Credit Analysis and Small Business Finance Case Studies
Finance5 Case StudiesUpdated Sep 2026
These five cases come from the lending and review work that small business finance teams handle every week. Each one turns on the same material: financial statements, ledgers, bank records, and the control evidence that shows whether those numbers can be trusted. The organizations are fictional, but the pressure points are ordinary ones, including records that disagree, forecasts that drift, borrowers asking for more debt, and audit findings that never quite close. Pay attention to the thresholds each team set, because those are the parts that carry over to the next file.
Case Study 1
Bank deposits refuse to match the ledger at close
Brackenridge Community Lending Partners, a regional lender funding working capital and equipment loans for small manufacturers and service firms.
The situation
Oakhollow Cabinet Works builds custom kitchen and office cabinetry and reported annual revenue of $4,180,000. It asked for a $650,000 equipment line to replace two aging panel saws. Priya, the credit analyst assigned to the file, pulled the general ledger, twelve months of bank statements, and the accounts receivable aging. Recorded customer deposits in the ledger ran $86,400 above what reached the operating account over the same period. The gap was not spread evenly. Nearly all of it fell in two quarters, and both quarters closed with a manual journal entry. The year-end statements and the tax summary both matched the ledger, which meant three documents agreed with each other while the bank did not.
The challenge
The committee would not approve credit on statements that rested on a ledger the bank contradicted. The question was not only where $86,400 went missing but whether the variance was a bookkeeping error, a timing issue, or something that changed how much cash the business actually produced. Priya also had to decide how much time to spend before the next committee date, which was two weeks away, and what the lender could reasonably ask for without making the owner feel accused.
Analysis
Agreement among the three documents told the team very little, since all of them drew from the same ledger. Priya rebuilt deposits from the bank side, matching each credit to an invoice number and a customer remittance where one existed. Roughly three-quarters of the variance traced to two manual journal entries that credited sales and debited an undeposited funds account, with no remittance behind them. That pattern mattered more than the dollar amount. A timing difference reverses within a month. These entries stayed open for most of a quarter, which points to a control that let revenue post before cash arrived and never checked whether the cash came.
The exposure question was narrower than whether the owner had been careless. If the $86,400 was revenue that never existed, the trailing earnings used to size the new line were overstated, and debt service coverage fell below the committee's 1.25 minimum once the overstatement was removed. Priya set a threshold before going further: any unreconciled variance above 2 percent of annual deposits, or any unsupported manual entry above $25,000, would stop the file from moving forward until the owner's accountant produced source documents. The variance was about 2.1 percent of deposits, and one entry alone exceeded the $25,000 line. The file was held, not declined.
Actions taken
- Require the owner's accountant to deliver bank-to-ledger reconciliations for each of the twelve months, signed and dated, with every reconciling item listed by amount and cause.
- Pull remittance advice and invoice numbers for every deposit over $5,000 and match each to a receivable that was open and then cleared. Any deposit without a match goes on an exceptions list.
- Set a control that blocks any manual journal entry touching revenue or undeposited funds unless a second approver with no posting rights signs it, and log every override for review by the credit committee.
- Add a covenant requiring monthly bank statements sent directly to the lender for the life of the line, so that deposit data reaches the file without passing through the borrower.
Results
The owner's accountant returned the reconciliations in the second week. Of the $86,400 gap, $71,900 traced to the two manual entries, which had been posted to close out customer accounts that had actually paid through a lockbox the bookkeeper never connected to the ledger. The remaining $14,500 was timing and unrecorded fees. The equipment line was approved at $650,000 with a monthly bank-statement covenant and a dual-approval rule on manual entries. Trailing coverage, restated on verified deposits, came in at 1.31, above the minimum, which kept the terms intact.
Questions for discussion
- Which of the three agreeing documents would you have trusted most before the bank statements came in, and what would have changed your answer?
