SBA 504 Method: The Job Test, the Project Split and What a Study Must Show the CDC

A 504 financing packs three decisions into one closing: how much the owner puts in, how much a bank lends on a first lien, and how much a Certified Development Company lends behind the bank on a debenture that SBA guarantees. Each layer has its own rule, and one of the rules has nothing to do with credit at all, because the Project must also create or keep jobs, or serve a stated public goal, in proportion to the guarantee. This article reads 13 CFR 120 subpart H and the statute behind it, 15 U.S.C. 695 to 697, one rule at a time, and for each rule sets out what a feasibility study has to put in front of the CDC so the test can be run on evidence instead of on the sponsor's word. Readers who need the 7(a) half of the same regulation will find it in the companion article on 7(a) tests.

How a 504 Project Is Financed, and Where the CDC Sits

Permanent 504 financing, as 120.801(c) lays it out, starts with owner equity of at least 10 percent of Project costs; above that the CDC lends up to 40 percent of Project costs, with some administrative costs added, from Debenture proceeds and behind a second lien ^[1]. A Third Party Loan from a bank or other lender supplies the balance and takes the first lien ^[2]. The Debenture is the CDC's own obligation, guaranteed 100 percent by SBA, and its proceeds fund the 504 loan ^[3]. SBA's notice of September 30, 2025 describes the same structure from the bank's side, with the private lender holding a senior lien that covers at least 50 percent of project cost ^[4].

Congress wrote the program's purpose into 15 U.S.C. 695(a): economic development, and jobs created or preserved, delivered through long-term finance for small firms ^[5]. That purpose explains why a 504 file carries a job count that a conventional mortgage never would, and why the job count is tied to the size of the guarantee rather than to the size of the building.

Three parties hold a stake in a 504 file. The bank holds the first lien ^[2]. The CDC is responsible for closing the 504 loan, shares the Debenture closing with SBA ^[6], and services the loan afterwards ^[7]. SBA approves each loan made from Debenture proceeds ^[8]. A feasibility study has a place in that file by regulation: among the loan conditions SBA may impose, 120.160(b) names a feasibility study next to professional appraisals and a survey ^[9]. SBA's own 504 page lists "a feasible business plan" among the general eligibility standards, beside qualified management expertise, good character and the ability to repay ^[10].

That gives a study a narrow and useful job. It does not decide whether a business is eligible, whether it is small, whether credit is available elsewhere, or whether the loan should be made. It supplies the evidence each of those decisions rests on: the market the Project will serve, the cost of what is being built or bought, the staffing plan behind the job estimate, the space the business will occupy, and the cash flow that has to carry two loans at once. The sections below take the rules one at a time.

The program is large. In FY2025 the program recorded 6,762 ^[11] approvals, a gross count that keeps approvals later cancelled, for $7.80 billion ^[12] of debentures; the dollar figure counts only the SBA and CDC layer, not the bank's first lien. Readers deciding between the two SBA programs will find that comparison in SBA 7(a) vs. 504 Loans for Real Estate Development, and the development companies that originate most of the program's volume are examined in Who Really Leads SBA 504 Lending?.

What the 504 Loan File Shows Before Any Project Is Tested

Before testing one Project it helps to know how approvals in this program end. The Loan Analytics dataset follows each 504 approval in SBA's public loan files from approval to its last reported status, and Table 1, at the end of this section, prints the program figures this article relies on, with the two case classes beside them.

Three readings matter for method. The first is that cancellation, not loss, is the common way a 504 approval fails to become a working loan. In the FY2017 to FY2024 approval cohort 13.60 percent ^[13] of approvals are recorded as cancelled, and that rate is a floor: counting every approval still undisbursed as cancelled lifts it to a ceiling of 16.90 percent ^[14]. The dataset does not say why so many approvals stay undisbursed. One structural reason that would lengthen the tail is that a Debenture is issued only after the interim lender certifies its disbursements and the CDC certifies completion to the final plans and specifications ^[15], with an exception for minor components finished from an escrow under 120.961 ^{[16] [17]}. At the data's cut-off of 2026-09-30, 17.30 percent ^[18] of FY2024 approvals were still undisbursed.

