County Business Patterns covers every county in the United States. All of them. A county of 4,000 residents appears in the same federal dataset as Los Angeles, updated on the same annual schedule, using the same methodology. Small county economic data gets dismissed because of one recurring frustration: a leader pulls up their county’s industry breakdown and finds letter codes where the employment numbers should be. That looks like the data isn’t there. It is there. The letters are telling you something specific, and once you know how to read them, small county economic data supports grant applications and state advocacy as well as it does anywhere else.
What Those Letter Codes Actually Mean
The Census Bureau protects individual businesses from being identified in published statistics. In a county where one employer accounts for most of an industry’s workforce, publishing exact employment would effectively publish that company’s payroll. The Bureau handles this in two ways, and the CBP methodology documentation explains both.
Since the reference year 2007, most cells receive noise infusion. A random multiplier perturbs each establishment’s data before the table is built. Every published value carries a flag showing how much distortion resulted:
- G (low noise): the value changed by less than 2 percent
- H (moderate noise): the value changed by at least 2 percent but less than 5 percent
- J (high noise): the value changed by 5 percent or more
A G or H flag means you have a usable number. Treat it as a close estimate rather than an exact count, and it will support any analysis you’d run on unflagged data.
Some cells are withheld entirely:
- S: withheld because the data do not meet publication standards. The underlying activity is still included in broader industry totals.
Older files use D for data withheld to avoid disclosing individual businesses. The Bureau retired that flag starting with reference year 2017, so expect the conventions to differ if you compare across years. There is also a hard floor. Beginning with reference year 2017, a cell is published only if it contains three or more establishments. Below that, the cell is dropped from the release entirely rather than flagged. If an industry you know exists in your county is missing from the table altogether, this threshold is usually why.
One caveat on all of the above. As of mid-2026, the Census Bureau has flagged its published CBP disclosure methodology as no longer current, and is evaluating alternatives following a Department of Commerce order prohibiting noise infusion. The flags described here apply to every CBP file currently in circulation. Expect the conventions to change in upcoming releases.
The Workaround Most Counties Miss
Even when employment for an industry is withheld, county files publish establishment counts by employment size class. Ranges start at 1-4 and 5-9 and rise through 1,000 or more, with additional higher brackets published at the county level. You can reconstruct a defensible employment estimate from them. Take the midpoint of each range, multiply by the number of establishments in that range, and sum across ranges. For the top range, analysts commonly assign a fixed value rather than an open-ended one. The result is less precise than a published employment figure. It also survives most disclosure rules, since county files publish establishment counts by size class wherever the industry cell itself appears. Industries with fewer than three establishments are dropped from publication entirely, so your smallest sectors may not appear at all. State the method in a footnote whenever you use it. Reviewers respect a transparent estimate far more than a number with no stated provenance.
What CBP Leaves Out
Before you build a county profile on CBP, know what is not in it. The series excludes crop and animal production (NAICS 111 and 112) entirely, along with rail transportation, postal service, private households, and public administration. It also excludes the self-employed and most government employees. For an agricultural county, that is a significant share of local activity sitting outside the dataset. CBP will tell you about the grain elevator, the equipment dealer, and the food processor. It will not tell you about the farms. Pair it with USDA Census of Agriculture data when farming is central to your economy, and note the exclusion whenever you present a CBP-based profile to a state agency or grant reviewer.
Where Small County Economic Data Delivers the Most Value
Grant Applications
Federal and state grant programs routinely ask applicants to characterize their local economy. Establishment counts by industry answer that question directly and come from a federal source the reviewer already trusts. A rural county applying for workforce funding can show how many employers operate in a target sector and what size those employers are. That’s a stronger foundation than a narrative description.
State Legislative Advocacy
When a county argues for a funding formula change or a program carve-out, CBP data lets that county show how its economic composition differs from the state average. Comparing your county’s sector mix against a statewide breakdown is straightforward once you pull both.
Economic Development Targeting
Establishment growth over time reveals which sectors are adding employers locally. A county tracking two consecutive years of CBP data can identify where new business formation is happening and direct recruitment effort accordingly.
When to Reach for QCEW Instead
The Bureau of Labor Statistics publishes the Quarterly Census of Employment and Wages, built from unemployment insurance records rather than the Census business register. It applies its own disclosure rules, so what is available for your county may differ between the two sources. Checking both takes a few minutes and occasionally fills a gap. QCEW also reports quarterly, so it detects a plant closure or a large hire months before the annual CBP release does.
The methodologies differ, so avoid mixing figures from both sources inside a single table. Pick the source that answers your question and cite it clearly. Both sit alongside several other federal data tools worth knowing. For a fuller picture of how CBP fits among them, our guide to County Business Patterns walks through the core metrics and how to access them.
A Note on the Reference Period
CBP employment reflects the week of March 12 in the reference year. Payroll appears in two forms: first-quarter payroll and annual payroll. For counties with pronounced seasonal economies such as agriculture or tourism, the March snapshot may sit at a seasonal low. That’s a limitation to disclose, not a reason to skip the data. Note the reference period when you present the numbers, and pair CBP with QCEW’s quarterly series if seasonality is central to the story you’re telling.
Frequently Asked Questions
Does County Business Patterns cover every county?
Yes. CBP publishes data for every county in the United States regardless of population, along with Puerto Rico and the Island Areas. Louisiana parishes, Alaska boroughs and census areas, and Connecticut planning regions are treated as county equivalents.
What does an “S” flag mean in CBP data?
The value was withheld because the data do not meet publication standards. The underlying activity is still included in broader industry totals. Files from before reference year 2017 use a “D” flag for a similar purpose.
Can I estimate employment when the figure is suppressed?
Yes. Use the published establishment counts by employment size class, apply the midpoint of each range, and sum. Disclose the method whenever you publish the result.
Is CBP data free?
Yes. All CBP data is available at no cost through data.census.gov. The Census Bureau’s API also requires no payment, though it does require a free API key.
The Takeaway
Small county economic data carries the same federal authority as data from the largest counties in the country. Suppression flags reflect a privacy rule, not a gap in coverage, and the size-class workaround fills most of what the flags hide. Pull your county’s most recent CBP file this quarter and read the flags before you read the numbers. Understanding what each code is telling you turns a table that looks incomplete into a usable economic profile.