
That means that the inverse is also true: when policing returns to “normal,” it’s very likely that reported crimes will go back to previous (historically low) levels. This should not cause alarm, although headlines will undoubtedly shout about a “spike” in crime when that happens. That “spike,” too, will be largely a reflection of police ramping enforcement back up to pre-pandemic levels. We should always be skeptical of short-term changes in crime data.
3. People may be less likely to report crimes, which will also lower crime rates
Not all crimes are reported to police by civilians, even under normal circumstances. The Bureau of Justice Statistics estimates that in 2018, less than half (43%) of violent crime victimizations were reported to police, and an even smaller share (34%) of property crime victimizations were reported. Now that people are avoiding in-person contact for fear of contracting or spreading the coronavirus – even avoiding hospitals when they need them – it’s very likely that even fewer crimes are being reported to police. Victims of rape and sexual assault (crimes that are always underreported) may be even less likely to go to the police and undergo forensic medical exams. And given the heightened risk of infection in prisons and jails, victims of intimate partner violence or other domestic abuse may be even more reluctant to seek police intervention, which is likely to result in an arrest of at least one person, if not both.
As with changes in police practices, when the number of infections falls and people become more comfortable with face-to-face contact, we can probably expect the number of reported crimes to go up again. When that happens, we should be cautious of interpreting that change as an “increase in crime,” when it may largely reflect a return to pre-pandemic crime reporting rates.
4. Local crime trends may vary a lot and appear dramatic: here’s why
Headlines that start with sweeping statements like “Crime Rates Plummet Around the World” obscure important differences in crime data across cities, states, and regions. Just as the impact of the coronavirus looks very different when we look at national versus local infection rates, crime trends become more pronounced, and less consistent, when we look at local data. Part of this is because local conditions affect local crime patterns: in neighborhoods where people go to drink, for example, we might expect to see more assaults at night. And in places under stricter “stay at home” orders, we are likely to see more dramatic changes in crime.
Another reason that crime trends appear more exaggerated at the local level is that when the number of crimes is small enough, even relatively small changes are noticeable in the data. For example, as Bloomberg reported, the number of homicides in Austin was up 25% compared to the same time last year – but this was because of one additional homicide, taking the total count from 4 to 5. So it’s important to look critically at trends described in “percent change” terms, to consider how much actual change it reflects, and to report about the data responsibly.
5. Where to find local crime data, and what questions to ask about it
Almost all law enforcement agencies report crime data to the FBI’s Uniform Crime Reporting (UCR) program. National data summaries have historically been published on an annual basis in the Crime in the United States reports, which also allow readers to access local agencies’ summary data. Now, the UCR program is shifting to quarterly data updates through the FBI’s online Crime Data Explorer tool, and will be phasing out the annual reports this year. The online tool provides more detailed local information (see Boston Police Department (link no longer available), for example), and allows you to look at trends over time. Currently, the UCR data is updated through 2018, but according to the FBI, data from January-March 2020 (the start of the pandemic in the U.S.) will be added in June 2020 to the Crime Data Explorer; it will be updated quarterly thereafter. This will be a key source of timely data that will offer more perspective on crime trends across the country.
In many places, local law enforcement agencies also maintain and publish their own data, which may be more useful and up-to-date for local researchers and journalists. Major cities like New York and Philadelphia, for example, frequently publish timely statistics. New York publishes updated crime and jail data for every county in the state, in fact. It’s worth checking with your local law enforcement agency(s) to see whether they publish crime data regularly, and with any state statistical or research agencies, which may provide insights that the FBI data does not.
When reviewing any local crime data, consider the following questions:
What exactly is being measured? Is it arrests, crimes reported to police, all incidents known to police, calls for service, or something else? Are traffic violations or other low-level offenses included? Are they relevant to the question you’re trying to answer?
What does the baseline – or historical data – look like? What were crime rates (or counts) at the same time last year, and in the years before? Arrests are more common at certain times of the year (i.e. summer), so comparing similar time frames will help narrow down factors that might affect changes in the data. And remember that “percent change” can look very dramatic when you’re starting with a small baseline amount of crime (such as homicides).
If crime has increased, what kinds of crimes have gone up? Noticing that crime rates overall have gone up is not necessarily reason to sound the alarms; this may reflect an increase in non-violent property offenses among people desperate for money in an economic crisis, or problems with substance abuse among people who are experiencing extreme distress. On the other hand, some upticks in reported crime may deserve more attention right now, such as an increase in domestic violence or online scams.
