Securities Trading and Derivatives Case Studies
Finance5 Case StudiesUpdated Sep 2026
Securities trading and derivatives work sits where ledgers, market data, and legal obligations meet, so small breaks between them can become large losses or regulatory findings. The five cases below come from the operations, risk, and finance teams of fictional firms. Each one follows a single problem from first signal to closure: a reconciliation gap, a forecast that stopped matching reality, a set of risk alerts, a funding request, and an audit report. Read them for the thresholds people set, the evidence they asked for, and the points where they chose to stop and escalate.
Case Study 1
The cash variance that would not close
Pinecrest Clearing Services, an introducing broker that routes equity and listed options orders for about 2,400 retail and small institutional accounts.
The situation
Pinecrest runs a small back office for equity and listed options trades. Each night, executions flow from the trade capture system to the general ledger, and the custodian's position and cash files arrive for matching. In the third week of the quarter, the cash difference on the custodian statement was 412,650 dollars above the internal ledger. The gap matched no single trade, though a corporate action on one held name fell near the break dates. Marisol, the operations lead, owned the reconciliation, and Dev, the controller, needed a figure he could sign before the filing package closed.
The challenge
The difference had grown over two nights, and matching by trade reference found nothing. One explanation was a missed dividend reinvestment at the custodian, the other a duplicate manual journal entry made to correct an earlier fee. Each predicted a different residual, and neither had documents attached. The team had to find the variance, decide whether it was a timing item or a loss, and tighten the process so it would not recur.
Analysis
Teams grounded in Securities Trading and Derivatives usually begin by separating the break into known items and unexplained residual before anyone proposes a cause. Here the custodian file was the external record and the general ledger the internal one, and each had to be rebuilt from the same cut-off time. Once the dividend and the journal entry were tested against the statement, the residual was small enough to classify. Pinecrest set the working threshold at a fixed amount per account, with any unexplained item above it escalated the same day regardless of its likely cause.
The more important question was exposure. A cash break of this size, left open into the filing period, could misstate the client money position that the firm reports, and a residual that persisted across two statement dates pointed to a control gap rather than a timing error. Dev asked for three pieces of evidence before signing: the custodian's original file with hash totals, the approval trail for the manual journal, and a list of corporate actions posted in the window. Only after all three were in hand did the team agree on the cause: a timing item plus one reversible duplicate, with the residual under the threshold.
- 1Cut-off and extractionFreeze both files at the same time stamp before any matching begins.
- 2Known item matchingClear dividends, fees, and settlement timing against statement lines.
- 3Residual classificationCompare what remains to the threshold and name an accountable owner.
- 4Evidence and correctionAttach source documents, reverse errors, and record both entries.
- 5Closure testTwo clean nights with no unexplained residual before sign-off.
Actions taken
- Match the custodian file to the ledger daily on a fixed cut-off, and hold any item above the escalation threshold for same-day review with a named owner.
- Require two-person approval and an attached source document for every manual journal entry that touches cash, regardless of amount or timing.
- Reverse the duplicate fee entry, record the dividend reinvestment at the correct date, and document both corrections with the statement page they rely on.
- Set a closure test: the residual must be zero or explained by a listed timing item for two consecutive nights before the break is marked resolved.
Results
The residual cleared within the same week. The dividend reinvestment had been booked to the prior settlement date, and the duplicate journal entry accounted for the rest of the variance. The filing package went out on schedule with the cash reconciliation signed by Dev and supported by a single evidence file. Over the following quarter, the same-day escalation rule caught four smaller breaks, none of which grew beyond the threshold. Marisol reported slower manual approvals, which the team accepted as the cost of a control that showed who had acted.
Questions for discussion
- Which of the three evidence items would you have asked for first, and what would you have accepted in place of the custodian file if it were delayed?
- How should the escalation threshold for unexplained cash differences be set when the firm handles both retail and institutional accounts?
- At what point does a recurring small break stop being an operations matter and become a question for the risk committee?
Case Study 2
A revenue forecast that stopped matching the book
Glenhaven Securities, a regional broker-dealer with a derivatives desk for hedging clients and about 380 active institutional accounts.