The second reading is that loss is rare and arrives slowly. For the FY2017 to FY2021 cohort the charge-off rate to date is 0.35 percent ^[19] of disbursed loans, and 0.17 percent ^[20] were charged off within five years of approval, a fixed horizon that lets cohorts be compared fairly. Both are approval-cohort floors, because SBA withholds the status of open loans and 22,353 ^[21] open loans sit in that denominator with no outcome shown. The 7(a) floor for the same cohort is 6.33 percent ^[22], a contrast the 7(a) article takes up.

The third reading follows from the first two. In these cohorts cancellation outnumbers charge-off by a wide margin, so a study that spends all its effort on the stabilized year has spent it in the wrong place. The loan file cannot say why any single approval was cancelled, and a study should not pretend to know; what it can do is narrow the risks that sit inside its own scope: cost, the source and timing of the contribution, the permits and the opening schedule. The articles on cancellation versus charge-off and on the approval-cohort charge-off floor explain how each rate is built and why neither can overstate.

Loan Data Behind This Article

Every figure in this table comes from the Loan Analytics loan dataset, analytics-loan-data freeze-3 (commit 360c856), and carries its key. SBA rates are approval-cohort floors and are withheld when fewer than 30 loans stand behind them.

Eligible Project Costs: A Cost Basis a Reviewer Can Rebuild

Nearly every number in a 504 structure is a share of Project cost, so the cost basis is the first figure a study has to get right, and it should be laid out so that a reviewer can rebuild it. The regulation lets 504 proceeds pay costs directly attributable to the Project, including professional fees that are essential to it, such as title work, architectural and engineering costs, appraisals, environmental studies and the legal costs of zoning and permits ^[25]. A construction budget may hold back up to 10 percent of construction cost as a contingency for overruns ^[26], and repaying the interim loan, with its points, fees and interest, also counts as a Project cost ^[27].

The exclusions shape the file as much as the inclusions. Debenture money cannot pay for anything that is not both directly attributable and necessary to the Project ^[28], and the regulation names the usual suspects, from working capital and franchise fees to organization costs, advertising and management or counseling fees ^[29]. SBA's 504 page repeats the bar on working capital and inventory ^[30]. Machinery and equipment qualify if they are long-lived, with at least 10 years of useful life left ^[31]. Short-term equipment, furniture and furnishings stay outside the Debenture unless they are essential and a minor part of the Project, and cars, trucks and airplanes stay out altogether ^[32].

Administrative costs sit in a category of their own. The guarantee fee, the funding fee, the CDC's processing and closing fees and similar charges are not Project costs, yet the 504 loan and the Debenture may finance them ^[33]. That distinction drives the arithmetic later in this article: the contribution percentages apply to Project cost with administrative costs left out ^[34], while the Debenture is grossed up to carry the financed fees.

Two further rules are easy to miss. If part of the building will be leased to another business, the cost of finishing the interior of that space is not an eligible Project cost ^[35], so tenant improvements need a separate source. And the contingency is not a cushion the borrower keeps: when the Project is complete, the Debenture is cut by the amount by which the unused contingency exceeds 2 percent of the anticipated Debenture ^[36].

A study that supports a 504 file therefore prints the cost basis in three groups a CDC can check line by line. The first holds eligible Project costs, each with its source: a contractor's bid, a purchase contract, an equipment quote, an architect's fee proposal. The second holds the administrative costs the Debenture will carry. The third holds the costs the 504 loan cannot pay, with a statement of how the borrower will fund them. Opening working capital, a franchise fee and the launch marketing budget belong in the third group, paid from cash or a separate loan, and the operating projection must still carry them. A contingency set at the cap with no explanation is a weaker answer than a smaller one tied to the stage of design and the way the contractor has priced the work.

The Project Split: Contribution, First Lien and Debenture

One sentence in the regulation sets the usual shape: equity from the borrower, Third Party Loans and the 504 loan make up the permanent financing, typically in shares of 10 percent, 50 percent and 40 percent ^[37]. The word typically does real work, because three rules move the shares.

The first rule is the contribution. The borrower's minimum rises to 15 percent where the business is in its first two years and to 15 percent for a Project involving a building or structure of limited or single purpose ^[34]; it is 20 percent when both conditions apply and 10 percent in all other circumstances ^[38]. Land the borrower already holds can count toward the contribution, with its buildings and site improvements, when it becomes part of the Project Property ^[39]. The cash can be borrowed from the CDC or a third party at a reasonable rate ^[40]; if that loan is secured by Project assets, it must sit behind the 504 liens and cannot be repaid faster than the 504 loan without SBA's prior written approval ^[41]. A study that finds a borrowed contribution should carry its payments in the debt service.

