Hospital Feasibility Study

The long record under 7(a) and 504

Since FY1991 SBA lenders have approved 703 7(a) loans to hospital projects for $281.0 million and 82 504 loans for $84.6 million. The FY2017 to FY2026 decade holds 231 of those 7(a) loans and $131.4 million of their dollars, and 7 of the 504 loans with $11.8 million. The decade claims a much bigger slice of the 7(a) dollars than of the 7(a) count, so the hospital loans of recent years run larger on average than those approved before FY2017. Across the whole record the 504 loans average far more per loan than the 7(a) loans, which matches the size bands further down, where the decade's 504 loans sit near the top. These are nominal approval dollars, not adjusted for inflation, so some of the rise in loan size may be price. The conclusion for a lender holds either way: a hospital request reaching an SBA lender today tends to carry more money per file than the long record suggests, and a study must defend the size of the request against the project itself, not against what hospital loans used to be.

The FY2006 to FY2008 peak and what followed

Through FY1998 the 7(a) hospital count never rose above 4 in a year. It reached 9 in FY1999 and 11 in FY2002, then climbed through 19, 22 and 29 in FY2003, FY2004 and FY2005 to the busiest run in the record: 47 loans in FY2006, 49 in FY2007 and 45 in FY2008. FY2009 fell back to 27 and FY2010 to 18. No later year has come near the peak; the highest since is FY2014's 36, and the decade itself moves between 13 loans in FY2020 and 34 in FY2025 without a trend in either direction. A study cannot point to a rising national flow of hospital approvals as support for a project, because there is none to point to. What the series does show is that SBA lenders have kept approving hospital loans at a steady modest rate for the whole period since FY2009, so the class is familiar to the program even where it is new to an individual credit officer.

Equal counts, unequal dollars

Where the year table prints dollars, it shows how little a year's count says about its money. FY2019 and FY2022 each recorded 23 7(a) hospital loans; FY2019's came to $20.8 million and FY2022's to $6.2 million. FY2023 approved 18 loans for $15.8 million, while FY2018's 15 loans totalled $3.3 million. The early printed years run the same way: FY2004's 22 loans came to $2.7 million, less than FY2003's $4.0 million on 19 loans, and FY2005's 29 loans to $5.3 million, against $16.3 million on the 33 loans of FY2011. Within the decade, swings of that size most likely come from the few loans in the largest occupied size band, which the loan-size section shows carrying most of the 7(a) dollars. A study should therefore read national hospital dollars as the trace of a few large projects laid over many small operating loans, and never as a demand series from which one project's prospects could be inferred.

Years when 504 carried more money than 7(a)

In FY2003 the 504 program's 7 hospital loans came to $6.0 million, more than the $4.0 million on that year's 19 7(a) loans, and in FY2004 its 8 loans came to $6.3 million against $2.7 million on 22. FY2005 and FY2011 reversed the order, with 5 504 loans each for $3.0 million and $6.1 million against 7(a) totals of $5.3 million and $16.3 million. FY2004's 8 loans are the highest 504 count in the record, and from FY2006 through FY2010 the program held at 4 a year. The decade is thinner: 2 loans in FY2017, 0 in FY2018, FY2019, FY2022 and FY2023, and 1 in each of FY2020, FY2021, FY2024, FY2025 and FY2026. A hospital sponsor proposing a 504 structure for a building is asking for something the class has done only 7 times in the decade, so the study has to carry the real estate case on its own: what the building costs, what it is worth as a licensed and operating hospital, what it would be worth if the license or the operator fell away, and why the operating projection supports the whole of the debt.

Blank cells are not zeros

The year table marks its blanks in different ways. Fewer than 5 loans marks a dollar cell resting on too few loans to print, which covers every 7(a) year through FY1998 and most 504 years. Withheld marks a dollar cell the table holds back although the year has 5 or more loans, as in most 7(a) years of the decade. Not measured marks a year when SBA's files code too few loans to an industry for the class to be counted; it applies to the 504 column in FY1991 and FY1993, so the early 504 hospital record is incomplete and its total counts only the loans that could be placed. None of these is a zero, and a study that quotes the table should leave each one as printed. The table also counts gross approvals with cancelled ones kept in; with a cancellation floor of 14.0% on the FY2017 to FY2024 approvals, the number of hospital loans that actually funded is lower than the approval count, by an amount this page does not give.

What the long record lets a study claim

Read whole, the record supports one claim and refuses another. It shows that SBA lenders approved at least 1 hospital 7(a) loan in every fiscal year from FY1991 to FY2026, 703 in all, so a hospital that is eligible to borrow through SBA is a known borrower type rather than an exception a lender has to be talked into. It does not show how those hospitals fared over the long run, because the page measures cancellations only on the FY2017 to FY2024 approvals and charge-offs only on the loans disbursed in FY2017 to FY2021. A study therefore cites the long record for eligibility and precedent and turns to the subject's own service area, payer contracts and medical staff for everything a lender will weigh on repayment.

The Huntington National Bank clears the listing floor

The lender table lists only lenders with at least 5 hospital loans in FY2017 to FY2026, and The Huntington National Bank is its only entry, with 28 loans for $8.0 million. Every other lender that approved an SBA hospital loan in the decade made fewer than 5. Set against the decade's 231 7(a) loans and 7 504 loans, that means hospital lending is spread across a long list of banks and other SBA lenders, each of which has handled a hospital file fewer than 5 times. For a sponsor, the lender reading the request is unlikely to hold a hospital book of its own to compare the project with, and the feasibility study becomes the reference the credit committee uses in place of that experience.

