Behavioral Health Feasibility Study

The count rank and the dollar rank diverge

California and Florida lead on both measures, 158 loans for $167.8 million and 99 loans for $86.0 million. Below them the two rankings split. Minnesota is third by count with 77 loans, but its $37.9 million equals Texas's total on 55 loans. Utah is fifth by count with 65 loans yet third in dollars at $69.9 million, and Arizona, sixth by count with 61 loans, is fourth in dollars at $57.7 million, ahead of Ohio's $41.1 million on 66. New Jersey's 37 loans carry $35.5 million, against $17.7 million for Pennsylvania's 37 and $10.2 million for Michigan's. New Hampshire's 20 loans carry $21.9 million, more than Virginia, Colorado or Maryland. A study's market section cannot assume that a state's loan count describes the projects its lenders fund; the dollar column says where the building-scale treatment loans went.

Median loan size separates property states from practice states

Medians run from $648,350 in California down to $100,000 in Maryland. California's median sits above even the national 504 median of $620,000, Arizona at $548,000 and Utah at $500,000 follow, and New Jersey's $421,000 is the only other median above $350,000. Virginia's $309,000, the $300,000 of Florida and Minnesota and Texas's $290,000 sit just above the 7(a) national median of $250,000; Pennsylvania at $242,000 and Ohio at $225,000 sit just under it, and Massachusetts at $182,500, North Carolina at $170,500, Michigan at $150,000 and Connecticut at $120,000 fall further below. The high medians line up with the states that use 504 most: California's 34 504 loans and Utah's 28 lead the table, while Maryland's $100,000 median rests on 44 7(a) loans and 1 under 504. The median sets the scale a lender expects: a building or a residential program in California, Arizona or Utah, a practice-scale loan in Maryland or Connecticut. A project that departs from its state's pattern should explain why.

Program split by state

The program split varies widely from state to state. Utah's record is 37 loans under 7(a) and 28 under 504, the most balanced split in the table. California pairs 124 7(a) loans with 34 under 504, Florida 80 with 19 and Arizona 48 with 13. At the other end Texas shows 53 7(a) loans and 2 under 504, and Maryland 44 and 1. Massachusetts, with 26 and 6, and Ohio, Virginia and North Carolina, each with 5 loans under 504, sit between. The split matters because it tells the lender how many property-backed treatment loans its state has already seen. A 504 request for a Texas treatment building arrives with almost no in-state precedent in this record, so the appraisal and the market section have to establish the property case from the building's licensed use and its alternative uses, while in Utah the lender has a deeper file of financed treatment buildings to compare against.

Cancellation floors

Cancellation floors are reported for the states whose FY2017 to FY2024 approvals reach 30 loans, and they spread widely. Texas has the highest at 15.4%, above both national program floors of 12.2% and 12.0%, followed by California at 13.6% and Minnesota at 13.1%. Florida sits at 11.5% and Ohio at 10.4%. Utah at 6.0% and Arizona at 5.0% have the lowest, and they are also the states with the highest medians after California, so loan size does not explain cancellation here: California, with the largest median, cancels at 13.6%. For a study in Texas, California or Minnesota the floor argues for settling the closing conditions inside the study: a license application filed and its status, the use permit, signed payer contracts and a staffing plan the lender can verify before funds move. From Maryland down the floor is not reported, and a study there should say so rather than borrow another state's rate.

Charge-off floors

Only California, Florida and Minnesota have a charge-off cohort large enough to report, and the spread between them is wide. Florida's floor is 7.1%, the highest in the table and above the national 7(a) floor of 3.1%. California's is 2.0%. Minnesota's is 0.00%, a floor on a modest cohort and not proof that its loans cannot fail. For every other state the column reads not reported, because the state's cohort holds fewer than 30 loans; that is an absence of measure, never a zero. A Florida study meets a lender that has seen this class lose money, and it should be written for that reader: debt service coverage tested at a lower census and a weaker payer mix than the base case, and the sponsor's liquidity set against a ramp that runs longer than planned. In California the lower floor sits beside the largest median in the table, which makes the size of each loan, rather than the frequency of loss, the lender's first concern.