The situation
Glenhaven built its annual plan around listed equity commissions and a fee stream from its options clearing relationship. By mid-year, commissions ran 18 percent below plan, and the fee stream had lost its largest contributor, a pension manager that moved execution to a larger venue. Costs had not come down with volume. Tomas, the finance lead, owned the forecast, and Ines, the chief financial officer, had to decide by the end of the quarter whether to cut hiring, reprice a service, or draw on the committed credit line. Covenant headroom on that line was the number everyone watched.
The challenge
The plan assumed a recovery in trading volume that had not arrived, and the team disagreed about how long the shortfall would last. A three-month view showed a manageable gap. A nine-month view, using the pension account's lost volume and a slower onboarding pace for new institutional clients, pushed the minimum liquidity buffer close to the covenant floor. Each assumption needed to be tested before any response was approved.
Analysis
A forecast is only as good as the records behind its drivers, so the first step was rebuilding the revenue line from trade-level data rather than from the prior plan. The team separated volume, average fee per ticket, and mix between retail and institutional flow, because each responded differently to the lost account. Several assumptions turned out to be inherited from a model nobody had reviewed in a long while. The revised file tied each driver to a source report and a named owner, which made it possible to test which numbers were stale and which were simply pessimistic.
Three scenarios were built with explicit triggers. The base case assumed partial recovery in volume by the next quarter. The downside assumed the pension loss persisted and onboarding stayed slow. The severe case added a fee repricing that clients would reject. For each, the team recorded the liquidity buffer, the covenant headroom, and the month in which the buffer would breach an agreed floor of 2,500,000 dollars. Both the downside and the severe case crossed that floor, but the downside did so late enough to leave room for a staged response. That timing, more than the size of the gap, shaped the recommendation.
- Week oneRebuild revenue drivers from trade-level records and assign owners.
- Week threeCirculate three scenarios with explicit liquidity triggers to the chief financial officer.
- Week sixRun the repricing pilot with the largest institutional accounts.
- End of quarterDecide on hiring freeze and credit line based on buffer position.
- Next quarterConfirm buffer stayed above floor and narrow forecast variance.
Actions taken
- Rebuild the revenue forecast from trade-level data and tie each driver to a source report with a named owner before any scenario is circulated.
- Set trigger points for action: if the liquidity buffer falls below the agreed floor in the downside case, begin the hiring freeze and the credit line review.
- Hold repricing until the largest institutional clients have responded to a pilot offer, and measure acceptance against the volume assumption in the model.
- Review the scenario file every two weeks against actual cash receipts, and document any change to an assumption with the evidence that prompted it.
Results
The downside case went to the committee with its trigger points, not a single number. Ines approved a hiring freeze for open roles outside client-facing functions and deferred the credit line draw. The pilot repricing was accepted by two of the five largest institutional accounts, enough to cover most of the expected fee recovery. By the end of the following quarter, the buffer had stayed above the floor and forecast variance had narrowed below five percent. Tomas noted that the model was now slower to update but much harder to argue with.
Questions for discussion
- Which assumption in the revenue model would you test first, and what evidence would change your view of it?
- How should a team decide between drawing on committed credit and cutting costs when both options carry different signaling risks to counterparties?
- When a pilot pricing change is accepted by only a few large clients, how much weight should the forecast give to the rest of the book?
Case Study 3
Concentration alerts on a single-name options position
Brackwater Partners, a proprietary trading firm with a small derivatives book and a compliance function of four staff.
The situation
Brackwater runs a proprietary book in listed equity options and single-name futures, with position limits set by the risk committee. A portfolio manager, Kwame, built a position in a mid-cap healthcare name over several sessions, mostly through calls and a short-dated spread. The aggregate delta exposure reached about 4,200 shares equivalent of the underlying, and the single-name limit in the risk policy was set at 3,000. The daily risk report flagged the breach, yet the trades had been split across two accounts and entered in small lots. Compliance received the alert late on the day the position was at its largest.
The challenge
The breach itself was not the hardest part. The harder question was whether the split accounts were a legitimate hedge structure or an attempt to stay under reporting thresholds. Documentation for the hedge existed only as a verbal description, and the rule under which the firm must report large positions applied at the combined level. Compliance had to decide whether to freeze the book, notify the regulator, or wait for the trader's explanation, with the position still open.