How have local routines changed, and could that affect crime patterns? Are most people staying home? Are businesses downtown closed? Are people leaving cars parked on the street for days on end, when they would normally be moving them frequently? These questions may get at some of the reasons behind any changes in crime patterns.
How have local police practices changed, and could those changes affect the data? Have there been changes in staffing, deployment, priorities, or directives to respond differently to low-level crimes?
The final question we should ask when looking at crime data now is: What can we learn about the criminal justice system from recent changes in crime, or responses to crime? One positive effect of this pandemic experience may be that we see how our criminal justice system might operate differently – now and in the future. The recent crime drop, for example, indicates that counties and states can safely release large numbers of people from prisons and jails without compromising public safety. That fact begs the question: Why were so many locked up in the first place? These are the urgently needed conversations we need to have, and looking critically at crime data in the context of social and policy changes can help us get there.
As cities attempt to reduce their jail populations, they should pay attention to the lesson of NYC’s slow decarceration: Even releasing "low-level offenders" is a complicated process liable to be bogged down by delays.
On March 21st, the Board of Correction of New York City recommended that the city decrease its in-custody population by at least 40%, or over 2,500 people.1 But a full month later, NYC has brought the jail population down from pre-pandemic levels by only 27%. And while NYC may be a model for cities that have done much less, like New Orleans, Miami, and St. Louis2, it is also a cautionary tale for doing so little. The infection rate on Rikers Island remains 5 times higher than that of New York City and over 7 times higher than that of New York State.
The city could bring down the infection rates in jails by releasing more people: A smaller jail population allows for physical space for social distancing and quarantining, reduces the burden on correctional healthcare staff, allows for accommodations for staff sick leave, and slows viral transmission both within the jail and to the community at large.
While all incarcerated people should be considered for release, it’s particularly shocking that NYC has been slow to release people held on “technical violations” of parole — that is, for infractions that are not even crimes. As of April 22nd, there were still 293 people held in NYC jails for technical parole violations:
The number of people held for technical violations has declined, but not as fast as it should have. The number of people held for technical violations on March 21st is from the Board of Correction’s letter to NYC DOC. The statistics for April 1st through April 22nd are published by the NYC DOC in daily reports to the Board of Correction. Data on the number of people held for technical violations of parole between March 21st and April 1st — when the DOC began their daily reports — are not available in a compatible format.
Date
Total held on a technical parole violation (with no open case)
Apr 22
293
Apr 21
306
Apr 20
305
Apr 19
303
Apr 18
303
Apr 17
328
Apr 16
339
Apr 15
336
Apr 14
357
Apr 13
391
Apr 12
(no data, weekend)
Apr 11
(no data, weekend)
Apr 10
427
Apr 9
454
Apr 8
486
Apr 7
494
Apr 6
493
Apr 5
(no data, weekend)
Apr 4
(no data, weekend)
Apr 3
490
Apr 2
544
Apr 1
547
Mar 22 – Mar 31
(data not available)
Mar 21
666
There are a few reasons to be surprised by the number of people still jailed for technical violations in New York City. First is that it seems to run contrary to Governor Andrew Cuomo’s March 27 order that 1,100 people incarcerated for technical violations in New York state — including 600 in NYC jails alone — be considered for release. Second is that the city is facing significant pressure to release people held on technical violations, not only from community advocates, but from authorities within the parole system. (In early March, a coalition of current and former community supervision executives called for the nationwide suspension of arrests for technical probation and parole violations.)
It’s not clear who is responsible for holding up the release of people held on technical violations — city officials, state parole officials, judges, or some combination of the three. But what is clear is that this failure to act quickly will make it much harder to reduce the dangerous conditions in New York City jails, leading to more coronavirus infections and deaths. These deaths will disproportionately be people of color, who are overrepresented among technical parole violation detainees in the city’s jails.
But as other state and local governments attempt to reduce the number of people held for these minor violations, they should pay attention to the lesson of NYC’s slow decarceration: Even releasing “low-level offenders” is a complicated process liable to be bogged down by delays. State parole officials, city correctional leaders, judges, prosecutors, and other decisionmakers all must cooperate in order for hundreds or thousands of people to be released in a short time frame. During a pandemic, any of these officials holding up the release process will mean more deaths.
