Reliance, Certification and Liability: Who May Rely on a Feasibility Study

A feasibility study is the one document in an SBA or USDA loan file that is written by someone who is neither the lender nor the borrower nor the agency, and that everyone in the transaction reads. That makes the reliance question unavoidable and the answer unintuitive. The regulator's remedy for a bad study runs against the lender, not the author. The secondary market purchaser who ultimately holds the guaranteed portion never receives the study and has no contractual claim on the author. The borrower who paid for the study is usually not its intended user. And the author's exposure, where it exists, is defined less by regulation than by two things the author controls: who the study says it is for, and what the author certifies.

This is a practitioner's reading of that architecture, current as of 7 September 2026, for lenders deciding what to require, for CDCs and secondary market participants deciding what they hold, and for sponsors deciding what they are buying. It is not legal advice, and the case law section in particular should be read with counsel in the relevant jurisdiction.

Whose report it is: the lender's file, the agency's remedy

Start with what the two federal programs actually say about reliance, because it is less than most people assume.

SBA. Nothing in 13 CFR Part 120 or SOP 50 10 8 makes SBA a party to the feasibility study. SBA "may require" one under 13 CFR 120.160(b), but it neither engages the author nor accepts the report. The study is an underwriting input the lender relies on, and SBA's review of it is derivative: it examines the lender's conduct. Under 13 CFR 120.524(a), SBA is released from its liability on a guaranty, "in whole or in part, within SBA's exclusive discretion," where the lender "has failed to comply materially with any Loan Program Requirement," has failed to make, close, service, or liquidate the loan in a prudent manner, or has placed SBA at risk through improper action or inaction, and purchase of the guaranty does not waive SBA's right to recover afterward. The Office of Inspector General's High Risk 7(a) Loan Review Program has used that clause for over a decade to recommend recoveries on early-defaulted loans where lenders "did not provide adequate documentation to substantiate financial projections." A deficient feasibility study is evidence in that review. The party who answers for it is the lender.

SOP 50 10 8.1, effective for applications receiving an SBA loan number on or after 1 October 2026, sharpens the point without changing it. Its quality of earnings requirement on acquisitions at $3 million and above must be obtained for the lender's benefit and may not be prepared by or for the borrower or seller. That is the reliance model SBA expects for third-party reports: the lender is the client, the lender is the intended user, and the lender carries the consequence.

USDA. Part 5001 is explicit. Under 7 CFR 5001.6(b), "lenders can contract for services, but such contracting does not relieve a lender from its responsibilities," and under 5001.6(c), "if a lender fails to comply with the requirements of this part, the Agency may reduce any loss payment in accordance with the lender's agreement and loan note guarantee." The definitions in 5001.3 supply the standard: negligent loan origination "means the failure of a lender to perform those services or actions that a reasonably prudent lender would perform in originating its own portfolio of loans that are not guaranteed." And 5001.521(d) supplies the remedy: negligent loan origination and negligent loan servicing "will result in a reduction of loss claims payable under the guarantee to the lender if any losses have occurred as the result of such negligence," and the reduction "could be a total reduction of the loss claims payable." USDA's own rulemaking preamble described the point of the provision as reinforcing "the concept of negligent loan origination throughout this part."

USDA does something SBA does not: it reviews the study itself, before commitment, against Appendix A, and can reject the author under the "acceptable to the Agency" test. But the consequence of a study that later proves wrong still runs through the lender's loss claim. The Agency does not sue the consultant. It reduces the guarantee.

Why it matters. Both programs place the feasibility study inside the lender's duty of prudent origination. The lender cannot outsource that duty to the author, and neither agency will look past the lender to the author when the loan fails. That is the first reason reliance has to be engineered in the engagement rather than assumed from the regulation.

Who relies, and how far

Five parties read a feasibility study. Their positions are not the same.

The lender. The client and the intended user. In SBA practice the borrower commissions and pays for the study; in USDA practice the cost is a project expense. In both, the study is written to the lender's underwriting question, and the lender is the party whose reliance is contemplated. Where the borrower is the contracting party, the engagement letter should name the lender as an intended user expressly, or the lender's reliance rests on nothing but custom.

