Market Analysis

The SBA Lender Report Card: Ranking the Most Active 7(a) Lenders by Charge-Off Rate, Volume and Loan Size

By David Jones, MAI, CCIM

Seventy-one institutions originated 500 or more SBA 7(a) loans between FY2021 and FY2025. Ranked by volume, the league table looks like the one the trade press publishes every October. Ranked by how those loans performed, it nearly inverts — and the variable that explains the inversion is not underwriting skill.

Dataset: the lender panel behind this analysis is available for download at the end of this article under an attribution licence.

The league table everyone publishes

Every October, the SBA’s lender activity report produces a familiar ranking, and the trade press reproduces it. Who wrote the most loans. Who wrote the most dollars. Institutions issue press releases about their placement, and the placement is real: volume is volume.

But a volume ranking answers a question no borrower is asking. It tells you which lender is busiest. It does not tell you whether the loans that lender wrote were good ones.

That question is answerable. The SBA’s own FOIA release carries the eventual status of every loan — paid in full, current, in liquidation, charged off — so it is possible to go back to a seasoned cohort, ask which institution originated each loan, and compute how that institution’s book actually performed.

This analysis does that. And the resulting order is close to the reverse of the volume table.

How this was measured

The performance cohort is FY2015–FY2019. Charge-offs take years to emerge. A lender’s FY2024 book will look immaculate because nothing has had time to fail. Using seasoned vintages is the only way to compare institutions honestly, and it is the single most common error in lender comparisons that circulate publicly.

The volume cohort is FY2021–FY2025. Current market position, five complete fiscal years.

Only 7(a) is included. The 504 programme uses a different status vocabulary and resolves defaults differently through the CDC structure; combining the two on a single charge-off measure produces a number that means nothing.

The floor is 500 loans for the volume table and 300 seasoned loans for the charge-off table. Below that, a handful of failures swings the rate by several points.

Cancelled approvals are excluded. Loans that never funded cannot default.

Lender names are as reported. Institutions were not consolidated across corporate families, subsidiaries or acquisitions, so a bank that bought an SBA platform mid-period may appear under two names. Concentration figures are therefore conservative — true group-level share is higher than reported here.

Charge-off rates among high-volume lenders

Sixty institutions cleared both thresholds. Their FY2015–FY2019 charge-off rates:

The national rate across all FY2015–FY2019 7(a) originations is 6.92%.

The lowest charge-off rates

The highest charge-off rates

Read the average loan column before drawing any conclusion about underwriting.

The variable that explains it

Every lender in the best-performing list except one has an average loan size above $399,000, and five are above $1 million. Every lender in the worst-performing list except two has an average below $700,000, and four are below $200,000.

This is not a coincidence and it is not primarily a story about credit skill. It is arithmetic about what small-dollar SBA lending is.

A $150,000 working-capital loan to a two-year-old business is a fundamentally different credit than a $1.4 million acquisition loan against an established operating company with a documented cash flow history. The small loan is cheaper to originate, faster to approve, frequently unsecured or thinly secured, and made to a borrower with less operating history. It fails more often. It was always going to fail more often.

Lenders that specialise in small-ticket SBA lending are not worse underwriters than lenders that specialise in seven-figure acquisitions. They are operating a different business, with different expected loss, priced and structured accordingly, and backed by a guarantee that exists precisely because that credit would not otherwise be extended.

The practical consequence is that a charge-off table cannot be read as a quality ranking, and any publication that presents it as one is misleading its readers. A lender at 15% is not necessarily doing a worse job than a lender at 2%. It is very likely doing a different job.

What the table does support is a narrower and more useful claim: it tells you what kind of book each institution is running, and therefore what kind of borrower it is set up to serve.

Where the table does carry a signal

Loan size explains most of the variation between lenders. It does not explain variation within a single lender over time. When one institution’s charge-off rate moves sharply across origination years while its peers hold flat, something changed inside that institution.

TD Bank is the clearest case in the data.

TD Bank began the period underwriting better than the market and ended it at more than three times the national rate, while the national rate itself was falling. The inflection lands between FY2016 and FY2017 and it is abrupt.

Compare Huntington across the same years — 6.65%, 6.67%, 5.70%, 5.22%, 4.60% — a steady improvement on the largest book in the country.

Two institutions, both high-volume, both small-ticket. One held its standard and improved. The other did not. That comparison is a quality signal, and it is the kind the volume league table can never show.

Market concentration

Two thousand and twenty-eight institutions originated at least one 7(a) loan in the period. Fifty of them wrote two-thirds of the volume.

The tail is longer than the concentration figures suggest and less meaningful than it looks. 950 lenders — 47% of all participating institutions — originated fewer than ten loans across five entire years, and together account for 1.2% of the market. These are banks with an SBA relationship on paper and effectively no SBA practice.

That distinction matters for borrowers. The presence of a participating lender in your county tells you very little. The relevant question is whether that lender writes SBA paper routinely or has written four loans since 2021.

What this means if you are a borrower

Match the lender to the size of your deal. The institutions with the lowest loss rates are the ones underwriting seven-figure acquisition credits, and they are set up for that borrower. If you need $180,000 of working capital, a lender whose average approval is $1.4 million is unlikely to be interested, and its low charge-off rate tells you nothing about your odds.

A high charge-off rate is not a warning about the lender. It usually means the lender serves borrowers like you. BayFirst and Northeast wrote thousands of small loans to businesses that would struggle to be financed anywhere else. That is the programme working as designed.

Ask how many loans the lender wrote last year. Not whether it is an SBA lender — nearly a thousand institutions technically are, while doing almost nothing. Volume is a proxy for whether the file will move.

Trend beats level. If you can see a lender’s performance moving in one direction across several years, that is worth more than its position in any single-year table.

What this means for lenders

For institutions, the useful frame is peer comparison within a size band rather than against the market. A lender averaging $150,000 approvals should be benchmarked against others in the same band, where the range still runs from roughly 8% to over 20% — a real spread, and one that reflects underwriting rather than product mix.

The secondary-market implication is worth noting. Guaranteed portions from small-ticket originators carry materially different loss experience from large-ticket paper, and cohort-level charge-off data by originator is public. Pricing that ignores it is pricing on incomplete information.

Limitations

Charge-off is not loss. The FOIA extract records charged-off amounts but not recoveries, so this measures write-off frequency and not net loss to the SBA or to the lender.

Names are unconsolidated. Mergers, acquisitions of SBA platforms and rebrandings mean some institutions appear under multiple names, and a book originated by one entity may be reported under its acquirer. Group-level concentration is higher than shown.

Cohort choice affects levels. FY2015–FY2019 originations were exposed to the pandemic during their term. Every lender in the comparison was exposed to it equally, so the relative ranking holds, but the absolute rates are higher than a non-pandemic cohort would produce.

Servicing transfers are invisible. The lender field reflects assignment at the time of the data snapshot. A loan sold or transferred may be attributed to an institution that did not originate it.

This measures outcomes, not decisions. A charge-off rate reflects the borrowers a lender chose to serve, the products it offered, the collateral it took and the economy those borrowers operated in. It does not isolate underwriting quality, and this analysis does not claim to.

Frequently asked questions