Research Note · Loan Reviews
What Federal Inspectors General Found in Loan File Reviews, 2012 to 2026
There is no inspector general or GAO report that grades a feasibility study. The reports audit lender and agency compliance after default: in the SBA Inspector General's High Risk 7(a) reviews, 16 of 35 early-defaulted loans had a material deficiency, 45.7 percent, with $17.2 million of $34.0 million purchased recommended for recovery 12. HUD's Inspector General found underwriting deficiencies in all 9 MAP loans it selected 3, and GAO found projected first-year revenue more than twice actual for 19 of 24 franchisees 4. The samples were chosen for trouble, so none is a prevalence rate.
What the inspector general and GAO loan file reviews show.
Federal inspectors general and the Government Accountability Office have reviewed lender and agency loan files for more than a decade, and none of those reports grades a feasibility study or a market study as a document. They ask a narrower question after a loan has defaulted or a portfolio has deteriorated: did the lender, and then the agency, follow the agency's own underwriting rules, and can the file prove it? The deficiencies they count are compliance and evidence failures, and every count on this page is a count of loans the reviewers chose to examine, not a rate across a loan program.
The largest body of counts comes from the program in which the SBA Inspector General reviews high-risk 7(a) loans. Across the two consolidated reports, 35 early-defaulted loans were examined and 16 had a material origination or closing deficiency, 45.7 percent, with $17.2 million of $34.0 million in guaranty purchases recommended for recovery 12. The most frequent loan-level deficiencies were program eligibility conditions such as credit elsewhere, size and use of proceeds (8 of the 35 loans), equity injection, working capital or debt refinance support (7), and unsupported revenue projections (5) 12.
The one audit that examines how an agency reviewed a market study, a revenue estimate and an appraisal is the HUD Inspector General's 2016 audit of Multifamily Accelerated Processing loans: of nine loans selected for strong indicators of problems, all nine had underwriting deficiencies, three on project revenue, three on appraisal and two on market studies 3. GAO's 2013 review of one franchise organization's loan applications found that projected first-year revenue exceeded twice the actual figure at 19 of 24 franchisees 4. The reviewers also failed: SBA's guaranty purchase center made an error or missed deficiencies on 6 of the 9 Recovery Act loans it reviewed 5, and SBA's credit risk office left 108 of 358 planned reviews of high-risk lenders undone 6.
What the reports do not contain matters as much. Nothing in these reports tallies unsupported demand, omitted competing supply or an absorption pace set too fast, so none can be cited as evidence that those failures are common or rare. Read the counts below as documented examples of what a lender must be able to prove from a file. The review failure taxonomy places each of them against the failure codes a study can be tested on, and the last section of this note maps the eight codes these audits touch.
What these reports are, and what they are not.
The note covers 23 reports that review loan files, lenders or an agency's own review, issued between March 2012 and April 2026 by the SBA Office of Inspector General, the HUD Office of Inspector General, the USDA Office of Inspector General and GAO. Three older items, a 2003 USDA case and two GAO reports from 2009, serve only as dated context and each is labelled as such. Three recent SBA Inspector General reports are used for context on agency oversight and carry no loan counts 789.
They audit compliance after the fact
An inspector general reviews a loan file against the agency's own rules. For SBA 7(a) loans the benchmark is the lending criterion in 13 CFR 120.150, which states the standard this way: loans "must be so sound as to reasonably assure repayment" 10. The reviewer asks whether the lender documented that credit elsewhere was unavailable, that repayment ability was shown, that the equity injection was made, that the business valuation was obtained and that affiliates were counted. It does not ask whether the market study's trade area was drawn correctly or whether the absorption schedule was realistic.
GAO explained the structural reason in 2009, outside this note's window and shown here as dated context. SBA's standard on-site review form assesses lenders' compliance with SBA policies and procedures "but not an assessment of lenders’ credit decisions", and SBA officials told GAO that it was not the agency's role to assess those decisions 11. A file review that tests rules and documents will find documentation failures. That is a design feature of the reviews, and it limits what any count can say about the analytical quality of the studies inside the files.
They are selected samples, not random draws
Each sample has its own selection rule, and the rule shapes the rate. The SBA Inspector General's sample is early defaults at or above $500,000, meaning a default inside 18 months of disbursement, drawn by an internal risk score, so the sample consists of early failures by construction 12. HUD's Inspector General noted that MAP-approved lenders underwrote and closed more than 8,600 loans between 2002 and 2010, and its review then covered nine loans with strong signs of trouble 3. GAO's franchise review followed one franchise organization's loans and one loan agent 4. Only the USDA audit drew a statistical sample, 53 loans from 26 states totaling $200 million, and it alone converts its counts into estimates for the whole audit universe 12.
Because the loans were chosen for trouble, a deficiency rate of 40 percent among them does not describe the loan program 1, and the absence of a tag for market demand partly reflects what the reviewers were asked to test. Where this note gives a rate it is a count of deficient loans over reviewed loans, computed from the report's own numbers, and the denominator and selection rule travel with it.
They sit after default
Most SBA samples are early defaults, and the audit asks whether SBA should have honored its guaranty. That is a different moment from a lender's approval decision. A study delivered before approval was read by a loan officer and a credit committee who did not yet know the outcome, and the audit reads the same file after the outcome is known. The distinction matters in the section on the consequence in regulation, where the guaranty's release depends on the lender's compliance and not on the study's accuracy. Readers who want the default economics behind these samples can use the SBA default economics note.
Every in-window report, its sample, its count and its dollars.