The states below the cohort line

From Illinois down, the table prints loans and dollars but no median and no floors, and several of those states carry more money than their counts suggest. Washington's 24 loans total $20.5 million and Colorado's 22 total $21.0 million, both more than Illinois's $14.0 million on 24 loans. Idaho's 22 loans reach $15.3 million. Tennessee shows only 9 loans but $12.2 million, more than Georgia's 20 loans at $9.2 million, and New York's 17 loans total just $3.7 million. A lender in one of these states has a short record of its own to compare against, so the study must assemble its case from the subject's market and from the national floors, and should name which national figure it uses. Small counts also mean that one large loan can set a state's dollar total, which is why the study reads these rows for presence, not for typical size.

Count leaders and dollar leaders are different banks

The lender table is ranked by count, and the count leader is not the dollar leader. The Huntington National Bank leads the count at 84 treatment loans, worth $14.6 million in all; Wells Fargo Bank National Association is next at 53 loans and $15.5 million. Live Oak Banking Company approved 35 for $48.8 million, the largest dollar total the table prints, and Celtic Bank Corporation's 42 loans total $28.5 million, the second largest printed. Newtek Bank, National Association follows with 39 loans and $22.9 million and U.S. Bank, National Association with 31 and $18.4 million. TD Bank, National Association approved 38 loans with its dollar total withheld, so no figure is given for it here, and the page does not estimate one.

Large-ticket lenders

Live Oak Banking Company, Celtic Bank Corporation and Newtek Bank, National Association carry far more money per loan than the count leaders, and a book of that weight is where a study has to carry a full property and program case. A study written for a lender of this kind is a credit document: the cost and value of the building, the licensed capacity of the competing programs in the service area, a payer mix built from contracts rather than assumptions, a staffing plan priced at local payroll levels, and debt service coverage tested from the first licensed year onward. The size band table shows how much of the class sits at that scale: its top 7(a) band holds 94 loans and $330.2 million, more 7(a) money than any other band.

High-volume, small-ticket lenders

Smaller tickets define the books of Manufacturers and Traders Trust Company, 39 loans for $5.5 million, Northeast Bank (ME), 27 for $4.6 million, and JPMorgan Chase Bank, National Association, 28 for $7.9 million. The Huntington National Bank's 84 loans for $14.6 million follow the same pattern at larger scale. Books of that weight sit toward the small end of the size table, where 427 7(a) loans carry only $27.9 million. A study for a lender working at that scale is narrower, but not lighter. The question is less the building than the practice: whether its insurer contracts transfer to a new owner, whether the clinicians who produce the billings will stay, and how much of the reported income has actually been collected.

Loss rates the table does not give

No lender row prints a charge-off rate. Each reads either withheld or not reported, the latter where the lender's cohort is under 30 loans, and neither is a zero. Lenders with fewer than 5 loans in the class are left out entirely. A borrower cannot choose a lender from this page by its loss experience, and a study should not imply otherwise. What the table does show is which lenders have a working book in treatment lending and of what size, which tells a sponsor whose credit officers have read files like this one before. A lender with 84 approvals in the class will compare a new study with its own portfolio, and the study prepared for it has to answer the questions that portfolio has taught it to ask.

More providers, smaller ones

County Business Patterns counts 25,390 establishments in the class's industry codes in 2023, up from 20,516 in 2018, with 545,433 employees against 489,510 and $28.33 billion of annual payroll in 2023. Establishments grew faster than employment, so the average establishment employed fewer people in 2023 than in 2018, which suggests growth through more, smaller programs rather than larger ones. For a study that makes a competitor count a weak measure of supply. What a lender needs is licensed capacity and its use: the beds, slots or treatment hours of the programs that draw from the same referral sources and payers, and how full they run.