- How would you set the holding threshold if the variance were 0.4 percent of deposits but concentrated in a single unsupported entry?
- What evidence would convince you that a lockbox account was missing from the ledger rather than being used to hide revenue?
Case Study 2
Juniper Street Catering misses plan by a wide margin
Juniper Street Catering Company, an event caterer running three commercial kitchens with a revolving credit line from its lender.
The situation
Juniper Street Catering Company runs three commercial kitchens and books corporate lunches, weddings, and conference catering. Its operating plan assumed $3,600,000 in revenue for the year, with bookings growing 14 percent over the prior season. Six months in, booked events were running 11 percent below plan, and two corporate accounts had moved their work to a competitor with a lower minimum order. Cash fell faster than revenue. Operating cash dropped to $212,000 against fixed monthly obligations of $58,000 including payroll and rent. The company had drawn $150,000 on its revolving line in the second quarter to cover a supplier backlog, and the lender's file showed the drawn balance rising without a matching rise in receivables.
The challenge
The relationship manager, Elena, had to decide what the bank should ask for before the next covenant test. The question was whether the shortfall was a timing problem that a few strong months could absorb or the start of a permanent drop in demand. The plan could not be used as written, but a revised forecast built on hope would be no better. She needed scenarios that showed where liquidity ran out under each assumption.
Analysis
Teams grounded in Credit Analysis and Small Business Finance usually begin with the assumptions behind the forecast rather than the forecast itself. Elena rebuilt the plan from booking-level data, separating events already under contract from events the sales team expected to close. Contracted revenue for the remaining months covered roughly 61 percent of plan. The rest depended on conversion rates from proposals, and those rates had dropped from the prior season. She also checked the receivables aging against the revenue line, because a catering business can book revenue on signed contracts long before deposits arrive. Two large corporate accounts were over sixty days past due, so cash lagged revenue even more.
Three scenarios came from the rebuilt model. The base case assumed conversion returned to last season's rate in the back half of the year, and cash stayed above the $150,000 floor the lender had set for the line. The downside case held conversion at its current level and assumed the two overdue accounts paid only half their balances. Cash touched the floor in the eleventh week after the review and reached $40,000 by quarter end. A severe case added the loss of one kitchen's wedding volume and pushed cash negative. Two assumptions decided the response: the floor itself, and how much of the overdue balance was collectible.
- 1Rebuild from contractsSeparate signed bookings from pipeline before changing any assumption.
- 2Test three scenariosBase, downside, and severe cases, each tied to the cash floor.
- 3Collect and reservePursue overdue balances, then reserve what remains uncollected.
- 4Cut and defer spendingTarget discretionary costs in the weakest kitchens first.
- 5Set borrowing triggersTie any further draws to weekly cash and a revised plan.
Actions taken
- Freeze new draws on the revolver until the overdue accounts are either collected or reserved, and require a signed weekly cash report from the owner that reconciles to the bank.
- Cut discretionary spend in the two kitchens with the weakest margins, targeting $22,000 a month in deferred purchases and overtime, and track the savings against the forecast each week.
- Renegotiate the two overdue accounts into a schedule with a deposit before each event, and reserve half of the remaining balance until payment clears.
- Set a trigger that tightens the revolver borrowing base if cash falls under $150,000 for two consecutive weekly reports, and require a revised plan within ten business days of any breach.
Results
Elena's team froze new draws, collected $64,000 from the overdue accounts, and reserved the remainder. Deferred purchases and overtime changes freed roughly $19,000 a month against a $22,000 target. Cash held above the floor through quarter end, though the margin was thin. The revised forecast, built on contracted revenue and the downside conversion rate, showed the business could meet its obligations without new borrowing if bookings recovered to within 6 percent of plan. The lender kept the line at its existing size with the tighter borrowing base trigger, and scheduled a review at the next quarter's close.
Questions for discussion
- If the overdue accounts had been a single client rather than two, how would the reserve policy change, and what concentration limit would you set?
