The SBA Lending Map: Why Your County Determines Your Odds of Getting a Government-Backed Loan

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

Introduction: the question nobody has measured

The Small Business Administration does not lend money. It guarantees loans that private institutions choose to originate. That distinction is stated in every introductory guide to the 7(a) and 504 programs, and it is almost always treated as a technicality — a footnote about mechanism rather than a fact with consequences.

It has consequences. If the government guarantees credit but private lenders decide where to deploy it, then access to a federally backed small business loan is not distributed according to where businesses are. It is distributed according to where lenders are, and more precisely according to where lenders have chosen to build an SBA origination capability. A guarantee that nobody is willing to underwrite in your county is not access.

That proposition is not new. The academic and policy literature on small business credit has documented the relationship between physical bank presence and local lending for two decades. The National Community Reinvestment Coalition found an exceptionally strong statistical association between the number of bank branches in a county and small business lending volume, strongest of all for banks with a branch in the same county where the loan was made (1). Research on branch closures following bank mergers has found that local lending falls by roughly a tenth, with the effect concentrated in small business loans, which depend more than any other category on soft information and relationship banking (2). The Federal Reserve’s own reporting to Congress has long framed the Community Reinvestment Act’s assessment areas around the geography of physical branch presence (3).

What has not been done is the measurement inside the SBA programs themselves, at county level, against the business base.

The SBA publishes county-level approval counts in its State and County Report. Those are raw totals, and raw totals tell you almost nothing: Los Angeles County has more SBA loans than Wyoming because Los Angeles County has more businesses than Wyoming. LendingTree has published a state-level analysis dividing SBA lending across the business population, which is the right method at the wrong resolution (4). The SBA’s Office of Advocacy maps small business lending per employer firm by county, but tabulates it from Community Reinvestment Act filings — a different universe of lending entirely, covering conventional bank credit rather than the guaranteed programs (5).

The gap is a county-level penetration measure for the guaranteed programs specifically. This analysis fills it.

How this was measured

The loan spine is the SBA’s own Freedom of Information Act release of 7(a) and 504 loan-level records, as of 31 March 2026 (6). The full extract contains 1,036,077 records from FY2010 forward.

Three decisions shape everything that follows, and each is stated here rather than buried.

Cancelled and never-funded approvals are excluded. Of the 1,036,077 records, 134,970 carry a status of cancelled or not funded. These are approvals that never became loans. Including them would inflate every geographic count by approximately 13% and would credit counties with credit they never received. The working set is 901,107 originated loans.

The window is FY2021 through FY2025. Five complete fiscal years. The denominator — County Business Patterns establishment counts — is a 2023 reference year, so the loan window is centred on the denominator rather than trailing it. FY2026 is excluded entirely because the FOIA extract covers only October through March, half a fiscal year, and a partial year silently depresses any rate calculated from it.

The denominator is businesses, not people. The SBA guarantees loans to firms. Measuring lending per resident produces a number that moves with demographics rather than with credit access. Establishment counts come from the Census Bureau’s County Business Patterns series for 2023, the most recent reference year available (7). Population figures, used only for context, come from the Vintage 2025 Population Estimates released in March 2026 (8).

A note on the denominator’s limitation, because it matters and it cuts one way. County Business Patterns counts establishments with paid employees. It excludes sole proprietors and other non-employer firms, which are a real share of small 7(a) borrowers. The effect is to understate the business base in counties with a high non-employer share — disproportionately rural and tourism-dependent counties — and therefore to overstate their lending penetration. Where this analysis finds rural counties underserved, the true gap is likely wider than reported, not narrower.

The national baseline

Between FY2021 and FY2025, lenders originated 293,896 SBA 7(a) and 504 loans in counties with a measurable business base, totalling $166.3 billion in gross approvals. Against 8,321,112 establishments, that is 35.3 loans per 1,000 businesses across five years — approximately one business in 28 receiving a government-guaranteed loan over the period, or one in 140 in any given year.