The table lists each report with loan, lender or portfolio counts, in date order within each agency. The rate is the report's deficient count divided by the number it reviewed, computed from the report's own figures and rounded to one decimal place. Counts are in different units (loans, lenders, planned reviews, nursing homes, franchisees, reporting requirements), so rows must not be added together, and the dollar column is the figure the report ties to that row, not an additive total.
| Report | Date | Scope and unit | Reviewed | Deficient or finding | Rate computed | Dollars the report ties to the row |
|---|---|---|---|---|---|---|
| SBA OIG 12-11R | March 23, 2012 | Recovery Act 7(a) early-defaulted loans, loans | 25 | 14 with material deficiencies | 56.0% | About $10.7 million questioned or recoverable 5 |
| SBA OIG 12-18 | August 16, 2012 | Repayment ability analysis at purchase review, limited scope audit | Not stated as a count | Analysis not effective | n/a | At least $43 million estimated savings over two years 13 |
| SBA OIG 15-09 | March 20, 2015 | High Risk 7(a) review, loans | 7 | 3, guaranty denial justified | 42.9% | $1.8 million 14 |
| SBA OIG 15-17 | September 30, 2015 | Memorandum on one prior-audit loan | 1 | Unverified revenue streams; seller-prepared statements | n/a | n/a 15 |
| SBA OIG 16-22 | September 30, 2016 | High Risk 7(a) review, FY2016 loans (cumulative 15) | 8 (15) | 3 (6) | 37.5% (40.0%) | $3.2 million ($5 million) 16 |
| SBA OIG 18-01 | October 12, 2017 | Management challenges report, loans with recoveries | n/a | 6 loans with recoveries recommended | n/a | Nearly $5 million 17 |
| SBA OIG 18-21 | August 15, 2018 | High Risk 7(a) advisory, loans | 2 | 2 | 100.0% | $1,351,565 purchased 18 |
| SBA OIG 18-23 | September 13, 2018 | High Risk 7(a) advisory, loans | 2 | 2 | 100.0% | $2,190,843 purchased 19 |
| SBA OIG 18-26 | September 27, 2018 | High Risk 7(a) consolidated, FY2018 loans (cumulative FY2014 to FY2018: 27) | 8 (27) | 4 (11) | 50.0% (40.7%) | Nearly $2.7 million (more than $8.5 million questioned) 1 |
| SBA OIG 19-15 | July 10, 2019 | High Risk 7(a) advisory, loans | 2 | 2 | 100.0% | $5,531,290 purchased 20 |
| SBA OIG 19-16 | August 14, 2019 | High Risk 7(a) advisory, loans | 2 | 2 | 100.0% | $2,100,448 purchased 21 |
| SBA OIG 19-19 | September 19, 2019 | High Risk 7(a) advisory, one loan | 1 | Delegated authority not appropriately used | n/a | $1,379,774 purchased 22 |
| SBA OIG 19-22 | September 26, 2019 | High Risk 7(a) consolidated, FY2019 loans | 8 | 5 | 62.5% | About $8.7 million of about $10.8 million purchased 2 |
| SBA OIG 20-03 | November 12, 2019 | Oversight of high-risk lenders, planned lender reviews | 358 | 108 not conducted | 30.2% | $13.3 million in 21 purchases questioned 6 |
| SBA OIG 20-18 | August 25, 2020 | High Risk 7(a) advisory, one loan | 1 | Equity injection not supported | n/a | $2,077,766 purchased 23 |
| SBA OIG 24-16 | May 15, 2024 | FY2023 payment integrity review, reporting requirements | 10 | 8 not met | 80.0% | n/a 24 |
| HUD OIG 2016-AT-0001 | May 20, 2016 | MAP underwriting, loans selected for problem indicators | 9 | 9 | 100.0% | Land value overstated by more than $15 million on three projects 3 |
| HUD OIG 2018-PH-0001 | August 10, 2018 | Section 232 loan approval, loan files | 3 | 1 area: management agent creditworthiness | n/a | Nearly $50 million, value of the three files 25 |
| HUD OIG 2018-BO-0001 | September 17, 2018 | Section 232 nursing home monitoring, nursing homes | 18 | 18 with ineffective monitoring | 100.0% | More than $32.1 million potential loss 26 |
| HUD OIG 2026-BO-0001 | April 2, 2026 | Section 232 residential care portfolios, loans | 84 | 58 on which lenders could claim | n/a (an outcome, not a deficiency) | More than $329.5 million 27 |
| USDA OIG 34703-0001-32 | March 29, 2013 | B&I Recovery Act loans, statistical sample of loans | 53 | 37 serviced incorrectly; 33 not adequately monitored by the agency | 69.8%; 62.3% | $200 million, value of the sample 12 |
| GAO-13-759 | September 2013 | 7(a) franchise loans, franchisees reviewed | 24 | 19 with projected first-year revenue more than twice actual | 79.2% | n/a 4 |
| GAO-18-421 | June 2018 | Credit elsewhere compliance, FY2016 on-site lender reviews | 40 | 17 found noncompliance | 42.5% | n/a 28 |
Source: SBA Office of Inspector General reports 12-11R through 24-16; HUD Office of Inspector General, audits 2016-AT-0001, 2018-PH-0001, 2018-BO-0001 and 2026-BO-0001; USDA Office of Inspector General audit number 34703-0001-32; U.S. Government Accountability Office, GAO-13-759 and GAO-18-421, report dates as shown, compiled by Feasibility Study Company from the inspectors' and auditors' published reports, retrieved November 9, 2026 1234561213141516171819202122232425262728.
Three features of the table deserve a second look. First, the SBA advisories 18-21, 18-23, 19-15, 19-16 and 19-19 are subsets of the consolidated reports 18-26 and 19-22, and the consolidated reports are the ones to count when a number is needed for the program 121819202122. Second, the 100 percent rows come from two- and nine-loan samples that the reviewers picked because the files already looked troubled 318192021. Third, the rows with no rate carry a finding without a count the report supports, and those are not blank by oversight: a rate computed from a number the report did not give would be invented.
SBA's High Risk 7(a) Loan Review Program, by cohort and by deficiency.