This graphic and data is from the Council of State Governments Justice Center’s invaluable report Confined and Costly: How supervision Violations Are Filling Prisons and Burdening Budgets. (In a technical note, the CSG explains that data is from 2017 exception for Virginia which is from 2016, and that technical breakdown was not available for CT, GA, MA, MD, MI, MN, NC, NH, NM, OK, and PA. Data on violations as a proportion of prison admissions was not available for DE and VT.) The report notes, helpfully, that “variation in these proportions across states is shaped by the overall size of each state’s supervision population, how violations are sanctioned, whether those sanctions are the result of incarceration paid for by the state or county, and how well state policy and funding enable probation and parole agencies to employ evidence-based practices to improve success on supervision.”
Footnotes
The NYC Board of Correction recommended over 2,500 people be reviewed for release in a March 21st letter to the city’s district attorneys and chief judge, the NYC DOC, and the NYS DOCCS. According to the Vera Institute for Justice, the NYC jail system held a total of 5,447 people on February 29th, 2020.
The table below shows the population reductions of the largest jails in the U.S. (those with pre-pandemic populations of over 500 people). This table was compiled using the Vera Institute for Justice’s jail population tool. Of note, we only compared large jails with pre-pandemic population data from January and February 2020.
Problems with data collection - and an unfortunate tendency to group Native Americans together with other ethnic and racial groups in data publications - have made it hard to understand the effect of mass incarceration on Native people.
The scarcity of data on Native Americans in the U.S. criminal justice system comes up a lot in our conversations with activists and reporters, who rightly wonder why Native populations are often excluded from comparisons with other racial and ethnic groups. While Census data reveals that Native populations are overrepresented in the criminal justice system, other information that could shed more light on the issue is sparse. So, we compiled the information that does exist — which is fractured and hard to locate — in one place below.
Preface: What the Census data says
We’ve previously used data from the 2010 Census to analyze incarcerated populations by race/ethnicity and sex for each state. In our analysis, data on prisons and jails were combined. We found that, in 2010, there were a total of 37,854 American Indian/Alaskan Natives in adult correctional facilities, including 32,524 men and 5,132 women (and 198 who were 17 or younger). That is equivalent to a total incarceration rate of 1,291 per 100,000 people, more than double that of white Americans (510 per 100,000). In states with large Native populations, such as North Dakota, American Indian/Alaskan Native incarceration rates can be up to 7 times that of whites. Once the 2020 Census data is released, we will update our analysis, since it is 10 years old now.
Other data on Native Americans in the criminal justice system
Prisons: In 2016, 19,790 Native men and 2,954 Native women (22,744 total) were incarcerated in U.S. state and federal prisons, according to the Bureau of Justice Statistics’ (BJS) National Prisoner Statistics (NPS) series. The NPS series reports the population of state and federal prisons – but not local jails – by race/ethnicity and sex, but the most recent data available with that level of detail is from 2016. However, other sources supplement these findings:
BJS reports an increase to 23,701, in Prisoners in 2017. Oklahoma tops the list as the state with the highest number of American Indian/Alaskan Natives incarcerated, followed by Arizona, Alaska, and California. However, this data is not broken down further by sex and race.
Limited state-level data is also available from some state Departments of Corrections, like Alaska’s, which identifies Alaskan Native populations in its annual Offender Profile. However, many other states, even those with large Native populations like California and Texas, group these populations into an “other” category when reporting demographics. (More on that in our discussion of data limitations below.)
Jails: The BJS annual report on jail inmates estimates 9,700 American Indian/Alaskan Native people – or 401 per 100,000 population – were held in local jails across the country as of late June, 2018. That’s almost twice the jail incarceration rates of both white and Hispanic people (187 and 185 per 100,000, respectively). Frustratingly, this data is also not reported by sex.
The 2016 BJS Jails in Indian Country report identifies 80 facilities operating on tribal lands, holding 2,540 people – 1,750 men and 620 women – in mid-2016. The number of inmates admitted to Indian country jails was 9,640 during the month of June 2016, giving us an idea of “jail churn” in facilities on tribal lands. Additionally, this report is one of the very few sources for this population’s offense data, although even here, about 35% of offenses are unhelpfully categorized as “other.”
