Car Wash Demand: Traffic Capture and Membership Market Sizing

The query that brings most readers to this page, car wash market analysis data, returns a page of paid reports: market sizes to the second decimal, subscription shares, growth rates compounded out to 2033. None of those figures can be checked by the reader, and none of them describes the intersection a lender is being asked to finance. This article builds the car wash demand model the other way round, from the public record outward: the vehicles the Census Bureau counts in every block group, the trips the Federal Highway Administration measures on every arterial, the commuting pattern that changed between 2019 and 2024, the 19,807 car wash establishments the Census Bureau counted in 2023 and the way their size distribution has shifted, and the one public revenue series that shows the subscription model arriving without ever naming it. Every number below carries a named source and a date, and every step can be rerun by a credit analyst with a browser.

The asset class deserves the care. Car washes sit on the Small Business Administration's list of special purpose properties, which raises the borrower's contribution on a 504 project from the standard 10% to 15% (504 Capital Corporation, guidance on SOP 50 10 8, 2025), and the express tunnel format that has driven the last decade of development is a single-use building on a high-traffic corner whose value depends almost entirely on the traffic it captures and the memberships it converts. A demand model that cannot show where the cars come from is not a model; it is a forecast borrowed from someone else's report.

What the public record can count: vehicles, trips and sites

Start with the denominator. The American Community Survey's 2024 one-year estimates, released on 11 September 2025, count 132,737,146 occupied housing units in the United States and record, for each, how many vehicles are kept at home. In 2024, 8.5% of households had no vehicle, 33.2% had one, 36.3% had two, 14.4% had three, 5.2% had four and 2.4% had five or more (U.S. Census Bureau, ACS 2024 1-year estimates, table B25044, 2025). Counting five or more as five, which understates the top class, the national fleet works out to at least 1.82 vehicles per household: 2.10 for the 86.6 million owner households and 1.29 for the 46.1 million renter households. Tenure is the first sorting variable in any car wash catchment. Owners are more than twice as likely as renters to keep three or more vehicles (28.8% against 9.2%), while 18.3% of renter households keep no vehicle at all against 3.3% of owners.

The same table exists for every block group in the country in the 2020 to 2024 five-year file released on 8 January 2026, and it is the table a catchment model should start from, because a car wash sells to vehicles, not to people. A three-mile ring around a suburban interchange in Utah, where the state's density of car wash establishments runs at 8.12 per 100,000 residents, holds a very different vehicle count from the same ring around a transit station in a Northeastern city, where the no-vehicle share of renter households can exceed half. The trade-area demographics article explains how the ACS is apportioned to rings and drive-time polygons and where the apportionment breaks; for car washes the rule is that the vehicle count, not the household count, is the unit of demand, and that the vehicle count is read from B25044 at the block-group level, never from a national average.

The second public count is trips. The Federal Highway Administration's Traffic Volume Trends series, compiled monthly from state highway agency counters, put cumulative travel on all roads and streets at 3,323.8 billion vehicle miles in 2025, up 0.9% on 2024 (FHWA, Traffic Volume Trends, December 2025, 2026). The third is sites. County Business Patterns counted 19,807 car wash establishments with paid employees in 2023, employing 172,644 people in the week of 12 March with an annual payroll of $5.07 billion (U.S. Census Bureau, County Business Patterns 2023, 2025). The Bureau of Labor Statistics' Quarterly Census of Employment and Wages, built from unemployment insurance records, put the 2024 figure at 19,736 private establishments on an annual average with 187,502 employees and $6.11 billion in wages, an average annual pay of $32,560 (BLS, QCEW 2024 annual averages, NAICS 811192, 2025). Vehicles, trips, sites: the three counts that the trade's paid reports assert and the public record publishes.

Traffic capture: reading AADT honestly

The traffic-capture method is the car wash industry's oldest demand model and its most abused. The arithmetic is simple: take the annual average daily traffic on the frontage road, apply a capture rate to estimate the share of passing vehicles that will stop, multiply by an average ticket or a membership conversion, and a revenue line appears. Each step deserves scrutiny, and the public record supplies it for the first step only. Annual average daily traffic, AADT, is defined in the Highway Performance Monitoring System's field manual as the total volume of vehicle traffic on a road segment for a year divided by 365, both directions combined, and every state publishes it for the segments it monitors (FHWA, HPMS Field Manual and public release data, 2022). What the figure does not say is as important as what it does: AADT is two-directional, so a site with access from one side captures from half the count at most; it is an annual average, so a corridor with a summer peak and a winter trough has a December figure well below the published number; and on arterials that carry a commute, the directional split at the evening peak is the number that decides whether the home-bound driver passes the site on the near side of the road or the far one.

