How Violence Rules Rikers

There is a Groundhog Day feel to the attempts to address the multiple tragedies and crises that beset the city’s jails. Except that unlike Groundhog Day, there is little comedy. From our very first publications, Vital City has used what data it could discover to offer recommendations on how to cure the conditions of violence in the jails. We are grateful to our partners at CUNY’s Institute for State and Local Government and to the many other organizations and individuals who have participated in this intensive work to offer solutions.

This is our third State of the Jails report, and the conclusion is roughly the same: The city’s jails remain the most richly resourced in money and staff in the nation, yet they are poorly run, resulting in appalling violence. Against everyone. The violence is dramatically above the levels that drove a federal court in 2015 to declare those conditions unconstitutional, and to put the City and the jails under the supervision of the federal court and a monitor (Nunez v. City of New York). And the failure to control the violence has compelled that court to contemplate seriously federal receivership, an extraordinary remedy that puts executive and operational power to run the jails in the hands of a person appointed by the court. Declines in some of the measures of violence in 2023 below where they were in 2022 might be encouraging. Except that the volatility of those metrics is like a bungee cord, rooted in no direction, rising and falling with apparently no control, resulting in variations that are unpredictable and cannot be relied upon as a durable trend.

What could control, and indeed better, the conditions inside in a durable way? Over eight years and thousands of pages, the Monitor has provided a detailed prescription. It boils down to this: fix management. This sounds antiseptic and bureaucratic in the face of the barbarous conditions of the jails. But it gets at almost everything that provides stability to the institution and life inside. There is an “if you give a mouse a cookie” aspect to this: To fix management means to train staff better; to train staff better means violence is not the first resort, thus the temperature of the facilities falls and with it the violence. With less violence, a decent life for both staff and those incarcerated is possible, one that includes uninterrupted programming, unimpeded medical visits, and even, in an important cycle, better relationships between officers and those in their care and custody. To fix management means fixing accountability, which means that adherence to a code of conduct ensuring decency is the norm not the exception, and so the entire culture of the jails shifts. 

Good management requires reliable metrics so that people — managers, of course, but also those incarcerated and the public as well — can see and have some confidence in whether things are improving. Metrics also provide a diagnostic as to where and why things are bad: Is violence worse in one facility? During a particular period of time, such as when people are transported to court or to medical visits? Do the incidents cluster among particular people in custody or particular officers on the watch? Does the violence increase the amount of time people stay in jail, thus pumping up the number of people in jail?

Over the course of the past year, the built-on-sand feature of any improvements was joined by another more malignant trend: hiding data and shading facts. Over the course of nine days in May 2023, one person who was incarcerated died; three suffered, as the monitor put it, “life-altering” injuries; and a fifth was mutilated. The intentional actions of the jail commissioner to hide information from the monitor obstructed efforts to understand and address the brutality of these incidents, and conclusively undermined the credibility of the department. The monitor only discovered the fact of the death and several of the other incidents either from the news or from tipsters. Neither contrition nor acknowledgement of fault was forthcoming from the correction commissioner, who instead asserted to the monitor that he was not required to report such information. It required a much litigated court order to reassert the basic norms of the consent decree.

Further examination of the data, and strange breaks in information and sharp departures in trendlines led the monitor to conclude that many of the key metrics — ones as basic as deaths and stabbings — could not be relied upon. In this report, we work with what is available but have “red-flagged” those charts where the monitor has deemed the data unreliable. Spoiler alert: There are many red flags.

Looking forward to 2024, we hope that the change in leadership in the department will also result in an injection of urgency, fed by reliable analytics, to address the multiple intersecting issues in the jails that threaten, daily, the lives and safety of all those housed and working there.  As Judge Laura Swain described in a December 20, 2023, court conference, “basic security lapses continue to put individuals at Rikers at grave risk of imminent harm…[T]here seem[s] to be a lack of any sense of urgency…about eliminating these dangerous conditions at the jails.”  We are encouraged that the court has ordered the parties to come up with metrics to measure progress. Meanwhile, we will continue to provide information that might shed light and, as always, welcome suggestions and information from our readers.