- Which assumption in the downside case would you test first with the sales team, and what evidence would move you off the current conversion rate?
- At what point does a revolving line stop being a bridge and start funding a structural gap, and how would you tell the difference from the cash report alone?
Case Study 3
Structured deposits trigger a portfolio review
Millbrook Valley Community Bank, a county lender with about 40 commercial relationships and a deposit base concentrated in one region.
The situation
Redlake Trading Partners distributes roofing and masonry supplies to contractors and reported annual revenue of $2,900,000. Its loan file had been quiet for a long stretch, with a $400,000 inventory line that was current and well collateralized. Then the bank's monitoring system flagged a pattern in the operating account. Over one quarter the company made 29 cash deposits in amounts between $9,400 and $9,900, totaling $276,000. Within two days of most deposits, funds moved by wire to a supplier account the company had never used. The business's customers mostly paid by check or ACH, so cash intake did not match its profile. The relationship manager, Guillermo, learned of the alert after the monitoring team had opened a case.
The challenge
Guillermo had to decide whether the pattern was a compliance matter that belonged to the bank's financial crime unit, a credit matter that affected the inventory line, or both. The company was current and its collateral was sound on paper, which made the activity easy to rationalize. The bank's obligations did not depend on whether the customer was honest. They depended on what the file showed, when it was documented, and who had been told.
Analysis
The file held three kinds of evidence, and each needed a different test. The first was account activity: deposit slips, teller records, and wire instructions showing cash in and funds out. The second was the onboarding profile, collected at account opening and last reviewed before the pattern began. That profile described a business paid by check and ACH, with a single operating account and modest cash sales. The third was the collateral record. The inventory appraisal was eight months old, and the lender had never requested a physical count. Stale records are weak evidence alone, but beside unexplained cash they point to a control gap.
Exposure had two parts. The credit exposure was the $400,000 inventory line, which was current, but a line secured by inventory that might be moving through the account unseen was not secure. The compliance exposure was larger. Under the bank's written procedures, structured cash deposits paired with rapid outbound wires required a suspicious activity review even with a plausible explanation. The threshold for a formal review was two of four monitoring indicators, and this file showed three. Guillermo did not judge the customer. He documented the indicators, froze new draws, and passed the file to the compliance officer, who owned the filing decision.
- Start of quarterFirst cash deposits posted in amounts just under the internal threshold.
- Within two daysOutbound wires to an unfamiliar supplier account followed most deposits.
- End of quarterTwenty-nine deposits totaled $276,000 in one operating account.
- Following weekMonitoring raised the alert and opened a case file.
- Within the next monthReview filed, draws frozen, and a physical collateral count ordered.
Actions taken
- Place a hold on new advances against the inventory line and order a physical count within the next two weeks, with the lender's own examiner present.
- Give the compliance officer the full deposit and wire history, the onboarding profile, and the monitoring indicators in a single file, and record the time each item was received.
- Coordinate with compliance before any customer contact, since outreach made without them could disrupt a review already in progress and expose the bank.
- Re-underwrite the full relationship with a fresh liquidity analysis, an updated collateral appraisal, and a documented decision on whether the line continues, reduces, or exits.
Results
Compliance filed its review within the required window, and the bank declined further outbound wires to the supplier account. The physical count put inventory at roughly 70 percent of the last appraisal, which cut the borrowing base and reduced the available line to $280,000. Redlake repaid the balance over the next quarter and moved its operating account elsewhere. The credit file now holds the timeline, the indicators, the count, and the decision, each item dated and initialed. Guillermo's lesson was that monitoring had worked, but the collateral review had not been connected to account activity until the alert arrived.
Questions for discussion
- What would have been the minimum evidence needed to escalate this file in the first week, before the monitoring case opened?
- How should a lender weigh a current loan against a collateral pool it has never physically verified?
- When a customer offers a plausible explanation for cash activity, what documentation would you need before it changes the escalation decision?
