Accounting, Audit and Financial Compliance Case Studies
Professional5 Case StudiesUpdated Sep 2026
Accounting, audit, and compliance work comes down to whether numbers agree with their records and whether someone with authority checked that they do. These five cases cover a reconciliation that would not close, a forecast that stopped matching cash, a payment pattern that tripped a monitoring rule, a credit request from a growing supplier, and an audit that found gaps in review. Each follows one thread through transaction records, ledger integrity, approval controls, and close deadlines. The goal is to show how practitioners narrow a problem, decide what evidence is enough, and build fixes that outlast their builders.
Case Study 1
The subledger gap that stalled the quarter close
Harbor Lane Medical Supply, a regional distributor of clinical consumables with four warehouses.
The situation
Harbor Lane Medical Supply runs four regional warehouses and carries roughly 1,200 active customer accounts. In the last week of the quarter, the accounts receivable subledger showed a balance $84,310 lower than the general ledger control account, and the cash application team could not explain the gap. The bank reconciliation had cleared, which made the problem harder to see. Remittance advice from three customers did not match the amounts posted to their accounts, and two incoming wires had been applied by hand to invoices that were already closed. Priya, the assistant controller, had until the fourth business day of the close calendar to deliver a clean trial balance to the parent company's group reporting team.
The challenge
The obvious explanation was a posting timing issue, and Priya's first pass confirmed it for part of the gap. But the remaining figure did not line up with any single batch. The team had to decide how far to trace before escalating, and whether to adjust the control account or the subledger first. Pressure from group reporting was real, but a plug entry would have buried the cause and left the same weakness for the next close.
Analysis
Practitioners trained in Accounting, Audit and Financial Compliance tend to start with the ledger itself, not the statement. The useful move was to pull every receipt posted to the control account in the final ten days and match each one to its bank line by value and reference. That left a short list of entries where the reference had been typed by hand, often truncated. Two of them pointed to one customer, and both had been posted twice, once from the lockbox file and once by hand. The subledger was duplicated rather than wrong, and the trail existed only because batch identifiers had been kept.
Approval controls had failed in a quieter way. Manual cash applications above a set threshold required a second signature, but the threshold was applied to the invoice amount rather than the cash amount, so a split wire could slip under it. The close calendar set no cutoff for manual cash entries, so adjustments could still land on the fourth day and pass as routine. The fix was narrower than a redesign: test the threshold against cash received, set a hard cutoff, and flag manual entries on closed invoices for controller review.
Actions taken
- Lock the lockbox import so each file hash is recorded at load and any re-import of the same batch is rejected automatically by the posting tool.
- Change the second-signature threshold to test cash received rather than invoice value, and record the change in the approval matrix with the reviewer who signed off.
- Set a hard cutoff for manual cash entries at the start of the close window, and route anything later through a labeled post-close adjustment queue.
- Run a weekly exception report of manual entries against closed invoices for the next two close cycles, and have the controller initial each item before the trial balance is released.
Results
The $84,310 gap resolved into three duplicate postings and one misapplied wire, which together accounted for all but $212 of the difference. That residual was a bank fee that had been netted in the wrong period and was corrected with a documented reclass rather than a plug. The trial balance went to group reporting within the deadline. Over the next two closes the exception report flagged eleven manual entries, nine of which were legitimate and two of which led to coaching for a single clerk. The lockbox duplicate check caught one re-import attempt in the following month.
Questions for discussion
- Which evidence would you require before accepting a plug entry as a temporary fix, and who should be allowed to approve it?
- How would you decide whether the cash application threshold should test invoice value or cash received when a single wire settles several invoices?
- What other posting paths into the subledger would you test for the same duplication risk once the lockbox import is locked?
Case Study 2
Rolling forecast after orders fell short of plan
Ostrander Pavilion Works, a mid-size maker of aluminum patio furniture with two plants and a wholesale channel.
The situation
Ostrander Pavilion Works makes aluminum patio furniture at two plants and sells mostly through regional wholesale accounts. The annual plan assumed steady reorders from its three largest dealers. By the end of the second quarter, those dealers had cut orders by roughly a fifth, and revenue sat $3,860,000 under plan. Gross margin held up better than volume, but operating cash flow had turned negative for two months running. The revolving credit line was drawn to $6,200,000 against a $9,000,000 facility, and the covenant test was due at the next quarter close.
