Is The NIBRS Transition To Blame For Our Current Crime Trends?
Is The NIBRS Transition To Blame For Our Current Crime Trends?
Short answer: No. Longer answer: Also no.
The NIBRS transition is an issue I’ve written about before, but I’ve started getting more questions about whether the 2021 switch is potentially contributing to the declining crime trend so it seemed worth a longer look. I can appreciate why it might make sense - if fewer agencies are reporting data then there would be less crime reported.
But a deep dive into the issue shows why the NIBRS transition was only really a big deal for our understanding of crime in 2021. The impact of the transition on our understanding of 2023 (or 2024) crime trends, however, is virtually nonexistent. Blaming NIBRS for what we are seeing in 2023 and 2024 would be akin to blaming NFL officiating for deciding who made the Super Bowl this year — it happened once1 but that was several years ago and doesn’t have any bearing on recent results.
At the heart of the issue is the FBI’s requirement that only NIBRS-compliant agencies could submit crime data beginning in 2021. NIBRS collects much more data in a more sensible manner, but it also requires major changes for law enforcement agencies making the switch. The result was that just 65 percent of the nation’s population was covered by a NIBRS-compliant agency when the 2021 data was released in 2022, so crime estimates for 2021 were based on less than two thirds of the nation’s population.
The FBI estimates crime counts to fill in gaps when estimating national figures when agencies fail to report. It's usually not a big deal when only 5 or so percent of crime data, but it's a big problem if you're estimating 35 percent. This led to large confidence intervals and a ton of questions about the nation’s crime figures for 2021. If you can see 90 or 95 percent of a picture then you can guess what that picture is with much better accuracy than if only 65 percent of the picture is visible.
Fortunately, the FBI backtracked in 2022 and allowed non-NIBRS agencies to submit data and coverage improved dramatically to just about 94 percent in 2022 - a level that was comparable to historical reporting norms.
I tried to grab population coverage over the last couple of decades to show the changing coverage over time. The FBI has published a table since 2017 on the share of the US population covered each year by NIBRS and the old system — SRS (see 2019’s participation table here). I was able to give a rough approximation of coverage prior to 2016 using agency participation data from the CDE. It’s not exact (nothing with crime data ever is) but it shows how big the drop was in 2021 but that 2022's coverage was roughly back to normal.
Participation in NIBRS has improved even more in 2023 with the FBI reporting 77 percent population coverage as of May 2023. It’s my understanding that more than 85 percent of the US population was covered by a NIBRS-reporting agency by the end of 2023 (including 9 of the 10 largest cities in the country). Adding in non-NIBRS agencies should mean normal coverage again for 2023 data.
So the NIBRS transition wasn’t great for 2021 crime data, but fortunately it is no longer 2021. It is also fortunate that understanding 2023 and 2024 national crime trends does not solely have to rely on NIBRS data published 10 months after a year ends.
Indeed, there are lots of data sources telling the same story about what is happening with murder and violent crime. The AH Datalytics murder dashboard takes data from over 200 cities and combines it to estimate the national murder trend. Groups like the Major Cities Chiefs Association and Council on Criminal Justice do the same thing.
The Gun Violence Archive publishes shooting data gathered from open sources. And then there’s the CDC’s provisional mortality statistics which has provisional data now available into the second half of 2023.
In some ways the NIBRS transition is making it easier to evaluate crime trends by automating data collection and reporting processes. There are over a dozen state UCR programs (all NIBRS compliant) that are publishing reasonably fresh crime data in easily accessible portals, so if you want to know the trend in those states then you can just go and grab the data.
All told, agencies are reporting data to the FBI roughly as frequently now as they did before the NIBRS switch. I would feel reasonably confident — albeit with a greater degree of uncertainty — if the FBI data was our only source for understanding the nation’s murder and violent crime trends. But it’s not! The FBI data is matched by other data sources which all tell the same story.
The transition to NIBRS was an annoyance in 2021, but it is no longer the albatross on crime data that it was a few years ago and shouldn’t be seen as an excuse for ignoring what the available evidence points to as our current national crime trends.
Yes, I’m still mad at Bill Vinovich.
The albatross seems to be the failure to report. Major crime centers getting tired of being the goat have stopped/slowed reporting.
Hi Jeff:
From the beginning of the NIBRS (many years ago), it was acknowledges that it would increase the number of crimes reported:
Rape: No effect.
Robbery: Increased 0.5 percent.
Aggravated Assault: Increased 0.6 percent.
Burglary: Increased 0.8 percent.
Larceny: Increased 3.1 percent.
Motor Vehicle Theft: Increased 2.8 percent.
Unlike the UCR, NIBRS data contains detailed information on every crime reported to the police. The UCR uses a Hierarchy Rule for reporting. This means that only the most serious crime in an incident is reported.
One of the reasons that police agencies were reluctant to engage in the NIBRS was the fear that the additional crimes counted would increase their crime statistics (far more than the estimates I described above) and if reported crimes contained a number of events (i.e., a robbery using a stolen vehicle) you would have two crimes instead of one under the old system.
Regardless of what I first offered as to a limited impact on crime statistics, NIBRS adoption was feared. That plus the law enforcement data dumps at the end of the year being inaccurate indicate that the FBI is still having issues with NIBRS "and" UCR compliance.
Social media is filled with references that agencies remain reluctant to record crimes. It's not a matter of NIBRS participation, it "may" be more of an issue of crime counting.
Then you get back to the issue of crime reporting. The National Crime Victimization Survey offers approximately 250,000 hate crimes a year where the FBI offers 11,000. That's just one example. The vast majority of reported crime is a severe undercount.
There are so many examples of problems with FBI data that it boggles the mind. It's not the FBI's fault. They just collect what's offered.
It's my understanding (and I have no direct knowledge of the issue) that many local law enforcement agencies are simply not complying with NIBRS dictates for a variety of reasons. I had conversations with people within the FBI who tell me that they don't want to disparage local police agencies but they are concerned about the data they are offering.
You state that "The FBI data is matched by other data sources which all tell the same story." I don't think that's correct or even near correct. There are 25 examples I listed today in an article on increasing crime from 2020-2024. The National Crime Victimization Survey in their latest reports on overall and juvenile crime record very big increases along with an array of other data showing increases, see https://www.crimeinamerica.net/25-examples-of-increasing-crime-is-record-fear-of-crime-justified/.
Even the FBI's preliminary statistics for 2023 break statistics down to metro and nonmetro areas and the reductions for metropolitan areas are slight except for homicides and rapes (tiny numbers when compared to overall violent crimes).
I would like to believe that the NIBRS is being faithfully followed because it would be a considerable improvement over the UCR system. But the more I use 2022 NIBRS data for articles, and contrast them with other data, the more concerned I become.
Best, Len.