The second rule holds the first lien up. In every Project the bank's loans have to match or exceed the 504 loan in amount, and in two cases they have to reach 50 percent of total Project cost ^[42]: where the business has two years of operation or fewer behind it, or where the Project buys, builds, converts or expands an asset of limited or single purpose ^[43]. The third rule caps the Debenture: "Generally, a 504 loan may not exceed 40 percent of total Project cost plus 100 percent of eligible administrative costs" ^[44]. SBA may raise the Project share to 50 percent for good cause shown ^[45], which is also the statutory ceiling on a loan funded by a guaranteed debenture ^[46].

Size limits frame the whole structure. No 504 loan may be smaller than $25,000 ^[47]. At the top, an ordinary Project faces an outstanding balance cap of $5,000,000 per borrower, counting affiliates ^[48]. The cap rises to $5,500,000 for each Project of a small manufacturer ^[49], meaning one with all of its production facilities in the United States ^[50], and for each Project that cuts the borrower's energy use by at least 10 percent or generates renewable energy or renewable fuels ^[51]. SBA's page gives $5.5 million as the program's maximum loan ^[52].

Table 2 lists the inputs for an illustrative $4,000,000 owner occupied Project, every one either sourced or labeled as an assumption, and Table 3 applies the three contribution rules to it.

Rounded as shown; lines may not foot by one unit in the last place.

Read across the columns, the first lien does not move. Because the Third Party Loans must still reach half of Project cost in the two special cases, the extra 5 or 10 points of equity come out of the 504 layer: the net 504 loan falls from $1,600,000 to $1,400,000 and then to $1,200,000, and its share of cost from 40.0 percent to 35.0 percent and 30.0 percent. In the standard column the net 504 loan sits exactly at the 40 percent ceiling, and the gross Debenture of $1,649,485 is larger only because it carries $49,485 of financed fees, which the ceiling allows.

Which column applies is not a matter of preference. Whether a business has operated two years or less is a question of when it began operating, which its own records show. The higher contribution for a limited or single purpose building is set without describing such a building in the sections cited here ^[34], so the classification turns on the facts of the asset, and that is where a study can help most, by describing the layout, the building systems and the fit-out, how much of the building could serve another occupant without major alteration, and how deep the market for that kind of property is. A tunnel car wash, with its equipment pits and water reclamation, lies toward one end of that range; a plain warehouse with a front office lies toward the other. A file that settles the classification late learns at the same moment that the borrower needs another 5 points of cost in cash, which on this example is $200,000.

The Job Test: What 120.861 Counts and How the Estimate Is Built

The regulation gives a Project two doors. It may count jobs under 120.861, or it may serve one of the community development and public policy goals in 120.862, and it has to pass through at least one of them ^[61]. Jobs are the default door. Under 120.861 each Project must produce one Job Opportunity, created or retained, for every so many dollars of 504 funding, with the dollar figure published and revised by SBA in the Federal Register ^[62], and the current notice sets it at $95,000 of SBA guarantee per job, except that "in the case of a Project of a small manufacturer or a project that meets an energy public policy goal" the figure is $150,000 ^[56]. The notice governs approvals dated October 1, 2025 and later ^[63]; the standard it replaced was $90,000 per job and $140,000 for a small manufacturing Project ^[64].

The statute puts a floor under the dollar amount. Its own figures are $65,000 of guarantee per job and $100,000 for a small manufacturer ^[65], and it bars the Administrator from adopting a lower amount by waiver ^[66]. SBA has set higher figures by notice, most recently moving from $90,000 to $95,000 per job ^{[64] [56]}, so a study should name the notice it tests against and the approval date it assumes, rather than quoting a dollar amount from memory.

What counts as a job is defined, and the definition in 120.802 has two halves. A created job must be permanent and full time or its equivalent, and it must be in place no later than two years after the 504 money arrives; a retained job is one kept in the community on account of the loan ^[67]. The statute measures created jobs from a different point, the completion of the project ^[68]. A study does not need to settle that difference. It needs to date every hire in its staffing plan so the CDC can count the jobs against whichever clock it applies, and it should flag any position that depends on a revenue level the projection reaches late.