-Elizabeth Glazer

Chart 1. Violence dipped in 2023 vs. 2022, but remains well above its 2015 levels

1. A short history of the jails' population, staff and budget

New York City has one of the lowest incarceration rates of any large city in the United States. At the same time, its jails are some of the most richly staffed in the nation. As the number of people in custody has declined, the ratio of staff to incarcerated people has grown, reaching an apex in FY 21. From the 1990s until the mid-teens, jailed people outnumbered the staff. But starting in 2015 (FY 16), that trend flipped, with staff well outnumbering incarcerated people. By FY 23, there were 6,299 uniformed staff and an average daily population (ADP) of 5,823 people in custody, a ratio of 1.1 uniformed staff for every person incarcerated. This is roughly twice the Department of Correction (DOC) staffing ratio in the 1990s and more than four times the national average. 

Chart 2. There are still more uniformed staff than people in custody

Despite these ratios, DOC continues to have difficulties adequately staffing its facilities, with unstaffed posts a frequent and dangerous state of affairs, ballooning overtime costs and making double tours — sometimes even triple tours — a common occurrence. Even with high levels of staff absent from the jails because they are out sick, because they are restricted from working for medical or psychological reasons, or because they are AWOL, the ratio of uniformed staff to incarcerated people is still high: an estimated 0.9 in FY 23. 

The staffing crisis caused by officers calling in sick or AWOL hit an all-time high in 2020.  

The good news in 2023 is that the crisis has abated. Nonetheless, DOC stands out, continuing to have the highest sick rates of any uniformed workforce, with limited repercussions for those abusing the system. For comparison, the total absence rate for uniformed staff at DOC in FY 23 was 15.6%, compared to 8.5% for the Fire Department, 4.5% for NYPD and 8.2% for the Department of Sanitation. 

According to data from the New York City Comptroller Department of Correction Dashboard, the average percentage of uniformed staff out sick was 14% in 2023. Applying that to the FY 23 staff totals would mean 5,417 uniformed staff available to work in the facilities.  

Chart 3. Though fewer correction officers are out sick and ineligible to work, rates for these personnel remain two to three times those of other uniformed forces

The increased number of staff returning to the facilities in 2023 has not made a difference in increasing the likelihood that people in custody will make it to their scheduled medical appointments. To the contrary, August 2023 saw the highest number of missed appointments since the Department began reporting them in June 2020, with an unusually high proportion classified as for “other” reasons — making them reasons not subject to public scrutiny.

Chart 4. Thousands of incarcerated people are missing medical appointments, often for unclear reasons

The failure to receive timely care is particularly troubling given the considerable proportion of the population with lengthy stays in the city’s jail facilities. Jails do not typically have the types of facilities and programming necessary for long-term stays. In New York City, however, increasingly high proportions of the population are being held for longer periods of time, with between a fifth and a third of the population staying a year or more over the past few years, and 1 in 20 staying two or more years (according to discharge data from NYC Open Data).

Chart 5. More than a fifth of the jails’ population is held for a year or more

In recent years, DOC spending, which is almost entirely made up of personnel costs, has increased significantly. In 2023, the City paid an average of $443,695 per person in custody, compared to $151,046 in FY 2013 (or $198,986 in 2023 dollars). After a sharp increase in 2020, DOC is now spending more money per incarcerated person than at any time in history, while many conditions have failed to meaningfully improve. 

What has that money bought? 

Chart 6. Cost per person in custody continues to hover near record highs

2. Death and violence continue to plague the jails

Since 2015, the jails on Rikers Island have been operating under a consent decree enacted to remedy unconstitutional levels of violence and other dangerous conditions, a settlement in the case of Nunez v. City of New York et al. Under the consent decree, a federal court and monitor reporting to it oversee the DOC’s attempts to make things safer and more humane for those incarcerated and those working inside.

