Investment Banking Analysis Case Studies
Finance5 Case StudiesUpdated Sep 2026
The five cases below come from the working life of several invented firms, including a bank, a broker dealer, and a private credit fund. Each opens with a problem an analyst could misread on first pass, such as a ledger that does not tie or a forecast that has drifted. The mechanics are concrete, including the thresholds people argued over and the evidence that settled each question. Read them in order. The same records and exposure measures reappear in different forms, so the habit of asking what would change the answer matters more than any single number.
Case Study 1
Unmatched Settlement Breaks During a Bond Syndicate Close
Kestrel Row Advisors, a boutique investment bank with two offices serving mid-sized industrial clients.
The situation
Kestrel Row acted as joint bookrunner on a corporate bond deal sized at $185,000,000 in par value. Allocations were finalized on the evening of the pricing day and entered into the firm's allocation ledger by the operations team. Three days later, during the monthly close, the middle office found that the ledger, the custodian's settlement statement, and the signed confirmations did not agree. The ledger carried 1,284 allocation lines across 37 accounts. The custodian had settled 1,262. Confirmations for two accounts still showed the original indicative sizes, and the total par value on the confirmations ran about $2,300,000 above the ledger. Nobody could say yet which record was wrong.
The challenge
The desk wanted the break cleared before the close sign-off due at week's end. Two accounts had already received allocation notices, and one had asked for a change to its settlement instructions the previous afternoon. The controller, Dana, needed to know whether the $2,300,000 gap was a booking error, a missed amendment, or a real shortfall in delivery. Each explanation pointed to a different fix and a different conversation with the custodian and the affected clients.
Analysis
The team rebuilt the trail record by record rather than trusting the summary totals. Teams grounded in Investment Banking Analysis usually start by separating breaks that change exposure from breaks that only change presentation. The 22 allocation lines missing from settlement mattered because they left an open delivery obligation of about $2,300,000 in par. Signed confirmations were the strongest evidence, but two of them had been superseded by amendments the ledger never received, so the team matched amendment timestamps against ledger entry logs. They agreed a threshold first: unmatched par above $250,000 would be escalated the same day, and smaller items would be logged and cleared in the normal cycle.
Once the gap was confirmed, the exposure question changed. The firm held no bonds, so the risk was operational and counterparty risk rather than market risk, but every open day carried a replacement cost. The controller asked for a daily mark on the unsettled quantity. The control failure was narrower than it looked: the amendment process existed on paper, yet nothing required operations to refresh the ledger after a settlement instruction changed. The corrective design was a hard stop, so no allocation notice could go out until the ledger, custodian feed, and confirmations agreed to zero or to a signed exception.
Actions taken
- Lock the allocation ledger after pricing, so any change to size or account must come through a logged amendment that operations can trace to a signed instruction.
- Run a daily three-way match of ledger, custodian statement, and confirmations until every open item is either cleared or formally excepted with an owner and a date.
- Escalate any unmatched par above $250,000 to the controller the same day, and record smaller residuals in a standing exceptions log reviewed weekly.
- Require a sign-off from the account manager on every amendment before it reaches operations, and keep the sign-off attached to the ledger entry for audit.
Results
The team traced the gap to two causes. One account changed its settlement instructions after the last ledger update, and a second confirmation was reissued at a larger size that was never keyed. Together they explained the full $2,300,000 difference in par. The custodian settled the open lines after the instruction was corrected, and the exposure closed before sign-off. Operations now refreshes the ledger on every amendment, and the same-day escalation threshold is written into the close checklist. The next cycle logged 14 residual items, all below threshold and cleared on time.
Questions for discussion
- Which of the three records would you have trusted first, and what would have changed that ranking?
- How would the escalation threshold change if the unmatched lines had sat on the firm's own book rather than with clients?
- What evidence would convince an auditor that an amendment process works in practice rather than only on paper?
Case Study 2
Revising a Forecast After Volume Falls Short
Greave Lantern Capital, a lending and advisory group that serves regional manufacturers and their sponsors.
The situation
Greave Lantern's leveraged finance group entered the year with a fee plan of $41,500,000, built on eleven advisory and underwriting mandates expected to close in the first two thirds of the year. By the midpoint review, four mandates had closed and fees recognized stood at $14,200,000, against a pro rata plan of $19,800,000. Two large mandates had moved their signing dates, and a third had been paused by its sponsor. Compensation accruals and the lease commitments on the Chicago office were fixed for the year, so the gap was flowing directly into operating cash flow. The head of the group, Tomas, asked for a revised forecast before the next capital committee meeting.
