Citation, reuse and doing this yourself

1. How to cite this

Quote the figure, name the page, and give the as-of date, because the underlying file is restated quarterly and a figure without a date cannot be checked. For example:

LMMMA, "Acquisitions by sector", https://lmmma.com/sectors, data as of 30 June 2026.

If you are quoting a headline figure, the honest form includes the qualifier. It is 40,612 SBA-financed business acquisitions, not 40,612 business acquisitions. The distinction is the difference between a true statement and a badly wrong one, for reasons set out in the traps below.

2. What you may reuse

The underlying SBA data is a work of the United States government and carries no copyright. You can do whatever you like with it.

The aggregations, the written material and the design on this site are ours. You may quote figures freely, including commercially, with attribution and a link. You may not republish the site in substance or build a service that substitutes for it. The terms set this out in full.

Journalists, researchers and students are welcome to use anything here. If a figure needs checking before publication, ask and we will show the working.

3. Doing it yourself

This is deliberately reproducible. Download the two 7(a) FOIA extracts from data.sba.gov/dataset/7a-504-foia. The files used here were stamped as of 30 June 2026 and retrieved 4 September 2026; the SBA rotates the filenames each quarter and stamps each with its own date.

Then, in whatever tool you prefer:

• Keep only rows where BusinessAge is exactly Change of Ownership.
• Keep only rows where ApprovalFY is 2019 or later.
• Keep only rows where LoanStatus, with spaces removed, is PIF, CHGOFF or EXEMPT.

That should leave 40,612 rows. If it does, the rest is arithmetic, and the methodology page defines every measure precisely enough to reproduce. If it does not, one of the traps below is probably why.

4. Four ways to get this wrong

Starting the series before FY2019. The most damaging error available with this file. The SBA did not use the change-of-ownership code before FY2018, so FY2010 through FY2017 contain exactly zero of them. A chart from FY2010 shows acquisitions exploding from nothing, which is a change in a form field, not in the market. FY2018 is a partial adoption year and is also unusable.

Counting approvals instead of loans. 3,965 approvals were cancelled before disbursement and 2,119 are approved but not yet funded. Counting all approvals inflates the total by roughly a seventh and, worse, inflates recent years more than old ones because the pending ones are all recent.

Comparing charge-off rates across approval years. Loans approved in FY2024 have barely had time to fail. A chart of charge-off rate by approval year slopes downward for no reason but the calendar. Fix a cohort, state the as-of date, and treat the result as a floor.

Calling loan amounts prices. A buyer combines the loan with their own equity and often with a note from the seller. The SBA records only its loan. Presenting the median loan as the median sale price understates what these businesses sold for by an unknown amount.

A fifth, less dangerous but worth stating: FY2026 is a partial year. The extract closes 30 June 2026 and the fiscal year ends 30 September, so the final column is never a decline.

5. How often this updates

The SBA restates these extracts quarterly. This site regenerates from the newest file and every page carries the as-of date it was built from. Because restatements can move historical figures slightly, a figure quoted from here should always carry its date.

The site currently publishes 19 sector, 147 industry group, 52 state, 315 county, 151 lender and 328 state-by-sector pages. A refresh can move a group across a reporting floor in either direction, so the set of pages is not fixed.

6. Asking for something else

If there is a cut of this data that would be useful and is not published, say so. The aggregation code can produce most things the file supports. Two constraints: nothing will be published at a grain that identifies an individual borrower, and nothing will be published where the row count is too small to be meaningful.

Ask through the contact form.