The estimate should come out of the operating projection, not sit beside it. The staffing plan that sets the payroll line in the cash flow is the same plan that sets the job count, position by position, with hours, start dates and the volume of business each hire depends on. That discipline catches the most common inconsistency in a job estimate, a count the payroll cannot pay for. Part time positions are converted to full time equivalents on a stated weekly basis; the regulation itself divides part time hours by 40 when it counts employees for one kind of refinancing ^[57], and the example here borrows that divisor as a labeled assumption.

Retained jobs need a causal account. Because the definition credits only jobs the loan keeps in place ^[67], an expansion that holds an existing workforce in the community has to show the alternative: the lease that is ending, the capacity the old building cannot provide, the move the business would otherwise make. Relocation cuts the other way. A relocation that shrinks the moving business's workforce by a third or more makes the Project ineligible unless the CDC can justify it ^[69], so a study for a business that is moving should count the jobs that leave as carefully as the jobs that arrive.

Table 4 runs the test on the illustrative Debenture. The example divides the gross Debenture, the amount SBA stands behind, by the dollars per job; a study should state which amount it divided and how it rounded, so the CDC can check the figure against its own count.

Under the standard rule the Debenture of $1,649,485 calls for 17.4 job opportunities, and a plan of 13 full time and 8 part time positions supplies 17.0, a shortfall of 0.4. There are three honest ways to close a gap like that. The plan can add a position the business genuinely needs, which must then appear in the payroll and in the debt service test. The borrower can contribute more: at a 20% contribution the Debenture falls to $1,237,113, the requirement falls to 13.0, and the same plan shows a surplus of 4.0. Or the Project can qualify on another objective, which the next section takes up. Rounding a part time role up to a full time one on paper is not a fourth way, because the count has to survive the CDC's annual report of the jobs each Project actually provided or is estimated to provide ^[70].

The third column shows how much the dollar amount matters. At $150,000 per job the same Debenture needs 11.0 job opportunities, so a borrower near the line has every reason to establish early whether it qualifies as a small manufacturer or whether its Project meets an energy goal. Both are determinations for the CDC and for SBA, which approves each loan ^[8]; the study's part is to describe the business and the Project accurately enough that the classification can be made on the record.

The Second Route: Community Development and Public Policy Goals

A Project with too few jobs of its own can still qualify on one of the 120.862 goals, provided the CDC's whole 504 book, counting the new loan, meets or beats the job average the CDC must hold; the application also has to explain how the Project serves the goal it names ^[71]. The statute sets up the same exchange, letting a project that qualifies on such a goal skip the job criteria when the development company's portfolio as a whole meets them ^[72].

The goals are specific. The community development list has five entries: improving, diversifying or stabilizing the local economy, stimulating other businesses and bringing in new income ^[73], assisting manufacturing firms ^[74], and assisting businesses in Labor Surplus Areas ^[75]. The public policy list is longer and includes rural development, productivity and competitiveness, and facility upgrades for health, safety and environmental compliance ^[76]. For Projects that qualify this way the current notice sets the portfolio average at one job opportunity for every $150,000 guaranteed ^[77].

This route shifts the burden of proof without removing it. A study has to show the goal on the kind of evidence the goal itself implies. For rural development, that is the Project's location against the definition the CDC applies. For bringing new income into a community, it is the share of demand that comes from outside the trade area, which a sound market chapter has already estimated. For stimulating other business development, it is the suppliers, contractors and services the Project will buy locally, with an order of magnitude attached. A general statement that a Project is good for the local economy proves nothing a reviewer can test.

What a study cannot show is the CDC's portfolio position. The average is measured across the CDC's own book ^[71], so whether it has room for a Project that brings few jobs is a question only the CDC can answer, and a sponsor who plans to rely on this route should ask the CDC at the outset rather than learn the answer when the file is nearly complete.

Occupancy: The Leasing Rules and the Growth Projection

The borrower has to use the Project Property itself; the one exception lets an Eligible Passive Company hold the property and rent it to the Operating Company ^[78]. Passive owners that do not occupy or use the financed assets are otherwise ineligible for SBA business loans ^[79], and SBA's 504 page names speculation or investment in rental real estate as something the program will not finance ^[80]. The 504 program finances a building for the business that works in it, not a building for its rent roll.

The regulation turns that principle into ratios. In a newly built Project no less than three fifths of the rentable space has to be in the operating business's own use from the start, and tenants on long-term leases may hold as much as 20 percent ^[81]. The remainder cannot sit idle by design: the borrower has to plan on growing into part of the unleased, unoccupied space inside three years, and into the whole of it inside ten ^[82]. In an existing building the general SBA threshold is lower: the operating business must hold a bare majority, 51 percent of the rentable property, and the rest may go to tenants ^[83].