The SBA Inspector General began detailed reviews of high-dollar early-defaulted 7(a) loans in fiscal 2014 and reports them in consolidated evaluations. By the FY2018 consolidation the reviews covered 27 loans with purchase amounts of about $23.2 million; the OIG recorded material deficiencies in lender origination and closing for 11 of them, with questioned costs of more than $8.5 million, and identified suspicious activity in 5 of the 27 1. The FY2019 consolidation reviewed another eight loans for which SBA had paid about $10.8 million 2. Added together, the 35 loans carry purchase amounts of $34,020,227 and recommended recoveries of $17,216,577, and the FY2014 to FY2018 loans alone account for $23,174,379 of the purchases and $8,554,433 of the recoveries 12. The cohort table lays the two reports side by side, ordered as the report's appendix orders its loans.
| Cohort | Loans reviewed | With material deficiency | Rate | Purchase dollars | Recommended recovery | Recovery over purchase |
|---|---|---|---|---|---|---|
| Appendix III loans 1 to 8 (the FY2016 sample of report 16-22) 116 | 8 | 4 | 50.0% | $8,468,522 | $3,995,733 | 47.2% |
| Loans 9 to 15 (reviews from July 2014 to February 2015, report 15-09) 114 | 7 | 3 | 42.9% | $4,039,701 | $1,785,784 | 44.2% |
| Loans 16 to 20 (FY2017 by order in the table) 1 | 5 | 1 | 20.0% | $5,244,611 | $917,107 | 17.5% |
| Loans 21 to 27 (FY2018) 1 | 7 | 3 | 42.9% | $5,421,545 | $1,855,809 | 34.2% |
| All 27 loans, report 18-26 (FY2014 to FY2018) 1 | 27 | 11 | 40.7% | $23,174,379 | $8,554,433 | 36.9% |
| All 8 loans, report 19-22 (FY2019) 2 | 8 | 5 | 62.5% | $10,845,848 | $8,662,144 | 79.9% |
| Both tables, 35 loans (FY2014 to FY2019) 12 | 35 | 16 | 45.7% | $34,020,227 | $17,216,577 | 50.6% |
Source: SBA Office of Inspector General, Appendix III of report 18-26 and report 19-22, report dates September 27, 2018 and September 26, 2019, compiled by Feasibility Study Company from the inspectors' and auditors' published reports, retrieved November 9, 2026 12. Loans 1 to 8 appear in report 16-22 as three deficient loans; the table shows four because loan 6, first referred to investigators, was returned to the evaluators and found deficient. Cohort boundaries follow the order of the appendix table.
What the tags say
The OIG's own summary of the FY2018 deficiencies is a list of items the lenders did not document: credit elsewhere, repayment ability, size standards, the execution of a franchise agreement, equity injection and debt refinance 1. The FY2019 report says the issues were similar to earlier evaluations, "such as business valuations, affiliates, equity injection, and repayment ability" 2. Read loan by loan from the tables in the two reports, with a loan counted once per category and a deficient loan able to carry more than one tag, regulatory and eligibility nonconformity appears on 8 of the 35 loans, unsupported equity injection, for working capital or for the sources and uses statement on 7, revenue and projections on 5, sponsors, management and affiliates on 4, valuation and collateral on 3, repayment ability on 2, and closing controls, documentation and independence on 1 each 12. The four loans sent to investigators count as reviewed, not as deficient, and the grouping of the OIG's wording into these categories is Feasibility Study Company's; the review failures page maps each group to its taxonomy code.
The ranking has a plain reading. The most common failures on these loans were evidence that the program's own conditions had been met, and the second most common was proof that the borrower had put in the equity and working capital the structure assumed. Revenue and projection tags come third, and they sit next to the two repayment ability tags: report 15-09 files unsupported projections under repayment ability deficiencies, so the two categories belong together, and seven of the 35 loans carry one or the other 114.
A projection tested against interim results
Advisory 18-21 shows what a revenue finding looks like in a file. The lender's projections suggested that net income before taxes would rise from a loss of $221,815 in 2010 to $594,897, an increase of 368 percent, in the first year after disbursement 18. Because the change from historical to projected statements was so large, the OIG annualized the borrower's three-month interim statement and calculated net income before taxes of a loss of $918,400, against the projected $594,897 18. The borrower had net losses and negative debt service coverage in 2008, 2010 and the first three months of 2011; the lender nonetheless used projections as the basis for repayment ability 18.
Memorandum 15-17 describes a second pattern: the projected revenue in the lender's analysis included unverified revenue streams that affected repayment ability, and the lender relied on seller-prepared financial statements in a change of ownership between close parties; the borrower made five interest-only payments, defaulted, and the business never opened 15. The same reviewers' FY2016 report lists the failure of repayment ability as unverified seller statements, liabilities not considered, affiliates not considered and unsupported projected sales, and lists inadequate business valuation as valuations not obtained as required 16.
Changes of ownership
One concentration stands out. Of the 15 loans reviewed through FY2016, 8 financed a change of ownership, and 4 of those 8 showed material lender deficiencies or signs of suspicious activity, with purchases totaling $2.8 million 16. In report 15-09, two of the three loans with material deficiencies financed changes of ownership 14. Acquisitions rest on the seller's historical statements and on a valuation of the business, which are the two items the OIG found missing or unverified most often, and the change of ownership practice area describes what SBA asks a lender to document.
The advisories and the agency's response
The advisories read alike. In 18-23, 19-15 and 19-16 the lenders did not provide adequate documentation to substantiate reasonable assurance that borrowers met requirements for items that varied by file: repayment ability, size eligibility and equity injection in one, repayment ability, eligibility, collateral, equity injection and debt refinance in another, eligibility, repayment ability and equity injection in the third 192021. Advisory 19-19 found that the lender did not appropriately use its delegated authority, and 20-18 found that the lender did not document the equity injection 2223. SBA agreed with the findings in each report. After the FY2018 consolidation SBA said it had improved its loan review with training, more review time and an updated checklist, and it recovered more than $4 million in guaranty payments on 5 of the 11 loans with recommended recoveries 1. For the FY2019 loans SBA agreed to contact the lenders for information that would bring the loans into compliance and, if that was not possible, to seek recovery 2.
HUD's MAP audit 2016-AT-0001: the market study, revenue and appraisal findings.
