USDA Feasibility Studies Under 7 CFR Part 5001: The Regulatory Spine

Michal Mohelsky, J.D. | MMCG Analytics | September 2026 Controlling text: 7 CFR Part 5001, current through 90 FR 57351 (Dec. 11, 2025); guarantee percentages and fees from the FY 2026 annual notice, 91 FR 11272 (Mar. 9, 2026) (1)(5)(24)

New to feasibility studies? Start with the overview.

Companion analysis: The 37 factors of 7 CFR Part 5001, enumerated in the blog.

This guide is written for the three people who read a feasibility study inside a USDA OneRD guaranteed loan file: the lender's underwriter who decides what the file needs and writes the credit evaluation, the appraiser whose value conclusion the study has to reconcile to, and the sponsor whose projections it tests. It is organised around the one question the Agency asks before it issues a conditional commitment, and the one question a loss reviewer asks after the borrower stops paying: can the project service its debt, and does the file prove it.

Two premises run through everything below, and both are stated up front because they are often misstated.

First, USDA does require a feasibility study, and that is the opposite of the SBA position, but the requirement is narrow, programme-specific and easy to overstate. It is not a blanket requirement across the four OneRD programmes and it is not attached to asset classes. Business and Industry requires an independent feasibility study for a guaranteed loan greater than $1,000,000 to a new business, under 5001.306(a)(3)(i), and only then (19). Community Facilities requires a financial feasibility report in every case and an independent feasibility study above $1,000,000 to a new entity or new activity, under 5001.304 (17). Water and Waste Disposal requires no feasibility study at all; it runs on engineering documentation under 5001.305 (18). REAP requires tiered technical reports scaled to project cost, not a feasibility study, for most projects (20). A file that assumes one programme's trigger governs another, or that presents the five-component feasibility framework as a WWD requirement, is built on a misreading.

Second, there are two feasibility appendices in Subpart D, not one, and they are different instruments. Appendix A, titled Feasibility Study Components, organises the full feasibility study under five components and thirty-seven enumerated factors and is the definitional anchor at 5001.3; it attaches to B&I and to REAP renewable-energy systems (7)(21). Appendix B, titled Financial Feasibility Reports, is the Community Facilities instrument, cited separately in 5001.304 (17)(21). A live dispute in the search results has one provider insisting the rule names five components and no thirty-seven factors; both numbers are correct and describe different levels of the same appendix. Section 4 settles it from the codified source.

A note on the text and its currency. Part 5001 is codified and changes only by rulemaking. It is current through the technical amendment at 90 FR 57351, effective December 11, 2025, and no 2026 document has amended the codified text (1)(5). The substantive 2024 rule, published at 89 FR 79698 and effective November 29, 2024, rewrote provisions across all six subparts but moved no feasibility threshold; its only feasibility-adjacent change was a terminology swap from new entity to new business at 5001.306, with the $1,000,000 trigger intact (3)(19). This guide cites the regulation by section and appendix, which are stable, cites the Federal Register by citation and page, and flags every figure that is set by annual notice rather than codified.

The trigger is the date the complete application is received. Under 5001.1(b), Part 5001 governs every complete application received on or after October 1, 2020; earlier applications were processed under the four predecessor programme regulations (1)(2)(27). A file received after that date is governed by the live Part 5001 text as amended, and a loss reviewer measures the file against the text in force when the loan was made.

The lineage matters because the most common error on this subject is to cite the 2020 rule as the live regulation, and the second most common is to treat the 2024 overhaul as having moved a feasibility threshold. Neither is right.

One CFR-amending action is pending, RIN 0572-AC66, OneRD Guarantee Loan Round 5, listed in the Unified Agenda in Final Rule Stage (38). The FY 2026 fee notice and any information-collection notices are notices, not codified amendments, and do not change the rule. On any deal in the queue while a rulemaking is pending, the lender should confirm the governing text before the third-party reports are ordered.

The common error is to start with a lender checklist or an agency handout. The feasibility study enters the chain at the regulation, is prescribed by an appendix, is consumed by the lender's credit evaluation, and is enforced through the loss claim.

The statute. The four programmes rest on three statutes: 7 U.S.C. 1926(a) for Community Facilities and Water and Waste Disposal, 7 U.S.C. 1932(a) for Business and Industry, and 7 U.S.C. 8107 for REAP (37). None of the three mentions a feasibility study. The statutes authorise the guarantees and set the outer limits; the feasibility requirement is a regulatory creation.