Case Study 4
Ferncliff Roasting Works seeks expansion term loan
Ferncliff Roasting Works, a specialty coffee roaster with two retail cafes and a wholesale book of restaurant and grocery accounts.
The situation
Ferncliff asked for a $750,000 seven-year term loan to buy a second roasting line, fit out a production room, and refinance a $180,000 equipment note carrying a higher rate. The owner, Desmond, presented three years of financial statements that showed revenue rising from $2,100,000 to $2,760,000, with the most recent year producing $312,000 in EBITDA. The wholesale channel had grown fastest, and two grocery accounts represented about 38 percent of wholesale revenue. The cafes were roughly break-even. The proposed loan would bring total funded debt to about $967,000, or 3.1 times EBITDA. The statements were prepared by an outside accounting firm and tied to tax returns for the two prior years.
The challenge
Desmond's case rested on a second roasting line that would double production capacity, but the statements showed a business whose growth depended on two grocery accounts and whose cafes only broke even. The credit committee wanted to know whether the company could carry $967,000 of debt if the wholesale channel softened, and what conditions would make the answer acceptable. The analyst, Naveen, had to separate a forecast the owner believed from one the bank could underwrite.
Analysis
The statements helped and hurt. They were prepared by an outside firm and tied to tax returns, which gave Naveen some confidence in revenue. But the firm had compiled rather than reviewed them, so the balances were untested. Naveen compared wholesale revenue against shipping records and the grocery accounts' payment history over eighteen months. Receivables had stretched from about 34 days to 51 days, and the two largest customers accounted for most of the increase. The inventory balance rose faster than sales, which could mean slow-moving green coffee or a valuation issue. Neither was disqualifying, but each changed how much reported EBITDA was cash.
The downside case set the terms. Naveen modeled three scenarios over the seven-year term. The base case put EBITDA at $380,000 with coverage above 1.5 times. The downside case cut grocery volume by a third and dropped wholesale margin two points, which brought EBITDA to about $240,000 and coverage to roughly 1.05 times. The committee's minimum was 1.30 times. A third case, in which a cafe closed and its lease stayed, fell below 1.0. The loan could proceed only if the downside case cleared 1.30 at a smaller amount or with more owner equity. Desmond's personal guarantee was the only secondary source, which made it a weak substitute for cash flow.
Actions taken
- Require a quality-of-earnings style review by an independent accountant, limited to receivables, inventory, and the wholesale revenue reconciliation, before the credit memo is finalized.
- Cap the loan at $600,000 unless the downside case clears 1.30 times, and require the owner to contribute at least $150,000 in equity toward the second roasting line.
- Set a covenant that no single account exceeds 25 percent of wholesale revenue, with a two-quarter cure period before a breach triggers default.
- Require quarterly statements and an aged receivables schedule within thirty days of each quarter end, and a standing review if coverage falls below 1.30 times on a trailing basis.
Results
The independent review confirmed the wholesale revenue but found $41,000 of receivables past sixty days with no reserve against them. Desmond agreed to a $600,000 facility with $150,000 of his own equity and the concentration covenant. Coverage in the downside case, recalculated on the smaller facility, came in at 1.36 times, just above the 1.30 threshold. The equipment note stayed with its current lender because the refinancing would have pushed coverage below the line. The second roasting line was installed within two quarters, and the first quarterly statements arrived on schedule.
Questions for discussion
- Which of the three scenarios would you weight most heavily for a seven-year term, and how would that weighting change if the loan ran for five years?
- How would you evaluate a guarantee from an owner whose home equity is the main collateral outside the business?
- If the independent review had found the receivables reserve was adequate, would the grocery concentration covenant still be necessary?
Case Study 5
Commercial loan review uncovers four control gaps
Calder Ridge Credit Union, a member-owned lender whose commercial unit has eleven underwriters and a loan review function reporting to the board audit committee.