Scale matters here. Of the 3,323.8 billion vehicle miles driven in 2025, 1,157.2 billion, or 34.8%, were driven on urban arterials other than interstates, the road class where nearly every express car wash sits; a further 587.7 billion were on urban interstates and 543.2 billion on other urban roads (FHWA, Traffic Volume Trends, December 2025, Table 2, 2026). Travel on urban other arterials grew 0.8% in 2025, slightly below the 0.9% national figure, and the month of December 2025 alone carried 265.8 billion vehicle miles across all systems, with the seasonally adjusted figure at 277.2 billion. The regional December figures show the geography of the traffic base: 61.3 billion miles in the South Atlantic region, 57.6 billion in the South Gulf, 57.5 billion in the North Central, 55.2 billion in the West and 34.2 billion in the Northeast, with the West up 1.2% and the Northeast down 1.0% on December 2024 (FHWA, Traffic Volume Trends, December 2025, 2026).

The capture rate is where the method leaves the public record. No federal series measures how many passing vehicles stop at a car wash, and the capture rates that circulate in the trade are proprietary rules of thumb whose provenance is the experience of the operator quoting them. The honest model therefore reports the AADT with its source, its count year and its directional split, states the capture rate as an assumption, and runs the revenue line at a range of capture rates rather than a single figure. The article on state DOT AADT counts covers the sources, the count-station coverage and the estimation methods state by state; the national traffic atlas shows how much of the network carries a real count against a modelled one. The difference between a counted AADT and a modelled one on a minor arterial can be a factor of two, which is larger than the spread between a good site and a bad one.

The commute that changed, and what it did to the arterial

The traffic-capture model carries a hidden assumption that stopped being true in 2020: that the arterial's volume is a commute, concentrated in two peaks, passing the site twice a day with the same drivers. The American Community Survey records the change. In 2019, roughly 76% of workers drove alone to work and 5.7% worked from home; by 2021 the drive-alone share had fallen to about 68% and the work-from-home share had tripled to 17.9%, while public transportation commuting fell by half, from 5% to 2.5% (U.S. Census Bureau, press release of 15 September 2022, 2022). The 2024 one-year estimates show a partial return: of 165,360,450 workers aged 16 and older, 69.2% drove alone, 9.2% carpooled, 3.7% used public transportation, 2.4% walked and 13.3% worked from home (U.S. Census Bureau, ACS 2024 1-year estimates, table B08301, 2025). The work-from-home share has settled at more than twice its 2019 level, and the drive-alone share, though recovering from 67.8% in 2021, remains about seven points below 2019.

Yet total driving did not fall. Cumulative travel dropped 13.2% in 2020, rebounded 11.2% in 2021 and has grown every year since: 0.9% in 2022, 2.1% in 2023, 1.0% in 2024 and 0.9% in 2025, so that 2025's 3,323.8 billion vehicle miles sit above the 3,269.1 billion preliminary figure for 2019 (FHWA, Traffic Volume Trends, December reports 2019 to 2025, 2020 to 2026). The reconciliation is the point for a car wash model: the miles came back, but fewer of them are commute miles. A driver who works from home two days a week makes fewer trips past the site on the commute axis and more on errands, at midday, on weekends, in directions the commute model never considered. The arterial's AADT may be back to 2019, but the peak-hour share and the directional split behind it have changed, and a site chosen in 2018 for its home-bound evening capture is now serving a flatter, more dispersed trip pattern. For a membership model this is good news, since the member washes on a schedule rather than on a commute; for a single-ticket model dependent on impulse capture at the evening peak, it is not.

The practical consequence is that the ACS commuting tables belong in the car wash model beside the AADT. Table B08301 is published for every county and place above 65,000 residents in the one-year file and for every tract in the five-year file, and a catchment where work-from-home runs at 25% behaves differently from one where it runs at 8%, even with identical traffic counts. The foot-traffic analytics article sets out what device-based visit data can and cannot add to this picture; the public tables give the structural pattern with a known sample and a published margin of error, which is the standard a credit file should be built on.