The 504 programme accounts for 9.7% of loan count. Borrowers reported 3,043,597 jobs created or retained, a self-reported figure the SBA explicitly does not audit and which should be read as a directional claim rather than a measurement.

That national average is the least interesting number in this study. It describes almost no actual county.

The distribution: a 5.6-fold spread

Restricting to the 2,445 counties with at least 200 establishments — a floor that removes the statistical noise of counties where three loans would double the rate, while still covering 99.5% of all lending — the distribution is as follows.

The ratio between the 90th and 10th percentile is 5.6 to 1. A business in a top-decile county is more than five times as likely to be operating in a place where SBA credit is actively deployed as an otherwise identical business in a bottom-decile county.

Sixty-seven counties run at more than twice the national rate. Seven hundred and sixty-one run at less than half of it.

The extremes are wider still. Covington County, Mississippi — population 18,041, with 328 establishments — recorded 96 SBA loans in five years. That is 292.7 loans per 1,000 businesses, more than eight times the national rate, and it implies that approximately one business in three and a half in Covington County took an SBA-guaranteed loan during the period.

The highest-penetration counties

Look at the state column. Eight of the top twenty-five counties are in Ohio. That is not a coincidence, and the explanation is the substance of this article.

The 240 counties with no SBA lending at all

Two hundred and forty US counties — 7.6% of the 3,144 in the Census universe — recorded not a single SBA 7(a) or 504 loan between FY2021 and FY2025. Not a low number. Zero.

Those counties contain 25,444 business establishments and 1.28 million residents. Their median population is 3,934, so most are genuinely small, and in a county with forty businesses a five-year run of zero loans is unremarkable arithmetic rather than evidence of anything.

But not all of them are small. Iroquois County, Illinois has 625 establishments and 26,135 residents, and received nothing. Cass County, Iowa: 444 establishments, nothing. Harlan County, Kentucky: 370 establishments, 25,306 residents, nothing. Concordia Parish, Louisiana: 363 establishments, nothing. Hampton County, South Carolina: 318 establishments, nothing.

At the national rate, a county with 625 businesses would expect roughly 22 SBA loans over five years. Iroquois County received none. That is not a rounding error in a small sample; it is the absence of a lending relationship.

The geographic clustering is pronounced:

Kentucky is the case that resists the small-county explanation. Eighteen Kentucky counties recorded zero SBA lending across five years, and several of them — Harlan, Hart, Larue, Monroe, Webster, Casey — have between 200 and 370 establishments apiece. Kentucky as a whole ranks 44th of 51 jurisdictions at 24.0 loans per 1,000 businesses, roughly two-thirds of the national rate. The state does not lack businesses. It lacks lenders originating SBA paper in its rural counties.

Size matters, but far less than expected

The intuitive explanation for uneven SBA access is urban concentration: big metros have banks, rural counties do not. The data support that story, but only weakly.

The gradient runs the expected direction. The smallest counties receive SBA credit at 19.8 loans per 1,000 businesses against 37.6 for the largest — a ratio of 1.9 to 1.

Set that against the 5.6 to 1 spread across the full distribution. County size explains perhaps a third of the observed variation. The majority of the gap is something else, and it operates within size bands: Ashtabula County, Ohio (96,845 residents) and Harlan County, Kentucky (25,306 residents) are both non-metropolitan counties in economically pressured regions of the eastern United States. Ashtabula received 199 SBA loans. Harlan received none.

The one place size dominates absolutely is the zero column. Of the 740 counties under 10,000 residents, 206 recorded no SBA lending — 28% of them. Below a certain scale, the fixed cost of an SBA origination relationship appears simply not to clear.

The state league table

State-level penetration ranges from 65.3 loans per 1,000 businesses in New Hampshire to 19.4 in Louisiana — a 3.4-fold spread between the best and worst-served states.