Youth: People under the age of 21 make up 42% of American Indian/Alaskan Native populations in the United States, so Native youth confinement is a special concern. With a detention rate of 255 per 100,000 in 2015, Native youth are approximately three times more likely to be confined than white youth (83 per 100,000). In Indian country jails, approximately 6% of the confined population was 17 or younger in 2016; unfortunately, the number of youth held in other adult prisons and jails is not broken down by race/ethnicity. The Census of Juveniles in Residential Placement reports data on Native youth in juvenile justice facilities across the U.S., most recently for 2017, including details about offense type, facility type, sex, age, and more.
Contributing to these confinement rates is disproportionate police contact: Native youth are arrested at a much higher rate than white youth. The 2018 arrest rate for Native youth was 2,251 per 100,000 while white youth were arrested at a rate of 1,793 per 100,000.
Data collection from Native populations suffers from a number of limitations
Data collection efforts in tribal communities face a number of problems that limit the data’s accuracy and comprehensiveness.
According to the National Institute of Justice, issues such as difficulty in outreach, overlapping jurisdictions, and differences between tribal justice systems make the collection of data from these communities especially challenging. U.S. government policies and priorities also limit the data it collects and reports about Native populations:
The DOJ has moved slowly: A Department of Justice (DOJ) oversight report in compliance with the 2010 Tribal Law and Order Act (TLOA) states that the “TLOA requires the Department’s BJS to collect data related to crimes in Indian country. However, 7 years after TLOA became law, its data collection and reporting efforts are still in development.”
Reporting is voluntary: According to the same report,“…because participation in the FBI’s Uniform Crime Reporting (UCR) Program is voluntary, not all tribes report crime statistics into the UCR database. As a result, Indian country crime statistics are so outdated and incomplete as to be virtually useless.” The BJS derives most of its crime data from the UCR program, which is especially incomplete when it comes to tribal jurisdictions’ data. The DOJ report found that while “207 tribes reported to the UCR in 2014, only 115 tribes submitted complete information that was included in the final UCR report.” It’s worth mentioning that there are, as of 2017, 226 tribal law enforcement agencies recognized by the federal government. Assuming the same number existed in 2014, that means 19 (8%) did not report crime data at all.
Data collection does not distinguish between tribes: According to the DOJ report, the National Crime Victimization Survey “does not allow the calculation of separate crime statistics for each American Indian tribe.” A report from the United States Sentencing Commission’s Tribal Issues Advisory Group also cites a lack of accurate databases in tribal courts, consistent and comparable disaggregation, and data sharing between federal and tribal entities.
Data aren’t used to help Native communities: The U.S. Sentencing Commission’s Report notes that the limited data that is collected has not been used to “evaluate and improve” law enforcement activities in Indian country. This adds to the strain caused by the general lack of cooperation between U.S. and tribal justice systems: According to a report by the National Tribal Judicial Center, federal and state correctional facilities “do not notify tribes of inmate release to parole or probation.” The report notes that tribal “protection orders are not validated by or enforced by state courts or state law enforcement. No outside agencies honor tribal court subpoenas.” This lack of reciprocity worsens the already countless issues with data collection and sharing.
Cultural and socioeconomic barriers lead to undercounting: More broadly, a “distrust of the U.S. government, a youth-heavy population, nontraditional addresses, low internet access, language and literacy barriers, weather and road access issues, and high rates of poverty and houselessness” create a deeply problematic undercounting of American Indian/Alaskan Native people. (A report by Rewire.News examines the consequences of this undercounting, including lower representation in Congress, funding deficits in health and human services, and a decline in tribal recognition and enrollment.)
“Other” data obscurities
Criminal justice data often uses racial and ethnic categories to break down the disproportionately high representation of Black and Hispanic populations in prisons and jails. Beyond these categories, however, lies the illusive “other” designation, which lumps together Asian Americans, Pacific Islanders, Native Hawaiians, and of course, American Indians and Alaskan Natives. However, as the Census data reveals, disproportionate incarceration rates for these groups are not negligible. This practice obscures differences between these groups and makes it difficult to determine how the justice system plays a role in Native communities. Specifically:
The Bureau of Justice Statistics categorizes American Indian/Alaskan Natives as “other” in their Felony Sentences in State Courts data series. According to research by the Native American Voting Rights Coalition, several Native women surveyed mentioned that their husbands/partners were ineligible to vote due to felony convictions, contributing to a variety of barriers that hinder Native American political participation. The lack of disaggregated data makes it difficult to determine the exact proportion of Natives who are disenfranchised.