The regulation. Part 5001 is the single regulation for all four programmes, and it is where the study is required, defined and prescribed (1). The definition sits at 5001.3: a feasibility study is a report including an opinion or finding by an independent qualified consultant evaluating the economic, market, technical, financial and management feasibility of the proposed project or operation in terms of its expectation for success, as outlined in Appendix A to Subpart D (7). The same section defines financial feasibility as the ability of a project to achieve sufficient income, credit and cash flow to financially sustain the project over the long term and meet all debt obligations, and defines qualified consultant as an independent third party with the knowledge, expertise and experience to perform the specific task required (7). The programme-specific triggers sit in Subpart D: 5001.304 for CF, 5001.305 for WWD, 5001.306 for B&I and 5001.307 for REAP (17)(18)(19)(20). The residual authority sits at 5001.303(c)(4): where the Agency cannot determine a basis for successful repayment from the lender's analysis, the business plan or other project information, or where the project will significantly affect existing operations, it may require an independent feasibility study using the applicable Appendix A elements (16).

The appendices. Appendix A to Subpart D, Feasibility Study Components, is the codified content standard for the full study; Appendix B, Financial Feasibility Reports, is the CF instrument (21). Both were codified as page images in the 2020 final rule at 85 FR 42518, which is why they do not extract as text from the eCFR and why every published transcription of them, including MMCG's own, has to be taken from the source graphics (2)(21)(39). Section 4 sets out what they contain.

The lender's credit evaluation. Under 5001.202 the lender prepares a written credit evaluation on the Five Cs of credit, and the Content list at 5001.202(b)(6) requires a separate written evaluation of the feasibility study, business plan, technical report and engineering and architectural reports as applicable (12). The study is evidence for that evaluation. It is not the evaluation, and the lender's judgment over it is non-delegable: 5001.6(b) provides that contracting for a service does not relieve the lender of its responsibilities, and 5001.201 provides that lender approval does not constitute Agency approval (11).

The Agency's review. Unlike the SBA model, the Agency reviews the file before the guarantee issues. An incomplete application is returned under 5001.303, the Agency sets the scope of any study it requires, and the conditional commitment under 5001.451 issues only on conditions the Agency states (16)(23)(36). The study is therefore judged twice: by the Agency on the way in, against the codified content, and by a loss reviewer on the way out, against a prudence standard.

The enforcement layer. What makes the requirement real is the layer of consequence. Section 5001.521(d)(1) provides that negligent loan origination and negligent loan servicing will result in a reduction of loss claims payable, and that the extent of the reduction, which could be a total reduction, depends on the extent of the losses caused (28). Section 5001.6(c) is the umbrella: where a lender fails to comply with the Part, the Agency may reduce any loss payment (11). Negligent loan origination is defined at 5001.3 as the failure to perform the services or actions a reasonably prudent lender would perform in originating its own unguaranteed portfolio (7). A thin credit evaluation resting on an inadequate feasibility analysis is exactly the origination lapse that clause reaches.

Put the layers together and the position is this. The USDA feasibility study is mandatory above programme-specific thresholds, prescribed to a codified list, prepared by an independent consultant, reviewed by the Agency before the guarantee issues, and then relied on by a loss reviewer who will reduce the claim if it did not support the repayment case a prudent lender relied on. That is a list plus a judgment, which is a harder standard than either alone: the list can be satisfied by a document that exists, and the judgment is satisfied only by a document that would have persuaded a competent underwriter at the time.

One mandate sits nearby and should not be confused with the study. For collateral over the thresholds in 5001.203, the lender must obtain a USPAP-compliant appraisal by an independent qualified appraiser, and where that appraisal contains any business or going-concern value, the lender must deduct it from the reconciled market value before discounting (13). That is an appraisal requirement answering a value question the feasibility study does not, and Section 7 returns to it.

The practical corollary for anyone writing or reading a study: state the trigger precisely. A B&I study is required by 5001.306(a)(3)(i) for a loan over $1,000,000 to a new business; a CF financial feasibility report is required by 5001.304 in every case; a study on any other file is required only where the Agency invokes 5001.303(c)(4) or the lender's own prudence calls for it. A credit memorandum that cites the wrong section, or none, invites the question a reviewer will ask first.

Because the mandate is programme-specific, the useful question is not "does USDA require a feasibility study" but "what does each programme require, at what threshold, and what sits behind the bright line." The answer differs across the four, and the residual Agency authority stands behind all of them.

