Charter School Feasibility Study

A thin start in the 1990s

The class barely registers in the early files. SBA approved 1 school loan under 7(a) in FY1991, 3 in FY1992 and 6 in FY1993, and the 504 column is not measured until FY1994, when it shows 1. Dollars for those thin years appear only where at least 5 loans stand behind them, and FY1996's 7(a) dollars are withheld outright on 22 loans, so the table gives no unbroken 7(a) dollar series from the start. From FY1995 through FY2008 the 7(a) count stays between 21 and 57 loans a year, with dollars from $6.5 million to $21.4 million, while 504 made no more than 7 school loans in any measured year before FY2001. For a lender this stretch is background rather than evidence: a trend fitted across it would mostly measure how seldom SBA financed schools in the 1990s, so the study takes its comparisons from the recent window that every other table here measures.

Matching loan counts, much larger loans

Set years with matching counts side by side and the growth in loan size is plain. Under 7(a), 57 school loans carried $20.5 million in FY2007 and $63.9 million in FY2021; 50 loans carried $10.6 million in FY2008 and $34.6 million in FY2023. Under 504, 20 loans carried $13.3 million in FY2004, $19.7 million in FY2007 and $29.9 million in FY2017. Between FY2010 and FY2019 the 504 ticket barely moved: 23 loans brought $19.1 million in the first of those years and $19.4 million in the second. It moved sharply late in the record, when 36 loans for $35.6 million in FY2022 were followed by 37 loans for $55.2 million in FY2023. A lender comparing a school project with this record therefore compares it with recent tickets, so the building, its conversion and its fit-out are priced in the study at current costs, not at the cost of the schools SBA financed in the early years of the series.

7(a): a step up in FY2015 and a count record in FY2025

Under 7(a), school dollars stepped up in FY2015, to $71.4 million on 65 loans from $36.0 million on 41 in FY2014. The 7(a) dollar record belongs to FY2018, $84.1 million on 74 loans, followed by FY2019's $74.4 million. The step did not hold in every year: FY2022 and FY2023 fell to $38.6 million and $34.6 million, the second of them the smallest 7(a) school total since FY2013's $34.3 million. FY2025 then set the count record, 84 loans, without a dollar record, since its $73.3 million falls short of both FY2018 and FY2019; the program financed more schools that year with less money per loan. FY2026 fell back to 48 loans and $46.8 million, close to FY2017's $47.7 million. The count swung from 74 in FY2018 to 43 in FY2020 and from 84 in FY2025 to 48 in FY2026, which is why a study should not read any single 7(a) year as the direction of school lending.

504 takes the dollar lead

The programs have traded the lead in school dollars, and in the latest years 504 has held it. 504 has never made more school loans than 7(a) in any measured year, yet it lent more money in FY2004, FY2005, FY2006, FY2008, FY2009, FY2010 and FY2012; in FY2006 its 15 loans carried $16.4 million against $7.6 million on 40 under 7(a). From FY2013 through FY2022, 7(a) led in dollars every year. 504 took the lead back in FY2023 and kept it through FY2025, and its FY2024 total of $97.7 million on 53 loans is the largest single-year figure for either program anywhere in the record, ahead of 7(a)'s $84.1 million in FY2018. FY2026 tipped narrowly back to 7(a), $46.8 million against $44.9 million, yet that $44.9 million is more 504 school money than any year before FY2023 produced. The rise of 504 is a rise in building loans, as the opening reading of its $1.2 million median on this page says, and a study read by a 504 lender has to carry the real estate case in full: the appraisal, the conversion budget and the building's value to another buyer.

Dips the record has absorbed

The series has dipped and recovered more than once, and a credit officer will ask where a new school loan falls in that pattern. 7(a) school approvals fell from 50 in FY2008 to 31 in FY2009 and 30 in FY2010, while 504 fell from 22 to 15 in FY2009; both had recovered past their FY2009 counts by FY2011. 7(a) fell from 71 in FY2019 to 43 in FY2020 and came back to 57 in FY2021. When 7(a) dollars sagged to $38.6 million and $34.6 million in FY2022 and FY2023, 504 was climbing toward its FY2024 record, which reads less like a retreat from schools than a shift in which program financed them. FY2026 is different: both programs fell, 7(a) to 48 loans from 84 and 504 to 30 from 47, a bigger one-year fall in 7(a) school loans than any earlier year shows. A study written now should not borrow the momentum of FY2024 and FY2025; it should show that the school's enrollment and cost case stands on its own, whatever the program totals do next.

