Census Business Formation Statistics (BFS)

Mapping & Data Analytics Team 2026-09-02

Overview

The Business Formation Statistics (BFS) track applications for Employer Identification Numbers (EINs) on IRS Form SS-4 as a leading indicator of entrepreneurial intent. An EIN application is a signal of business initiation; it is not proof that a new employer business was created. BFS captures the earliest observable administrative step in the business formation process, before applications become establishments, before establishments create jobs, and before those jobs show up in employment and wage data.

The program’s primary county-level series is the Business Application Series (BA), which captures a broad subset of EIN applications associated with business initiation — excluding tax liens, estates, trusts, certain financial filings, and applications in industries with low transition rates to employer businesses. At the national and state level, BFS also publishes a High-Propensity Business Applications (HBA) series — restricted to applications more likely to become employer businesses — and Projected Business Formations (PBF), which estimates conversion rates within four and eight quarters. These finer series are not available at the county level.

What this data tells us about rural America

Business formation statistics — tracking EIN applications as a leading indicator of entrepreneurial intent — provide an early signal of economic momentum that establishment counts and employment data can only confirm after the fact. For rural communities, this leading-indicator quality is particularly valuable: because business applications occur before employer business formation and employment, changes in application activity can provide earlier visibility into changing entrepreneurial conditions than establishment and employment measures alone.

The pandemic-era surge in rural business applications also illustrates why formation activity needs to be evaluated alongside subsequent business formation and survival. A rise in applications does not necessarily translate into durable establishment growth. Comparing BFS with BDS can help identify where strong application activity was followed by actual business formation and job creation — and where it was not.

NOTE

Coupling BFS with other datasets is a primary step for its use: BFS shows where people are initiating businesses while Business Dymanics Statistics reveals whether those applications become employer establishments. Other datasets, such as the Quarterly Census of Employment and Wages provides the critical employment and wage picture. Together, these data describe the full progression of rural economic development, from entrepreneurial intent to labor market outcome.

Coverage

• Geography: County, State, National

• Time range: Monthly data begin July 2004; annual county Business Application data begin 2005

• Update frequency:

  • Monthly national and state data
  • Annual county data (typically released approximately six months after the reference year)

• Universe: EIN applications on IRS Form SS-4, excluding tax liens, estates, trusts, agriculture (NAICS 11), and public administration (NAICS 92)

NOTE

Only Business Applications are available for County data: The Census county-level file contains the Business Application (BA) series only. High-propensity (HBA) and Projected Business Formation (PBF) series, which estimate how many applications become employer businesses, are available at the national and state level but not at the county level. County-level BFS tells us where people are applying to start businesses. It does not provide county-level estimates of which applications ultimately become employer businesses.

NOTE

Mind the noise: County-level BFS application counts are subject to disclosure-avoidance procedures including differential privacy, with geometric noise added to counts and subsequent calibration to state totals. Small year-to-year changes in county application counts should be interpreted with caution, particularly in very small counties. See the BFS methodology page for details on the disclosure avoidance approach applied to county data.

Key variables

Counts and rates answer different questions. Counts show the scale of business application activity in a county; rates allow comparison of entrepreneurial activity across communities of very different sizes.

• Total business applications (BA) — count of EIN applications meeting BFS coverage criteria; the primary county-level measure of business initiation activity

• Applications per 1,000 population — population-normalized rate for comparing application intensity across counties

Accessing the data

Via CORI R package

The cori.data.bfs package provides tidy annual county-level access to BFS business application counts, with population normalization via cori.data.pep.

library(cori.data.bfs)

# All counties, all years
get_business_applications(geography = "county")

# Specific counties, recent years
get_business_applications(
  geoids = c("33009", "33011"),
  years  = 2010:2024
)

Direct from source

Monthly national and state data, and annual county files, are available from the BFS data page. See the BFS methodology page for documentation of the EIN application filtering process, the high-propensity classification methodology, and the projected formations estimation approach.

Related

• Package: cori.data.bfs — CORI’s tidy county-level access layer
• Package: cori.data — meta-package including cori.data.bfs
• Dataset: Census Business Dynamics Statistics — provides observed establishment births, deaths, and job flows; pairing BFS with BDS helps assess whether increases in application activity are followed by actual employer business formation and job creation
• Dataset: QCEW Employment & Wages — provides current employment and wage levels, allowing researchers to connect earlier application trends to changes in the local labor market
• Project: Rural Economic Outlook — BFS is an input to CORI’s multi-dataset economic analysis