The challenge
The controller, Malik, had two forecasts on the table: a recovery case built on dealer commitments that were not yet signed, and a downside case that assumed no reorders until the next season. The board wanted one number for the lender conversation. The team had to decide which assumptions could be defended with evidence on hand and which were hopeful. Choosing the wrong base case would either trigger an unnecessary covenant waiver request or leave the lender surprised later.
Analysis
The forecast was only as sound as the order book underneath it, so the first step was to tie dealer commitments to transaction records. Malik pulled every purchase order and matched it to a signed confirmation. About a third of the recovery volume sat in entered but unconfirmed orders, and several carried prices off the current list. Returns from two dealers had also been netted against revenue in the wrong period, which flattered the trend. Correcting those entries before modeling moved the starting point more than any assumption change would.
Approval controls shaped the forecast more than the market did. Two dealers had received concessions through a standing exception that no one had reviewed, and the discount matrix required sign off above a set percentage. The covenant certificate needed a reconciled revenue figure, which meant unapproved concessions could not be counted. So the team split the forecast into two layers: a base case built only from confirmed, approved volume, and an upside layer labeled as management intent with each assumption tagged to its approval status.
Actions taken
- Rebuild the revenue base from confirmed purchase orders only, and show unconfirmed dealer volume as a separate line so the lender sees what is contingent.
- Run a cash bridge from operating cash flow to revolver draw for each month of the forecast, using actual payment terms from the receivables aging rather than standard terms.
- Stress the base case with a reorder delay of one full cycle and a five point margin decline, then record which covenant ratio breaks first and by how much.
- Ask the sales director to re-approve or retire the standing concession exception before the next close, and attach the approval trail to the forecast file.
Results
The base case showed the revolver peaking at $7,450,000 before recovering in the later months, which kept the leverage covenant inside its limit but with little room. The downside stress broke the covenant by a narrow margin in the second month of the stress period. Management presented the base case with the stress result attached and asked the lender for a standby waiver conditional on dealer orders being confirmed at a defined checkpoint. The lender accepted the framework and requested monthly reporting on the concession exception, which the sales team has since closed out.
Questions for discussion
- Which assumption in the recovery case would you be least willing to defend to a lender, and what evidence would change your view?
- How would you sequence the revised forecast and the covenant conversation if the base case still passes but only with a narrow margin?
- When does a standing discount exception stop being a commercial choice and become a control failure, and who should decide that?
Case Study 3
Clustered settlements and an incomplete merchant file
Greyline Pay Services, a payments processor for about 2,300 small merchants, with a six-person compliance team.
The situation
Greyline Pay Services processes card and ACH payments for about 2,300 small merchants and has a compliance team of six. In the third week of the month, the monitoring system flagged a restaurant supply merchant whose settlement pattern had changed sharply. Daily settlements that had averaged $18,400 were now arriving as a run of transfers each just under the internal review threshold of $9,500, spread across four linked bank accounts. The merchant's onboarding file listed a single owner, but the bank accounts had been opened by two different individuals, and the business address on the application matched a residential property in a neighboring county.
The challenge
Nothing in the pattern proved wrongdoing. Restaurant supply businesses do have lumpy cash cycles, and the owner had recently signed a supplier contract that could explain larger payments. The compliance lead, Terrence, had to decide whether the alert was a documentation problem, a customer behavior problem, or a matter requiring escalation, and he had to make that call before the next settlement batch released funds. Waiting for certainty was not an option, and neither was filing on a hunch.
Analysis
Start with the transaction records rather than the alert summary. Terrence exported every settlement for the prior four months and rebuilt daily totals from the processor ledger, not the merchant portal, which showed rounded figures. The rebuilt series matched the monitoring system's view, which mattered because a mismatch would have pointed to a data fault rather than behavior. He then checked the four receiving accounts against the settlement references. Three carried blank memo fields, and one payment had been reversed and reissued at a different amount, a detail a quick review would miss.