Supply: 19,807 establishments and a format shift the count alone hides

County Business Patterns gives the supply series its spine. Car wash establishments with paid employees rose from 16,388 in 2017 to 16,697 in 2018, 16,976 in 2019, 17,584 in 2020, 18,520 in 2021, 19,463 in 2022 and 19,807 in 2023, a gain of 20.9% in six years that did not pause for the pandemic (U.S. Census Bureau, County Business Patterns 2017 to 2023, national files, 2019 to 2025). Employment moved far less: 163,101 in 2017, a dip to 156,419 in 2021, and 172,644 in 2023, up 5.9% over the period. Payroll, by contrast, rose 57.5%, from $3.22 billion to $5.07 billion. Three series, three different slopes, and the divergence is the first sign that the count of sites is the least informative of the three.

The size-class columns explain it, and they carry this article's one finding that the paid reports do not show. Between 2017 and 2023 the number of car wash establishments with fewer than five employees grew 20.7%, from 8,580 to 10,352; those with five to nine employees grew 20.1%, from 3,051 to 3,665; and those with ten to nineteen employees grew 59.6%, from 2,335 to 3,727. Every larger class shrank: establishments with 20 to 49 employees fell 13.5%, from 1,961 to 1,697; those with 50 to 99 fell 21.8%, from 412 to 322; and those with 100 or more fell from 46 to 40 (U.S. Census Bureau, County Business Patterns 2017 and 2023, 2019 and 2025). The 10-to-19 class is the labor footprint of the express exterior tunnel: a handful of attendants at the pay stations and the vacuum lot, no interior service, a conveyor doing the work that a full-service wash did with a crew of 30. The full-service format, the 20-to-49 and 50-to-99 classes, is in retreat, and the self-serve bay format in the smallest class is growing at the same pace as the industry. The story the establishment count tells, steady growth, is true; the story the size classes tell, a format replacing another, is the one that decides whether a new site competes with the washes already in its catchment or takes their customers.

Two caveats belong beside the series. First, County Business Patterns counts only establishments with payroll, and the self-serve bay operated by an owner without employees appears instead in the Nonemployer Statistics, whose 2023 edition was released on 15 May 2025 (U.S. Census Bureau, Nonemployer Statistics 2023, 2025); a catchment inventory that stops at the employer count misses exactly the low-price competitor that sets the floor under single-ticket pricing. Second, the establishment is a location, not a brand; a chain of 40 express tunnels is 40 establishments, and the public record cannot tell a new entrant from a new unit of an existing operator. What it can do is date the growth by county through the annual county file, which the County Business Patterns article covers in detail, and that dating is what a lender needs to know whether the three competing tunnels in a catchment opened in 2016 or in 2023.

Where density is high and low: the state record

Dividing the 2023 establishment count by the July 2023 population gives 5.83 car wash establishments with paid employees per 100,000 residents nationally, and the state figures around that average are not where a reader steeped in the trade's Sun Belt narrative would expect them. The densest states are the Plains and the Mountain West: Wyoming at 10.26 per 100,000, Iowa at 9.39, Montana at 9.38, Nebraska at 8.40, Utah at 8.12 and North Dakota at 7.62, followed by Florida at 7.27, Kansas at 7.09, Idaho at 7.05 and New Jersey at 7.04 (U.S. Census Bureau, County Business Patterns 2023 state file, 2025; Vintage 2025 Population Estimates, 2026; MMCG computation). The least dense are Mississippi at 3.53, Vermont at 3.86, Minnesota at 4.03, Hawaii at 4.18, Wisconsin at 4.40, West Virginia at 4.52, Louisiana at 4.63, Maryland at 4.64 and Pennsylvania at 4.83, with the District of Columbia, a city rather than a state, at 1.32. Among the ten most populous states the spread is narrower: Florida 7.27, Michigan 6.64, Georgia 6.25, Illinois 6.20, Ohio 6.06, Texas 6.05, California 5.69, North Carolina 5.66, New York 4.87 and Pennsylvania 4.83.

The pattern says more about format than about demand. The Plains states' density is a density of small establishments: Wyoming's 60 car washes employ 346 people, 5.8 per site, and Iowa's 302 employ 2,185, 7.2 per site, against 11.8 per site in Texas, 10.5 in California and 10.4 in Maryland (U.S. Census Bureau, County Business Patterns 2023 state file, 2025). Road salt, gravel and a dispersed rural population produce many self-serve bays and small automatics; the dense, high-income coastal states produce fewer, larger tunnels. A per-capita figure that does not carry the size distribution beside it compares a Wyoming bay with a New Jersey tunnel as if they were the same thing, and the same error at the catchment level, counting a four-bay self-serve as a competitor to a 150-foot express tunnel, is the most common mistake in car wash supply inventories. Establishments per 100,000 residents is the screen; employees per establishment is the correction; and the drive-time inventory by format, built from the address-level file and a look at each site, is the answer.