A book of small loans

The listed lender's 28 loans come to $8.0 million, so its hospital loans are small on average, closer to the class's median of $200,000 than to the large loans that carry most of the decade's dollars. The 18 7(a) loans in the largest occupied size band alone total $70.5 million, far more than the listed lender's whole book, so most of those large hospital loans were made by lenders with fewer than 5 hospital loans each. A sponsor seeking a hospital loan of that size will most likely deal with a lender whose hospital experience runs to a handful of files, and the study has to be thorough enough to replace the comparison that lender cannot make from its own portfolio: the market, the medical staff, the payer mix, the cost and the coverage, each documented.

A withheld rate, not a clean one

The lender's charge-off floor is withheld. The page gives no number for it, and a reader should treat it neither as zero nor as a sign that the lender's hospital loans have done well or badly. The class floors, 4.2% of the FY2017 to FY2021 disbursement cohort charged off to date and 14.0% of FY2017 to FY2024 approvals cancelled, are the only loss measures this page publishes for hospital lending. A study that discusses lender experience should cite those, name the lender only as the table names it, and leave the withheld rate blank.

More establishments, slower growth in jobs

Nationally, County Business Patterns counts 6,906 hospital establishments in 2018 and 7,523 in 2023, with employment rising from 6,055,138 to 6,344,178 and a 2023 payroll of $511.81 billion. Employment rose more slowly than the establishment count, which leaves the average hospital establishment of 2023 smaller than in 2018. California posted the largest gains of any state in both establishments, from 500 to 702, and employees, from 566,860 to 620,841. Florida added the next most establishments, from 304 to 405, with employment up from 331,397 to 369,683, and Texas, which counts the most hospital establishments of any state in both years, went from 629 to 712, with employment up from 410,696 to 460,981. California, Florida and Texas also sit among the leaders of the SBA hospital ranking, so where SBA hospital lending concentrates, new hospital supply concentrates too. A study in those states has to keep its competitive inventory current, name the facilities that opened since 2018, and test the subject's volumes with them in the market.

Where the count fell

North Carolina lost the most establishments, from 217 to 194, and its employment fell from 197,396 to 186,641. Connecticut went from 50 establishments to 44 and from 66,208 employees to 55,879. Elsewhere fewer establishments came with more staff: New York fell from 285 to 275 while employment rose from 468,726 to 480,148, Pennsylvania from 314 to 307 with employment up from 299,268 to 310,311, and Ohio from 270 to 267 with employment up from 272,919 to 292,786. New York, Pennsylvania and Ohio moved toward fewer but larger establishments; North Carolina and Connecticut shrank on both counts. A study in a shrinking market has to say whether the subject fills a gap left by closures or competes for a falling volume, and a lender in North Carolina, with 8 SBA hospital loans for $7.6 million in the decade, will want that answered with service-area evidence rather than with the state trend.

More establishments sharing fewer staff

In another group of states the count of establishments rose while employment fell. Mississippi went from 109 to 120 establishments as employees fell from 69,839 to 61,790; Alabama from 118 to 126 as employees fell from 98,355 to 95,784; Illinois from 224 to 255 against a fall from 241,593 to 237,688; Minnesota from 152 to 160 against 123,651 to 121,246; and Wisconsin from 161 to 173 with employment nearly unchanged at 115,714 and 115,597. Clinical staffing is usually the largest expense in a hospital budget, as this page's note on study scope says, and in these states a new hospital recruits from a workforce that has not grown while the number of employers competing for it has. The staffing plan in the study has to show that the positions can be filled at the wages assumed, and the sensitivity runs should test a slower hiring pace and higher pay.

Payroll per employee by state

Payroll per employee differs widely between states, and a cost budget has to follow the state it is in. California's $65.43 billion payroll on 620,841 employees works out well above the national ratio of $511.81 billion on 6,344,178, and Washington, at $12.63 billion on 132,802, and New York, at $45.10 billion on 480,148, also run high. Alabama's $6.14 billion on 95,784 employees, Utah's $3.32 billion on 50,309 and Mississippi's $4.16 billion on 61,790 sit well below the national ratio. A study that applies a national staffing cost to a California hospital understates its largest expense, and one that applies it in Alabama overstates it. Loan Analytics prices staffing from the state's own figures and the subject's labor market, then shows the lender how coverage moves if pay runs above plan.

Many small hospitals or a few large ones

Employees per establishment sort the states into different competitive settings. Kansas counts 164 hospital establishments in 2023 with 62,140 employees, while Maryland counts 69 with 107,820; Nebraska has 107 establishments and 44,712 employees, and Louisiana 237 and 105,719. In Kansas, Nebraska and Louisiana the competitive set is a long list of small facilities; in Maryland, as in New York with 275 establishments and 480,148 employees, it is a short list of large ones. The competition section of the study changes with the setting. Where facilities are many and small it maps the nearby hospitals and the services each offers, and asks which of them the subject would draw patients from; where they are few and large it analyzes how a new entrant wins patients and physicians from established institutions with far greater scale.

States the table cannot sum

For Alaska, Hawaii, Montana, New Hampshire, North Dakota, Rhode Island, South Dakota, Vermont and Wyoming the table prints no total for either year: Census withholds one or more of the class's industry codes there. Oregon has no total in either year for the same reason, Delaware, Iowa and West Virginia have a total for 2023 only, and Maine for 2018 only. Those cells are unpublished, not zero. Hawaii, Montana, New Hampshire, Oregon, South Dakota and Wyoming each have SBA hospital loans in the decade and no industry total here, so a study in any of them builds its supply count from the state's own facility records rather than from this table. Further limits apply everywhere. The source note says most government activity falls outside these counts, so public hospitals are largely missing from them; and nothing in the Census counts separates the nonprofit hospitals that, as the archived note says, cannot borrow through SBA, so the establishments listed are a wider population than the one the SBA record can draw from.