Three observations on the table.

The first is that USDA mandates the broad document and prescribes it, which is the inverse of the SBA design. The SBA compels the narrow, credentialed documents, the appraisal, the valuation, the completion statement, and leaves the unbounded market question to lender judgment because no template guarantees a good answer to it. USDA can afford a list because the Agency reviews the file before the guarantee issues; the list tells the Agency what to look for, and the Agency's acceptance of the study is itself a gate.

The second is that the bright lines under-describe when a study will actually be required. The B&I trigger turns on the definition of new business at 5001.3, which reaches a business in operation less than one full year, or one that has not achieved full operational capacity or stable operations as determined by the Administrator; that last clause is a discretionary hinge, and a two-year-old business that has not stabilised can still draw the requirement (7). The words acceptable to the Agency in 5001.306 mean the study's preparer and content are subject to Agency judgment, not merely the lender's. And the residual authority at 5001.303(c)(4) means any file whose repayment case the Agency cannot otherwise support can be sent for a study. A lender that treats the $1,000,000 line as the whole test will be surprised.

The third is that a feasibility study satisfies none of the other mandated documents. It does not establish collateral value, which is the appraisal's job under 5001.203; it does not supply the financial statements 5001.9 requires; it is not the engineering documentation a WWD file runs on; and it is not the technical report a REAP file needs at its cost tier. And a CF financial feasibility report to Appendix B is not a B&I feasibility study to Appendix A. A file that supplies one where the programme calls for the other is deficient on the regulation's own terms.

This section is set out from the codified source rather than from any secondary transcription.

Appendix A to Subpart D is titled Feasibility Study Components (21). It is organised under the five feasibility dimensions the 5001.3 definition names, economic, market, technical, financial and management, wrapped by three framing sections: an Executive Summary describing the nature and scope of the project, its purpose, location, design features, capacity and estimated capital costs, together with a summary of the feasibility determination on each component; a Recommendation; and the consultant's Qualifications (21). Each of the five component headings carries a one-line statement of what the component is, followed by an enumerated list of factors to consider. By direct count of those enumerated factors there are thirty-seven: five under economic (the minimum inputs to operate successfully, contracts in place and to be negotiated, environmental risks, cost relative to the increase in revenues or benefits, and overall economic impact), six under market (competition, service or product or commodity type, new versus established target market, captive versus competitive end users, by-product revenue streams, and industry risk), nine under technical, twelve under financial, and five under management (21)(39). The five-component structure and the thirty-seven-factor enumeration are not competing claims. The factors live inside the components, and a provider who says the count is five is counting headings while a provider who says thirty-seven is counting the lines under them.

Appendix B is a different instrument. Titled Financial Feasibility Reports, it is cited separately in the regulation, attaches to Community Facilities under 5001.304, and specifies the content of the CF financial feasibility analysis and of the financial feasibility study with examination opinion (17)(21). Appendix A anchors the definition at 5001.3 and attaches to B&I under 5001.306 and to REAP renewable-energy systems under 5001.307; WWD requires neither (7)(18)(19)(20). Treating Appendix B as a subset of Appendix A is a misreading of 5001.304, which cites only Appendix B, and of the definition at 5001.3, which cites only Appendix A.

Neither appendix has been amended. Both carry the single source line of the 2020 rule, the eCFR change timeline for each shows one entry at October 1, 2020, and the amendments at 86 FR 70354 and 89 FR 79704 touched the citing sections, 5001.304 and 5001.306, not the appendices (3)(5)(21). The appendices were codified as page images, which is why a search snippet cannot resolve the count and why the transcription MMCG published in July 2026 was taken from the Federal Register graphics rather than from any prior transcription (39).

The practical consequence for a preparer is that a B&I or REAP study is written to Appendix A and a CF report is written to Appendix B, and the two are not interchangeable. The consequence for a lender is that the written evaluation of the study required by 5001.202(b)(6)(ii) should walk the components the applicable appendix names, because that is what the Agency will check the study against (12).

The feasibility study is prepared for the lender's credit evaluation, not for the borrower, and it is evidence that discharges specific credit-evaluation duties rather than a substitute for the lender's own analysis. Section 5001.202 requires the lender to prepare a written credit evaluation using credit documentation procedures and an underwriting process consistent with generally accepted prudent lending practices and with the lender's own policies, addressing the Five Cs at 5001.202(b)(1) through (5), and a Content list at 5001.202(b)(6) that specifies what the written evaluation must physically contain (12).