The situation
Calder Ridge's internal audit team reviewed a sample of 84 commercial loans made over the prior 12 months, drawn from a population of 412 originations. The sample was weighted toward larger credits and loans with covenant exceptions. Testers found that spreadsheets used to calculate debt service coverage lacked version control, that three approved loans had financial statements older than the credit policy allowed, that collateral appraisals were not checked against the lending limit for two properties, and that exceptions to policy were approved by the same officer who originated the loan. Total commitments in the sampled loans came to $31,800,000. The review was routine, but the findings were serious enough that the audit committee asked for a ranked response.
The challenge
Tamsin Okafor, the chief credit officer, needed a response the board could act on within one meeting. Not every finding carried the same weight. A stale spreadsheet and an officer approving his own exceptions both broke control, but they did not expose the institution the same way. Each finding needed a materiality call, a named owner who could actually fix it, and a closure test that someone outside the lending unit could run. The board had seen remediation plans closed by memo. It wanted evidence.
Analysis
Ranking began with exposure, not with how embarrassing each finding looked. The officer approving his own exceptions ranked first. It touched the authority structure, and the sample held nine exceptions averaging $410,000, several carried past their review dates. The lending-limit gap ranked second because it could place the credit union above its internal limit to a single borrower group if the two properties were pledged to one relationship. Stale financial statements ranked third. They were serious but bounded, since the three loans were performing. The spreadsheet issue ranked last in dollar terms but made every coverage figure unverifiable, so it stayed on the list.
A finding closes when a test passes on work done after the fix, not when the owner reports it fixed. For the self-approved exceptions, the closure test was to pull every exception approved in the following quarter and confirm that a second approver without origination duties signed each one above $100,000, with no failures in a sample of at least 25. For the lending-limit gap, the test was a daily collateral-to-limit report with any unexplained difference above $5,000 escalated that day. Owners were named by role: Tamsin for exceptions, the controller for spreadsheets, and the compliance manager for limits, so no one held two fixes.
- Rank by materiality Weigh exposure and dollars before assigning any owner or deadline.
- Assign owners Owners are named by role and carry no conflicting fix.
- Run closure test Retest on work performed after the fix, using a fresh sample.
- Report to committee Attach test results, not status memos, before any finding is marked closed.
Actions taken
- Reassign approval authority for all policy exceptions so that the originating officer cannot approve, and require a second signature from a credit officer outside the originating team.
- Rebuild the coverage spreadsheets in a controlled template with locked formulas, a version log, and a named preparer and reviewer on every tab before the next audit cycle.
- Pull current financial statements for every commercial borrower whose statements are older than the policy window, and restrict new advances until fresh statements are on file.
- Run the collateral-to-limit reconciliation daily for the two pledged properties and the related borrower group, and report any unexplained difference to the audit committee each month.
Results
Tamsin's team closed the stale-statement and spreadsheet findings within the first quarter, each on a retest of files opened after the fix. The exception finding took longer. Of the 25 exceptions sampled after the reassignment, two lacked a second approver, and the closure test failed until the approval queue was rebuilt. It passed in the following quarter with no failures. The lending-limit reconciliation found one collateral mismatch of $12,000, which was resolved the same day. The audit committee accepted all four closures with the test evidence attached, and the loan review function reported to it for the next two cycles.
Questions for discussion
- Would you rank the stale statements above the self-approved exceptions if the stale loans carried the larger balances, and what would change your materiality view?
- What minimum sample size would you accept for a closure test when the population of fixed items is small?
- How should an audit committee treat a finding that is closed on paper but whose owner has since left the institution?
Learning Resources
Finance CASE STUDIES
China, India, United States, Indonesia, Pakistan, Nigeria, Brazil, Bangladesh, Russia, Japan, Mexico, Philippines, Vietnam, Egypt, Turkey, Germany, United Kingdom, France, Italy, Canada, South Korea, Australia, Spain.