Is FBI Crime Data Unreliable? Fact-Checking the 'Missing Cities' Narrative

In recent political cycles, a major debate has emerged regarding the reliability of the United States' national crime statistics. Following declarations by federal law enforcement that violent crime has experienced historic declines since its post-pandemic peak, conservative commentators and political candidates have argued that the data is incomplete and misleading. They allege that the FBI’s reports are "fake" or "meaningless" because thousands of local police departments, including those in the nation’s largest metropolitan areas like New York City and Los Angeles, fail to report their statistics [4]. Conversely, criminologists and federal statisticians argue that while initial reporting gaps existed during a major system transition in 2021, participation has since rebounded to historical norms [1]. To evaluate these claims, we examine the mechanics of federal crime reporting, the current participation rates of major cities, and how federal data compares with independent crime indices.

The Origin of the Reporting Gap: The NIBRS Switch

The skepticism surrounding federal crime statistics originates from a fundamental change in how the FBI collects data. For nearly a century, the FBI relied on the Summary Reporting System (SRS) to compile national crime numbers under its Uniform Crime Reporting (UCR) program [6]. SRS was a relatively simple tally of the most severe offense in a given incident. On January 1, 2021, the FBI officially retired SRS and transitioned exclusively to the National Incident-Based Reporting System (NIBRS) [4][6].

NIBRS is a far more robust system. Instead of counting only the single most severe crime in an incident, it collects detailed, incident-level data on up to ten distinct offenses per event, including the demographics of victims and offenders, the relationships between them, and weapon use [6]. However, implementing NIBRS required local law enforcement agencies to make complex and expensive upgrades to their Records Management Systems (RMS) and retrain personnel [4]. Because participating in the FBI’s crime reporting program is entirely voluntary, many police departments were not prepared when the deadline arrived. In 2021, over 40% of agencies nationwide—including the NYPD, LAPD, and the Chicago Police Department—failed to submit any data to the FBI [4][7]. This left massive holes in the national database, forcing the FBI to rely on statistical estimation models to estimate national trends, which understandably invited scrutiny from researchers and political commentators alike [4].

The Transition Rebound: Closing the Reporting Gaps

Recognizing that the lack of participation was undermining the credibility of national crime numbers, the FBI adjusted its approach. In 2022, the agency began allowing law enforcement departments that had not yet transitioned to NIBRS to submit their data in the legacy SRS format, or assisted them in writing conversion scripts to map SRS data onto NIBRS requirements [1][5].

This policy change, combined with local technical progress, led to a dramatic rebound in participation. By the 2022 and 2023 reporting cycles, population coverage rose back to approximately 94% [1]. By the time the FBI published its finalized 2024 annual crime report on August 5, 2025, participation had stabilized [1]. A total of 16,675 law enforcement agencies submitted data, representing 95.6% of the U.S. population [1]. Crucially, all large-city agencies serving populations of 1,000,000 or more, including the NYPD, contributed a full 12 months of data [1]. Furthermore, on March 7, 2024, the LAPD officially launched its NIBRS-compliant Records Management System, allowing it to submit granular incident-level data alongside other major cities [2]. Preliminary figures for 2025 indicate NIBRS-specific reporting alone covers 90% of the population, with overall coverage (including legacy submissions) sitting at 96% [1].

What the Data Shows: Cross-Referencing Independent Sources

Because the voluntary nature of the FBI's reporting can result in administrative lags, researchers often cross-reference federal statistics with independent, non-governmental databases. One of the most prominent sources is the Major Cities Chiefs Association (MCCA) Violent Crime Survey, which collects data directly from 69 of the largest U.S. law enforcement agencies, bypasses federal reporting pipelines, and publishes comparative reports [2].

The MCCA’s Year-End Comparison Report for 2024, published on February 4, 2025, corroborates the downward trend documented by the FBI [2]. The MCCA report tracked a combined 6,520 homicides across major cities in 2024, compared to 7,797 in 2023—representing a 16.4% decline [2]. Rapes fell by 5.6%, robberies by 9.2%, and aggravated assaults by 4.6% [2]. These major-city trends closely track the FBI’s nationwide numbers, which recorded a 14.9% decline in homicides and a 4.5% decline in overall violent crime for the same period [1].

Additionally, the Real-Time Crime Index (RTCI), compiled by AH Datalytics, automatically pulls crime data directly from the public-facing dashboards of hundreds of police agencies nationwide [5]. The RTCI also recorded a double-digit decline in homicides in 2024 that has persisted into 2025, providing immediate, bottom-up validation of the federal numbers without depending on the FBI's processing timeline [5].

The Full Picture: Why Public Perception Diverges (UCR vs. NCVS)

While the data confirms that reported crime has fallen, public concern regarding public safety remains elevated. A key explanation for this divergence lies in the difference between two federal programs: the FBI’s UCR (which measures reported crimes) and the Bureau of Justice Statistics’ National Crime Victimization Survey (NCVS) [3][6].

Unlike the FBI data, which relies on police logs, the NCVS conducts annual surveys of approximately 240,000 individuals in U.S. households to ask if they have been victims of crime, regardless of whether they reported it [3]. According to the NCVS, only about 45% of violent crimes and 30% of property crimes are ever reported to the police [3]. Consequently, if reporting rates change, UCR statistics can fluctuate even if the actual level of crime remains constant.

The NCVS report for 2024 (released in September 2025) found that the rate of violent victimization was 23.3 per 1,000 persons age 12 or older, which was not statistically different from the 2023 rate of 22.5 per 1,000 [3]. However, critics point out that these rates remain higher than the pandemic-era low of 16.4 per 1,000 recorded in 2020 [3]. Criminologists explain that this increase is driven primarily by simple assault (which is the most common victimization and rarely reported to police), whereas severe, lethal offenses like homicide—which cannot be measured by household surveys—have demonstrably fallen. Homicides rose by nearly 30% in 2020 but have since plummeted, a trend validated across police logs, hospital emergency room data, and CDC mortality statistics.

Timeline of the Crime Data Controversy

Conclusion

An objective examination of public safety data shows that the narrative of "missing cities" hiding rising crime is outdated. While the transition to NIBRS in 2021 created legitimate reporting gaps that temporary statistics had to account for, by 2024 and 2025, participation rebounded to cover 95.6% to 96% of the U.S. population, including full reporting from New York and Los Angeles. Independent data from major city police chiefs and real-time scrapers confirm that violent crime, particularly homicides, has declined significantly from its pandemic-era peak. However, the BJS victimization survey illustrates the complexity of public perception: because most non-lethal crimes are never reported to police, victimization rates remain higher than their 2020 lows, keeping public concern elevated even as the most severe violent offenses continue to decline.