Membership sizing from the public record

The subscription model has reorganised the car wash industry in a decade, and no public dataset measures memberships. That is the honest starting point, and it is not the end of the analysis, because the public record measures the model's consequences with some precision. The Census Bureau's Service Annual Survey reports the revenue of employer car wash firms every year: $9,724 million in 2017, $10,935 million in 2018, $11,487 million in 2019, $11,401 million in 2020, $14,678 million in 2021 and $16,268 million in 2022 (U.S. Census Bureau, Service Annual Survey, NAICS 811192, via FRED series REVEF811192ALLEST, updated 31 January 2024). Revenue grew 67.3% over five years in which the establishment count grew 18.8% and employment grew 2.0%. Set the series against County Business Patterns and the subscription model appears in arithmetic: revenue per establishment rose from about $593,000 in 2017 to about $836,000 in 2022, a gain of 41%; revenue per employee rose from about $59,600 to about $97,800, a gain of 64%; and payroll fell from 33.1% of revenue to 28.5%. The 2020 figures are the clearest single exhibit. Traffic fell 13.2% that year, yet car wash revenue fell 0.7%, because a membership billed monthly does not fall with the miles driven.

The sizing method a credit file can defend therefore has two layers. The first layer is the denominator, which is entirely public: the vehicles kept at home in the catchment from ACS table B25044 at the block-group level, weighted by tenure; the vehicles passing the site from the state DOT's AADT, with the directional split and the count year stated; and the commuting structure from B08301, which governs how many of those passing vehicles pass on a schedule. The second layer is the conversion, which is not public: the share of catchment vehicles that will hold a membership at a given site, and the share of passing vehicles that will stop. The defensible memo reports the first layer with sources and treats the second as a sensitivity, running the revenue line at a range of penetration rates and showing which rate the loan needs. A memo that presents a single penetration rate as a fact is presenting someone's opinion as data, and the public revenue series is the check: if the pro forma implies revenue per site far above the $836,000 national average of 2022, adjusted for the catchment's vehicle density and for the years since, the memo owes the reader an explanation of why this site is different.

The comparison with self-storage, the asset class covered in the previous article in this series, is instructive. Both are single-purpose buildings whose demand comes from a transition, the move in one case and the accumulation of dirt in the other; both have adopted a recurring-revenue model that public data cannot observe directly; and both show the model's fingerprint in revenue per site. The difference is that storage demand is counted in households and car wash demand in vehicles, and the vehicle count is a sharper instrument, because the ACS records it by tenure and the block group records it by address.

What the lender reads

Car washes are named examples of special purpose property in SBA lending guidance, and the classification carries a cost: the borrower's contribution on a 504 project rises from 10% to 15%, and to 20% when the business is also new (504 Capital Corporation, guidance on SOP 50 10 8, 2025; NAGGL, SBA Procedural Notice 5000-872764 revising SOP 50 10 8, 2025). The reasoning is alternative use. A tunnel building with its conveyor, its water reclamation and its pay-station canopy is worth a great deal as a car wash and very little as anything else, so the lender's exposure on failure is the land and the traffic count, not the improvements. The article on special purpose property risk works through what the public loan data show for that class; the public SBA 7(a) and 504 datasets, which MMCG Analytics' SBA layer is built on, carry the industry's lending history by NAICS code, subject to the rule that no performance rate is shown for any cohort of fewer than ten loans. MMCG Analytics supplies the data and the analysis; the credit decision rests with the lender.

What the demand model contributes to that decision is the answer to the lender's real question, which is not whether car washes are a good business but whether this site's traffic and vehicle base can carry the debt if the membership conversion disappoints. The public record answers it in layers: the AADT with its count year and directional split, the catchment's vehicles by tenure, the commuting structure, the inventory of competing sites by format and opening year, the state density as context, and the national revenue per site as a reasonableness check on the pro forma. A memo carrying those six items with sources can be tested line by line; a memo carrying a market size from a paid report and a capture rate from the borrower's consultant cannot.

