The NCCS Core Data Series is a panel of nonprofit organizations derived from annual IRS Form 990 filings. It is the flagship product for tracking nonprofit financials, governance, and program activity over time, and serves as a primary research dataset alongside the Business Master File.

Need to download specific files? Browse the Core Data Catalog →

Looking for blank IRS forms? Blank IRS Forms Archive → — every year of Form 990 / 990-EZ / 990-PF, their schedules, and IRS instruction booklets.

What's in (and what's missing from) the Core Series

The Core Series harmonizes annual Form 990 filings into year-by-form-type files with stable variable names. Each year/file pools 501(c)(3) charities and all other 501(c) types into a single product. Use the is_501c3 column (boolean, TRUE for 501(c)(3) public charities) to filter when needed; the raw subsection_cd is also available for finer-grained 501(c) subclassification.

Files are organized by filing form rather than organization type. The four products are:

• 990 — full-form filers (the larger nonprofits, with the richest variable set)
• 990EZ — short-form filers (mid-sized nonprofits, fewer disclosed variables)
• 990 Combined — 990 and 990EZ filers stacked together using the shared variable subset
• 990PF — private foundations (a separate filing form with its own variables)

The Core Series does not include 990N ePostcard filers — the smallest nonprofits (gross receipts < $50k) that file a minimal information return. These organizations make up a large share of the registered nonprofit population but disclose almost no financial or programmatic detail. If they are relevant to your research, consult the 990N Postcard dataset separately.

Other scope notes:

• The is_501c3 column (boolean) distinguishes 501(c)(3) public charities from all other 501(c) types within each file; filter on this rather than expecting separate files. subsection_cd is also available for finer 501(c) subclass breakdowns.
• Data dictionaries are per year, per form type. The variable set evolves with the underlying 990 form, so dictionary files are not interchangeable across years.
• Variables in the Core Series are financial and governance fields disclosed on the 990 itself. For organizational attributes like NTEE code, 501c type, or address, merge against the Business Master File using EIN.

Which dataset should I use?

Each catalog section exposes the underlying S3 prefix for programmatic access and links per-year dictionaries and quality reports.

Documentation & support

Primary sources

• IRS Form 990 series — the upstream IRS filing forms (990, 990EZ, 990PF) and their published instructions.
• Nonprofit Open Data Collective concordance file — the variable-name harmonization mapping that underlies Core Series naming.

NCCS resources

• Core Data Catalog — every published file with downloads, dictionaries, and quality reports.
• Blank IRS Forms Archive — every year of blank Form 990 / 990-EZ / 990-PF and their schedules, plus IRS instruction booklets.
• CORE Pipeline Guide — architecture, harmonization logic, known upstream-IRS data quirks, and per-year/per-form quality reports.
• Business Master File — merge organizational attributes (NTEE, 501c type, address, lat/lon) onto Core rows via EIN.
• Census Crosswalk — aggregate Core data to tracts, counties, or metro areas using standard FIPS codes.
• Sector in Brief dashboard — request full panels with BMF attributes pre-joined.
• Contact / Get Help — questions, corrections, data requests.