HUD's Multifamily Accelerated Processing program lets approved lenders underwrite FHA-insured multifamily loans that HUD staff then review. The audit counted what that review missed. Between 2002 and 2010 MAP-approved lenders underwrote and closed more than 8,600 loans; the Inspector General reviewed nine of the loans that had strong indicators of problems and found underwriting deficiencies in all nine 3. HUD had approved the nine loans from six MAP lenders, which exposed the FHA insurance fund to unnecessary risk 3. This is the only audit in the window that examines how an agency reviewed a lender's market study, revenue estimate and appraisal, and the sample is nine loans picked for trouble. The MAP Guide is the HUD document that governs FHA multifamily underwriting; a reader needing current requirements should check the current edition, and this note states only what the audit reports.
| Deficiency in HUD's review | Loans | Share of 9 | Category in this note |
|---|---|---|---|
| Principals and contractors 3 | 7 | 77.8% | Sponsor, management and affiliates |
| Appraisal 3 | 3 | 33.3% | Valuation, appraisal and collateral |
| Project revenue 3 | 3 | 33.3% | Revenue and projections |
| Market studies 3 | 2 | 22.2% | Market area and demand |
| Project financial background 3 | 2 | 22.2% | Documentation and currency |
| Critical repairs not completed before closing 3 | 1 | 11.1% | Closing and construction controls |
| Project repair cost and scope of work 3 | 1 | 11.1% | Closing and construction controls |
| Technical staff conclusions improperly considered 3 | 1 | 11.1% | Reviewer and agency control |
| Prepaid cost 3 | 1 | 11.1% | Equity, working capital and sources and uses |
| MAP timelines for site inspections not followed 3 | 1 | 11.1% | Closing and construction controls |
| FHA loan limits not enforced 3 | 1 | 11.1% | Regulatory and eligibility nonconformity |
Source: HUD Office of Inspector General, audit 2016-AT-0001, May 20, 2016, retrieved November 9, 2026; computed by Feasibility Study Company from the published file 3. A loan can carry more than one deficiency, so the shares do not sum to 100 percent.
The market study findings
Two loans had inadequate review of market studies 3. In one, the Alafia market study included outdated statistics such as unemployment rates, census data, employment trend analysis and building permits, covering January 2000 to May 2007, about 15 months before the effective date of the July 2008 market study report 3. The finding is about currency: the data the study used had a date, and the date was older than the study's own effective date by more than a year. In the other, the market study described the Amaranth property as superior to its competitors in the market and as warranting higher rents; the Inspector General reviewed the comparable property data and determined that other properties offered the same or similar amenities and in some instances were located in superior areas 3. That finding is about advocacy: a claim of superiority made in a study meant to support the rent, and not supported by the comparable data in the same file.
The revenue findings
Three loans had inadequate review of project revenue 3. The audit found that a lender overstated the project revenue estimate because it failed to use available up-to-date market data and relied on optimistic indicators, a violation of HUD requirements, and that the overstatement affected the project's ability to meet its obligations 3. In a related finding on appraisals, HUD did not question the lender's unreasonable project expense estimate or its unsupported capitalization rates 3. The revenue finding is the HUD counterpart of the SBA projection findings: in both agencies the reviewer tested the estimate against current data, and the estimate did not hold.
The appraisal findings
Three loans had inadequate review of the appraisal 3. On one project the lender overstated the property value by more than $11 million: the subject had previously sold for $16 million, and the appraiser valued it at $28.6 million less than a year later with only $1.4 million in repairs 3. Because HUD did not enforce MAP appraisal and Uniform Standards of Professional Appraisal Practice requirements, the lender's appraisers overstated three projects' combined land value by more than $15 million 3. An appraisal is not a feasibility study, but a lender reads the two together, and the audit treats an unreconciled jump from a recent sale price as a finding about the reviewer.
Principals, contractors and the response
Seven of the nine loans had inadequate review of principals and contractors 3. Five lenders inappropriately approved multiple principals and contractors that may have lacked the ability to complete and run the projects; in one file the borrower's net worth calculation included $6.9 million in real property 3. HUD agreed to revise its memorandum of understanding with lenders, update the MAP Guide and train staff 3.
Projected first-year revenue against actual first-year revenue.
GAO-13-759 is the only report in the window that sets a projection against what happened. GAO reviewed 7(a) loans to franchisees of one franchise organization and compared the first-year revenue projected on loan applications involving a loan agent or her employer with the franchisees' actual first-year revenue. At 19 of the 24 franchisees reviewed, 79.2 percent, the average projection was more than double the actual first-year figure 4.
The limits are in the report. GAO was unable to conclusively determine whether the loan agent intentionally provided exaggerated projections to help franchisees qualify, and it made no recommendations; SBA debarred the agent and took initial steps on loan data and agent oversight 4. The sample is 24 franchisees tied to one agent and one franchise organization, and the finding describes those applications. It is not an estimate of how far projections exceed actuals across 7(a) loans or across franchise systems.
The report adds three context points that explain why the gap could arise. Interviewed franchisees noted difficulties meeting anticipated revenue estimates and limited access to information that would aid in business planning 4. Federal regulations do not require franchise organizations to provide actual first-year average revenues for start-up businesses in their disclosure document 4. And the loans themselves performed badly: of the total population of 170 loans, 74 defaulted, 55 of which, 74 percent, originated from four lenders with the highest loan volume and default rates on loans to the franchisees, and of the 88 loans reviewed from those four lenders 55, 63 percent, defaulted 4.
For a study the lesson is specific. A first-year revenue figure with no actual first-year comparable behind it is a projection a reviewer can test against what similar units achieved, and a franchisor's disclosure document is not required to supply that comparable for a start-up. The evidence has to come from somewhere else in the file.
Purchase review, skipped lender reviews and the credit elsewhere test.
The same reports count the reviewers' misses, and the counts are as exact as the lender counts. They explain why a file can pass SBA's purchase review and still be found deficient later.
Guaranty purchase review
The 2012 audit found that on 6 of the 9 loans SBA's National Guaranty Purchase Center reviewed, the center made a purchase error or failed to identify deficiencies in its guaranty purchase reviews, which produced improper payments of about $3.3 million; the OIG recommended a specialized early-default purchase review unit 5. The center's quality control process did not evaluate lender compliance with SBA eligibility and use of loan proceeds requirements 29. A limited-scope audit that August found that SBA's way of assessing delegated lender underwriting for early-defaulted loans did not work in identifying whether lenders had been clearly negligent in judging repayment ability, and it estimated that better repayment ability reviews could save SBA at least $43 million over two years 13. By the Fall 2013 semiannual report management had not responded to one recommendation of that report 30.