The long series beside the recent years

Over the full series, FY1991 to FY2026, schools received 1,477 7(a) loans for $1.02 billion and 685 504 loans for $728.1 million. The window from FY2017 to FY2026 holds 603 of those 7(a) loans and $565.8 million of the 7(a) dollars, and 315 of the 504 loans with $439.5 million of the 504 dollars: in both programs the recent years weigh more in money than in loan count. The counts carry a further caution. In any year where too many of SBA's records lack a usable industry code, the printed figure is a floor, because the school class is counted by its code, so a year's true total can be higher than printed but not lower. The study uses the long series for direction and scale, and tells the lender that its school comparisons come from the recent, larger and more building-heavy book rather than from the averages of the whole record.

Dallas-Fort Worth outweighs most states

Dallas-Fort Worth-Arlington leads the metro table on every measure it prints: 57 loans, $89.0 million and a median of $1.3 million, the largest median of any metro and above Texas's own $1.1 million. Its dollars exceed those of every state except California, Texas and Florida, and Los Angeles-Long Beach-Anaheim's $69.3 million does as well. Houston-Pasadena-The Woodlands ranks sixth by loans, 35, but third by dollars, $48.4 million, more than Georgia lends to schools statewide. A school study in Dallas-Fort Worth is therefore read against the largest school tickets on this page, and the lender will want the cost of the campus, the enrollment needed to carry that cost and the competing seats in its part of the metro proven at that scale, not at the scale of the national 7(a) median.

Texas concentrates; California and Florida spread

Texas's record sits mostly in Dallas-Fort Worth and Houston: their 57 and 35 loans make up most of the state's 131, and most of its $191.4 million. California and Florida run the other way. Los Angeles holds 51 of California's 138 loans, and the Miami metro 41 of Florida's 107, so most school lending in both states happens outside Los Angeles and Miami. For the study, that changes where the comparison comes from. A Dallas or Houston school can be set against its own metro row; a school in a California or Florida market outside those metros has no metro row to stand on and is compared with its state row, which in Florida carries a median of $976,600 against Miami's $908,000, and in California $980,000 against Los Angeles's $1.0 million.

Chicago is Illinois; Phoenix and Atlanta nearly so

Chicago-Naperville-Elgin matches Illinois exactly: 30 loans, $36.4 million and a median of $924,500 in both rows, so on these figures the state's school lending in this window is the Chicago metro's. Phoenix-Mesa-Chandler holds 38 of Arizona's 45 loans and $44.7 million of its $53.0 million, and Atlanta-Sandy Springs-Roswell 33 of Georgia's 40 and $38.2 million of $46.6 million. In Arizona the state median, $896,000, sits well above Phoenix's $660,000, which means the loans outside Phoenix run larger than those inside it. A study for an Illinois, Arizona or Georgia school is in practice read against its main metro's record, and for an Arizona school outside Phoenix the closest comparison is a few larger loans rather than the metro's typical one.

New York and Boston: metro and state diverge

New York-Newark-Jersey City has 42 loans, more than New York State's 33, because the metro reaches into New Jersey. Its $41.6 million exceeds the whole state's $23.3 million, and its median, $487,200, sits well above the state's $250,000, so a large part of the metro's school lending lies outside New York State, beside a New Jersey row of 16 loans and $20.0 million that prints no median. For a school in northern New Jersey the metro row is the better yardstick. Boston-Cambridge-Newton, with 25 loans and $25.4 million, has the smallest metro median on the page, $200,000, beside a Massachusetts median of $150,000. A building loan in greater Boston is therefore far above the local typical school loan, and the study for it draws its comparisons from the national 504 book rather than from the metro.