Where the study lands inside that structure is fixed by what it proves. The Conditions factor at 5001.202(b)(5) is the study's home: the lender is to consider the availability and depth of the resource or feedstock market, the analysis of current and future market potential and off-take agreements, competition and project type, infrastructure and environmental considerations, technical feasibility and demonstrated performance of the technology, construction complexity and the contractor's strength, contracts and permits, counterparty creditworthiness, and industry public-policy issues, and when applicable the lender should submit supporting documentation, a feasibility study, a market study, preliminary architectural or engineering reports, in accordance with 5001.304 through 5001.307 (12). The Capacity factor at 5001.202(b)(2) is the second home: the borrower's ability to produce sufficient cash to repay, including the feasibility and likelihood of the project and borrower to produce sufficient revenues to service the debt over the life of the loan, supported by a cash flow analysis (12). The Content list then requires the separate written evaluation of the study at (b)(6)(ii) and the spreadsheets and ratio analysis of the financial statements at (b)(6)(iii) (12).

Two features of the numeric regime are routinely misstated. First, the debt service coverage ratio is defined at 5001.3 as earnings before interest, taxes, depreciation and amortization less reasonably expected replacement capital expenditures, divided by annual debt service, and USDA nowhere further defines reasonably expected replacement capital expenditures (7). That is not the conventional market formula. A study or credit memorandum that runs plain EBITDA over debt service and calls it the USDA ratio has used the wrong numerator, and on a capital-intensive asset the difference is material. Second, there is no minimum debt service coverage ratio anywhere in the credit-evaluation regime; 5001.202 relies on the lender's own prudent-lending standard and states no floor (12). The only numeric coverage minimums in the entire Part are eligibility tests for refinancing at 5001.102(d): 1.1 times historical where the majority of the loan refinances existing debt, and 1:1 on current income to demonstrate recovery (8). A file that cites a "1.25x USDA requirement" as if it were codified invites the question of which section, because none states it.

Financial information carries its own rule that the study must respect. Under 5001.9, financial statements must be prepared in accordance with accounting practices acceptable to the Agency, and a tax return is not an acceptable financial statement for underwriting, though tax-return information may be used to prepare statements and to determine REAP eligibility (26). A study whose historical base is a tax return rather than a statement has built on a document the regulation does not accept.

For the feasibility study the division of labour is clean. The lender computes the coverage test on the 5001.3 basis, verifies the equity, structures the debt and evaluates the study in writing. The appraiser supplies the collateral market value. The study supplies the demand, revenue and cost assumptions that feed projected cash flow, and tests them against the trade area. It does not certify coverage, value or injection, and a study that purports to has overreached in a way the Agency and a reviewer will both notice.

Independence is built into the study at the definitional level, which is the second sharp difference from the SBA position. The 5001.3 definition requires an independent qualified consultant, and defines qualified consultant as an independent third party with the knowledge, expertise and experience to perform the specific task required (7). The B&I section adds that the consultant must be acceptable to the Agency (19). For Community Facilities the financial feasibility study with examination opinion must be prepared under AICPA attestation standards by a preparer carrying professional liability insurance, which is the one place in the Part where a professional standard and an insurance requirement attach to feasibility work by name (17).

What the text does not require is as important as what it does. It names no credential: no MAI, no CPA for the B&I study, no accreditation body. It sets no experience threshold in years or engagements, in contrast to the energy auditor, energy assessor and inspector definitions at 5001.3, which carry explicit numeric experience tests (7). The asymmetry is deliberate: the Agency reserves the acceptability judgment to itself under 5001.306 rather than delegating it to a credential. A preparer or a lender who asserts that a particular designation is required has imported a rule the regulation does not contain.

The conflict-of-interest regime is minimalist on its face and substantive in its definitions. The entire operative text of 5001.208 is one sentence: the lender must report all conflicts of interest, in writing, to the Agency (15). The catalogue of what counts as a conflict lives at 5001.3, which names a person acting as compensated agent of both borrower and lender on the same loan, distributions of loan funds to a borrower's owner or immediate family, and refinancing debt owned by a loan packager, broker or referral agent or its affiliates (7). For a feasibility preparer the practical bar is the definitional one: a consultant with a brokerage, development, financing or equity interest in the project is not an independent third party, and a lender that uses one has a study that fails the definition before anyone reads it.