The statute explains what the plans are for. Where the project is a new facility, it allows up to 33 percent of the total project to be leased if "reasonable projections of growth demonstrate" a need for that space within three years and its full use within ten ^[84]. That clause is, in effect, a description of a feasibility study's growth chapter, and a study for a building with space held for future use should supply it.

A study for a 504 building should therefore carry a space program: the rentable area and how it is measured, the area the operating business occupies at opening, any area leased to others and on what terms, and the area held for growth with the dates the business expects to fill it. The growth projection behind the held space has to be the same projection that drives revenue and staffing. A business that projects a flat headcount cannot also claim it will need another bay in three years, and a study that lets those two chapters disagree has given the CDC a question it should have answered itself. Where space is leased to tenants, the tenant income belongs in the cash flow at market terms, and the cost of finishing that space has to come from outside the 504 loan ^[35].

Maturities, Rates and Fees: Pricing Two Loans

A 504 Project carries two debt service lines, and they follow different rules. SBA sets the rate on the 504 loan and its Debenture, subject to the Treasury Secretary's approval ^[85], and SBA's page says the rate is set at a margin over what 10-year Treasury issues currently yield ^[58]. SBA publishes the available maturities in the Federal Register ^[86], and SBA's page lists 10, 20 and 25 year terms ^[59].

The first lien has limits of its own. Its term is tied to the Debenture's, because a 10 year 504 loan needs a first lien running 7 years or longer and a 20 year 504 loan needs one of at least 10 years ^[87], and SBA publishes each quarter a maximum legal interest rate for the commercial loans inside a 504 Project ^[88].

The fees come in layers. For processing, the CDC "may charge up to 1.5 percent of the net Debenture proceeds to process the financing" ^[54]. Its monthly servicing fee falls between 0.625 percent and 2 percent a year of the unpaid balance ^[89], and a fee above 1 percent, or above 1.5 percent in a rural area, needs SBA's prior written approval ^[90]. On the SBA side there is a guarantee fee of 0.5 percent of the Debenture and a yearly fee capped at 0.9375 percent of the unpaid principal ^[55], and the bank's participation itself draws a one-time charge of 50 basis points when the bank sits senior to SBA ^[91]. Which fees the Debenture may finance is a matter for the regulation, and 120.883 lists them among the administrative costs ^[33]. SBA's 504 page adds, under its interest rate heading, a line that reads "Totals approximately 3% of the debt, rate may be financed with the loan" without saying what the total covers ^[53], so the example below treats its 3% financed share as an assumption.

Before the permanent structure exists, there is interim financing, and it may fund every Project cost except the borrower's contribution ^[92]. The rule says nothing about timing, but because no interim money can stand in for the contribution, a study should show where that cash sits and when it goes in, preferably before construction starts. The permanent money arrives last: the Debenture follows the completion certifications described earlier, with the escrow exception for minor components ^[16], and its proceeds can then retire the interim loan together with the points, fees and interest it carried ^[27].

For the study, the consequences are practical. A study's debt service test should price each loan separately, at a rate stated as of a date and tied to a published benchmark. SBA sets the Debenture rate with the Treasury's approval ^[85] and the Debenture is issued only after completion is certified ^[15], so at the study date its rate can only be estimated. The ongoing SBA and CDC fees on the unpaid balance are charged every year ^[55], so a debenture rate that leaves them out understates debt service. The construction period needs its own lines: interest on the interim loan, which the cost basis should carry as a Project cost, and the months of operation before the business reaches the occupancy and revenue the stabilized year assumes. The illustration below moves both rates together by a point to show how much of the cushion depends on the rate at sale.

Coverage on Both Liens

Conventions. This article follows the conventions of the Loan Analytics worked examples. Coverage is tested against the program floor the site's program page states, where the program has one, and against a stated 1.25x lender assumption, and the higher of the two binds. Equity is measured at the loan the cash flow supports, with fees recomputed on that loan. An acquisition is tested on the seller's historical earnings, not on projections.

The credit test for a 504 loan is the general one in 120.150: a creditworthy applicant, and a loan sound enough that repayment is reasonably assured ^[93]. Collateral is part of that judgment but not the deciding part. The statute treats the borrower's collateral, usually a junior lien on the financed property, as one factor among several in the credit decision ^[94], and the CDC usually takes a second lien on the Project Property, sometimes with additional collateral ^[95]. Repayment from operations carries the case.