Top ten states by SBA penetration, FY2021–FY2025

Bottom ten states

California, with the largest business base in the country at 1,028,026 establishments, sits at 35.2 — almost exactly the national rate. Texas sits at 32.1 and carries 30 zero-lending counties, more than any other state. Neither of the two largest state economies is remarkable on this measure. Size of economy does not predict penetration.

Note the average loan column. The high-penetration states cluster toward smaller average loans — Ohio at $331,119, New Hampshire at $325,170, against a national average of $523,606. The low-penetration states cluster toward larger ones — Tennessee at $655,402, Alabama at $678,970, Louisiana at $660,731.

That pattern is the tell. States with high SBA penetration are not places where borrowers need more capital. They are places where lenders are writing a high volume of comparatively small loans.

The explanation: SBA access is a lender-coverage phenomenon

Ohio ranks second in the nation at 65.1 loans per 1,000 businesses, 84% above the national rate, with zero counties recording no SBA lending across all 88 of them.

Ohio is not a state with an unusual density of entrepreneurs, an unusual concentration of SBA-eligible industries, or an unusually favourable regulatory environment. What Ohio has is Huntington.

Between FY2021 and FY2025, The Huntington National Bank originated 62.4% of every SBA 7(a) loan approved in the state of Ohio — 9,932 of 15,924. No other lender reached 4%.

Nationally, Huntington originated 33,667 7(a) loans over the same period. That is 12.6% of all SBA 7(a) lending in the United States, from a single institution, in a market with 2,028 distinct participating lenders. The bank has publicly announced its position as the country's largest 7(a) originator by loan count for seven consecutive years and identifies itself as the leading originator in Colorado, Illinois, Indiana, Kentucky, Michigan, Ohio, Pennsylvania, Texas and West Virginia (9). In this dataset it is the top lender by count in eleven states.

The mechanism is visible in the average loan size. Huntington's average 7(a) approval is $254,027 against a national average of $523,606 — less than half. This is a high-volume, small-ticket origination model, and it is exactly the model that produces high penetration rates: many businesses reached, each with a modest loan.

Now look back at the state table. Six of the top ten penetration states — Ohio, Michigan, Colorado, and by extension the Huntington footprint states of Indiana, Illinois and Pennsylvania in the upper half — overlap with the bank's stated origination territory. The correlation between one institution's branch map and the national geography of SBA access is not subtle.

Concentration at county level

The national market does not look concentrated. The top lender holds 12.6% share, the top five hold 27.9%, the top ten hold 41.5%, and 2,028 institutions participate.

Top ten 7(a) lenders by loan count, FY2021–FY2025

National concentration is the wrong frame. Borrowers do not shop a national market; they shop whoever will take their call locally.

Across the 1,133 counties with at least 25 SBA 7(a) loans in the period, the median county's largest lender holds 22.2% of local originations — a competitive market by any standard. But the distribution has a long tail:

In 166 counties (14.7%), one lender originated more than half of all SBA 7(a) loans.

In 31 counties (2.7%), one lender originated more than three-quarters.

And single-lender dominance is precisely what characterises the highest-penetration counties:

Every county in the top fifteen by penetration is dominated by a single institution, and in most cases the top three lenders account for 80–97% of all local SBA lending.

This is the finding. High SBA penetration is not a signal of a competitive local credit market. It is the signature of one committed originator. Covington County, Mississippi does not have eight times the national rate of SBA lending because its businesses are eight times as creditworthy or eight times as ambitious. It has eight times the national rate because Peoples Bank decided to run an SBA desk and worked its territory.

The corollary is the uncomfortable part. If a county's SBA access depends on one institution choosing to serve it, then that access is contingent, and it can be withdrawn by a single credit-committee decision, a merger, or a change of strategy — none of which have anything to do with the businesses in that county.

What this means if you are a borrower

The practical translation is uncomfortable but useful.