The later High Risk reviews repeat the pattern at small scale. In report 15-09 the quality control team reviewed one of the three deficient loans and identified no material deficiency, and the loan was also in SBA's FY2014 improper payments review without being identified 14. In 16-22 three of the five loans the OIG formally reported or referred for potential fraud were in SBA's FY2015 improper payments review, and SBA did not identify the improper payments 16. The FY2018 consolidation counts four loans, with $4.5 million in improper payments, that were reviewed for lender compliance by loan specialists and still not reported 1. In FY2019, SBA staff completed purchase and quality control reviews on all eight loans and did not identify or fully address the material deficiencies 2.
Time and depth
The FY2018 management challenges report gives one reason: loan specialists generally had 6 hours to finish a loan file review, and in some cases they questioned lenders on material compliance issues but ultimately accepted explanations and documentation that did not cure the deficiencies 17. It summarized the program's results to date as purchase and quality control operations that continued to miss material lender underwriting deficiencies, with recoveries of nearly $5 million recommended on six loans 17. SBA's response to the FY2018 consolidation, more training, longer review time and an updated checklist, addressed time and depth directly 1.
Skipped lender reviews
Audit 20-03 examined SBA's oversight of high-risk lenders. The credit risk office did not always oversee effectively: of 358 reviews planned for high-risk lenders in fiscal years 2015 to 2017 it left 108, or 30 percent, undone 6. For 21 defaulted loans that SBA purchased, the office had identified material deficiencies; the OIG saw no sign that SBA checked if those lenders had fixed the problems, and it questioned the purchases, which totaled $13.3 million 6. Five lenders that did not receive planned reviews had an average default rate of 19 percent for loans approved and disbursed in those years 6. SBA management agreed with the finding and recommendations 6.
Credit elsewhere
GAO tested one eligibility rule across SBA's own lender reviews. Of the on-site lender reviews performed in fiscal year 2016, over 40 percent, 17 of 40, found lenders out of compliance with the credit elsewhere requirement 28. Most on-site reviewers did not document their assessment of lenders' policies or procedures because SBA did not require them to, so SBA lacked information that could explain the high rate 28. GAO recommended that reviewers document the assessment, that SBA collect information on the criteria lenders use and analyze it for trends, and SBA generally agreed 28. In the 2009 report, shown here as dated context outside the window, SBA's reviews of 97 lenders between October 2006 and March 2008 found 31 that had failed to consistently document that borrowers met the credit elsewhere requirement or the personal resources test 31. A second 2009 GAO report found that 97 percent of the lenders SBA's risk rating identified as high risk in 2008 were not reviewed on site 11.
Recent context
Three recent reports describe the same agency's oversight without loan counts. KPMG auditors found that SBA was not compliant with 8 of the 10 reporting requirements under the Payment Integrity Information Act for fiscal 2023 24. The audit of SBA's oversight of non-bank lenders in the Paycheck Protection Program found that SBA performed limited oversight of those lenders and did not execute its risk-based review plan; the report is cited here only for that oversight finding 7. The FY2026 management challenges report says SBA's credit risk office manages guaranteed portfolios of about $163 billion as of June 2025, that the agency lowered underwriting standards, and that in 2025 SBA reversed many of those changes 8. The FY2025 payment integrity review found no improvement in payment integrity for 7(a) guaranty approvals and 504 approvals between fiscal 2024 and 2025 9.
USDA: the in-window Business and Industry audit and one dated case.
The USDA audit in the window examines servicing and agency monitoring and does not test underwriting or a feasibility study. USDA's Inspector General drew a statistical sample of 53 Business and Industry guaranteed loans, made in 26 states and totaling $200 million, under the American Recovery and Reinvestment Act 12. Based on issues with 37 of the 53 loans, it estimated that lenders did not correctly service 72 percent of the loans in the audit universe, and because Rural Development did not adequately monitor the lenders of 33 of the 53 loans it estimated 64 percent for agency monitoring 12. Rural Development generally agreed with the recommendations 12.
The sample is the one statistically valid design among the reports, so its estimates carry to the audit universe of Recovery Act loans. The audit is also the one least connected to study quality, because servicing is a post-closing function. A reader who wants USDA's study standard should read the rule itself and not look for it in this audit.
The 2003 Georgia sawmill, as dated context only
One older USDA audit comes closest to a feasibility finding. It is outside the window, concerns one borrower in 2003, and is excluded from every count on this page. The lender did not fulfil its responsibilities for ensuring that the facility was designed using accepted architectural and engineering practices, was completed with available funds, and produced the quality and quantity of lumber called for in the application 32. It allowed an owner without experience in architectural and engineering design, construction and construction cost estimating to design the facility, estimate the cost and serve as general contractor 32. The loan collateral, originally appraised at $10.9 million, was liquidated for $1,186,174, and although the sawmill had not been constructed and had no production or income history it was appraised as if it were a going concern for sale 32. The case is the clearest documented example of the unbuilt project problem described in the special purpose property risk note, and it remains a single case.
HUD's Section 232 audits: approval, monitoring and portfolio risk.
Three HUD audits cover Section 232, which insures loans on nursing homes and other residential care facilities. They test approval and monitoring, which makes them evidence about coverage and occupancy only indirectly.
The first, from August 2018, found that for the three loan files reviewed, valued at nearly $50 million, HUD kept documentation showing it had adequately reviewed loan applications for program eligibility, the operator's financial operations, the creditworthiness of the borrower, operator and operator parent, professional liability insurance and the project capital needs assessment; it did not adequately evaluate the creditworthiness of the management agents as required, and it took corrective action during the audit 25. A clean result belongs in the record: an audit that tests a defined list of review steps can find that nearly all of them were done.
The second audit, in September 2018, found HUD's monitoring was not effective in addressing problems at any of the 18 financially challenged nursing homes reviewed 26. In one case the underwriter had recommended that HUD require a marketing plan with strategies, timelines and goals, and HUD provided no evidence that it ensured a plan was implemented at the 2012 refinance 26. On the defaulted mortgages HUD faced a possible loss above $32.1 million, and it owed more than $10 million in carrying costs 26. This is the one finding in the window about a ramp a lender assumed and an agency did not follow up.