For 504, the site's SBA program page puts the floor at 1.15x, measured on history or on projections two years out ^[60]. The site's worked examples set that floor beside a stated 1.25x lender assumption and test against whichever is higher. The 1.25x figure is an assumption, not a program rule, and a study for a particular file should replace it with the minimum that file's lender states; it is printed here so that the example says plainly what it is testing.

A study should print two ratios from one cash flow. The first measures the first lien alone, the loan that holds the senior position on the Project Property ^[2]. The second measures both liens together, because the Debenture stands behind the first lien in collateral ^[95] and both loans are paid from the same operating income. Printing both on the same net operating income lets the bank, the CDC and SBA each read the ratio that bears on its own position without rebuilding the other's number.

In the base column of Table 5 the stabilized net operating income of $410,000 covers the first lien 2.36x and both liens 1.33x, which clears the floor the site's SBA page states and the 1.25x lender assumption. Take a tenth off net operating income and combined coverage falls to 1.20x: still above the 1.15x floor, but below the lender assumption, so the higher test binds and the income is $14,904 short of it. The table prints the cure the conventions call for, the loan the cash flow supports, which here means $142,737 of added borrower cash applied against the Debenture.

A one point rise in both rates does nearly the same damage, taking combined coverage to 1.22x with $82,066 of added cash as the cure. The last column shows what a higher contribution buys. At a 20% contribution the Debenture and its debt service shrink, and combined coverage rises to 1.50x: the extra equity a both conditions file must bring also gives it the strongest coverage of the four columns.

Two cautions keep the test honest. First, the stabilized year is not the first year. A new business, the same case that raises the contribution, reaches stabilized income only after a ramp, and the study should show the months of shortfall and the reserve that funds them, because debt service starts before the income does. Second, a coverage ratio is only as good as the net operating income beneath it, which in a study comes from the market and operating chapters and never from the lender's target. A projection solved backward from the required ratio is easy to spot: its revenue climbs exactly as far as the debt needs and no further.

Car Wash and Self Storage in the 504 File

The 504 worked examples on this site, listed with the related pages below, cover an express car wash in Maryland, a storage yard for RVs and boats in New Jersey and an owner occupied retail building in Michigan. The first two fall in asset classes the dataset can measure nationally within the 504 program, and their figures read differently from the program's.

Car washes drew 997 ^[23] approvals under 504 from FY2017 to FY2026, with a median Debenture of $1,094,000 ^[23]. In the FY2017 to FY2024 cohort the class's cancellation rate is 16.46 percent ^[23], above the program's 13.60 percent ^[13], and its ceiling is 21.91 percent ^[23] once approvals still undisbursed are counted as cancelled. The charge-off rate to date for the FY2017 to FY2021 cohort is 1.00 percent ^[23], a floor with 199 ^[23] of the loans in its denominator still open and their status withheld. The class shares one industry code between car washes and truck washes, so the dataset cannot separate the two.

Self storage drew 946 ^[24] approvals under 504 over the same years, with a median Debenture of $890,000 ^[24]. Its cancellation floor is higher again, 19.15 percent ^[24], and counting its undisbursed approvals as cancelled lifts that to 23.76 percent ^[24]. Its charge-off rate to date is 0.00 percent ^[24]: no charge-off is recorded among the disbursed loans of the FY2017 to FY2021 cohort, but 267 ^[24] of them are open with their status withheld, so the figure is a floor and not a finding that the class cannot lose money. The dataset has no separate class for RV and boat storage, because no industry code isolates it; such facilities appear under self storage when the borrower used that code, and that is the class the New Jersey example reads.

These are national figures for a program and a class, not a forecast for any Project. What they tell a study is where a 504 file in these classes is most likely to stop. Both classes are cancelled more often than the program as a whole; car washes also carry a charge-off floor above the program's 0.35 percent ^[19], while self storage has none recorded. Either way, cancellation is the larger risk in both files, so the evidence a study should put first is the evidence that the Project will close and open as planned: a construction budget priced by a contractor, a site with its approvals in hand or a dated path to them, a contribution that is already in the bank, and an opening ramp the reserve can carry. The 504 data pages for car washes and self storage, and the asset class index, carry the full cells.