Analysis
In a risk alert, the first step is a record review rather than a verdict on the trader. The team rebuilt the exposure across both accounts on a single netting basis, using the same price source the risk system used at the close. Once the combined delta, gamma, and days-to-liquidate sat on one page, the picture changed: the spread the trader described as a hedge did not offset the call exposure in the way the documentation implied. Compliance also pulled the order timestamps, because small lots entered in rapid succession are often the first evidence that a split was deliberate.
The escalation question turned on two thresholds: an internal single-name limit, and the reporting threshold that applies to the combined position. Breaching the internal limit required notice to the risk committee within the same session. The reporting obligation carried a filing deadline measured in business days from the trigger date. Because the hedge documentation was verbal, the firm could not show that the two accounts were economically linked, which made the aggregate figure the one to report. Control evidence, not intent, decided the escalation path.
Actions taken
- Net both accounts on a single basis at the close and record the aggregate delta, gamma, and liquidation time in the breach file.
- Freeze new entries in the name until the risk committee accepts a written hedge rationale, signed by the portfolio manager and the head of trading.
- Pull the reporting determination forward, file within the deadline the compliance manual sets, and keep a copy of the filing with the trade records.
- Add a rule that flags split orders in the same name across linked accounts within a session, and test it against last quarter's trades before relying on it.
Results
The combined position was reported within the deadline, and the book was frozen in the name for the rest of the session. The written hedge rationale arrived two days later, but it described a spread that did not match the order records, so the committee cut the trader's limit and required a second approver. Kwame stayed on the desk with the reduced limit. The split-order rule flagged three further cases in the following quarter, two of which turned out to be legitimate and one of which led to a second filing.
Questions for discussion
- What evidence would have been enough to treat the split accounts as a documented hedge, and who should have been responsible for collecting it?
- Should a firm report on the aggregate basis when the trader's own description of the economics points the other way?
- How should a new split-order rule be tuned so it catches deliberate splitting without burying the desk in false alerts?
Case Study 4
Approving a margin facility for a long-short fund
Larkspur Prime Services, a prime brokerage unit that finances client equity and derivatives positions for about 60 funds.
The situation
Larkspur's prime brokerage unit finances client equity and derivatives positions under agreed haircut schedules. Verdant Ridge, a long-short equity fund with about 310,000,000 dollars in net assets, asked for an 85,000,000 dollar financing line to support small-cap longs and total return swaps. The manager, Oksana, supplied audited statements, twenty-four months of monthly returns, and a position list. Idris, the credit officer, had to recommend approval, a reduced line, or a decline to the unit's credit committee. Two of the largest longs traded thinly, and the swap counterparties were also the fund's other lenders.
The challenge
The return history looked strong, but it was smoothed by month-end marks on the thin positions, and the fund's leverage moved sharply around quarter-end. The downside was the question. If the two illiquid longs had to be sold into a stressed market, the collateral shortfall could exceed the facility's cushion within days. The committee wanted to know what line size and haircut would survive that scenario, and what conditions would make approval acceptable.
Analysis
Start with the records, because a return series that looks smooth often reflects how prices were marked rather than how positions performed. The team requested the valuation policy, the independent price sources for each illiquid name, and the month-end marks against executed trade prices for the prior two years. Where marks were set by the manager, the analysts applied a discount to match realized sales, which lowered the apparent return by a visible margin. The leverage record, reconstructed from the prime account and the fund's administrator statements, showed exposure peaking in the final week of each quarter.
The stress test used the fund's actual positions. The mandate permitted small-cap exposure, so suitability was not the obstacle; liquidity was. Under a scenario where the illiquid longs fell by a third and could be sold only over several sessions, the collateral surplus turned negative in the second session. Shared counterparties meant one margin call could trigger others. Larkspur's conditions were a line of 40,000,000 dollars, a haircut schedule on illiquid names set from the stressed case, daily collateral reporting, and a reduction right on any breach of the leverage ceiling.
- Evidence review Rebuild returns from marks, trades, and administrator statements.
- Downside testing Run stressed liquidation and counterparty margin scenarios on actual positions.
- Terms and conditions Size the line, set haircuts, and write in reduction rights.
- Monitoring and review Check collateral daily and reassess when leverage or illiquidity rises.