On whether one firm may perform both the appraisal under 5001.203 and the feasibility study on the same project, the regulation is silent, and the silence should be reported as silence rather than filled with an inferred prohibition (13). The nearest constraints are the independence qualifier in each instrument, the independent qualified appraiser of 5001.203(d)(1) and the independent qualified consultant of 5001.3, and the reporting duty at 5001.208 (7)(13)(15). Nothing categorically bars the dual role, but each engagement must independently satisfy its own independence standard, the lender must report the arrangement in writing, and role separation is the conservative default because a firm that both values the collateral and opines on the project's viability creates an evident appearance problem even where no bright-line rule is broken.

The appraisal and the feasibility study are different instruments serving different credit factors, and collapsing them is the error this section exists to prevent. The appraisal discharges the collateral-value duty at 5001.202(b)(4)(i), under which the lender must determine the market value of collateral as established by an appraisal in accordance with 5001.203; the feasibility study addresses repayment viability under the capacity and conditions factors and is triggered separately under 5001.303(c)(4) (12)(13)(16). Neither reaches the other's conclusion, and a study that opined on collateral market value would not satisfy 5001.203, because it is not a USPAP Standards 1 and 2 appraisal by a qualified appraiser and would not survive the appraisal review report the lender must file under 5001.203(g) (13).

Two provisions of 5001.203 bear directly on how a study is written. The first is the going-concern deduction the 2024 rule added: where an appraisal contains any value attributed to business valuation or a going concern, that value must be deducted from the reconciled market value before discounting (3)(13). On a special-purpose asset, a hotel, a care facility, a fuel-and-convenience site, which is frequently appraised on a going-concern premise, the appraisal's income approach assumes a stabilised operation, and the study tests whether the trade area supports that stabilised operation. In a clean file the study's stabilised net operating cash flow is consistent with the income-approach inputs the appraiser used; a study projecting revenue materially above the appraiser's stabilised revenue is a file with two inconsistent documents, and the reviewer will read the lower one. Note the boundary against SBA practice: USDA requires the deduction and locates it before discounting, but it does not require the appraiser to allocate value across land, building, equipment and intangibles, which SBA expressly requires on special-purpose property; a preparer working both programmes should not port the SBA allocation mandate into a USDA file (13). The second is the construction rule: for a construction project the lender must obtain both an as-is market value and a prospective market value as of the date of completion, and the study's stabilisation curve should reconcile to the as-completed premise, not the as-is one (13).

The dependency runs one way. The study consumes the appraisal's value conclusion; the appraisal does not depend on the study. A study commissioned before the appraisal has a number is a study that will be revised, or worse, that contradicts the appraisal in the file.

The application process runs through 5001.301 to 5001.303: the lender begins the application, may request a preliminary eligibility review under 5001.302, and submits the complete application under 5001.303, which the Agency returns if incomplete (16). The conditional commitment under 5001.451 is the Agency's conditional undertaking to issue the loan note guarantee on stated conditions, not an approval of the loan, and 5001.408 separately bars a lender from representing a conditional commitment as a guarantee; the loan note guarantee itself issues under 5001.453 after the conditions precedent are met and the loan closes (23). The study must already be in the file at the application stage under 5001.303, which is well before the conditional commitment and long before the guarantee issues (16).

Two lender-ordered reports gate the study. The environmental review runs under 5001.207, which makes the lender responsible for ensuring the Agency's environmental requirements under 7 CFR part 1970 are met, and a project that violates part 1970 is ineligible under 5001.115 (9)(14)(35). The appraisal runs under 5001.203 and must be acceptable to the Agency approval official before the guarantee issues (13). The study consumes both: contamination, remediation cost and engineering controls affect marketability and cash flow, and the appraisal's stabilised income and as-completed value are what the study's revenue build must reconcile to.

The calendar, term sheet to guarantee, with the dependencies that matter for the study:

The sequencing point most often got wrong is step 4. A study commissioned at step 1, before the eligibility position is settled and before the appraisal has a number, is a fee spent on a deal that may not be eligible and a document that will have to be reconciled after the fact.

Everything above is enforced through one process, and it starts after the borrower stops paying. The lender files a report of loss under 5001.521, the Agency audits it and must be satisfied the lender has maximised collections, and the claim is paid up to the guarantee percentage of eligible loss (28). The guarantee percentage is a ceiling of 90 percent under 5001.407 and an operative rate set each fiscal year by notice under 5001.10; for FY 2026 the notice at 91 FR 11272 sets 85 percent for B&I under $5 million, 80 percent for B&I of $5 million or more and for CF and REAP, and 90 percent for WWD and for isolated rural Alaska B&I (22)(24)(27). Those are notice-set figures and will change.