The CDC. On a 504 loan the CDC is a second lender with its own credit decision and its own certification to SBA. Where the third-party lender orders the study, the CDC should be added as a named intended user or obtain a reliance letter. A CDC relying on a report addressed to another lender is relying on a document that does not name it.

The agencies. SBA reads the study at guaranty purchase, derivatively, to judge the lender. USDA reads it before commitment to judge the project and, in practice, the author. Neither is a client, and neither is an intended user in the private-law sense, but each is a foreseeable reader, and the certification should be written knowing that.

The secondary market. Under 13 CFR 120.620, SBA guarantees to a Registered Holder the timely payment of principal and interest on a Pool Certificate, and that guarantee "is backed by the full faith and credit of the United States." The holder's protection is the guarantee, not the file. Third-party reports, feasibility studies among them, stay in the originating lender's loan file and do not travel to the pool or the certificate holder. A secondary market purchaser has no privity with the study author and, in the ordinary case, is not a named intended user, which leaves it with no contractual claim and a weak tort claim. Its remedy for a bad origination is SBA's guarantee; SBA's remedy is the lender.

The borrower. The party that most often pays and least often relies, in the legal sense. A study prepared for a lender's underwriting is not a business plan for the sponsor, and a certification that names the lender as the intended user says so. Sponsors who want to rely on the study for their own investment decision should be told that plainly, and the engagement should reflect it.

Reliance letters: the appraisal model, borrowed

Feasibility practice has no codified reliance mechanics. The Uniform Standards of Professional Appraisal Practice do, and lenders apply them by analogy because the same credit officers handle both reports.

Under USPAP, the client is the party who engages the appraiser, and an intended user is the client and any other party identified by name or type as a user of the report. The report must state the intended use and identify the intended users, and a party not identified cannot properly rely on it. Adding an intended user after delivery is a new or amended assignment, not a readdressal, which is why the Interagency Appraisal and Evaluation Guidelines of December 2010 caution banks about accepting appraisals prepared for another party. A reliance letter, issued by the original appraiser, extends reliance to a named third party subject to the same assumptions and limiting conditions, without expanding the scope of work or the duty of care, and it is routinely priced as a modest fee against the original engagement. The ASTM E1527 standard for Phase I environmental site assessments carries the same concept in its user-reliance provisions, and SBA's environmental appendix builds on it with a prescribed reliance letter.

For feasibility studies the practical translation is this. The study should state its client, its intended users by name or type, and its intended use, in the report and in the certification. Reliance by a party not named, a participating lender, a CDC, a bond purchaser, a franchisor, is extended by a reliance letter from the author, on the same terms and assumptions, for a fee, or not at all. A study that is silent on intended users is a study that invites everyone to rely and protects no one.

The certification: what the author actually warrants

Neither program prescribes a certification paragraph for the feasibility author. That is a gap the author fills, and the way it is filled determines what the signature means.

SBA. SOP 50 10 8 and 8.1 require signed certifications from appraisers under USPAP and from business valuators as Qualified Sources, and they require the QoE provider to act for the lender's benefit. They require nothing specific of the feasibility author. Lenders read the appraisal and valuation standards across.

USDA. Part 5001 gets closer. The 5001.3 definition makes the study "a report including an opinion or finding conducted by an independent qualified consultant(s)." Appendix A requires an executive summary that includes "a summary of the feasibility determinations made for each applicable component," a recommendation section that must "conclude with an opinion and recommendation presented by the consultant," and a qualifications section with "a resume or statement of qualifications of the author of the feasibility study, including prior experience." That is a signed opinion with a stated basis and a named author. The Community Facilities examination-opinion report goes further: prepared under AICPA attestation standards, with "the requisite professional liability insurance in place."

The USPAP model. Standards Rule 2-3 requires the appraiser to certify, among other things, that the statements of fact are true and correct, that the analyses and conclusions are the appraiser's own and limited only by the stated assumptions and limiting conditions, that the appraiser has no present or prospective interest in the property and no bias toward the parties, and that compensation is not contingent on a predetermined result. No feasibility program replicates that, and every reviewer who reads appraisals expects it.