Although DOC jails have more money and higher staffing ratios than in past years, by many measures, violence inside them has sharply increased over the unconstitutional levels that led to the settlement in 2015. In 2022, the rate of death among incarcerated people was the highest it’s been since 1996. Violence has followed a similar trend. From FY 99 through FY 14, rates of stabbings and stashings remained below 10 per 1,000 average daily population (ADP). But from 2015 to the present, violence increased steadily, reaching a peak in FY 2022 of 88 per 1,000 ADP. 

Following FY 2022’s high-water mark, there were modest reductions in death rates and violence over the course of 2023. Understanding what is driving the moderation in levels of harm is critical to sustaining these trends, especially considering the variation in serious injury rates and other measures of violence month-to-month.  

Here, there is a disturbing red flag. If the data are unreliable or partial or otherwise tainted, then it is a fool’s errand to build a working management system guided by them. The federal monitor has raised questions about the reliability of the data, even with respect to information as critical as in-custody deaths. In one of his most recent reports, the monitor highlighted a 49% drop in stabbings and slashings from January to February 2023 following the issuance of a DOC memo that suggested to staff that not all serious violence needed to be reported at the discretion of staff. The monitor further identified the failure to classify a number of stabbings and slashings as such. 

Stabbings and slashings have historically been used as markers of violence in part because they were, in the past, considered less subject to discretion — each stabbing and slashing must be recorded by DOC and/or medical staff. That these too may now be unreliable amplifies existing concerns about the extent to which levels of violence can be meaningfully assessed in the facilities.  

With all these cautions, Vital City is presenting the data we have as context for conditions on Rikers. Where there is documented concern with the accuracy of the numbers, the reader will see those charts marked with a red flag. 

A. Deaths

Chart 7. Deaths in custody in 2023 dropped substantially from 2021 and 2022 highs

In 2022, The in-custody death rate reached a disturbing, historic high of more than 3 deaths per 1,000 people in custody — 19 in total. While the death rate for 2023 was lower, in total there were 28 known deaths of individuals in custody or shortly following release in the first two years of the current mayoral administration.

B. Injuries to incarcerated people and staff

Since the start of the federal monitoring period in 2015, violence rates have been elevated across all available metrics: among people in custody, against people in custody by staff and against staff by people in custody. The high rates of violence among people in custody are notable by both the most serious metric — stabbings and slashings — and the broadest, fights. 

Even by the Department’s own reports — which have themselves previously been shown to be undercounts — rates of serious injury by all causes in FY 23 were almost six times higher than they were in FY 17, the first year for which comparable data is available. 

Chart 8. Serious injuries to incarcerated people dropped — but the numbers are in question

i. Stabbings and slashings

When stabbings and slashings reached their lowest point in FY 08, there were roughly two uniformed staff people for every three people in custody. By contrast, since FY 16, despite more than one uniformed staff person per person in custody, stabbings and slashings have been consistently elevated, increasing to an alarming high of 88 per 1,000 ADP in FY 22. While the rate decreased in FY 23, it was still well above the high-water mark of the mid-1990s, 4,650% higher than FY 08’s low and 396% above the rate when the consent decree was put in place.  

Chart 9. Stabbing and slashing rates dropped, but they remain elevated — and again the numbers are in question

While a consent decree might be expected to decrease violence, the opposite has happened. Even accounting for the modest decreases in stabbings and slashings since FY 22’s peak, the FY 23 rate of stabbings and slashings was 396% higher than it was in FY 16, the start of the monitoring period. 

Moreover, the monthly rates were again on the rise in the latter half of the 2023 calendar year, with particularly elevated rates in August through October — a troubling uptick, even if the integrity of the data was not in question.

Chart 10. Viewed by month, stabbings and slashings remain well above 2016-2020 norms — and the numbers are in question

Looking at violence by facility is critical for understanding among whom violence is concentrated and the extent to which initiatives intended to curb violence are having an effect. (Facility census numbers are not available, so we are limited to showing total stabbing and slashing incidents rather than rates.) For example, the Commissioner’s Violence Reduction Plan and other efforts to reduce violence in the Robert N. Davoren Center (RNDC), initiated in early 2022, appeared to yield initial results, yet violence seems to be on the rise again in the facility. 