The reduction chain is the proof that the guarantee is conditional. Section 5001.521(d)(1) provides that negligent loan origination and negligent loan servicing will result in a reduction of loss claims payable, and that the extent of the reduction, which could be a total reduction, depends on the extent of the losses caused; 5001.521(d)(2) attaches the same total-reduction exposure to non-compliance with the design requirements at 5001.205(a) and the WWD engineering requirements at 5001.305(a); and 5001.6(c) is the umbrella authority to reduce any loss payment for non-compliance (11)(14)(18)(28). Negligent origination is the reasonably-prudent-lender standard at 5001.3 (7). Read against a feasibility file, the exposure is specific: a credit evaluation that relied on a study which adopted the sponsor's revenue and called it tested, or on a study by a preparer who was not an independent third party, or on no study where the programme required one, is an origination the Agency can characterise as negligent, and the claim can be reduced to zero if that origination caused the loss.

What the public record shows is thinner than the SBA's, and it should be reported as thin. USDA has no published equivalent of the SBA Inspector General's loan-level review programme naming unsupported projections as a recurring deficiency. It does not publish a granular programme-level series of realized default rates, delinquency, loss claims paid and recoveries for the OneRD cohort, and it does not publish how often it invokes 5001.521(d) (29). There is no public loan-level finding attributing a loss-claim reduction to an inadequate feasibility study. The requirement exists in the regulation; the precedent is not published.

What is published points the same way. The nearest quantified loss expectation is the credit subsidy rate in the OMB Federal Credit Supplement, a net present value budgetary cost per dollar of guarantee that must not be read as a realized default rate: for B&I it ran 2.29 percent in FY 2024 and 0.20 percent in FY 2025, with estimates of 1.09 percent for FY 2026 and 1.93 percent for FY 2027, while CF, WWD and REAP guarantees carried negative rates in FY 2024 and FY 2025, meaning they were scored as net moneymakers for the government (31). Portfolio delinquency was exposed publicly at scale for the first time in 2026, when the Lender Lens dashboard on the Rural Data Gateway, launched January 19, 2026, made the commercial guaranteed portfolio downloadable to the individual loan with a named holding lender and a delinquency category, and on May 12, 2026 Rural Development removed ten lenders from the programme citing roughly $620 million in delinquent loans, about 47 percent of the agency's delinquencies, identified through desk audits (29)(30). Per-lender figures circulating in secondary reporting are provisional and subject to appeal and should be treated as unverified at the institution level (30). One data-hygiene rule governs any figure drawn from this record: direct loans dwarf guarantees in three of the four programmes, so a number that mixes direct with guaranteed, or an appropriated allotment with actual obligations, overstates the guaranteed record by multiples.

What a study causes, and what it cures. Read against that structure, the feasibility study is two-sided. Done badly, it is the source of the exposure: a study that adopts the sponsor's revenue is exactly the document that carries a thin credit evaluation past the Agency and into a loss claim the Agency can reduce, and a study prepared by a party with a fee interest in the closing converts the study itself into an independence defect. Done properly, the study is the cure: it supplies the trade-area test, the supply inventory and the stressed coverage that a reasonably prudent lender would have required, it puts a current independent document in the file where 5001.306 or 5001.304 demands one, and it gives the written evaluation under 5001.202(b)(6)(ii) something to evaluate.

USDA prescribes the study's structure through Appendix A, but the appendix is a floor. Two standards actually judge a study, the codified content on the way in and the prudence standard on the way out, and a study that satisfies the list can still be a poor one. The following is the standard MMCG builds to, set out as method so that a lender ordering a study from anyone can hold it to something.

The Appendix A structure, in full. An Executive Summary describing the nature and scope of the project, its purpose, location, design features, capacity and estimated capital costs, and summarising the feasibility determination on each component. The five components, each addressed under its own heading with the enumerated factors walked and answered rather than listed: economic, market, technical, financial and management. A Recommendation. The consultant's Qualifications (21). For a CF file, the Appendix B content in place of, or alongside, Appendix A, and where the section calls for it the financial feasibility study with examination opinion under AICPA attestation standards from an insured preparer (17)(21). A study that omits a component the appendix names, or that answers the factors with a heading and a sentence, will not survive the Agency's check against the appendix.