The same certification is what separates an authored study from a generated one; see Can AI Write a Lender-Grade Feasibility Study?.

What a lender-grade feasibility certification should therefore contain, and what the author is warranting by signing it:

1. Identity of the client, the intended users by name or type, and the intended use, with reliance limited to those users for that use.
2. That the author is an independent third party with no ownership, employment, or financial interest in the borrower, the seller, the franchisor, the developer, the contractor, or any vendor to the project, and no other role on the transaction.
3. That the fee is fixed and not contingent on loan approval, closing, loan amount, or the conclusion reached.
4. That the statements of fact are, to the author's knowledge, true and correct, and that the sources relied on are identified in the report.
5. That the analyses, opinions, and conclusions are the author's own, developed independently, and limited only by the assumptions and limiting conditions stated in the report.
6. That the projections are estimates based on stated assumptions, that actual results will differ, and that the author does not guarantee them.
7. The author's qualifications and relevant experience, and a statement of professional liability insurance in force.
8. A signature, a date, and, where the study is updated, the date of the data and the date of the conclusion.

The signature warrants the process and the independence. It does not warrant the outcome. That distinction is the whole of the liability question, and it is why item six belongs in the certification rather than in a footer.

Liability: where the cases put the author

The legal exposure of a feasibility study author is not defined by SBA or USDA. It is defined by state law on negligent misrepresentation, and in most states by Section 552 of the Restatement (Second) of Torts, which imposes liability on one who, in the course of business, supplies false information for the guidance of others in their business transactions and fails to exercise reasonable care, but only to a limited group of persons for whose benefit and guidance the information was intended, and only in a transaction the supplier intended the information to influence.

Three cases show the range.

Bily v. Arthur Young and Company (California Supreme Court, 1992) is the leading statement of the limited-group rule. The court rejected liability to all foreseeable readers of a professional report and adopted the Restatement approach, under which a supplier is liable to a third party only where it "has undertaken to inform and guide a third party with respect to an identified transaction or type of transaction." A supplier who "merely knows of the ever-present possibility of repetition to anyone" bears no responsibility to those readers. The intended-user statement in the report is the evidence a court looks at.

Bilt-Rite Contractors v. The Architectural Studio (Pennsylvania Supreme Court, 2005) shows the other edge. The court held that a design professional who supplies information for the foreseeable use of third parties can be liable in negligent misrepresentation to a party with which it has no contract, on the reasoning that such professionals are "in the business of supplying information" for the guidance of others and are paid for exactly that. States that follow the Pennsylvania reading reach further than California does.

Hardin County Savings Bank v. Housing and Redevelopment Authority of the City of Brainerd (Minnesota Supreme Court, 19 September 2012) is the case on point. Six banks bought $3.3 million of revenue bonds financing a residential lot development, relying on an offering memorandum that included an appraisal and a feasibility study prepared by a consulting firm. The feasibility study projected that all lots would sell within seven years at about fourteen lots a year and called that "a very achievable plan." Three lots sold in three years; the issuer defaulted. The banks sued the consultant for negligent misrepresentation, alleging that the study restated an appraisal value that ignored $1,085,000 of special assessments, that the absorption schedule was "a virtual impossibility" at the true lot cost, and that the study "was intended to be relied upon by Bondholders." The trial court dismissed and the court of appeals affirmed; the Supreme Court reversed and remanded, holding that the banks had pleaded every element of the claim, including that the consultant "intended to supply the information for the benefit and guidance of Bondholders in their business transactions" and "intended the information to influence the transaction for which the information was supplied."

Read together, the cases establish four things a feasibility author should take as settled. First, the author's exposure runs to the parties the study says it is for, and to parties the author knew the client intended to reach. Second, an absorption schedule, a capture rate, or an occupancy projection is a representation for these purposes when it is presented as achievable rather than as an estimate under stated assumptions. Third, an error carried forward from another report, the appraisal in Hardin County, is the author's error once it is restated as the author's own. Fourth, the defenses that work, no duty to unnamed readers, stated assumptions, disclosed limitations, non-contingent compensation, are the same items that belong in the certification. The New York line under Ultramares and Credit Alliance requires something close to privity before a third party can recover; most states do not.