The challenge
The revised forecast had to answer three questions for the capital committee: how much of the shortfall was timing and how much was lost, whether the cash position could cover fixed costs through the year, and what the group should do about hiring plans already approved. Assumptions about close probability were inconsistent across the pipeline, with some deals carrying a probability of 90 percent despite having no signed term sheet.
Analysis
The team built three scenarios instead of one revised figure. The base case moved each delayed mandate to its latest signing date and weighted it by documented stage, so a deal with a signed term sheet carried a far higher probability than one still in early discussion. The downside case dropped the paused mandate and pushed one delayed deal beyond the year. The upside case assumed both delayed deals signed in the final stretch. Each scenario was tested against the same fixed cost base, because the team wanted to see how quickly the cash position would breach the minimum liquidity covenant the group had agreed with its bank.
Assumption testing mattered more than the scenario math. Two deals carried probabilities of 90 percent without signed documents, and when the team asked for the last email exchange with the sponsor, one of them had no contact in more than a month. Those probabilities came out of the forecast and were replaced with stage-based values. The team set decision thresholds for the response: if liquidity headroom fell below $6,000,000 in any scenario, discretionary hiring would pause and the bonus accrual would be reviewed. If the downside case stayed above that line, no structural action was needed. The committee approved the threshold rather than the midpoint.
- Refresh pipeline Re-date every delayed mandate from signed documents and latest sponsor contact.
- Reweight probabilities Assign close probability from stage evidence, not from banker judgment.
- Test scenarios Run base, downside, and upside against the fixed cost base and covenant.
- Compare to thresholds Trigger hiring pause or bonus review only when headroom breaches the line.
Actions taken
- Re-date every delayed mandate only from a signed document or a dated sponsor communication, and drop any deal that has gone silent past the agreed follow-up window.
- Assign close probabilities by documented stage using a fixed table, and require the banker to explain any move outside that table in writing.
- Model the base, downside, and upside cases against the fixed cost base and the liquidity covenant, and show the month in which headroom first falls below the line.
- Pause discretionary hiring and review the bonus accrual automatically if headroom breaches the threshold in any scenario, and report the trigger at each committee meeting.
Results
The group's leadership signed off on the revised forecast with a downside case as the planning base. Fee revenue for the year was set at $36,400,000, about $5,100,000 below the original plan, with the two delayed deals supplying most of the recovery in the final stretch. Discretionary hiring paused for one quarter, which removed four open requisitions. Even the downside case stayed above the covenant once the pause took effect. The group now reviews the probability table each month, and the sponsor contact field is required before any deal can carry a probability above 50 percent.
Questions for discussion
- Which assumption in the pipeline would you test first, and what document would count as enough evidence?
- How should the team treat a deal that has slipped twice but still has an active and engaged sponsor?
- What would make you reject the downside case as too pessimistic, or the upside case as too generous?
Case Study 3
Warning Flags on a Concentrated Note Position
Ashgrove Lane Securities, a broker dealer whose structured products desk sells notes to private and advised clients.
The situation
Ashgrove Lane's structured products desk had built a position of $58,000,000 in notes referencing a basket of regional utilities. By the time the daily risk report was reviewed, 61 percent of the desk's inventory of these notes sat with one reference issuer, and that issuer's credit spread had widened by a margin the desk's stress test had assigned to a much rarer move. Separately, the surveillance team flagged three client accounts that bought a combined $9,400,000 of the notes within two days, all under the same introducing adviser, and at prices above the prevailing mid. None of the three had an updated suitability questionnaire on file.
The challenge
The question was whether this was a concentration that had grown over time, a sales pattern that needed escalation to compliance, or both. The desk head, Ruth, wanted to keep the position open because the notes were marked on a model that had been left unchanged for some time. The compliance officer, Bilal, wanted the three accounts frozen to new purchases until suitability was documented. Both positions had some merit, and the disagreement needed a decision rule rather than a debate.
Analysis
The exposure review started with the model, not the market. The notes were marked on a spread curve that had not been recalibrated recently, so reported value understated the loss already implied by the widening. Recomputing value on observed spreads produced a gap large enough to breach the desk's risk limit. Concentration was the sharper issue, since one reference name drove most of the inventory and the basket structure offered little diversification against its default. Documentation was sorted into complete files, files missing only a signature, and files with no suitability evidence at all.