Independence, stated and structural. The study is engaged by, or at the direction of, the lender; it is prepared for the lender's reliance; the fee is fixed and not contingent on the finding; and the firm has no brokerage, development, financing or equity interest in the project. The 5001.3 definition requires an independent third party, and the reviewer will test it (7).

Every claim cites a document or is derived. Nothing rests on a conversation. A franchisor's number comes from the disclosure document; a traffic count from the state transportation department; a lease term from the executed lease; a construction cost from the executed contract or the signed budget; a management assertion from a signed management representation. A study that cites "per the operator" for a material assumption has an unsupported projection at its core with a citation attached.

The trade area is drawn, not assumed. The primary trade area is defined by drive time or distance against the asset's actual demand geography, mapped and defended. A ten-mile radius is not an analysis.

Demand is estimated from primary data. Population, households, income and employment from the Census Bureau's decennial and ACS releases; traffic from state and local counts; industry demand drivers from the federal series appropriate to the class. Each figure carries its source, vintage and geography.

Supply is inventoried, not summarised. Every competing facility in the trade area is identified by name, location, size and, where observable, rate and occupancy; the pipeline is identified from permits, planning approvals and announced projects. The capture rate is derived from the inventory, not asserted.

The revenue build reconciles to the appraisal. Stabilised revenue reconciles to the appraisal's stabilised income and, on a special-purpose asset, to the going-concern appraisal's value net of the business value the lender must deduct under 5001.203 (13). The two documents meet or the difference is explained.

Expenses are benchmarked. Operating expense ratios are benchmarked against published industry data for the class and reconciled to the sponsor's budget line by line, with departures explained.

The stabilisation curve is explicit. Month-by-month or quarter-by-quarter ramp to stabilisation, with the absorption assumption stated and supported from the supply inventory and the demand estimate, and reconciled to the as-completed premise on a construction deal.

Coverage is stressed on the USDA basis. Debt service coverage is computed as EBITDA less reasonably expected replacement capital expenditures over annual debt service, the 5001.3 definition, under the base case and under downside cases, with the revenue shortfall that breaches coverage stated as a number (7). A study that reports coverage at stabilisation and nowhere else has not tested anything, and a study that runs plain EBITDA has used the wrong ratio.

The conclusion is one of three. Feasible as proposed; feasible subject to stated conditions; or not feasible. A study that cannot reach a conclusion is not a study, and a study that reaches feasible on every engagement is not independent.

The register is an appraisal report's. Short, precise, sourced. No narration of what an exhibit already shows.

A lender who receives a study meeting that standard has a document that will pass the Agency's check against the appendix and carry the credit evaluation through a loss review. A lender who receives a formatted business plan with a market section has not.

The single most common factual error on this subject is to present the feasibility study as an SBA mandate for whole asset classes, and to attribute USDA's five-component framework to the SBA. The two programmes are built on opposite principles, and the difference decides how a study has to be written.

The consequence for a preparer working both programmes is that a USDA study is written to a list and an SBA study is written to a judgment. The USDA list is a floor, and a study that satisfies it may still be a poor one; the SBA judgment has no floor, and a study that would satisfy the USDA list is usually the safest way to meet it. The companion MMCG spine on SBA feasibility studies under SOP 50 10 8.1 sets out the SBA side in full (40).

Does USDA require a feasibility study? Sometimes, and it depends on the programme. B&I requires an independent study for a guaranteed loan over $1,000,000 to a new business under 5001.306(a)(3)(i); CF requires a financial feasibility report in every case and an independent study over $1,000,000 to a new entity or new activity under 5001.304; WWD requires none; REAP requires tiered technical reports (17)(18)(19)(20). The Agency can require a study on any file under 5001.303(c)(4) (16).

Is it true the rule names five components, not thirty-seven factors? Both are correct and describe different levels of Appendix A. The study is organised under five components, and thirty-seven enumerated factors sit inside them: five economic, six market, nine technical, twelve financial and five management (21)(39).

What is the difference between Appendix A and Appendix B? They are separate instruments. Appendix A, Feasibility Study Components, is the full study for B&I and REAP renewable-energy systems and the anchor of the 5001.3 definition. Appendix B, Financial Feasibility Reports, is the CF instrument cited in 5001.304. Appendix B is not a subset of Appendix A (17)(21).