The third audit, April 2026, covers residential care portfolios. As of June 2024, 167 of the 3,670 HUD-insured Section 232 borrowers, nearly 5 percent, had defaulted on their mortgages 27. As of July 2025, lenders could make insurance claims for 58 of 84 loans in the reviewed portfolios, totaling more than $329.5 million 27. Borrowers' audited financial statements disclosed withdrawals of funds when properties had no available surplus cash, cash deficiencies, and properties whose cash flow was too small to pay current debts, and the Inspector General asked HUD to quantify unauthorized distributions in a portfolio of 35 loans 27. The report also records that HUD's staff for the function fell from 61 employees in October 2024 to 38 in May 2025 27. These counts describe portfolios reviewed because they were troubled, and the 167 defaults are an outcome and not a deficiency.
13 CFR 120.524: when a deficient lender file costs the guaranty.
The consequence of a deficient file is written into SBA's rules. Under 13 CFR 120.524, SBA's liability on its guarantee can be released, wholly or partly, where, among other conditions, the lender "failed to comply materially with any Loan Program Requirement for 7(a) loans" or "failed to make, close, service, or liquidate a loan in a prudent manner" 33. The text was current as of November 5, 2026 33. The High Risk reports find deficiencies that "justified denial of the guaranty" 14, which is the audit-side statement of that rule, and the recoveries SBA pursued are its practical effect: in the FY2018 consolidation, 5 of the 11 loans carrying recommended recoveries had already returned over $4 million in guaranty payments to SBA 1.
Two implications follow. The party exposed by a deficiency is the lender, so a lender that reads a study asks whether each assumption can be filed against a program requirement. And the test applied at default is the lender's compliance with program requirements and prudence, not the correctness of every figure in a study, which is why the audit counts documentation failures. A study that supplies a source, a date and a method for each figure gives the lender what it needs to show prudence if the file is reopened. The audit evidence also shows that a file can clear the first review: the purchase center and quality control reviewed loans on which the OIG later found material deficiencies 25.
How to read the counts, and where they stop.
The counts are the reports' own, and the rates are computed from them. Several limits apply to every row.
- Selection. The SBA samples are loans of at least $500,000 that defaulted early and were chosen by an internal score, the HUD sample is nine loans chosen for problem indicators, the GAO franchise review follows one organization's loans, and GAO called its 2009 samples not generalizable 13431. Only the USDA sample is statistical.
- No sampling error. None of the non-statistical samples supports a confidence interval, and none is computed here. A rate describes the reviewed loans only.
- Units. Loans, lenders, planned reviews, nursing homes, franchisees and reporting requirements are different units. A row's count cannot be added to another row's, and the consolidated SBA reports already contain the advisories, which are not added a second time 12.
- Dollars. Dollar figures are what each report ties to its finding: questioned costs, purchase amounts, estimated savings or value of the sample. They are not additive and not a loss rate.
- Outcome is not deficiency. The 167 Section 232 defaults and the 74 franchise defaults in GAO-13-759 are outcomes, and the audits do not attribute them to the deficiencies counted elsewhere 427.
- Mapping. The category names in the tag tables are Feasibility Study Company's grouping of the reports' wording. Report 15-09 files unsupported projections under repayment ability while the tables here file them under revenue, so the two categories are read together 14.
- Currency. Agency responses and open recommendations are as each report states them on its date. A recommendation listed as unanswered in 2013 may have been closed since, and the Fall 2013 listing is cited only for what it said then 30.
- Window. The in-window reports run from March 2012 to April 2026. A 2003 USDA case and two 2009 GAO reports are labelled as dated context and appear in no count.
The central limit is the one stated first. These are reviews of lender and agency compliance with agency rules after default, and none grades a study. Absence of a finding for market demand, supply or absorption is partly a product of what the reviewers were asked to test, and presence of a finding for revenue or valuation does not show that studies are generally weak on those points. The counts support one claim: when these reviewers looked at a file, these are the items they found missing, unverified or unsupported.
Mapping the findings to review failure codes.
A lender reading a study before approval is deciding whether it could defend the loan after default, and the audit findings list the points at which lenders and agencies could not. The table maps the findings to eight of the codes in the review failure taxonomy, with the count that evidences each and the evidence the file needs to carry. The other codes have little or no support in these reports. The only count for the market study codes is HUD's 2 of 9, and no report counts omitted supply or an absorption pace set without a benchmark, so this note does not use these reports to rank those failures against the ones below. Related reading is the note on why studies fail lender review, and the analytical framework sets out the method a study should follow.
| Code | Failure | Documented count and report | Evidence a file should carry |
|---|---|---|---|
| F5c | Projections far above actual or interim results, or revenue streams never verified | 19 of 24 franchisees, GAO-13-759; interim statement annualized to a $918,400 loss against a $594,897 projected profit, SBA OIG 18-21; 3 of 9 on revenue, HUD 2016-AT-0001; 5 of 35 revenue tags, SBA 18-26 and 19-22 123418 | Each revenue stream tied to a verified source and a comparable actual, with the interim year compared to the projection |
| F7e | Repayment ability stated but not demonstrated; coverage not computed from the verified cash flow the program uses | All 3 deficient loans in SBA OIG 15-09; 2 of 35 repayment ability tags; purchase review not effective, SBA OIG 12-18 11314 | Coverage computed on the verified historical cash flow, with projections shown as projections |
| F8d | Program eligibility not met or not documented (the credit elsewhere test, size standards, how proceeds were used, loan limits) | 8 of 35 SBA loans; 17 of 40 on-site reviews, GAO-18-421; 1 of 9 on loan limits, HUD 1328 | The eligibility facts the study relies on, stated with their source, so the lender's documentation can cite them |
| F9c | Seller or sponsor prepared figures used without independent verification | Seller-prepared statements in a change of ownership between close parties, SBA OIG 15-17; unverified seller statements, SBA OIG 16-22 1516 | Seller figures reconciled to tax returns or bank records, with the reconciliation shown |
| F10a | Study, appraisal or data staler than program rules permit | Market study data dated 15 months before its effective date, one of 2 market study findings in HUD 2016-AT-0001 3 | Every dataset dated, with the date of the data and the effective date of the study stated |
| F11 | Sponsor, management and capital: equity injection, working capital, sources and uses (F11a); sponsor and affiliate capacity (F11b) | 7 of 35 equity, working capital or refinance tags; 4 of 35 affiliate tags; 7 of 9 principals and contractors, HUD; equity injection, SBA OIG 20-18; management agent creditworthiness, HUD 2018-PH-0001 132325 | A sources and uses table with the evidence of each equity source, and a list of affiliates with their obligations |
| F12 | Valuation and collateral: appraisal or business valuation unreliable or not obtained (F12a); an unbuilt project valued as a going concern (F12b) | 3 of 35 valuation tags; 3 of 9 appraisal findings, HUD; 2003 USDA case, dated context only 1332 | A valuation obtained as the program requires, reconciled to recent sales of the same property |
| F13 | Costs, plans, repairs or site inspections left unverified before or during construction | 3 of 9 in HUD 2016-AT-0001 (critical repairs, repair cost and scope, site inspection timelines); 1 SBA loan, final plans not obtained 13 | Construction costs tied to final plans and a contractor's bid, with inspection dates |
Source: SBA Office of Inspector General audits and reports 12-18, 15-09, 15-17, 16-22, 18-21, 18-26, 19-22 and 20-18; HUD Office of Inspector General, audits 2016-AT-0001 and 2018-PH-0001; U.S. Government Accountability Office, GAO-13-759 and GAO-18-421; USDA Office of Inspector General, 2003 audit number 34601-4-At (dated context); compiled by Feasibility Study Company from the inspectors' and auditors' published reports, retrieved November 9, 2026 1234131415161823252832. The code mapping and the evidence column are Feasibility Study Company's.
