Bad Watchdog Season 2 launches June 20.

Policy Letter

POGO Calls for Garland to Improve Justice Department’s Implementation of Death in Custody Reporting Act

The Honorable Merrick Garland
Attorney General
U.S. Department of Justice
950 Pennsylvania Avenue
Washington, DC 20530-0001

Dear Attorney General Garland:

The Project On Government Oversight (POGO) is a nonpartisan independent watchdog that investigates and exposes waste, corruption, abuse of power, and when the government fails to serve the public or silences those who report wrongdoing. We champion reforms to achieve a more effective, ethical, and accountable federal government that safeguards constitutional principles. We write to urge you to improve the Department of Justice’s implementation of the Death in Custody Reporting Act of 2013 (DCRA). Seven years after the act became law and over a year since the murder of George Floyd demonstrated once again the urgent need to address violence by law enforcement officials, the Justice Department is still not fulfilling its statutory obligation to provide policymakers and the public with basic facts about deaths in custody. Ensuring that law enforcement agencies are properly reporting this data and that offices within the Justice Department are promptly and adequately disclosing their findings to the public are essential steps in meeting your pledge to pursue equal justice as attorney general.1

The act, which became law in the wake of the 2014 killing of Michael Brown by a police officer in Ferguson, Missouri, is meant to ensure the federal government has the information necessary to oversee law enforcement and corrections agencies.2 The act requires the Justice Department to collect information on instances in which people die while being detained, arrested, or incarcerated by local, state, or federal authorities. State and federal law enforcement agencies must report this information to the Justice Department. The act also requires the attorney general to prepare a study within two years of its passage that provides recommendations on how to reduce deaths in custody, and allows the attorney general to reduce noncompliant states’ criminal legal funding allocations under the Edward Byrne Memorial Justice Assistance Grant program.

We are particularly concerned about delays in data collection and reporting, inadequate data collection, and a lack of data transparency.

Delays in Collecting and Reporting Data

The department has made the Bureau of Justice Assistance (BJA) responsible for collecting state and local data, and tasked the Bureau of Justice Statistics (BJS) with collecting federal data. But BJA did not begin collecting information from state agencies until 2020.3 While we are encouraged that collection is now underway, this delay has cost the department and the country years of vital insights into law enforcement violence. In the meantime, a patchwork of independent research and media organizations have attempted to track how many people have been killed by law enforcement officials sworn to protect them.4 The findings emerging from these independent monitoring efforts have been profoundly unsettling. For example, a report released in May 2021 by the Raza Database Project and UnidosUS found that people of color, who comprise less than 40% of the U.S. population, comprised more than 60% of all people who were killed by law enforcement officers or died in their custody, and that the deaths of people of color are profoundly undercounted in many databases—perhaps by as much as 50%.5

The Bureau of Justice Statistics has collected data from federal law enforcement and corrections agencies in a timelier manner.6 However, its analysis of that data has still taken years to reach the public. The bureau released its first report on federal deaths in custody, covering fiscal years 2016 and 2017, in December 2020, more than three years after the end of the period the data covers.7 While we understand that analyses often take time, a reporting lag of this magnitude seriously undermines the usefulness of the reports in informing current public policy responses intended to minimize deaths in custody. At a time when ending violence by law enforcement and corrections officials against minority communities is one of the most pressing issues facing the country, the absence of comprehensive data on officer-involved homicides is deeply troubling.

Lastly, delays in collection and reporting have made it impossible for the Department of Justice to complete its statutorily mandated analysis of how to reduce deaths in custody. In 2016, then-Attorney General Loretta Lynch submitted a report to Congress on the department’s planned implementation of the act.8 However, as meaningful data had not yet been collected, it did not contain any analysis or recommendations for how to reduce deaths in custody. Nevertheless, as of 2018, the department’s position was that the 2016 report satisfied the act’s requirement and no further report was planned.9

Data Collection Problems

Both the Bureau of Justice Assistance and the Bureau of Justice Statistics collect death-in-custody data by soliciting responses from relevant agencies using standardized collection forms. However, both bureaus’ collection methodologies lower the quality of the data they receive.

The Bureau of Justice Assistance’s data collection methodology does not provide a clear mechanism for improving data quality over time, nor is it sufficiently robust to provide a detailed and comprehensive picture of deaths in custody.