According to the American Indian and Alaskan Natives in Local Jails report, there were 56,400 individuals in jails – in addition to those categorized as “single race” American Indian/Alaskan Native – who identified as American Indian/Alaskan Native and another race(s) or ethnicity, suggesting higher rates of incarceration nationwide if multi-racial individuals were included in Native population counts or rates.
Rewire.News’s report also highlights how gender categorization of Native populations can often obscure those who identify as Two Spirit, non-binary, or transgender.
As it stands, there are many more questions than answers about Native Americans in the criminal justice system. Until criminal justice agencies overcome the limitations on data collection — and until the offices that publish the data are willing to list Native Americans as a distinct demographic group, rather than a member of an “Other” category — informational gaps will continue to make it difficult to understand how overcriminalization has impacted Native populations.
Many local jails and pretrial systems are taking action to reduce their populations in advance of the COVID-19 pandemic, but state prison systems are not, raising the question: Are state prisons prepared to handle a pandemic within their walls? We set out to survey prison systems on the capacity of their health facilities, their plans for any necessary external hospitalizations, their levels of equipment, their staffing levels and their general priorities.
Unfortunately, our April 3-10 survey shows that state prisons are still largely unprepared for a global pandemic that can reasonably be expected to hit their entire state prison system — and their supporting state government — all at the same time.
Most prisons are still aiming to keep the virus out of their facilities, rather than focusing on how to minimize the harm to incarcerated people, to their staff and to society as a whole. Containment might be a reasonable goal when it comes to outbreaks of flu, tuberculosis, or MRSA – diseases that prison systems know how to guard against by vaccinating people, screening, and so on. But COVID-19 is different both in terms of how it spreads and by the fact that it is already stressing the public hospital system that state prisons historically rely on for back-up support.
Given the number of large number of staff required to run a facility1 and the apparent ease with which asymptomatic people can infect others, no combination of security restrictions — such as suspending family visitation, checking the temperature of incoming staff, or confining the entire population to their cells — can keep out the virus that causes COVID-19 for long.2 And once the virus enters a facility, the density and lack of sanitation will allow it to spread quickly to all incarcerated people and staff, and will accelerate the spread to the surrounding community.
Ideally, state prisons’ first response to the pandemic should have been to do like many jails and reduce the number of people incarcerated. But at the very least, we expected to see them developing plans that acknowledged the inevitability of a COVID-19 outbreak and its unique challenges. Their plans should anticipate the need to isolate vulnerable people, work around staffing shortages, and navigate shortages of medical supplies and hospital beds. But except for a few notable exceptions — particularly North Dakota — most states have not even gotten that far.
The good news is that in some states, the spread of the virus is several weeks behind other states, so some of the states that are the least ready still have the potential to learn from other states and improve their planning.
The state of pandemic planning in state prisons
In our survey of state Departments of Corrections, we sought to gather more information about how states are preparing for the pandemic to breach the prison gates. We asked about five major topics of pandemic preparedness:
the capacity for isolating particularly vulnerable individuals and quarantining people with suspected cases of COVID-19 within facilities,
protocols for people requiring hospitalization,
equipment (including ventilators, medical-grade PPE, COVID-19 tests) accessible to facilities,
anticipated staffing changes and availability of healthcare staff within facilities, and
what the most immediate priorities are in planning for a COVID-19 outbreak in correctional facilities.
We hoped to hear from each state that steps were being taken to prepare facilities to navigate the inevitable: a positive COVID-19 test among the hundreds of people living in close proximity to one another in prison. We expected each facility to have designated cells, units, or wings that could be easily isolated from the rest of the facility to allow either isolation of people who are particularly vulnerable to complications from COVID-19, or quarantine of people who test positive for COVID-19.
Given the recommendation from expert Dr. Homer Venters that facilities establish a plan for hospitalizations that recognizes that staff will be in short supply, and thus does not require the usual 2:1 ratio of correctional officers to patients, we expected correctional departments to have established new pathways and protocols for hospitalization.
We also expected that because prisons have finite resources for respiratory support (such as oxygen), facilities would have realistic plans to transfer people to hospitals while protecting staff from exposure. We hoped prisons could tell us how they would secure sufficient equipment and supplies, and how they will respond when supplies run out or when hospitals refuse to admit. (For example, there is already a shortage of COVID-19 tests, restrictions on who can be tested, and personal protective equipment (PPE) for medical professionals is already in short supply.) This planning, unfortunately, was largely absent.
