← Back to Blog
DataAnalysisCrime TrendsDeveloper

Eighty-Eight Means None

📅 November 16, 2026·⏱ 10 min read·By SpotCrime

On page seventy-four of the 2017 edition of the FBI's NIBRS user manual, below the codes for anti-Sikh and anti-transgender bias, there is a line that reads “88 = None (no bias).” Type it on every offense for a year and your agency has reported zero hate crimes. In 2024, about thirteen thousand agencies did.

That figure comes from the Southern Poverty Law Center, which reported that 81 percent of the 16,419 agencies sending hate crime data to the FBI for 2024 “affirmatively reported zero.” More than fifty of them serve populations over 100,000. Almost three thousand other agencies sent nothing at all.

This post is about that zero. Where it comes from, what it is made of, and why you should treat it differently from every other zero in a crime feed.

A zero built one offense at a time

Under the old Summary Reporting System, hate crime was its own paperwork. An agency filed incident reports when it had them and a quarterly form when it didn't, and the quarterly form had a place to say none. NIBRS folded all of that into a single field. Data Element 8A, Bias Motivation, sits on every offense in every incident. The manual lets an officer enter up to five bias codes per offense, or one of two values that exclude all the others: 88 for none, 99 for unknown.

The manual is specific about the difference. Offenses with no facts pointing to bias get 88. Offenses with “ambiguous facts” get 99, and the agency is supposed to go back and update the record once the investigation settles it. The bar for a real code is high on purpose. An officer should report bias only when the facts would lead “a reasonable and prudent person” to conclude the offender was motivated by it.

So an agency's annual zero isn't a single decision by anybody. It's the sum of thousands of small ones, made by officers filling out a field at the end of a call, mostly on offenses where bias was never a question. A burglary of a garage gets 88. So does a shoplifting. So, in some departments, does a beating that came with a slur, because nobody wrote the slur down.

That last case is the one that matters, and the field can't tell you how often it happens. A correctly coded 88 and a missed hate crime look identical in the data. Both are the number eighty-eight.

81%
Participating agencies that reported zero hate crimes for 2024, per SPLC.
10,881
Hate crime incidents reported to the FBI for 2025, per the Justice Department.
~246,900
Average annual hate crime victimizations, 2005–2019, per the BJS victim survey.

The two counts

The country measures hate crime twice, and the two numbers are not close. The FBI count is built from those 8A codes. For 2025, according to the Justice Department's summary, agencies reported 10,881 incidents involving 13,438 victims, down from 11,679 incidents in 2024.

The other count comes from the National Crime Victimization Survey, which asks people directly. In a 2021 report, Bureau of Justice Statistics statisticians Grace Kena and Alexandra Thompson estimated about 246,900 hate crime victimizations a year between 2005 and 2019. For 2010 through 2019 the annual average was 243,770. Over the same decade, the FBI program recorded an average of 7,830 hate crime victims a year.

BJS is careful here, and so should we be. The survey counts a crime as hate-motivated if the victim perceived it that way, for instance because the offender used hate language. The FBI requires police to find sufficient objective facts. The report says plainly that “complete congruity should not be expected” between the two. A thirty-to-one ratio is not a measured undercount.

The survey does tell you where the gap opens, though. About 44 percent of the hate crime victimizations it captured from 2010 to 2019 were reported to police. Of those, victims said only 13 percent were confirmed by investigators as hate crimes. The rest met the survey's definition because of slurs or symbols left at the scene. Some of that gap is victims who never call. Some is the field at the end of the report.

The list of big zeros

When the FBI released its 2025 numbers in August, the American Jewish Committee put out a statement noting that at least fifty localities of 100,000 or more had reported zero hate crimes. It named some: Elizabeth, New Jersey. Hollywood, Florida. Little Rock. Montgomery. Orange, California. Savannah. Suffolk, Virginia. Toledo. Waco. (AJC didn't publish its method or a full list, so treat the names as its snapshot, not ours.)

Now the other end of the table. Jeff Asher, working from the FBI's preliminary monthly files in a May post, found that the agency with the highest reported hate crime rate in 2025, among those serving at least ten thousand people, was Malden, Massachusetts. Population about sixty-six thousand. Reported hate crimes: sixty-six. Forty-six of them were theft offenses coded with an anti-LGBT bias.

Asher also noticed that Teaneck, New Jersey, a township of about forty-two thousand, reported as many 2024 “Cyberspace” hate crimes as California, New York and Texas combined. Eight.

Toledo has roughly four times Malden's population. We are not going to tell you which of those two numbers is closer to the truth. We don't know, and neither does the data. What the pair shows is that a single department's coding habits can put it at the top of a national ranking, and another department's can keep a city of a quarter million off the table entirely.