Three habits follow from the audits. First, state the standard and the date for every figure, because the file will be read against both after default. Second, separate the verified from the projected: the SBA findings that mattered were projections without a bridge from the interim results, and the HUD findings were estimates that ignored current market data. Third, do not argue superiority without comparables. The HUD Amaranth finding shows how an unsupported claim of a better property is tested against the comparables already in the file.
In regulation, 13 CFR 120.160(b): SBA may require professional appraisals, a survey, or a feasibility study 34, and the audits above count what lenders could not document, not whether a study existed. A study therefore protects the lender only to the extent that its figures are traceable, current, independent of the seller and tied to the verified cash flow the program tests.
Inspector general loan file review questions.
Do inspectors general review feasibility studies?
No. The reports in this note audit whether lenders and agencies followed the agency's own underwriting rules after a loan defaulted or a portfolio deteriorated. SBA's lending criterion requires loans sound enough to reasonably assure repayment 10, and the reviewers test the file's documents against program requirements. No report in this note tallies failures of demand, supply or absorption 13.
How many SBA loans did the Inspector General find deficient?
In the two consolidated High Risk 7(a) reports, 11 of 27 loans reviewed from fiscal 2014 to 2018 and 5 of 8 in fiscal 2019 had material origination or closing deficiencies, 16 of 35 in all, or 45.7 percent 12. The loans were early defaults at or above $500,000, picked from an internal risk ranking, so the rate describes only that sample and not the 7(a) program.
What did HUD's Inspector General find about market studies?
Audit 2016-AT-0001 recorded inadequate market study review on 2 of the 9 MAP loans examined 3. One market study used statistics dated about 15 months before its effective date, and another claimed a property was superior to justify higher rents while comparables offered the same or similar amenities 3. The nine loans were chosen for strong indicators of problems.
How often do projected revenues exceed actual revenue?
The reports give one test. GAO compared first-year revenue projected on loan applications tied to one loan agent or her employer with actual first-year revenue: on average the projection ran above double the actual figure at 19 of the 24 franchisees in the review, and GAO could not determine intent 4. The finding concerns one agent and one franchise organization and does not measure projection error across 7(a) loans.
Why do SBA reviews miss deficiencies that auditors later find?
The Inspector General found that SBA's purchase center erred or missed deficiencies on 6 of 9 Recovery Act loans it reviewed, that loan specialists were generally required to complete file reviews within 6 hours, and that 108 of 358 planned high-risk lender reviews were not conducted 5617. Reviewers also accepted explanations that did not cure deficiencies 17.
Can SBA refuse to honor a guaranty because of a deficient loan file?
Yes, in defined cases. 13 CFR 120.524 relieves SBA of its guaranty obligation where, among other conditions, the lender did not materially comply with a Loan Program Requirement for 7(a) loans or did not make, close, service or liquidate a loan prudently 33. Through fiscal 2018, SBA had recouped over $4 million on 5 of the 11 loans where recovery was recommended 1.
Do these reports show which study failure is most common?
No. The SBA loan tags put eligibility conditions first (8 of 35 loans), then equity and working capital support (7) and revenue projections (5) 2. Those tags sit on a risk-selected sample of lender files. None of the reports has a category for market demand, supply or absorption, so the tags cannot rank study failures against one another 1.
Where this research is applied.
Loan programs
Loan program pages
Related research notes
- Minimum Debt Service Coverage Ratio by Loan Program and Lender
- Why Feasibility Studies Fail Lender Review: A Taxonomy of Rejection
- Franchise or Independent? Underwriting the Brand, Not the Binary
- Gas Station Demand Per Capita: Why a Site's Gallons No Longer Set Its Value
- SBA Default Economics: Loss Rates by Industry, and Why the Headline Number Misleads
- SOP 50 10 8.1: What Changed for Feasibility Studies
Preparing a study a lender must defend? Make each figure traceable.
A study that gives the source, date and method for every figure lets a lender show the file was prudent. Feasibility Study Company prepares independent studies built to that standard.
Request a methodology briefingNumbered references.
Primary sources first: the inspector general, GAO and Federal Register documents cited in order of first citation, each with its report date and the date it was retrieved.
- U.S. Small Business Administration Office of Inspector General, SBA OIG Report 18-26: Consolidated Results of the OIG High Risk 7(a) Loan Review Program, September 27, 2018; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Small Business Administration Office of Inspector General, SBA OIG Report 19-22: Consolidated Results of the OIG High Risk 7(a) Loan Review Program (FY2019), September 26, 2019; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Department of Housing and Urban Development Office of Inspector General, HUD OIG Audit 2016-AT-0001: HUD Did Not Enforce and Sufficiently Revise Its Underwriting Requirements for Multifamily Accelerated Processing Loans, May 20, 2016; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Government Accountability Office, GAO-13-759: Small Business Administration, Review of 7(a) Guaranteed Loans to Select Franchisees, September 2013; retrieved November 9, 2026.