Given the rate of infection and the length of hospitalization for severe cases of COVID-19, it is reasonable to assume facilities will encounter staffing shortages and that plans need to be in place for making sure the essential services behind bars continue in the face of staff calling out sick (i.e. food, medical care, telephone access, etc.). Given that most correctional staff are already stretched thin, departments should have begun planning for a staffing shortage weeks ago by reducing the number of people incarcerated and reducing the burden on staff. Most departments did not, but we still expected that they were preparing for the medical challenge created by that inaction.
The responses we received were largely disappointing. Rather than developing plans to mitigate the harm of an inevitable outbreak, most states are still focusing on restricting the movements of incarcerated people within facilities — in other words, attempting to “contain” the virus, which is all but impossible with COVID-19.
Even worse, some state departments of corrections informed us that no changes to their existing medical capacity were needed, suggesting that their established medical facilities and staff were sufficient to combat a disease that is overwhelming entire city infrastructures across the nation. Other states notified us that because no positive cases had occurred in their prisons, many changes have not yet been implemented. A number of states referred us back to their websites, which although providing up-to-date information on some aspects of their COVID-19 response, did not answer the specific questions about policy and planning changes in the past few weeks.
The good news is that many states still have time to do a far better job. At this point, on April 10, state Departments of Corrections need to be focused on mitigating the disastrous consequences of the COVID-19 pandemic entering prisons, rather than sticking to the belief that their high prison walls can effectively keep a global virus at bay.
Footnotes
Every facility is different, but outside observers can roughly estimate the number of staff that go in and out of a facility by dividing the incarcerated population by 4.6 for jails and 4.7 for prisons. (The typical jail has one staff member for every 3.3 incarcerated people, and the typical prison has one staff member for every 3.4 incarcerated people. If you assume that the staff work 5 days out of 7, you can divide the incarcerated population by 4.6 or 4.7 to get an estimate of how many staff enter and leave the facility every 24 hours.)
We are separately tracking press releases and updates from state departments of corrections on visitation suspension, changes in communication costs, screening policies for staff and new admissions, and facility lockdowns. ↩
To protect the American public from COVID-19, schools have closed, non-essential stores have been shuttered, people with desk jobs have started working from home, and public gatherings have been prohibited. But the prison system continues to hum along as though nothing has changed: Prisons have done virtually nothing to reduce the population density that puts both incarcerated people and staff at grave risk.
To justify their lack of action, criminal justice officials and elected leaders imply that saving the lives of people behind bars is not worth the inevitable public safety cost of releasing them. This talking point is as old as time. It’s also out of step with history.
Large-scale releases have been common throughout U.S. and international history for a variety of legal, political and health reasons. Below is a partial and non-exhaustive summary of some notable examples in U.S. and international history. (These examples were originally collected for a different project with Leah Sakala in 2014.)
If the places where these releases took place became hotbeds of crime, we’d know about it already. But they didn’t. In fact, in many cases, the inverse happened — and the academic literature about these experiences prove it.
SELECTED HISTORICAL DECARCERATION EXAMPLES
U.S. examples
California (adults, 1968 – 1972) Between 1968 and 1972, while Ronald Reagan was the tough-on-crime Governor of California, the state’s incarceration rate dropped from 146 to 96 per 100,000. The historical record suggests that the decrease was largely due to a state program to incentivize local probation departments to decrease commitments to state facilities, as well as an increased use of parole.
California (youth, 1996 – 2009) Although California is currently struggling politically with reducing its adult population, that state is a national leader on reducing its incarceration of kids. Previously, the Youth Authority was a “catch-all” for even the lowest-level offenders. Among other reforms, the state has created financial disincentives for counties to send kids to the state system while rewarding them if they kept the kids in local programs.
California (currently) Beginning in 2006 and accelerating in 2009, the California prison population has been dropping. Spurred in part by the Supreme Court’s order in Plata, major changes are underway (although far less than most of us hoped and far less than most of our opponents feared.) Some of the drop in the prison population is the illusory result of “Realignment,” a legislative change that sends people who would previously have gone to state prison to local jails. The California prison population drop is still notable because the state’s prison population is dropping faster than the jail population is increasing, but the actual decline in the number of people incarcerated in California is not as large or as quick as the Supreme Court ordered.