Method: the five numbers a car wash memo should carry

First, the vehicle base: households by vehicles available in the drive-time catchment, from ACS table B25044 in the 2020 to 2024 five-year file apportioned to block groups, reported as a vehicle count with the tenure split, and scaled to the current year with the Vintage 2025 county estimate. Second, the traffic base: AADT on each frontage and each approach, from the state DOT with the count year, the station type (counted or estimated) and the directional split, with the figure halved where the site is accessible from one direction only. Third, the commuting structure: the drive-alone, carpool and work-from-home shares from B08301 for the catchment's tracts, read against the 2019 and 2024 national benchmarks of roughly 76% and 69.2% driving alone, which decide how much of the AADT is schedule traffic and how much is errand traffic. Fourth, the inventory: every car wash establishment in the catchment's ZIP codes from County Business Patterns, every nonemployer filing from the Nonemployer Statistics, reconciled by address and classified by format (self-serve bay, in-bay automatic, express exterior tunnel, full-service) with the opening year from the permit record. Fifth, the reasonableness check: the pro forma's revenue per site against the Service Annual Survey national average of about $836,000 per employer establishment in 2022, with the catchment's vehicle density and the years since stated as the adjustments.

None of the five is a forecast. Each is a measurement with a source and a date, and the forecast, the membership penetration and the capture rate, is shown as a range on top of them. This is the provenance standard the whole library is built on, and the reason the asset-class pillar treats car washes, travel centers and self-storage with the same discipline: the public record supplies the denominator for every property type, and the honest model says where the denominator ends.

A worked sequence for one corner

The sequence, run for a proposed express tunnel on a suburban arterial, goes as follows. The analyst pulls the state DOT's AADT for the frontage segment and for the cross street, notes whether each is a counted or an estimated station and in which year, and records the directional split where the state publishes one. A ten-minute drive-time polygon is built on the road network, intersected with block groups, and the B25044 vehicle counts are apportioned and summed by tenure; the result is a vehicle base, say 60,000 vehicles in 28,000 households, of which the owner households hold 70%. The B08301 shares for the same tracts give the commuting structure, and the travel-center article shows the same AADT discipline applied to a fuel site, where the truck share of the count does the work the directional split does here.

On the supply side the analyst lists every establishment coded 811192 in the catchment's ZIP codes, adds the nonemployer filings for the county, visits each address in the aerial imagery and the permit record, and classifies it by format with its opening year. A catchment with two express tunnels that opened in 2022 and 2023, three aging in-bay automatics at fuel stations and one self-serve bay is a different market from one with the same count of sites built before 2010. The pipeline column carries permitted but unbuilt sites, which for this asset class the county permit record usually shows a year or more ahead of opening.

The output is three lines and a range. Vehicles per competing site in the catchment, with the format mix stated; AADT on the frontage with its count year and directional split; and the revenue per site the pro forma implies, set against the national $836,000 of 2022 and the state's density and size-class profile. Beneath them, the membership penetration and the capture rate the loan requires, stated as the assumptions they are, and the penetration at which debt service fails. A lender reading that page knows exactly which number is measured and which is assumed, and the public sources let the lender check the measured ones in an afternoon.

Frequently asked questions

What public data measures car wash demand?

Three federal series: the American Community Survey's vehicles-available table (B25044), which counts the vehicles kept at home in every block group; state DOT annual average daily traffic, which counts the vehicles passing a site; and the ACS commuting table (B08301), which shows how much of that traffic is schedule traffic. County Business Patterns and the Service Annual Survey measure the supply side and its revenue.

How is a traffic-capture model built from AADT?

AADT from the state DOT is the total annual volume on a segment divided by 365, both directions combined. The model records the count year, whether the station is counted or estimated, and the directional split, halves the figure where the site is accessible from one side, and applies the capture rate as a stated range rather than a fact, since no public series measures it.

How many car washes are there in the United States?

County Business Patterns counted 19,807 car wash establishments with paid employees in 2023, up from 16,388 in 2017, and the BLS Quarterly Census of Employment and Wages counted 19,736 private establishments on a 2024 annual average. Self-serve washes without payroll appear separately in the Nonemployer Statistics.

Can membership or subscription demand be measured from public data?

Not directly. No federal dataset counts car wash memberships. The Census Bureau's Service Annual Survey shows the model's effect: revenue of employer car wash firms rose from $9,724 million in 2017 to $16,268 million in 2022, a 67% gain against a 19% gain in establishments, and revenue held in 2020 while traffic fell 13.2%. A defensible model reports the vehicle and traffic base from public data and treats penetration as a sensitivity.

Which states have the most car washes per capita?

