Violent Crime Trends, 1990-2021

Since the 1930s, the Federal Bureau of Investigation's (FBI's) Uniform Crime Reporting (UCR) program has collected data on reported crimes from federal, state, local, and tribal law enforcement agencies and published nationwide crime statistics using these data. UCR compiles and publishes the data and provides datasets and some trend analyses in a standardized format that law enforcement agencies, researchers, and policymakers can access to better understand crime in the United States.

Shifting Crime Data Collection Tools

UCR previously collected crime data through the Summary Reporting System (SRS); however, on January 1, 2021, the FBI retired SRS in favor of the National Incident Based-Reporting System (NIBRS). NIBRS was developed to address several shortcomings of SRS. For example, NIBRS expands the number of crimes for which data are collected and captures details about these crimes that were not previously measured in SRS (e.g., the relationship between the victim and offender).

While NIBRS has recently became the FBI's primary crime data collection tool, it was first made available to law enforcement agencies beginning in the late 1980s. In the ensuing years, agencies could report their crime data in either the SRS or NIBRS formats. In 2016, the FBI announced their intention to retire SRS as of January 1, 2021. This meant that 2021 would be the first year for which crime data were collected solely through NIBRS. Despite the 2016 announcement, many law enforcement agencies did not adopt NIBRS; as a result, participation rates in the UCR program dropped in 2021 (when 62.7% of law enforcement agencies submitted NIBRS data to the FBI). These lower participation rates required the FBI to adjust their data analysis and reporting procedures.

Because there is only a single year of NIBRS data and new procedures had to be adopted to account for the low participation rates, the FBI stated that the 2021 data cannot be integrated into longitudinal analyses that mix NIBRS and SRS methodologies. Therefore, crime data from 2021 are discussed separately from earlier crime data in this In Focus.

Violent Crime Trends, 1990-2020

UCR data on violent crime includes homicide (i.e., murder/non-negligent manslaughter), rape, robbery, and aggravated assault. Data are published about each offense individually as well as in a composite measure—the violent crime index.

This section focuses on the violent crime and homicide rates from 1990 to 2020. As demonstrated in Figure 1, the national violent crime rate (blue line) peaked in 1991 at 758.2 violent crimes per 100,000 people. The violent crime rate has generally decreased since 1991. There was an increase in this rate in 2005 and 2006, after which it continued decreasing until 2014. The violent crime rate increased from 2014 to 2016, from 361.6 to 397.5 per 100,000; decreased between 2017 and 2019; and increased again from 2019 to 2020. During the most recent increase, the violent crime rate rose from 380.8 in 2019 to 398.5 per 100,000 in 2020. The rate observed in 2020 was still below the peaks observed in the early 1990s and the increase in the mid-2000s, but was close to the rate of 397.5 per 100,000 measured in 2016.

Figure 1. Violent Crime and Homicide Rates, 1990-2020 Rate per 100,000 People

Source: UCR Program.

Similarly, during the years considered, the national homicide rate (red line) peaked during 1991 at 9.8 homicides per 100,000 people. The rate generally decreased from there, with small increases in some years during the 2000s but never exceeding more than an 0.2% uptick in consecutive years. Similar to the violent crime rate, the national homicide rate began to show an increase beginning in 2015, growing from 4.9 per 100,000 in 2014 to 5.4 in 2015. The homicide rate remained in a range of 5.0 to 5.4 per 100,000 until 2020 when it increased to 6.5 homicides per 100,000 people.

For the other offenses included in the violent crime index, data indicated the following changes from 2019 to 2020:

• The rate of aggravated assault increased from 250.4 to 279.4 offenses per 100,000 people.
• The rape and robbery rates both decreased (from 43.6 to 38.4 and from 81.8 to 73.9 per 100,000 people, respectively).

Violent Crime, 2021

The FBI released nationwide crime data in 2021 even though participation in NIBRS was low. These data were generated using new estimation procedures developed by the FBI in partnership with the Bureau of Justice Statistics (BJS) to account for missing data from nonreporting law enforcement agencies and those that reported incomplete annual data. The FBI has used estimation procedures to account for missing data since the 1960s. An August 2022 report from the FBI stated that in the past, about 95% of the population was covered in SRS data. This level of coverage was not achieved in the 2021 NIBRS data, and as a result new estimation procedures were necessary to account for the lower participation rate.