The escalation analysis turned on whether the sales pattern was a suitability breach or a surveillance event. Three accounts buying within two days through one adviser at prices above mid could reflect genuine client demand, but it could also indicate that the adviser was steering purchases. The applicable rule set required a documented review of any pattern where a supervised person's recommendations concentrated in a single product, and the desk had no such review on file. The team set two thresholds: any account without current suitability evidence would be restricted to closing trades, and any adviser-linked pattern would go to compliance the same day, whatever the position's merit.
Actions taken
- Restrict the three accounts to closing trades until current suitability questionnaires are signed, and document the restriction in the client file on the same day.
- Replace the stale spread curve with observed levels, rerun the valuation, and bring the mark change to the valuation committee before the next reporting cycle.
- Set a single-name concentration limit for the structured inventory and sell down the excess through counterparties with suitable mandates, not through the original adviser.
- Route any pattern of purchases linked to one introducing adviser to compliance the same day, with the rule and the threshold written into the surveillance procedure.
Results
Compliance restricted the three accounts to closing trades pending updated questionnaires, and the desk head agreed to a revised reference model that moved the position to a fair value about $1,100,000 below the prior mark. The adviser was placed under heightened review, and two of the three clients later confirmed interest but asked for smaller allocations. The desk reduced concentration by selling down part of the inventory to a counterparty with suitable mandates. Residual exposure to the single issuer stayed inside a tightened limit, and adviser-linked sales now generate a same-day alert.
Questions for discussion
- At what concentration level would you have pushed for a limit breach rather than a model review?
- How should a compliance officer weigh a desk head's view of a position when the documentation is incomplete?
- What evidence would separate genuine client demand from adviser steering in a pattern like this?
Case Study 4
Credit Request for an Acquisition Term Loan
Marrow Fen Capital, a private credit fund that lends to owner-operated industrial companies.
The situation
A regional packaging manufacturer, Dunmore Flats Holdings, asked Marrow Fen for a $48,000,000 senior term loan to acquire a smaller competitor. The target reported trailing EBITDA of $11,200,000, but its figures had been prepared by an outside bookkeeper and never audited. The borrower's own business had shown steady volume for three years, with one customer accounting for 34 percent of revenue. Pro forma leverage at close would sit at about 4.6 times, above the fund's internal ceiling of 4.0 times for this sector. The request arrived with a sponsor timetable that left little room for a full diligence cycle.
The challenge
The investment committee needed a recommendation that would hold under a weaker year, not just the borrower's management case. The question was whether the combined business could service the loan through a downturn in volume with the largest customer at risk, and whether the unaudited target figures were reliable enough to underwrite. The private equity sponsor also wanted a faster answer than the normal diligence window allowed.
Analysis
The first pass tested the management case against the borrower's own history. Revenue had grown in the two most recent periods, but margin had compressed as raw material costs rose, which the management case treated as temporary. The team rebuilt EBITDA for the target from bank statements and tax filings rather than the seller's summary, and the rebuilt figure came in about $1,600,000 lower. Debt service coverage on the rebuilt numbers stayed above 1.2 times only if the largest customer renewed at current pricing. A downside case cut a third of that customer's volume and erased the margin recovery, pushing coverage near 1.0 times.
Suitability depended on structure as much as on the ratio. At 4.6 times the loan breached the fund's ceiling, so the team tested a subordinated seller note that brought senior leverage to 3.8 times. That worked on paper, but the seller note carried a cash interest toggle the fund had not priced. A customer at 34 percent of revenue also called for a reporting trigger on its payment behavior, not only a leverage test. Approval conditions became a short list: audited target financials, a key customer contract running beyond the loan's first year, and a seller note barred from cash interest while senior coverage sits below 1.5 times.
Actions taken
- Require audited financials for the target covering the two most recent periods, with a bridge from audited EBITDA to the figures used in the credit memo.
- Obtain a renewal or extension of the largest customer contract that runs beyond the loan's first year, or size the senior debt to hold coverage above 1.2 times without it.
- Price the seller note as subordinated equity, and bar any cash interest payment while senior coverage sits below 1.5 times, so the structure protects the senior lenders.