On 2023 County Business Patterns counts and July 2023 population, Wyoming (10.26 per 100,000 residents), Iowa (9.39), Montana (9.38), Nebraska (8.40) and Utah (8.12) lead; Mississippi (3.53), Vermont (3.86) and Minnesota (4.03) trail; the national figure is 5.83. The high-density states are dense in small establishments, so the figure must be read with employees per establishment beside it.

Why does SBA lending treat car washes as special purpose property?

Lender guidance on SOP 50 10 8 lists car washes among special purpose properties because the improvements have little alternative use. The borrower's contribution on a 504 project rises from 10% to 15%, and to 20% for a new business. The classification reflects collateral, not demand; the public SBA loan datasets carry the industry's lending history by NAICS code.

Sources

1. U.S. Census Bureau, County Business Patterns 2023, national and state files (cbp23us, cbp23st), NAICS 2017 code 811192, released 26 June 2025. https://www2.census.gov/programs-surveys/cbp/datasets/2023/
2. U.S. Census Bureau, County Business Patterns 2017 to 2022, national files (cbp17us to cbp22us), NAICS 811192 by employment size class, released 2019 to 2024. https://www2.census.gov/programs-surveys/cbp/datasets/
3. U.S. Census Bureau, Service Annual Survey, total revenue for car washes, all establishments, employer firms (NAICS 811192), 1998 to 2022, as published by the Federal Reserve Bank of St. Louis, FRED series REVEF811192ALLEST, updated 31 January 2024. https://fred.stlouisfed.org/series/REVEF811192ALLEST
4. U.S. Bureau of Labor Statistics, Quarterly Census of Employment and Wages, 2024 annual averages, NAICS 811192 Car Washes, private, United States, 2025. https://data.bls.gov/cew/data/api/2024/a/industry/811192.csv
5. U.S. Census Bureau, American Community Survey 2024 1-year estimates, table B25044 Tenure by Vehicles Available, released 11 September 2025 (table-based summary file). https://www2.census.gov/programs-surveys/acs/summary_file/2024/table-based-SF/data/1YRData/acsdt1y2024-b25044.dat
6. U.S. Census Bureau, American Community Survey 2024 1-year estimates, table B08301 Means of Transportation to Work, 2025; and 2021 1-year estimates, table B08301, 2022. https://www2.census.gov/programs-surveys/acs/summary_file/2024/table-based-SF/data/1YRData/acsdt1y2024-b08301.dat
7. U.S. Census Bureau, The Number of People Primarily Working From Home Tripled Between 2019 and 2021, press release CB22-155, 15 September 2022. https://www.census.gov/newsroom/press-releases/2022/people-working-from-home.html
8. Federal Highway Administration, Office of Highway Policy Information, Traffic Volume Trends, December 2025 (compiled with data on hand as of 28 January 2026), 2026. https://www.fhwa.dot.gov/policyinformation/travel_monitoring/25dectvt/25dectvt.pdf
9. Federal Highway Administration, Traffic Volume Trends, December 2019 through December 2024 reports (cumulative annual travel and percent change, preliminary), published 2020 to 2025. https://www.fhwa.dot.gov/policyinformation/travel_monitoring/tvt.cfm
10. Federal Highway Administration, Highway Performance Monitoring System Field Manual and HPMS Public Release (AADT definition and data items), page updated 13 September 2022. https://www.fhwa.dot.gov/policyinformation/hpms/shapefiles.cfm
11. U.S. Census Bureau, Vintage 2025 Population Estimates, state totals (NST-EST2025-ALLDATA), July 2023 population used as the density denominator, 2026. https://www2.census.gov/programs-surveys/popest/datasets/2020-2025/state/totals/NST-EST2025-ALLDATA.csv
12. U.S. Census Bureau, Nonemployer Statistics 2023, released 15 May 2025. https://www.census.gov/newsroom/press-releases/2025/2023-nonemployer-statistics.html
13. 504 Capital Corporation, SBA 504 Loans for Special Purpose Properties and Real Estate (guidance on SOP 50 10 8 classifications), 2025. https://504capital.com/blog/financing-special-purpose-properties-sba-504-loans/
14. National Association of Government Guaranteed Lenders, SBA Notice Revising SOP 50 10 8 (Procedural Notice 5000-872764), 2025. https://www.naggl.org/sba-notice-revising-sop-50-10-8/
15. MMCG Research, SBA 7(a) Performance Series: MMCG analysis of the public SBA 7(a) loan register, 2026. https://mmcganalytics.com/sba-default-rates/