What a department that looks for it produces

The NYPD runs a dedicated Hate Crime Task Force, and the city publishes confirmed hate crime complaints as their own open-data file, one row per complaint with a bias motive column. Here is the query we ran against it:

curl -G "https://data.cityofnewyork.us/resource/bqiq-cu78.json" \
  --data-urlencode "\$select=complaint_year_number,count(*)" \
  --data-urlencode "\$group=complaint_year_number" \
  --data-urlencode "\$order=complaint_year_number"

The file lists 606 hate crime complaints for 2025, down from 759 in 2023 and 710 in 2024. Through September 2026 it has 506. Seventy-two of the city's seventy-seven precincts logged at least one in 2025. The most common charge was aggravated harassment in the first degree, with 186. Felonies outnumbered misdemeanors, 345 to 255.

Two notes on that. Precinct 90, which covers much of Williamsburg, had 46, more than double any other precinct. And the city's file starts in earnest in 2019, so it can't tell you anything about the years before. We aren't holding New York up as the correct number. We're pointing out that a department with a unit whose job is to find bias finds it in nearly every precinct, while dozens of cities of over 100,000 report finding it nowhere.

Most of the city incident files we ingest carry no bias field at all. When one does, it usually arrives the way New York's does, as a separate file of confirmed cases. That means the absence of a hate crime in a general feed is rarely a zero. It's a column that was never there. For more on how a single true-or-false field can carry an entire policy argument, see our post on Chicago's domestic flag.

Participation went up, and that complicates the trend

According to the ADL's summary of the FBI release, 16,791 agencies took part in hate crime reporting for 2025, or 85.9 percent of those enrolled, up from 16,419 (84.9 percent) the year before. That's 372 more agencies. The FBI also publishes a like-for-like comparison, limited to agencies that reported at least six common months in both years, precisely because the raw total moves when the roster does.

We didn't find a published zero-reporting share for 2025, so we won't guess one. Our hunch, and it is only a hunch, is that late joiners skew small, and small agencies are the ones most likely to report zero. If so, more participation can mean a bigger national total and a bigger share of zeros in the same year. In our view that is a reason to read the like-for-like series and ignore the raw total when you are talking about trend.

Asher's own caveat is the right one to keep in your pocket. Hate crime data, he wrote, frequently reflects changes in reporting patterns, and a decline in reported hate crimes doesn't by itself mean fewer of them happened. We'd add that the reverse holds too. A department that trains its officers on bias indicators should expect its count to rise. We have written before about offenses that exist in the data only when somebody goes looking for them. Hate crime is that problem with an extra step, because the offense was reported but the motive has to be found.

One state, up close

Indiana is a useful scale model. WFYI reported that 240 Indiana agencies submitted data for 2024, and 181 of them did not report a single hate crime. Only a quarter reported at least one. The state total was 149 incidents, down 17 percent from the year before.

A 17 percent drop sounds like a finding. Spread over fifty-nine reporting agencies, it is a few dozen incidents moving between a handful of departments. That's the small-numbers problem again, and it shows up at the state level here because the counts are so thin.

What to do with a hate crime field

If you build anything on top of this data, here is what we would do.

  1. Keep three states, not two.“Reported zero,” “did not report,” and “reported a number” are different facts. Store non-participation as null. Never coalesce it to 0.
  2. Never rank places by hate crime rate. Malden and Toledo are the whole argument. The FBI itself warns against ranking with its data, and this is the series where the warning matters most.
  3. Flag long runs of zeros in large jurisdictions. One zero year in a town of four thousand is expected. Five consecutive zeros in a city of two hundred thousand is a reporting-practice signal. Label it as one.
  4. Carry 99 separately if your source has it.“Unknown” is supposed to be temporary. A pile of old 99s tells you an agency isn't finishing the job.
  5. Use the like-for-like series for trend. Raw national totals mix real change with changes in who reports.
  6. Put the victim survey next to the police count.Not as a correction factor, because BJS says the two aren't directly comparable, but as context. The same idea runs through our post on the dark figure of crime.

Where we come down.We don't think the FBI's hate crime count is useless. It is the only national record of incidents police confirmed, and the like-for-like series is a fair read on direction. We think a zero in that count is a statement about an agency's paperwork, and it should be displayed that way. For a parallel case where a participation threshold decides what the FBI may publish at all, see the use-of-force number.

The manual lists thirty-four bias codes, each a two-digit number, from 11 for anti-White to 85 for anti-Sikh. Most agencies in the country went through 2024 using none of them. They used the thirty-fifth.

Access Address-Level Crime Data

Real-time incidents · free A–F Crime Grade · rolling 90-day window · 22,000+ US cities. Normalized and verified — because raw data isn't enough.