How to Use ATF FFL Data for Insurance Underwriting

Federal Firearms License (FFL) data is one of the cleanest underwriting signals available for any insurance product touching the firearms industry. The ATF publishes monthly snapshots of every active FFL holder in the United States, and the file structure has not changed in years. For underwriters writing general liability, product liability, commercial property, or specialty crime coverage on dealers, manufacturers, importers, gunsmiths, and ranges, this is a free dataset that most carriers underuse.

Here is how to actually pull underwriting signal out of it.

What ATF Publishes

The ATF Federal Firearms Licensee (FFL) listing is released monthly as a public file. It includes every active license, the holder's business name, premises address, license type (01 through 11), expiration date, and the date the license was first issued.

The official source is the ATF FFL eZ Check and FFL Listings page at atf.gov. The file is free, comes out around the first of each month, and is structured consistently enough to be ingested directly into a pricing model.

What it does not include: revenue, employee count, claims history, or NAICS code. You bring those in via match. The FFL file is the spine; the rest is enrichment.

License Type Tells You the Hazard Class

The single most important field for underwriting is license type. The ATF defines eleven types, but five matter for most commercial policies:

• Type 01 — Dealer in firearms other than destructive devices. The largest category. Includes brick-and-mortar gun shops and many home-based dealers.
• Type 02 — Pawnbroker dealing in firearms. Higher loss ratio historically because of inventory turnover and theft exposure.
• Type 07 — Manufacturer of firearms other than destructive devices. Often includes Type 01 activity by default and carries product liability exposure.
• Type 08 — Importer of firearms. Smaller population, customs and chain-of-custody exposure.
• Type 09, 10, 11 — Destructive devices and ammunition for destructive devices. Specialty risk, almost always underwritten separately.

A carrier writing a base general liability program for Type 01 dealers should price differently than one writing Type 07 manufacturers, and license type alone explains a large share of that variance. Pulling the ATF file gives you the population denominator for any segment you are pricing.

Tenure Is a Better Predictor Than Most People Realize

The first issue date on the license is underused. A FFL that has been continuously renewed for fifteen years has a very different loss profile than one issued six months ago. Established dealers have established compliance practices, established record-keeping (the bound book), and established relationships with their ATF Industry Operations Investigator.

For underwriting, the rule of thumb that survives empirical testing:

• Tenure under 2 years — higher claims frequency, less compliance maturity
• Tenure 2 to 5 years — improving but still elevated
• Tenure 5 to 10 years — stabilized, near population average
• Tenure 10+ years — best-in-class loss experience

This signal is free, sitting in plain text in every monthly file. Most carriers either ignore it or only use it for new business at quote, not for portfolio segmentation.

ZIP-Level Density Predicts Theft Exposure

The third signal is geographic density. Aggregate active FFLs by ZIP code or county. High-density ZIPs (urban areas with many dealers) and very-low-density ZIPs (rural single-dealer markets) both carry elevated risk, for different reasons.

Urban high-density areas show elevated burglary and inventory-loss frequency. Rural low-density areas show elevated severity because a single dealer often holds significantly larger inventory relative to local norms and is the only target within range.

Cross-referencing the FFL density map against FBI Uniform Crime Reporting data at the agency level (UCR is published at fbi.gov) lets you build a county-level theft hazard score that beats most off-the-shelf commercial scores for this class of business.

Combining FFL Data with NAICS and Claims Data

The ATF file does not include NAICS. You match the business name and premises address against open business registry data to attach NAICS 451110 (Sporting Goods Stores), 332994 (Small Arms, Ordnance, and Accessories Manufacturing), or 423920 (Toy and Hobby Goods and Supplies Merchant Wholesalers).

Once NAICS is attached, you can pull aggregate loss ratios from NCCI or your own portfolio for the same NAICS segment and use FFL-specific signals (type, tenure, density) as relativities against that baseline.

For carriers without sufficient internal data volume, ISO and Verisk publish loss costs by class code that can serve as the baseline against which FFL relativities are applied. ISO advisory loss costs are available through their standard licensing.

What Most Carriers Get Wrong

Three common mistakes show up in firearms-industry underwriting:

Treating all FFLs the same. A Type 01 dealer operating from a 3,000 sq ft retail storefront in Houston is not the same risk as a Type 03 Collector of Curios and Relics operating from a basement in Vermont. The license type field is right there in the file.

Ignoring renewal patterns. FFLs renew every three years. If a dealer has skipped a renewal cycle in the past, the gap is visible in the data once you stitch monthly snapshots together. Gaps correlate with compliance issues.

Using stale data. Carriers that pull the ATF file once a year miss exits, new entrants, and address changes. The file is monthly. Use it monthly.

A Practical Workflow

For a carrier or MGA underwriting firearms-industry GL:

1. Ingest the most recent ATF FFL listing.
2. Stitch the last 36 monthly snapshots together to compute continuous tenure and detect renewal gaps.
3. Aggregate by ZIP and county for density signal.
4. Match against business registry data for NAICS attachment.
5. Cross-reference against FBI UCR for local theft hazard.
6. Build relativities for license type, tenure band, density quintile, and crime score against your portfolio baseline.

This is not exotic data science. It is straightforward joins on public data, run monthly, that most carriers in this segment do not bother to run.

What This Looks Like for a Small MGA or Captive

You do not need a data engineering team. A single analyst with SQL can stand up the full pipeline in a week. The monthly refresh runs as a scheduled job. Once it is built, the marginal cost of running it forever is essentially zero.

If you are evaluating whether to license commercial firearms-industry data from a broker or build this yourself, the build path is realistic for any team with one technical resource. The ATF data is free and the rest of the joins are against data you almost certainly already license.

For other public-data underwriting plays, our writeup on detecting fraud signals in public data covers a similar approach applied to different verticals. For broader data product strategy, see the Palavir data products catalog.

Get the Data

If you want a clean, deduplicated, ready-to-load version of the April 2026 ATF FFL snapshot with all 77,000+ active licenses, license types decoded, and tenure pre-calculated, we publish it as a free Kaggle dataset.

Pull the file here: US FFL License Directory — April 2026 on Kaggle. Load it into your underwriting environment, run the playbook above, and you will see signal within a week.