What the Study Does Not Decide

A study is evidence for decisions other people make, and a 504 file has more of those decisions than most. The borrower test sits in the regulation, which has the Lender or CDC "certify or otherwise show that the desired credit is unavailable to the applicant on reasonable terms and conditions" ^[96]. The statute adds a separate test at the CDC's level, allowing SBA to guarantee a Debenture only where the necessary funds are "not available to such company from private sources on reasonable terms", the company being the development company ^[97]. SBA approves each loan made from Debenture proceeds ^[8], so size, the eligibility of the business and of the Project, and whether a goal other than jobs is met are determinations for the CDC and SBA.

The decisions continue after approval. SBA may decline to close the Debenture if it finds an unremedied material adverse change, such as a deterioration in the borrower's financial condition ^[98], and the CDC keeps reporting the jobs each Project provided or is estimated to provide ^[70]. A study that set expectations it could not meet becomes, at that point, part of the record the borrower has to explain.

Three things in particular lie outside a study's reach. It does not value the property: any appraisal the file needs is a separate document, and a study that borrowed an appraisal's conclusion would be repeating it, not testing it. It does not certify the job count: it supplies the staffing plan and the basis for each position, and the CDC decides what it reports. It does not approve the credit: its task is to test whether demand, cost and operating income can carry the structure the sponsor has proposed, and to say so plainly when they cannot, because a study willing to report a shortfall is the only kind whose favorable conclusion carries weight.

A Checklist for the CDC, the Lender and the Borrower

A reviewer can test a 504 study against the rules above in a short list of questions.

• Cost basis. Are eligible Project costs, financed administrative costs and excluded costs shown separately, each line sourced, with any construction contingency inside the cap ^[26] and the funding of every excluded item stated?
• Split. Does the study name the contribution rule that applies and the reason, show the first lien at or above its floor and the 504 loan at or below its ceiling ^[44], and recompute all three from its own cost basis?
• Jobs. Does the staffing plan match the payroll line, date each position, convert part time hours on a stated basis, explain every retained job, and test the count against the amount in force for the approval date ^[63]?
• Other objective. If the Project qualifies through a 120.862 goal instead of its own job count, does the study prove that goal on evidence and record what the CDC has said about its portfolio average ^[71]?
• Occupancy. Is there a space program with occupied, leased and held areas, and does the growth projection behind the held space match the staffing plan ^[84]?
• Debt service. Is each loan priced separately against a dated benchmark, with the ongoing Debenture fees included and a rate shock shown ^[58]?
• Coverage. Are first lien and combined coverage computed on the same income and tested against the floor the site's SBA program page states and the lender's own minimum, with the ramp years shown ^[60]?
• Closing risk. Does the study show why the Project will reach completion, with priced construction, approvals, the contribution in hand and a reserve for the ramp, given that the Debenture generally follows the completion certification, with an escrow exception for minor components ^[16]?

Read this way, the 504 rules are less a list of hurdles than a specification for the study. Each rule names a number, and each number needs evidence that a reviewer at the CDC, the bank or SBA can trace to its source.

Related Pages on Loan Analytics

• The Loan Analytics methodology series (all articles)
• How a feasibility study is done: the Loan Analytics methodology
• SBA feasibility studies (7(a) and 504)
• Loan data by state
• Loan data by asset class
• Worked examples (case studies)
• Worked Example: SBA 504 Express Car Wash Feasibility Study in Maryland
• Worked Example: SBA 504 RV and Boat Storage Feasibility Study in New Jersey
• Worked Example: SBA 504 Owner-Occupied Retail Feasibility Study in Michigan
• SBA 7(a) vs. 504 Loans for Real Estate Development: A Comprehensive Guide
• Who Really Leads SBA 504 Lending? An Audit of the Nation's Certified Development Companies
• SBA 7(a) Method: What 13 CFR 120 Tests and How a Feasibility Study Answers Each Test
• The Approval-Cohort Charge-Off Floor: Why a Published SBA Loss Rate Can Only Understate

Commission a Feasibility Study

Loan Analytics prepares lender-grade feasibility studies for SBA 7(a), SBA 504, USDA Business and Industry, conventional and EB-5 financing, built on the methods this series sets out. A study is not an appraisal: it tests whether a project's market, costs and cash flow can carry the debt a lender is asked to approve. To discuss a study for your project, contact Loan Analytics.

Daniel Smith, MAI, Loan Analytics