Arizona's build-out

Arizona's build-out is among the sharpest in the table. Its establishments rose from 585 to 1,031 and its employees from 16,334 to 24,512, and its 2023 payroll of $1.40 billion exceeds Ohio's $1.23 billion although Ohio has more establishments, 1,062, and more employees, 25,802. Arizona's SBA record follows the same arc: 61 loans for $57.7 million, a median of $548,000 and the lowest cancellation floor in the state table, 5.0%, with Phoenix-Mesa-Chandler holding 56 of the loans. A study for an Arizona program enters a market that has added providers quickly, and a lender will want the newest competitors counted by name and capacity, not inferred from older directories.

Large industries with thin SBA records

The Census ranking and the SBA ranking do not match. New York counts 1,278 establishments, 33,516 employees and $1.79 billion of payroll, the second largest payroll after California's $3.16 billion, yet SBA approved only 17 treatment loans there for $3.7 million. Pennsylvania has 1,141 establishments and $1.46 billion of payroll against 37 loans for $17.7 million, and Tennessee $976.8 million of payroll against 9 loans. Missouri's 514 establishments and $659.3 million of payroll stand beside 6 SBA loans. The Census count is not limited to the businesses SBA can lend to, so a thin SBA record says little about the size of the market; it says the lender has few SBA precedents in the state. A study there has to bring the market evidence the loan record cannot, and its comparison set will come from the state's licensed providers, not from SBA files.

New York and Pennsylvania shrank while Florida grew

New York's establishments fell from 1,364 to 1,278 and its employees from 35,317 to 33,516, and Pennsylvania's fell from 1,146 to 1,141 establishments and from 32,876 to 29,546 employees. Florida moved the other way, from 1,071 to 1,278 establishments and from 28,826 to 31,586 employees, so by 2023 Florida and New York count the same 1,278 establishments. Florida pairs that growth with 99 SBA loans and a 7.1% charge-off floor, the highest the state table reports. Growth in providers and losses on loans sit in the same state, and a Florida study has to address both: whether the subject's referral base is deep enough in a market that keeps adding programs, and whether its debt holds if census arrives slower than planned.

Where SBA lending runs ahead of the industry

Utah has 457 establishments and $445.1 million of payroll, smaller than Ohio, Pennsylvania or New York on every Census measure, yet its 65 SBA loans carry $69.9 million, more than any of those states. Minnesota's 764 establishments and $892.0 million of payroll sit behind 77 loans, third in the state ranking. New Hampshire's 126 establishments and $188.5 million of payroll stand beside 20 loans for $21.9 million, and Idaho's 189 establishments beside 22 loans for $15.3 million. In these states lenders have financed a large volume of treatment projects against a modest industry base, so the risk a study must test is less whether the market exists than whether it is already served: how many recently financed programs draw on the same referral sources, and whether the subject's census survives sharing them.

Payroll sets the staffing line

Clinical staffing, more than the building, drives a treatment budget, and the payroll column shows how much that cost moves between states. Connecticut's 12,862 employees drew $811.5 million of payroll in 2023, more than Texas paid its 16,342. Tennessee's 16,023 employees drew $976.8 million, more than Minnesota's 17,167 at $892.0 million. Connecticut's industry workforce also grew faster than any other state's in the table, from 8,343 to 12,862, alongside establishments rising from 308 to 430. These are March employment counts against annual payroll, so they are not wage rates, but the ordering is clear enough to warn against importing a staffing budget from another state. A study prices the clinical roster at the subject state's own pay level and checks it against the payroll the Census reports for the industry there.

Workforces that shrank

Employment in the industry fell between 2018 and 2023 in several states, even where establishments rose. Arkansas added establishments, 210 to 324, while its employees fell from 5,977 to 5,612. Delaware went from 1,593 employees to 1,183, Wyoming from 1,050 to 853, South Carolina from 2,632 to 2,413, Vermont from 4,741 to 4,500 and Hawaii from 2,096 to 1,835, and Alabama lost both establishments, 221 to 210, and employees, 4,632 to 4,554. A falling headcount does not prove a shortage of clinicians, but it is a reason for the study to show where the subject's staff will come from, at what pay, and how long hiring will take before a licensed program can open at the census the pro forma assumes.