Which version of the rule governs my file? The codified Part 5001 in force when the complete application was received, currently through 90 FR 57351, effective December 11, 2025, with no 2026 amendment. The 2024 rule moved no feasibility threshold (1)(3)(5).

What is a new business for the B&I trigger? Under 5001.3, a business in operation less than one full year, or one that has not achieved full operational capacity or stable operations as determined by the Administrator. The second limb is discretionary (7).

What coverage ratio does USDA require? None is stated in the credit evaluation. The ratio is defined at 5001.3 as EBITDA less reasonably expected replacement capital expenditures over annual debt service, and the only numeric coverage minimums in the Part are the refinancing eligibility tests at 5001.102(d), 1.1 times historical and 1:1 current (7)(8)(12).

Who can prepare a USDA feasibility study? An independent qualified consultant, defined at 5001.3 as an independent third party with the knowledge, expertise and experience for the specific task, and on B&I acceptable to the Agency (7)(19). No designation is named. The CF financial feasibility study with examination opinion must be prepared under AICPA attestation standards by an insured preparer (17).

Is a feasibility study the same as a business plan? No. A business plan is the sponsor's document and states the sponsor's case. A feasibility study is an independent test of that case against primary data on demand and supply, prepared for the lender's reliance to the Appendix A components, and it may conclude that the project is not feasible.

Is a feasibility study the same as a market study? A market study is the market component of Appendix A. The feasibility study adds the other four components and the financial test: the revenue build, the expense benchmark, the stabilisation curve and the stressed coverage (21).

Does the study satisfy the appraisal requirement? No. Collateral market value is established by an appraisal under 5001.203, by an independent qualified appraiser under USPAP Standards 1 and 2, with any going-concern value deducted before discounting. The study addresses repayment viability, not value (12)(13).

Can one firm do both the appraisal and the study? The regulation is silent. Nothing bars it categorically, but each engagement must independently meet its own independence standard, the lender must report the arrangement in writing under 5001.208, and role separation is the conservative default (7)(13)(15).

When does the study have to be in the file? At the application stage under 5001.303, before the conditional commitment under 5001.451 and long before the loan note guarantee under 5001.453 (16)(23).

What happens if the study was inadequate and the loan defaults? The loss claim can be reduced, up to a total reduction, under 5001.521(d)(1) where the loss traces to negligent origination, defined as the failure to act as a reasonably prudent lender would in its own unguaranteed portfolio (7)(28).

Does the SBA use the same framework? No. Under the SBA programmes the study is discretionary, unprescribed and judged in hindsight. Attributing the five-component USDA framework to the SBA is the single most common error on this subject (40).

MMCG Analytics is an underwriting data platform for the people who build USDA and SBA files: state, county and place pages that carry the rurality position, the demographic, traffic, parcel, flood and zoning data a credit evaluation starts from, and the guaranteed-lending obligation and delinquency figures for the trade area; and asset-class pages that carry the class-specific demand drivers. This guide is the regulatory reference those pages hang from, and it is the reference the MMCG Invest programme pages and state pages cite for the USDA feasibility requirement. Where a page states a rule, it is the rule stated here, cited to the section, and it is revised when the section is. The neutral posture is deliberate: the platform is a reference and data layer, and it points to the practice rather than competing with it.

The feasibility practice at MMCG Invest prepares lender-facing studies for USDA B&I and Community Facilities, for SBA 7(a) and 504, and for conventional files, built on the platform's data and to the standard in Section 10, and relied on by USDA and SBA lenders and by banks. For a lender with a file the programme sends for a study, or a sponsor whose lender has asked for independent support, the conversation starts with the programme, the rurality and eligibility position, and the applicable threshold, in that order, because those decide whether a study is the right document before any question of what it should say.

Every Part 5001 citation is to the section, subpart or appendix, which are stable. Figures set by annual notice (the guarantee percentages and fee rates) are quoted from the FY 2026 notice at 91 FR 11272 and are replaced when the next fiscal year's notice publishes; confirm the current-year notice before pricing. The two feasibility appendices are codified as page images at 85 FR 42518 and are described and counted from that source rather than from any secondary transcription. The public record on loss-claim reductions for feasibility deficiencies is thin and is reported as thin; per-lender delinquency figures referenced in Section 9 are secondary-source and provisional and are treated as unverified at the institution level. Nothing here is legal or underwriting advice for a particular loan; the operative rule for any file is the codified text in force on the date its complete application is received.