The Justice Department’s original proposal for this program contained several measures to ensure that data quality would improve over time. One such mechanism was that BJA would conduct open-source reviews for mentions of deaths in custody, and cross-reference those findings with what the states were reporting to the bureau.10 In addition, states would have had to file yearly plans outlining how they planned to improve their compliance with the act. This proposed methodology incorporated lessons from an earlier program the Bureau of Justice Statistics used to collect data on arrest-related deaths.11 However, in 2018, the department revised the proposed methodology, eliminating the open-source review and improvement plan provisions.12 POGO joined a broad coalition of organizations condemning the decision at the time, and we remain concerned that the 2018 methodology will produce inaccurate results.13 The Justice Department inspector general raised similar concerns in a 2018 review of the department’s implementation of the act.14

Even if the information the Bureau of Justice Assistance receives from state agencies is accurate, the office may not be requesting enough information from these agencies for the resulting analysis to be useful to policymakers. While the collection form has standardized fields for demographic information about individuals who die in custody and for the cause of death, it relies on a text field for any additional information about the incident leading to a death. The form provides minimal prompts for the type of information that should be included in this field. This may result in responses that lack sufficient detail. Additionally, the qualitative data from this text field cannot be easily aggregated or compared across incidents. In contrast, the Bureau of Justice Statistics collection form for federal agencies prompts agencies to include more detailed information, and includes several additional standardized data fields to aid with aggregation and comparison.15

However, BJS’s federal data collection also appears to have gaps. The 2018 inspector general review found that at least three federal agencies that are required to report data had not done so, and at the time of the review, the bureau did not even know the total number of federal agencies that were required to report.16

Lack of Data Transparency

Even if the department collects accurate data and analyzes it in a timely manner, that information will only be useful for informing public policy if the public has meaningful access to it.

Attorney General Lynch’s 2016 report did commit to releasing information collected under the act to the public, once personally identifying information had been removed.17 However, guidance released during the Trump administration indicated that those plans had been abandoned, and that data collected from states would not be released except where required by the Freedom of Information Act.18 As of April 2021, the department’s guidance had been updated to suggest that non-personally identifying information could be released, and that the department does “periodically publish general data related to deaths in custody.”19 While an improvement, this still falls short of a commitment to release anonymized data to the public.

When the department does release data and analysis, it must ensure that it provides meaningful information. The Bureau of Justice Statistics’ first report on federal deaths falls short in this regard. While it does contain a variety of useful details, it obscures one of the most important numbers. The report lists homicides, a category it notes includes but is not limited to instances in which a federal officer committed the homicide, but does not provide the specific number of each type of homicide.20 Notably, the 2016 collection form did ask for data that distinguished between homicides by federal officials and other types of homicides while in custody (such as an altercation with a fellow detainee), but it appears the report that was ultimately released to the public purposefully obscured this detail. Without more detailed reporting, it will remain impossible for policymakers and the public to know how many people have been killed by law enforcement and corrections officials.

There are a number of reforms that will address these problems.

Improve collection methodologies. Both BJA and BJS have problems with their collection methods. BJA should adopt the methodology originally proposed by Attorney General Lynch in 2016 and consider modeling its collection form on the one BJS uses to ensure it receives information that allows for meaningful analysis. For its part, BJS must ensure it is collecting data from all relevant federal agencies.

Perform analyses in a timely manner. The longer it takes the department to analyze and report on data, the less useful it will be, and more people will die at the hands of officials in the meantime. The department must make its analyses under the act a high priority.

Commit to performing the statutorily mandated study. Now that BJA is collecting data, the department must perform the study the act requires, including analyzing the data and developing responsive strategies to reduce deaths in custody.

Commit to releasing data publicly. Now, more than ever, it is essential for the public to know how many people die in law enforcement custody. While we are encouraged that the department has apparently revised its position that death-in-custody data would not be proactively released at all, it is essential to recommit to the same level of transparency promised by Attorney General Lynch in 2016: the release of all anonymized data to the public.

We appreciate your attention to this critical matter. If you have any questions, please contact Brandon Brockmyer at [email protected]. We would welcome an opportunity to discuss any obstacles to the department’s full implementation of the Death in Custody Reporting Act.


Danielle Brian
Executive Director

Brandon Brockmyer
Director of Research