Two consequences follow for how a study is written. Projections should be presented as conclusions under assumptions the reader can test, with the sensitivity that breaks them shown, not as promises. And figures adopted from another professional's report, the appraisal, the engineer's cost estimate, the franchisor's evaluation, should be identified as such and tested, not restated as the author's finding.

Insurance: what an E&O policy covers, and what it excludes

Feasibility and market study firms are insured under miscellaneous professional liability or consultants' errors and omissions forms, written on a claims-made-and-reported basis with a retroactive date, so that a claim arising from a study delivered before the retroactive date is not covered. Limits for small and mid-sized firms are commonly written at $1 million per claim with a $1 million or $2 million aggregate, and lenders that ask should expect to see a certificate at that level or above for the larger engagements.

The exclusions matter more than the limits. Standard forms exclude guaranteed, warranted, or promised results; contractual liability assumed beyond what would exist absent the contract; dishonest or intentional acts; fee disputes; and bodily injury and property damage. Some manuscript forms exclude liability for guarantees of financial projections or investment outcomes. A certification that guarantees a projection, or an engagement letter that indemnifies the lender for the projection's accuracy, can therefore take the author outside its own coverage. The engagement letter and the policy should be read together, and the certification should be drafted to stay inside the policy.

USDA is the only program that requires insurance in terms, and only for the Community Facilities examination-opinion report. The Interagency Guidance on Third-Party Relationships: Risk Management issued by the Federal Reserve, FDIC, and OCC in June 2023 does the rest: it expects banks to address performance, liability, information rights, and insurance in their third-party contracts, and a regulated lender's vendor management function will increasingly ask a feasibility consultant for the same documentation it asks of an appraisal management company.

The engagement letter: allocating what the regulation does not

Because neither program prescribes the reliance terms, the engagement letter does. The clauses that matter, and what each one is doing:

Client and intended users. Names the party engaging the author and the parties entitled to rely, by name or type. This is the clause a court reads first under Section 552.

Intended use. States the underwriting purpose and the program. A study prepared for an SBA 7(a) credit decision is not automatically a study for a USDA State Office review, a bond offering, or an equity investment.

Reliance by additional parties. Provides that reliance by any party not named is extended only by written reliance letter from the author, on the same terms, for a stated fee.

Independence and compensation. Confirms no financial interest and a fixed, non-contingent fee. This clause is also the author's Form 159 answer under 13 CFR Part 103: the author is not representing the applicant to SBA.

Data and assumptions. Identifies the client-supplied and third-party data the author relies on without independent verification, and the assumptions the projections rest on.

No guarantee of results. States that projections are estimates, that actual results will vary, and that the author does not warrant them. Drafted to match the E&O exclusions.

Limitation of liability. Caps the author's liability to the client at the fee paid or a stated multiple, and waives consequential and indirect damages. Enforceability varies by state and does not bind non-parties, which is why the intended-user clause carries more weight than the cap.

Indemnity. Where present, mutual and limited to third-party claims arising from each party's own negligence. A one-way indemnity from the author for the accuracy of projections is the clause most likely to be uninsured.

Delivery, updates, and shelf life. States the data date, the delivery date, and the period after which the conclusions should not be relied on without an update, which matters under USDA's 90-day currency rule and SBA's 180-day financial statement convention.

What a lender should require

Lenders and CDCs that want the reliance question settled before the loan closes, rather than litigated after it defaults, can document it in six items.

1. An engagement letter that names the lender, and any CDC or participant, as an intended user, with the intended use stated.
2. A signed certification in the study covering the eight elements above.
3. A certificate of professional liability insurance in force at delivery, at a limit proportionate to the loan.
4. A statement in the study of which figures are the author's own and which are adopted from other reports, with the source and date of each.
5. A reliance letter, on the author's letterhead, for any party added after delivery.
6. For USDA files, confirmation that the author is acceptable to the State Office before the study is commissioned, and a study built to Appendix A with the recommendation and qualifications sections the appendix requires.

None of this shifts the agency's remedy off the lender; nothing can. What it does is put the lender in the position the regulations assume: a prudent originator who relied on a qualified, independent, insured author whose warranty is in writing.