Florida (1963 – 1965) On the heels of the Supreme Court’s Gideon v. Wainwright decision, Florida had to give thousands of incarcerated people new trials, this time with court-appointed lawyers. For some people, the evidence was too flimsy or dated to withstand a proper legal defense, so over 1,000 people were released in a very short time period.
Illinois (1980 – 1983) Concerned that the 1978 legislative switch to “determinate” sentencing would lead to prison overcrowding, the Department of Corrections instituted a special program of the parole board awarding extra good time credits. In sum, over 21,000 people, or 60% of all prison releases, were released an average of 105 days early.
Massachusetts (youth, 1969) Massachusetts, under Republican Governor Frank Sargent and newly-appointed Department of Youth Services Commissioner Jerome Miller, closed its training schools for kids and decarcerated nearly 900 children. The state paroled some children directly home while a new system of community-based alternative programs were developed.
New York & New Jersey (~1999 – present) A mix of reforms — including policing, sentencing reform and parole — have allowed these two states to radically reduce both the number of people entering prison and how long they are incarcerated. Governors of both parties implemented these reforms at a time when the prison population was still rising nationally. In fact, much of the national prison drop in recent years is the result of these two states plus California.
Washington State (1979-1984) Over 1,600 people were released early in six different periods over the course of five years.
International examples
Czech Republic (2013) Outgoing President Václav Klaus gave a mass amnesty/pardon to over 6,000 people, approximately one third of the incarcerated population, as a way to both respond to an overcrowding crisis and to mark the anniversary of Czech Independence. “The president pardoned all convicts with prison terms under one year. The amnesty… also includes people sentenced for non-violent crimes to up to two years in jail, and seniors aged at least 70 whose prison terms do not exceed three years and those aged at least 75 with terms of up to 10 years.”
Finland (1976 – present) Finland used to have one of the highest incarceration rates in Europe. Finland made a long series of policy changes — including decreasing sentence lengths — to radically lower their use of the prison, and that country now has one of the lowest incarceration rates in the entire world.
Source: Nils Christie, Crime Control as Industry.
Israel (1967) The Israeli Knesset passed an Amnesty Law that released 501 incarcerated people and closed 15,376 criminal investigations.
Italy (2006 and 1990) In 2006, to respond to prison overcrowding, the Italian government released 22,000 people, generally those serving three years or less, except for those convicted of Mafia-related crimes, terrorism, sexual violence or usury. An earlier mass pardon in 1990 released 8,451 people out of the total incarcerated population of 26,000.
Russia (numerous, late 1990s through present) Russia has repeatedly issued large-scale amnesties, used both to manage the populations and to celebrate key events like the 20th anniversary of the constitution. Some amnesties also applied to people with pending charges. One notable and major large-scale amnesty was in 1999, when incarcerated people were released to help control a tuberculosis epidemic that was incubating in the prisons and then spreading to the rest of the country.
To accompany our work on what the criminal justice system is doing — and should be doing — to respond to the COVID-19 pandemic, the Prison Policy Initiative today released a spreadsheet showing what each state Department of Corrections has told the public about its virus response plan. Prepared for internal use, we’re sharing it in the hopes it will save other advocates, journalists and policymakers time looking up the same information.
The spreadsheet includes:
Links to each state’s COVID-19 page or its archive of press releases
Links to the infection and fatality trackers for each state (about half of all states have this)
Notes on whether each state is addressing 15 separate topics, including the suspension of visits, changes in telephone policies, increased access to hygiene materials, employee screening, staffing changes, isolation plans, etc.
This spreadsheet is a useful view into what the state prison systems see as important to communicate to the public, although it is not necessarily the definitive statement on what state prisons are doing. (For example, some states may think that suspending unaffordable medical copays during a global pandemic was too obvious to announce, and other states may be planning to accelerate parole releases but are choosing to be quiet about it. By the same token, states that are stubbornly refusing to change their copay policies,1 for example, are predictably not going to trumpet that fact on their websites.)
Other advocates looking for information on what state prison systems are doing should look at:
Professor Sharon Dolovich at the UCLA School of Law has shared a growing comprehensive spreadsheet including results from a state-by-state survey of changes in visitor policies, requests for population reduction, and actions taken to reduce the incarcerated population.
We will continue to update the spreadsheet as long as we find it useful internally, and the spreadsheet will always have the last modified date at the top.