This section summarizes the findings in two FBI reports, 2021 Crime in the Nation Summary and The Transition to the National Incident-Based Reporting System (NIBRS): A Comparison of 2020 and 2021 NIBRS Estimates, both of which were released in October 2022. These reports used estimates generated from NIBRS data collected during 2020 and 2021, which makes it appropriate to compare these data to each other. As discussed earlier, SRS data should not be analyzed or graphed alongside NIBRS data; however, trends (e.g., increase in homicide rate) observed in SRS data can be discussed relative to those seen in NIBRS data.

The FBI's analyses indicate that the violent crime index remained relatively stable between 2020 and 2021. The number of violent crimes decreased by 1%, but this change was not statistically significant. The FBI advised that the stability between 2020 and 2021 should be considered in the context of the increase observed between 2019 and 2020. In other words, while violent crime may not have significantly changed between 2020 and 2021, it remains situated within the upward trend that began in 2019.

The number of homicides increased by 4.3% between 2020 and 2021; however, this increase was not statistically significant. Here again, the 2021 homicide count should be placed in the context of the prior increase. In this case, the non-statistically significant increase in homicide from 2020 to 2021 should be considered in the context of the upward general trend that was observed in the SRS data beginning in 2015.

The number of aggravated assaults decreased by 0.1% between 2020 and 2021. This change was not statistically significant. The aggravated assault rate increased between 2019 and 2020, and thus the FBI interpreted the 2021 data to mean that after the prior year's increase the aggravated assault rate remained stable in 2021.

The number of rapes increased by 3.4% between 2020 and 2021; however, this increase was not statistically significant. The rate of rapes decreased between 2019 and 2020, so this nonstatistically significant increase can be interpreted to indicate that after the prior year's decrease the rate of rape remained stable in 2021.

The number of robberies decreased by 8.2% between 2020 and 2021. Although this decrease was not statistically significant, it is consistent with the downward trend observed from 2019 to 2020. Notably, the FBI determined that the diminished number of robberies was the main driver of the decrease in the overall violent crime index.

In summary, the increase in violent crime observed from 2019 to 2020 did not continue in 2021, nor was it reversed. There were non-statistically significant increases observed in the number of homicides and rapes in 2021, and non-statistically significant decreases observed in the number of robberies and aggravated assaults as well as in the overall violent crime index.

Policy Considerations

The coming years may be indicative of how UCR reporting could change during the NIBRS era. If participation rates remain relatively low, the FBI could try to adopt a long-term strategy to address crime data gaps or the federal government may consider whether to invest greater resources in an effort to convince law enforcement agencies to adopt NIBRS.

Beyond concerns about data quality, low rates of NIBRS adoption may also affect law enforcement agencies' access to federal funding. Several federal grants, most notably the Edward Byrne Memorial Justice Assistance Grant (JAG) program, use UCR crime data to calculate allocations for state, local, and tribal governments. JAG funds are allocated using a statutorily defined formula under which 50% of a state's JAG allocation is based on its share of the population and the other half on the number of violent crimes (i.e., homicide, rape, robbery, and aggravated assault) reported to the FBI. Any state that does not adopt NIBRS may lose out on the funds determined by the number of crimes reported to UCR.

Although estimation procedures can be used to approximate national crime trends, gathering data from a larger proportion of law enforcement agencies would result in more precise federal crime data. Congress could consider offering incentives or leveraging penalties to encourage law enforcement agencies to adopt NIBRS. Congress may target these efforts at smaller agencies that struggle to meet the increased technological requirements for NIBRS or at larger agencies that have yet to implement the program.

For more information on the adoption of NIBRS, see CRS Report R46668, The National Incident-Based Reporting System (NIBRS): Benefits and Issues; and CRS Insight IN11936, NIBRS Participation Rates and Federal Crime Data Quality. For more information on the JAG program, see CRS In Focus IF10691, The Edward Byrne Memorial Justice Assistance Grant (JAG) Program.

This product was originally authored by former CRS Analyst in Social Policy Emily J. Hanson.