SBA 7(a) FOIA Data
The U.S. Small Business Administration publishes its complete 7(a) loan portfolio under FOIA (Freedom of Information Act) requests. This dataset covers every SBA-guaranteed loan since fiscal year 1991 — including the borrower name, loan amount, approval date, lender, franchise code, and the ultimate disposition of the loan (paid in full, charged off, cancelled, or still active).
The SBA also publishes a Franchise Directory that maps franchise brands to unique identifier codes (format: S####). We use this directory to link FranScout franchise brands to their SBA loan records.
The dataset is updated quarterly, approximately one month after each quarter ends. Our current analysis uses data as of December 31, 2025.
2014–2023 approvals, 2-year buffer
We focus on loans approved between January 1, 2014 and December 31, 2023. This 10-year cohort gives us enough historical volume to compute statistically meaningful rates while keeping the data relevant to the current franchise landscape.
We apply a 2-year buffer — excluding loans approved after 2023. Loans made in 2024 and 2025 are still in early repayment, and charge-offs from these vintages are largely unresolved. Including them would artificially inflate repayment rates for newer loans and compress rates for established brands.
Pre-2014 loans are excluded because the franchise landscape and SBA program terms changed significantly over time, and including very old loans would dilute the signal from the current operating environment.
Charge-off rate → repayment rate
A loan is considered charged off when the SBA data has a non-empty ChargeOffDate field. We do not rely on the LoanStatus field alone because status codes can be ambiguous (e.g., “EXEMPT” for COVID/disaster loans may or may not reflect repayment outcomes).
Charge-off rate = charged-off loans ÷ total loans in cohort × 100
Repayment rate = 100 − charge-off rate
We intentionally use a simple charge-off rate rather than loss severity (dollar-weighted) because it reflects systemic risk at the borrower level — the probability that an SBA loan to a franchise borrower results in default, regardless of loan size.
25 loans = meaningful signal
A brand needs at least 25 SBA loans in the cohort before we compute or display a rate. Below that threshold, a single charge-off moves the rate by 4+ points, making it statistically unreliable.
Brands below 25 loans display as “Insufficient Data” (—) in the browse table and in the brand detail page. This is not a negative signal — many excellent franchise systems simply don't rely heavily on SBA financing.
Only 7.4% of SBA loans are franchise-coded
The SBA data has a well-known coverage gap: the FranchiseCode field is blank for approximately 92.6% of all 7(a) loans — including many loans that were in fact made to franchise borrowers. Lenders are not required to populate this field consistently.
This means our loan counts are undercounts for most brands. Jersey Mike's Subs shows 310 loans in our cohort — the actual number of SBA loans made to Jersey Mike's franchisees is certainly higher. We present the franchise-coded loans as a representative sample, not a complete census.
As a result, the SBA Repayment Rate should be read as a directional signal — a brand with a 0.6% charge-off rate over 310 coded loans is almost certainly a strong performer. But the absence of data, or a low loan count, does not imply the opposite.
Four tiers, one consistent rubric
Every brand with sufficient data is assigned one of four tiers based on its repayment rate:
Historical charge-off rate suggests the system works. SBA borrowers repaid at or above the top-tier benchmark.
Repayment rate is in line with the broader franchise universe. Not a warning sign, but not a standout either.
Charge-off rates are elevated relative to peers. This warrants closer scrutiny of the business model and unit economics before committing capital.
Too few SBA loans in the cohort to compute a meaningful rate, or we could not find a matching SBA franchise code. Not a red flag by itself.
How we connect brands to SBA codes
The SBA Franchise Directory uses its own naming conventions, which often differ from the brand names in our database. “Orange Theory Fitness” vs. “ORANGETHEORY® Fitness.” “SportClips” vs. “Sport Clips.” “McDonalds” vs. “McDonald's.”
We use a combination of exact matching (after normalization) and fuzzy matching to link each FranScout brand to its SBA franchise code. Fuzzy matches with a confidence score above 85% are accepted. The match type and confidence score are stored in the methodology notes for each brand's record.
Some brands have multiple SBA codes (e.g., co-branded locations, rebrands, regional variants). In these cases, we sum loan records across all applicable codes before computing the rate.