The documentation gaps were the second layer. The file held a beneficial owner declaration for only one of the two people with signing authority, and approval had come from a sales manager, which policy allows only below a set exposure limit. Terrence did not treat either gap as proof of anything. He recorded each one, placed a temporary hold on settlements to the linked accounts, and sent a written escalation to the designated reporting officer, who alone decides on any external filing.
- 1Pattern detectedSettlements cluster below internal review threshold across linked accounts.
- 2Source ledger rebuiltDaily totals rebuilt from processor records and matched to alert data.
- 3Documentation gaps loggedMissing ownership forms and out-of-policy approval recorded with evidence.
- 4Temporary hold appliedLinked accounts restricted while the case is assessed, not closed.
- 5Reporting officer decidesDesignated officer determines whether any external filing is warranted.
Actions taken
- Freeze new onboarding approvals for merchants with more than one signing party until beneficial ownership forms are complete for each person with authority.
- Add a rule that routes settlements clustering under the internal review threshold to a queue with a named reviewer and a same-day clearance deadline.
- Require sales managers to attach the exposure calculation to any approval, and have compliance countersign approvals above the policy limit within the same batch cycle.
- Run a quarterly sample of closed alerts and confirm each has a complete escalation memo, a documented disposition, and evidence that holds were lifted or extended on purpose.
Results
The hold lasted four days while Terrence collected the missing ownership forms. Two of the three signing parties supplied them, and the third declined, which led to the account being restricted rather than closed. The reporting officer reviewed the memo and found that the facts did not meet the threshold for an external filing, but recorded the reasoning and set a review point. Over the following quarter the team cleared 412 alerts, and the rule change reduced the number of settlements waiting more than a day from roughly one in twenty to fewer than one in fifty.
Questions for discussion
- What would have to be true about the owner and the signing parties for the hold to be lifted without an escalation memo?
- How would you test whether the rule that flags settlements just below the review threshold is catching structuring behavior or simply catching seasonal cash cycles?
- Who in your organization should own the decision to restrict an account, and how should that decision be documented?
Case Study 4
Term loan request for a fourth cold storage depot
Quillen Cold Storage Holdings, a family-held refrigerated logistics company with three depots and leased trucks.
The situation
Quillen Cold Storage Holdings wants a $4,200,000 term loan to build a fourth depot near a grocery distribution hub, and the owners have asked for a decision within the quarter. The company has run three depots for nine years, but most revenue comes from two grocery accounts whose contracts renew within the next year. Audited statements show revenue of $31,800,000 and an EBITDA margin that has slipped from the mid-teens to about eleven percent over two years. The owners would fund a third of the project from equity and secure the rest on the new building and the existing equipment.
The challenge
The credit analyst, Hana, had a model that looked healthy on the base case and fragile on almost anything else. The committee would see the base case first. The question was whether the cash flow could carry the debt if one grocery contract was lost or repriced, and whether the owners' equity contribution was real money or a contribution in kind. Hana also needed to decide which conditions would make approval acceptable, since a flat decline was not the only way this could go wrong.
Analysis
Hana began with the ledger, since the projections were only as reliable as the revenue history beneath them. She reconciled three years of invoices to the general ledger and to bank receipts, then compared the two grocery accounts' payment records against contract terms. Revenue tied out, but the receivables aging showed one grocer paying well beyond terms, and some of that delay had been absorbed through credit notes posted near period end. Those notes reversed in the following period, which smoothed the monthly trend and made the margin history look steadier than the cash suggested.
The downside work mattered more than the base case, because the committee's real question was survival. Hana repriced the larger grocery contract at renewal and ran a lost-contract case that cut volume by a third, testing debt service coverage under each. The base case cleared the coverage minimum with narrow room, and both stressed cases fell below it. Suitability also needed attention, since the equity was partly a deferred payment to a related entity that policy does not count as cash. The committee would need to see it converted before funding.
- Week oneRelated-party equity treated as unverified until bank evidence arrives.
- Week threeDownside cases run, and coverage falls below minimum under both stresses.
- End of quarterCommittee reviews conditional recommendation with the lost-contract analysis attached.
- Second week after approvalEquity converted to cash, and first tranche released against the depot deposit.
- Two quarters laterCoverage tests at minimum, and a covenant review opens with the lender.
Actions taken
- Require the related-party equity payment to be converted to cash and deposited into a controlled account before the first draw, with the bank statement as evidence.