- U.S. Small Business Administration, SBA OIG Report 12-11R: High-Dollar Early-Defaulted Loans Require an Increased Degree of Scrutiny and Improved Quality Control at the National Guaranty Purchase Center, March 23, 2012; retrieved November 9, 2026. SBA's document page for the Office of Inspector General report.
- U.S. Small Business Administration Office of Inspector General, SBA OIG Report 20-03: Audit of SBA's Oversight of High-Risk Lenders, November 12, 2019; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Small Business Administration Office of Inspector General, SBA OIG Report 25-04: SBA's Oversight of Non-Bank Lenders and Third-Party Service Providers Associated with PPP Loans, November 13, 2024; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Small Business Administration Office of Inspector General, SBA OIG Report 26-01: 2026 Top Management and Performance Challenges Facing SBA in Fiscal Year 2026, December 18, 2025; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Small Business Administration Office of Inspector General, SBA OIG Report 26-09: Independent Auditors' Report on SBA's FY 2025 Compliance with the Payment Integrity Information Act, May 15, 2026; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- Office of the Federal Register and U.S. Government Publishing Office (eCFR), 13 CFR 120.150, What are SBA's lending criteria?, eCFR current, up to date as of November 5, 2026; retrieved November 9, 2026.
- U.S. Government Accountability Office, GAO-10-53: Small Business Administration, Actions Needed to Improve the Usefulness of the Agency's Lender Risk Rating System, November 2009; retrieved November 9, 2026.
- U.S. Department of Agriculture Office of Inspector General, USDA OIG Audit 34703-0001-32: American Recovery and Reinvestment Act, Business and Industry Guaranteed Loans, Phase 3, March 29, 2013; retrieved November 9, 2026.
- U.S. Small Business Administration, SBA OIG Report 12-18: A Detailed Repayment Ability Analysis is Needed on High-Dollar Early-Defaulted Loans to Prevent Future Improper Payments, August 16, 2012; retrieved November 9, 2026. SBA's document page for the Office of Inspector General report.
- U.S. Small Business Administration Office of Inspector General, SBA OIG Report 15-09: The OIG High Risk 7(a) Loan Review Program Recommends $1.8 Million in Recoveries, March 20, 2015; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Small Business Administration Office of Inspector General, SBA OIG Management Information Memorandum 15-17: The OIG High-Risk 7(a) Loan Review Program, September 30, 2015; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Small Business Administration Office of Inspector General, SBA OIG Report 16-22: The OIG High Risk 7(a) Loan Review Program Recommends $3.2 Million in Recoveries, September 30, 2016; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Small Business Administration Office of Inspector General, SBA OIG Report 18-01: Report on the Most Serious Management and Performance Challenges Facing SBA in Fiscal Year 2018, October 12, 2017; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Small Business Administration Office of Inspector General, SBA OIG Management Advisory 18-21: OIG High Risk 7(a) Loan Review Program (two loans), August 15, 2018; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Small Business Administration Office of Inspector General, SBA OIG Management Advisory 18-23: OIG High Risk 7(a) Loan Review Program (two loans), September 13, 2018; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Small Business Administration Office of Inspector General, SBA OIG Management Advisory 19-15: OIG High Risk 7(a) Loan Review Program (two loans), July 10, 2019; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Small Business Administration Office of Inspector General, SBA OIG Management Advisory 19-16: OIG High Risk 7(a) Loan Review Program (two loans), August 14, 2019; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Small Business Administration Office of Inspector General, SBA OIG Management Advisory 19-19: OIG High Risk 7(a) Loan Review Program (one loan), September 19, 2019; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Small Business Administration Office of Inspector General, SBA OIG Management Advisory 20-18: High Risk 7(a) Loan Review Program (one loan), August 25, 2020; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Small Business Administration Office of Inspector General, SBA OIG Report 24-16: Independent Auditors' Report on SBA's Fiscal Year 2023 Compliance with the Payment Integrity Information Act of 2019, May 15, 2024; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Department of Housing and Urban Development Office of Inspector General, HUD OIG Audit 2018-PH-0001: HUD's Office of Healthcare Programs Generally Approved Section 232 FHA-Insured Loans in Accordance With HUD Requirements, August 10, 2018; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Department of Housing and Urban Development Office of Inspector General, HUD OIG Audit 2018-BO-0001: HUD’s Monitoring of the Financial Performance of Section 232 Nursing Homes, September 17, 2018; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Department of Housing and Urban Development Office of Inspector General, HUD OIG Audit 2026-BO-0001: HUD Did Not Always Address Risks Reported in Borrowers' Audited Financial Statements for Section 232 Residential Care Facility Portfolios, April 2, 2026; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Government Accountability Office, GAO-18-421: Small Business Loans, Additional Actions Needed to Improve Compliance with the Credit Elsewhere Requirement, June 2018; retrieved November 9, 2026.
- U.S. Small Business Administration Office of Inspector General, SBA OIG Semiannual Report to Congress, Spring 2012 (Oct 1 2011 to Mar 31 2012), Spring 2012 (October 1, 2011 to March 31, 2012), dated May 31, 2012; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Small Business Administration Office of Inspector General, SBA OIG Semiannual Report to Congress, Fall 2013, Fall 2013, dated December 2013; retrieved November 9, 2026. Report as published on oversight.gov, the report portal of the inspector general community.
- U.S. Government Accountability Office, GAO-09-228: Small Business Administration, Additional Guidance on Documenting Credit Elsewhere Decisions Could Improve 7(a) Program Oversight, February 2009; retrieved November 9, 2026.
- U.S. Department of Agriculture Office of Inspector General, USDA OIG Audit 34601-4-At: Rural Development Lender Servicing of Business and Industry Guaranteed Loans in Georgia, January 2003; retrieved November 9, 2026.
- Office of the Federal Register and U.S. Government Publishing Office (eCFR), 13 CFR 120.524, When is SBA released from liability on its guarantee?, eCFR current, up to date as of November 5, 2026; retrieved November 9, 2026.
- Office of the Federal Register and U.S. Government Publishing Office (eCFR), 13 CFR 120.160, Loan conditions (feasibility study provision), eCFR current, up to date as of November 5, 2026; retrieved November 9, 2026.