Your county's historical SBA volume is information about your odds. Not because the SBA applies a geographic test — it does not — but because the presence of an active local originator determines whether there is anyone nearby who processes these loans routinely, knows the paperwork, and has a credit committee that has seen your industry before. In a county running at 9 loans per 1,000 businesses, you are likely to be the first SBA file that branch has handled in months.

If your county is thin, look outward rather than giving up. Roughly two-thirds of the national market is held by lenders outside the top ten, and a significant number of the largest originators — Live Oak, Readycap, Newtek, BayFirst — operate on a national rather than a branch-footprint basis. Their average loan sizes are larger, which tells you something about the deal sizes they prefer, but they are not constrained by geography in the way a regional bank is.

Expect a different loan profile in a high-penetration county. The states and counties with the highest penetration are dominated by small-ticket lenders. If your project requires $2 million, the local high-volume originator whose average approval is $254,000 may not be the right counterparty regardless of how much SBA paper it writes.

Lender selection is a bigger variable than most borrowers assume. The data show a market in which the identity of the institution changes not only your pricing but your basic probability of finding anyone to underwrite the deal. That is worth as much diligence as the property or the business itself.

What this means for lenders and policymakers

For lenders, the map is a market-entry document. Two hundred and forty counties with 25,444 businesses have no incumbent SBA relationship at all, and 761 counties run at less than half the national penetration rate. In a programme where the government guarantees 75–85% of the exposure, the absence of an originator in a county with several hundred businesses is a distribution gap rather than a credit judgment.

For policymakers, the finding sharpens a long-running debate. The SBA has expanded its Small Business Lending Company licensing specifically to widen the distribution network into underserved markets, and has framed the objective in terms of geographic and demographic capital access gaps (10). This analysis suggests the geographic gap is measurable, substantial, and concentrated: not diffusely spread across rural America, but clustered in specific states — Kentucky, Louisiana, Alabama, Tennessee, Arkansas — where no high-volume originator has established a franchise.

It also suggests a metric. Loans per 1,000 establishments at county level is computable from public data on a quarterly cadence, and would give the agency a direct read on whether distribution interventions are reaching the counties that lack coverage.

Limitations

Stated plainly, because a study that hides its weaknesses invites someone else to find them.

The denominator excludes non-employer firms. County Business Patterns counts establishments with payroll. Sole proprietorships without employees are outside it, and they are a meaningful share of small 7(a) borrowers. This inflates measured penetration in counties with a high non-employer share, which skews rural. The rural gap reported here is therefore conservative.

Reference-year mismatch. The establishment count is 2023; the loan window is FY2021–FY2025. Business counts move slowly enough that this is acceptable for a density measure, but it is not a perfect alignment, and counties with rapid business formation or loss during the window will be measured against a mid-period base.

Project county is the SBA's field, not ours. Loans are placed by the project county recorded in the FOIA extract, which can differ from the borrower's headquarters. For multi-site borrowers this may attribute a loan to the location of the financed project rather than the operating business.

Connecticut required reconstruction. Connecticut replaced its eight counties with nine planning regions as county equivalents for federal statistical purposes in 2022 (11). The SBA's extract carries both geographies simultaneously — some records coded to planning regions, others still to legacy counties, within the same fiscal years. Legacy-coded records were reallocated to planning regions through the borrower's town using the CT Data Collaborative crosswalk (12). Two hundred and sixty-eight records could not be resolved and are excluded from Connecticut figures.

Territories are excluded. Puerto Rico municipios, Guam and the US Virgin Islands lack a matching denominator in the Census county file used here and are not included in the national or state figures.

This measures deployment, not demand or denial. The FOIA extract contains approved loans only. It does not record applications that were declined, or businesses that never applied because they knew of no lender to approach. A low penetration rate is consistent with weak demand, restrictive underwriting, or absent distribution, and this dataset alone cannot separate them. The lender-concentration evidence points strongly toward distribution, and the academic literature on branch presence supports that reading, but the causal claim is an inference rather than a measurement.