- Set a minimum debt service coverage covenant with a cure period, and test it quarterly against the lost-contract case so the margin for error is visible in the agreement.
- Limit the first draw to the depot shell and equipment deposits, and release the remaining tranches only after the second grocery contract is renewed in writing.
- Obtain the receivables aging and credit note register each month for the first year and reconcile them to the general ledger before each tranche is released.
Results
The committee approved the loan at a reduced amount of $3,500,000 with the conditions above. The equity was converted to cash in the second week after approval, which removed the largest open question. The second grocery contract renewed at a lower rate than the base case assumed but above the repricing case, and the first tranche was released against the depot deposit. Two quarters later, coverage tested at the minimum with less room than expected, and the covenant conversation has begun. The owners declined to pledge personal assets beyond the building.
Questions for discussion
- Which downside scenario would you add to the model before the committee sees it, and why that one rather than a general haircut?
- If the related-party equity had been converted to cash before approval, would the conditions still need to include the receivables reconciliation? Why?
- How should a lender weigh a borrower's reluctance to pledge personal assets against the strength of the operating cash flow?
Case Study 5
Ranking audit findings at a regional health network
Thornbury Regional Health Network, a nonprofit operating four hospitals and a physician group.
The situation
Thornbury Regional Health Network runs four hospitals and a physician group with about 6,000 staff and a shared services center that handles accounts payable and payroll for all entities. The internal audit team completed a review of the procure-to-pay and payroll cycles in the second half of the year. Sampling covered 340 vendor payments and 2,100 payroll change records. Auditors found that vendor bank details could be changed by a single user without a second approval, that nineteen payments above the delegated limit had been approved by a supervisor who lacked authority, and that the monthly payroll exception report was reviewed but not signed. Management responses were due within the quarter.
The challenge
Eight findings were open, and the audit committee wanted them ranked by materiality rather than by the order auditors wrote them. Simone, the chief financial officer, asked whether the vendor bank change issue was a fraud risk or a documentation gap, since the answer determined whether the matter went to the board's fraud reporting line. Owners for each finding had not been assigned, and two of the responsible managers had recently changed roles.
Analysis
The vendor bank change was the highest priority, and the transaction records showed why. Eleven of 340 sampled payments went to bank accounts changed in the same week, and in four cases one user ID entered both. The vendor master log recorded change timestamps, but the approver field was blank across the full change population. That is a ledger integrity problem as much as a control gap, because a reviewer cannot tell whether a change was authorized or simply convenient.
Ranking by materiality meant looking past dollar size. The nineteen out-of-authority payments totaled a large sum but were recoverable and sat in a single category, while the bank change weakness touched every vendor in the master file and could redirect funds without a clear trail. Payroll exception reviews were unsigned, which is a documentation finding with low direct loss but would weaken any reliance on the control in the next audit cycle. The ranking put the bank change first, the delegated authority breaches second, and the payroll review third, with owners and closure tests assigned in that order.
Actions taken
- Require a second approver on any change to vendor bank details, enforced in the system, and block the change until the approver's user ID differs from the requester's.
- Pull a full population report of vendor bank changes and payments within the following five days of each change, and send any overlap to the fraud review group.
- Reset delegated payment limits in the approval matrix and remove the supervisor's override, then have the controller recertify all active approvers before the next close.
- Require a signed exception report for each payroll cycle, with the reviewer's name and the time of review recorded in the payroll system's audit trail.
Results
Closure tests were written to need evidence rather than assurance. For the bank change finding, the test required a complete population report with approver IDs, and zero overlaps unexplained. The team reran it at the end of the quarter and found two overlaps, both traced to legitimate changes made by an accounts payable supervisor with documented vendor calls. The delegated limit finding closed once the matrix was reissued and a sample of new payments confirmed the new limits in use. The payroll finding stayed open for one more cycle because two of the signed reports were missing reviewer timestamps.
Questions for discussion
- How would you rank a finding that has low dollar exposure but touches every record in a master file?
- What makes an owner accountable for closure in practice, rather than just named in the tracker?
- When a closure test depends on a report the control owner produces, how would you verify that report independently?
Learning Resources
Professional 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.