These numbers alone cannot tell us what is behind increases or decreases in violence. While a decrease in stabbings and slashings in one facility might signify meaningful improvements in violence, they could be the result of any number of other factors, such as increased lockdowns or moving known inciters of violence to other facilities. 

Chart 11. Rates of violence vary dramatically across jail facilities

ii. Other injuries to incarcerated people

Injuries to people in custody are not limited to stabbings and slashings. One of the core findings leading to the consent decree was that “the City has engaged in a pattern and practice…of subjecting these inmates to excessive and unnecessary use of force.” Accordingly, curbing use of force by staff against people in custody and improving tracking and reporting practices were central to the consent judgment and have been closely watched by the monitor. Despite this, the use of force has continued to proliferate in the facilities, and there are real questions about the reliability and completeness of the data describing it. 

While reduced from the apex in FY 21, the use of force rate in FY 23 was 374% higher than in FY 13 and 145% higher than in FY 16, the start of the monitoring period. The rate of use of force resulting in serious injury was similarly up 349% from FY 13, the first year of available data, and a staggering 526% from FY 16. While the total use of force rate to date in FY 24 is lower than the past few years, it continues to be elevated. 

The rate of use of force resulting in serious injury was noticeably lower in the recent months compared to the past two years, largely due to a precipitous drop in November 2022. After the numbers started to trend upwards again, the DOC reported another significant drop in use of force with serious injury in the past few months, with September and October reporting the lowest rates since FY 17. 

Educated observers should have serious concerns about the accuracy of those numbers. The overall use of force rate has not shown a similarly sharp decrease, and in fact October saw the highest monthly total use of force rate since December 2022. It is unclear, then, whether people in custody are being seriously injured less often by staff or whether instead their injuries are going unreported or being reported as less severe; the monitor has expressed considerable skepticism about the Department’s general reporting of serious injuries. 

Chart 12. Use of force rates have dipped slightly in previous years, but they remain high — and the numbers are in question

Monthly trends show an overall rate of force that appears to once again be on the rise in recent months. These month to month indicators also underscore the sharp drop in reported force resulting in serious injury to people in custody, and how aberrant a consistent drop to these levels — six or fewer incidents per month starting in August — is. 

Chart 13. Use of force rates viewed by month remain well over their 2016-2018 norms — and the numbers are in question

Fights showed a slightly different trend than the more serious forms of violence, with rates peaking during the height of COVID, decreasing in 2022 and then rising again in 2023. While not always resulting in serious injury, this increase in fights reflects the difficulty that DOC is having in controlling chaos in the facilities. 

Chart 14. Reported fights in custody went up after a dip in 2022

iii. Self-harm

The rate of self harm is another crucial indicator of conditions inside, and one that has been raised as an area of considerable concern by the federal monitor and the court. While data on self-harm is limited, what information is available suggests considerable fluctuation month-to-month, yet a generally elevated rate for most specific types of harm starting mid-2020 and continuing through September 2023, the last month for which data is available. The incidence raises serious questions not only about the Department’s ability to prevent self-harm, but to recognize and report it as such — a lapse that may further impede prevention efforts.

Chart 15. Reported rates of self harm among incarcerated people remain elevated

iv. Injuries to staff

The violence in the facilities is a concern not only for the people housed there but also for those working there, and Rikers remains for many an unsafe place of work. However, this is an area where the Department seems to be making some headway. Similar to fights, assault rates against staff peaked during the height of COVID in 2020 and 2021 and decreased in 2022. By contrast, however, here there have been more sustained decreases that continued into 2023, resulting in the lowest annual rate since 2017.

Chart 16. Assaults on staff are substantially down from a 2021 peak

Assaults on staff resulting in serious injury remain a relatively infrequent occurrence, and here, too, there are signs of improvement. The trends have generally mirrored those of overall assaults on staff, with the important exception that declines did not continue from 2022 to 2023; instead, rates reported through October 2023, the latest information available, are almost identical to those at the same time in 2022. 

Chart 17. Fewer assaults on staff are resulting in serious injury, though the data for 2023 is incomplete