Footnotes
We are looking at you, Delaware, Hawaii, and Nevada. ↩
The report includes an interactive map showing where people convicted of violence have been "carved out" of recent criminal justice reform laws.
April 7, 2020
As the threat of a COVID-19 disaster in U.S. prisons looms, people serving time for violent crimes may be most at risk, as states like California and Georgia exclude them from opportunities for rapid release. “Violent offenders” — even those who are old and frail — are being categorically denied protection in a pandemic.
Letting people convicted of violence apply for life-saving opportunities requires political courage, just as it has for decades. But denying relief to people based exclusively on their crime of conviction is as ineffective as it is unjust. In a new report,Reforms Without Results, we review the existing research on violent crime, explaining six major reasons why states should include people convicted of violence in criminal justice reforms:
Long sentences do not deter violent crime.
Most victims of violence, when asked, say they prefer holding people accountable through means other than prison, such as rehabilitative programs.
People convicted of violent offenses have among the lowest rates of recidivism — belying the notion that they are “inherently” violent and a threat to public safety.
People who commit violent crimes are often themselves victims of violence, and carry trauma that a prison sentence does nothing to address.
People age out of violence, so decades-long sentences are not necessary for public safety.
The health of a person’s community dramatically impacts their likelihood of eventually committing a violent crime — and community well-being can be improved through social investments rather than incarceration.
Demonstrating how common it is for people convicted of violence to be left behind, our report includes an interactive U.S. map showing 75 examples of state criminal justice reform laws that have excluded them. The map reveals that:
At least 16 states have passed laws excluding people convicted of violent crimes from veterans’ courts, mental health courts, diversion programs, and other alternatives to incarceration.
In at least 10 states, people convicted of violent crimes have been “carved out” of laws designed to ease the reentry process.
At least 20 states have passed laws that expand parole, good time, and other mechanisms for early release — but offer no relief to people convicted of violent offenses.
Unless states are willing to change how they respond to violence, reducing U.S. incarceration rates to pre-1970s levels will be impossible: Over 40% of people in prison and jail are there because of a violent offense. Lawmakers serious about ending mass incarceration — or limiting the toll COVID-19 takes behind bars — can no longer afford to ignore people serving time for violent crimes. In Reforms Without Results, we provide the data and arguments they will need to craft more courageous and effective criminal justice reforms.
Jails and prisons are often overcrowded, and their residents are disproportionately likely to have chronic health conditions that make them especially vulnerable to viral infections. So as the COVID-19 pandemic unfolds, we’ve been asked: Is social distancing (as recommended by the CDC and other public health agencies) even possible behind bars? Can incarcerated people maintain 6 feet from each other, and from correctional officers and other staff?
In short, the answer is no.
To answer this question, we looked at how the physical space of jails and prisons compare to that of cruise ships and nursing homes, two of the most prominent incubators of the virus.
The Grand Princess and Diamond Princess, two cruise ships implicated in the outbreak of COVID-19 in the United States, have typical cabins that range from 73 to 79 square feet per person (with furnishings like beds, dressers, chairs, desks, and tables).
And generally, any shared bedroom in a nursing home is required to have 80 square feet of space per resident (including necessary furnishings, like a bed, dresser, table, and chair).
We found that incarcerated people are living in quarters that are similarly sized, if not smaller. According to the American Correctional Association (ACA), cells in correctional facilities should have at least 25 feet of space per person in each cell that are “unencumbered,” meaning they are not taken up by the bunk, desk, or other furnishings.
That’s a 5X5-foot space for each person, leaving almost no room for maneuvering while maintaining the recommended 6 feet of distance between people. And we know that in some facilities, beds can be as close as 3 feet apart.
COVID-19 is hammering cruise ships and nursing homes because social distancing is impossible. Incarcerated people are living in comparable if not smaller quarters, but with a notable difference: On cruises and in nursing homes, people have in-room access to the necessary hygiene products and water – something that is often missing in correctional facilities.
We’re already seeing the appalling result in city and county jails nationwide, most notably on Rikers Island in New York City, where the coronavirus infection rate is already nearly 8 times higher than the rest of the city.
Incarcerated people are disproportionately affected by underlying health conditions known to exacerbate COVID-19, and social distancing is impossible. There is no time to waste: State and local governments must take swift action to reduce prison and jail populations.
For our virus response tracking and other jurisdiction-specific information, see our virus response pages.