- Set a walk-away point in writing before the sponsor call, so the committee can decline quickly if the audit or the renewal does not arrive in time.
Results
The committee declined the loan as structured and countered at $38,000,000 of senior debt, with the balance covered by equity and a seller note. Audited target financials landed close to the rebuilt figures, which eased one concern but left customer risk untouched. The borrower accepted, and the loan closed with a renewal covenant on the largest customer and a quarterly reporting trigger. In the first year, coverage ran at about 1.4 times in the base case, and the fund has since kept the sector limit unchanged.
Questions for discussion
- Which of the five scores would you revisit first if the largest customer signed a longer renewal?
- Should a fund accept unaudited target figures when a rebuilt set lands close to the seller's numbers? Explain the threshold you would use.
- How would you set the point at which the committee walks away when the sponsor is pressing for speed?
Case Study 5
Ranking Control Gaps After a Fee Review
Pembroke Vale Capital, an investment bank with about 210 staff, a retail wealth arm, and a small advisory practice.
The situation
Pembroke Vale's internal audit team reviewed how advisory fees were recorded, billed, and recognized over the most recent fiscal period. The review sampled 64 engagements out of 212 active mandates. In 9 of the 64 files, the fee invoice and the engagement letter disagreed on the percentage or the milestone trigger. Three engagements had fees recognized before the milestone was met, and one invoice had been issued to the wrong legal entity. Separately, the fee approval workflow allowed a single senior banker to both approve a fee change and release the invoice, with no second signature recorded in the system. Management had not yet linked any of these items to the revenue close.
The challenge
The audit committee wanted the findings ranked by materiality rather than by how many files each one touched. Management had grouped everything into one plan with a shared deadline, and the head of advisory wanted the visible items closed first, the wrong-entity invoices. Each finding needed a named owner, a clear view of the revenue at risk, and a test that an outside reviewer could repeat without relying on management's word.
Analysis
Materiality was measured against revenue recognized, not against the count of exceptions. The three early-recognition engagements carried $4,300,000 of fees booked ahead of their milestones, which made them the most significant finding even though they were the fewest. The approval segregation gap affected every fee change in the period, so its exposure was broader but harder to quantify until the team pulled the change log. The percentage mismatches in nine files were real errors, but most were small and did not change recognized revenue by more than a few thousand dollars in aggregate.
Each finding received one accountable owner, and the owner was someone who could change the underlying process. The early recognition finding went to Priya, the head of finance operations, since only that function could hold revenue until a milestone was evidenced. Approval segregation went to Owen, the chief operating officer, since the workflow configuration sat with technology under his direction. Closure tests were written so an outside reviewer could repeat them: for early recognition, a sample of every engagement with a recognized fee above $250,000, matched to signed milestone evidence. Segregation was tested with a system report showing no fee change released by its approver.
- 1Rank by materialityOrder by revenue at risk and effect on close, not exception count.
- 2Assign single ownersEach owner must be able to change the process, not only report on it.
- 3Fix root causesHold early recognition until milestones are evidenced; split approve and release roles.
- 4Correct the sampleRe-perform the nine percentage mismatches and reissue wrong-entity invoices.
- 5Test closure evidenceOutside reviewer repeats the sample and system report before any finding closes.
Actions taken
- Rank the findings by revenue at risk and by effect on the close, and present the ranking to the audit committee with the basis for each rating written beside it.
- Give each finding one named owner with authority over the process involved, and set interim checkpoints that the owner reports against each month.
- Hold revenue on any engagement lacking signed milestone evidence, and separate the approval and release roles in the fee workflow before the next billing run.
- Define closure tests in advance: an outside reviewer re-performs the sample and pulls the system report, and a finding closes only when both show no exceptions.
Results
Management accepted the ranking with one change. Segregation moved to the top once the change log showed 38 fee changes approved and released by the same banker. The early recognition engagements were reversed from revenue, and the head of finance operations now holds recognition until milestone evidence is attached. The closure test for segregation passed on the first repeat, while the early recognition test found one further engagement missing from the original sample. That file joined the remediation list, and the finding stayed open until the sample was re-run.
Questions for discussion
- If the change log had shown no approvals released by the same banker, how would you rank the segregation finding?
- What evidence would you require before accepting that a revenue milestone was met?
- Who should own a finding when the fix depends on both technology and a business unit, and how would you record that split?
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.