Actions taken
- Require the valuation policy and independent prices for every position above a set share of net assets before the committee sees the return history.
- Set the line at 40,000,000 dollars with a haircut schedule built from the stressed scenario, and review it whenever the illiquid share of the book rises.
- Receive collateral and exposure reports daily, with a separate intraday check during the final week of each quarter when leverage peaks.
- Write a reduction right into the agreement that cuts the line automatically if leverage exceeds the ceiling for two consecutive reporting days.
Results
The committee approved a reduced line of 40,000,000 dollars with the haircut schedule and reporting conditions. The manager pushed back on haircuts for two liquid names, and the committee accepted a narrower schedule for those. Four months into the facility, the fund's leverage ceiling was breached during a quarter-end window, the reduction clause triggered, and the line fell to 25,000,000 dollars with no loss to Larkspur. Idris noted that the same request at the original size would have required a margin call during that window.
Questions for discussion
- Which of the stress assumptions would you challenge first, and what evidence would you need to change it?
- How should a lender weigh a strong historical return series against valuation practices that rely on the manager's own marks?
- Where should the line between a reduced facility and a decline sit when the borrower's strategy is permitted but its liquidity is thin?
Case Study 5
Ranking control gaps after a trade settlement review
Tallow Creek Financial, a mid-sized futures commission merchant with about 14,000 client accounts and an internal audit group of six.
The situation
Tallow Creek's internal audit team reviewed the settlement and margin process for listed futures and cleared derivatives over a three-month window. The review produced 11 observations across margin calls, confirmations, client money segregation, and system access. Three were tied to the daily margin run, where a manual override was used 37 times without documented approval. Hannah, the chief risk officer, and Pedro, head of operations, each received a draft that listed findings in no particular order. The audit committee wanted a ranked list with owners and dates before its next meeting.
The challenge
Ranking the findings was harder than listing them. Materiality here depended on the amount of client money exposed, the frequency of the override, and whether the weakness could have led to an unnoticed shortfall. Some observations were procedural and easy to fix, while others touched controls the firm relies on to demonstrate segregation to its regulator. Each owner needed a clear test that would show the fix worked, not just that a document had been updated.
Analysis
Materiality was set by exposure, not by how embarrassing a finding looked. The team recalculated the client money that passed through each affected process over the review window, then tested how many margin calls had been satisfied by overrides rather than by the system's own run. The override path touched about 6,300,000 dollars of client margin across the 37 occasions, which put it at the top of the list even though the individual items were small. Confirmation exceptions ranked lower because they were matched by the counterparty within a day, and the residual unmatched count stayed below the threshold the firm had set for escalation.
Owners were assigned to the person who controlled the process rather than the person who would report on it. The override finding went to Pedro, with Hannah holding the closure sign-off so that the operations team could not close its own work. Each closure test required a sample and a result. For the override control, the test was that every manual margin change in the following quarter had a dated approval, a reason code, and a second reviewer, with zero exceptions in a sample of at least 200 entries. Findings below the materiality threshold were closed on a documented walk-through, while those above it required re-performance by internal audit.
Actions taken
- Disable the manual margin override in the position system, and route any exception through a request form that requires a named approver and a reason code.
- Assign the override finding to the operations head, with closure sign-off held by the chief risk officer so the owner cannot approve their own fix.
- Set a closure test for each finding: a defined sample, a stated pass rate, and re-performance by internal audit for anything above the materiality threshold.
- Report the ranked list and closure status to the audit committee at each meeting until every high-rank item has passed its test on two consecutive reviews.
Results
Within the quarter, the override function was disabled and replaced by a request form, and the first 40 exception requests all carried approvals. The override closure test was passed in the following review period, with no exceptions in a sample of 260 entries. Two lower-ranked confirmation items were closed on walk-through rather than re-performance, which the committee accepted because their residual exposure was small. When the external examiners asked for test evidence, the file contained it without a further request for documents.
Questions for discussion
- How would you weigh a small procedural finding that touches a large number of accounts against a rare but high-value control gap?
- Should closure sign-off ever sit with the same function that ran the process, and under what conditions would you accept that?
- What sample size and pass rate would convince you that a manual control has truly been fixed, rather than only documented?
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.