Metro charge-off floors are withheld

Every metro charge-off floor on the page is withheld except Phoenix's, which goes unreported since fewer than 30 loans sit in its cohort. Withheld is not zero, and a study never treats it as zero. When a lender asks how school loans have performed in Dallas-Fort Worth, Los Angeles or Miami, the answer comes from the floors the page does print: 5.2% in Texas and 0.00% in California and Florida, and nationally 2.5% for 7(a) and 0.00% for 504. Each is a floor, a minimum that can rise as the FY2017 to FY2021 cohort ages, and the study states it that way rather than presenting a metro as loss-free.

Where the program quartiles meet

The quartiles show 7(a) and 504 serving different school loans, with a narrow overlap. Under 7(a) the lower quartile is $140,000, the median $469,000 and the upper quartile $1.2 million; under 504 they are $589,000, $1.2 million and $1.9 million. The 504 median equals the 7(a) upper quartile, and the 504 lower quartile sits above the 7(a) median, so even the smaller 504 school loans are larger than the typical 7(a) one. A school project sized near $1.2 million is a large 7(a) loan and a typical 504 loan at the same time, and the study for it should be written so that either lender can read it: the real estate case a 504 lender expects, and the operating case for the school beside it.

7(a): many small loans, most dollars in large ones

Under 7(a) the smallest band, $150,000 or less, holds more school loans than any other, 187, but only $15.2 million. The second-highest band holds 90 loans and $302.2 million of the program's $565.8 million, more money than all the other 7(a) bands put together, and the top band holds 0 7(a) loans and $0. The 7(a) school book splits into distinct parts: a large number of small loans and a smaller number of loans near the top of the range this record shows for 7(a), which carry most of its money. A study for a 7(a) school loan in the second-highest band is read against the part of the book that holds the money and will be tested as a building loan; a study for a small 7(a) loan is a narrower scope, centered on whether enrollment and tuition carry the payment.

504: clustered around its median

Under 504 the loans cluster around the band that holds the 504 median of $1.2 million, the program's busiest, with 108 loans for $155.2 million; the band above it carries fewer loans, 73, but more money, $205.1 million. The band holding the 7(a) median has 92 504 loans and $59.0 million. At the edges the page suppresses dollars: the smallest band holds 5 504 loans with their dollars withheld, and the top band 2 loans with too few behind them for a dollar figure. Neither dollar figure is given here, and neither is zero. For a 504 school study the comparison set is this middle stretch of the table; a project in the top band, where the page cannot print dollars at all, has almost no SBA school record to be measured against, and the study must build its case from the subject's own enrollment and cost evidence.

Head to head at the top

In the band that holds the 504 median, 504 leads: 108 loans and $155.2 million against 89 7(a) loans and $129.6 million. In the band above it, 7(a) leads: 90 loans and $302.2 million against 73 504 loans and $205.1 million. Short of the top band, the largest school loans in this record are more often 7(a) loans than 504 loans. That runs against the easy assumption that a large school building goes to 504 by default, and it means a large school study cannot be written to one program's habits. The lender chooses the program to fit the deal, and the study should be written so that the debt service test, the collateral case and the sponsor's equity read correctly under the program the lender actually proposes.

The program floors

The program table adds the outcome floors. For school loans approved in FY2017 to FY2024, cancellation stands at a floor of 11.3% in 7(a) and 10.1% in 504, rates close enough that cancellation looks like a risk of the class rather than of either program. Funded loans part ways. The 7(a) charge-off floor is 2.5%, with 2.2% charged off within 5 years, while 504 stands at 0.00% on both measures for the FY2017 to FY2021 cohort. Most 7(a) school losses so far came early, in the years when a school is still filling seats. A 504 floor of 0.00% is a count of recorded losses in a cohort still seasoning, not a promise that a school building cannot lose money, and the study does not present it as one.

Sizing the study to the loan

Taken together, the size and program tables tell a study where to put its weight. A loan in the smallest band, where 187 7(a) loans carry $15.2 million, is a modest credit, and the study for it concentrates on enrollment and the revenue that carries the payment. A loan in the band holding $302.2 million of 7(a) money, or in the 504 bands where most 504 dollars sit, is a building credit: the study adds the cost of conversion or construction, life safety and accessibility work, coverage through the enrollment ramp and the building's value to a buyer that is not a school. The fee follows the scope, so a buyer should tell us the loan size and the program at the outset.