Insurance Underwriting, Claims and Reinsurance Case Studies
Finance5 Case StudiesUpdated Sep 2026
Insurance underwriting, claims and reinsurance sit where premium, promises to pay and borrowed capacity meet, so a small error in one ledger can travel far. The five cases here involve fictional carriers, managing general agents and reinsurance intermediaries. Each opens with a problem a finance team would recognize: records that refuse to agree, a forecast that drifted from plan, a portfolio with warning signs, a funding request, and an audit that found controls on paper only. Read them for the evidence each team gathered, the thresholds they set, and the point where they stopped waiting for more data.
Case Study 1
Reconciling a quarter close with three disagreeing ledgers
Harrowgate Mutual Indemnity, a regional property and casualty carrier writing homeowners and small commercial package policies across six states.
The situation
Harrowgate Mutual closed its third quarter with a paid loss balance that did not match. The general ledger showed $4,286,450 in paid losses for the period, while the claims administration system listed payments totaling $4,127,900. Separately, the reinsurance recoverable schedule prepared by Dimitri, the reinsurance accountant, carried $318,600 more than the ceded amounts the treaty team had confirmed with two reinsurers. Nobody had flagged the gap during the monthly review because each balance had been signed off against its own source document. The controller, Shauna, asked for a trace before the statutory filing draft was locked, which left about nine working days.
The challenge
The variance on paid losses was $158,550, and the recoverable gap was $318,600. Neither figure was large against total reserves, but the team could not explain either one from the records they had. Some of the difference might be timing, since checks issued on the last business day had cleared the bank after the system cut off. Some might be duplicate entries or reversals keyed by hand. The team had to separate those explanations, quantify each one, and decide how much variance could stand before the filing deadline.
Analysis
Teams grounded in Insurance Underwriting, Claims and Reinsurance usually begin by refusing to treat each balance as correct until its source has been matched to the next one down the chain. Here the three figures were not three independent facts. The payment register fed the ledger, the ledger fed the filing draft, and the recoverable schedule depended on claim-level cessions that the treaty team tracked separately. Tracing the paid loss gap meant pulling a matched sample of claim numbers from each side, comparing payment dates, amounts and reversal codes, and marking every item as timing, duplication, classification or unexplained. Only the unexplained bucket carried real risk.
The team set a threshold before reconciling the last dollar. Any unexplained variance above $25,000 on paid losses, or any unexplained recoverable difference above $10,000, would hold the filing and go to the chief financial officer. Anything below those levels would be booked with a documented reason and a named owner. That mattered because the actuarial team built reserve estimates on the same paid triangles. The control evidence that came out of the exercise was practical: a signed matching sheet, a list of manual journal entries with preparer and reviewer initials, and a rerun of the system extract with a fixed cutoff.
- First working dayPayment register and ledger totals pulled and compared line by line.
- Second working dayBank clearing dates matched, isolating timing items from true breaks.
- Third working dayRecoverable schedule traced to treaty cession confirmations and endorsements.
- Before filing lockUnexplained items tested against thresholds, then booked with owners.
Actions taken
- Rebuild the payment register extract with a fixed cutoff time, then rerun the paid loss match so that bank clearing items fall into the same period on both sides.
- Require two-person sign off on every manual journal entry above $5,000, with the preparer's note stating the source document and the reversing period.
- Set the recoverable schedule to reconcile to treaty-level cession confirmations each month, with any difference above $10,000 escalated before the filing draft is circulated.
- Add a quarterly control test that samples twenty-five paid claims from the system and traces each one to the ledger, retaining the matched sheet as audit evidence.
Results
The timing items accounted for $112,300 of the paid loss variance once the bank clearing dates were matched. Duplicate reversals explained another $31,850, leaving $14,400 unexplained, which was below the $25,000 threshold and was booked with a named owner and a written reason. The recoverable gap resolved almost entirely: $301,200 was a treaty endorsement that had been ceded in the wrong layer, and $17,400 was a scheduling lag that cleared in the following week. Shauna approved the filing on time. The rerun of the control test in the next quarter found no unexplained items above threshold.
Questions for discussion
- Which of the three reconciling categories would you expect to grow first if the payment cutoff moved by two days, and how would you measure that growth?
- How should the $25,000 and $10,000 thresholds change if the filing window were shorter, and who should be allowed to waive them?
- What evidence would convince an outside reviewer that the manual journal entries were controlled rather than merely signed?
Case Study 2
Recalibrating a specialty auto forecast after loss deterioration
Corrin Ridge Specialty Auto, a surplus lines writer that insures commercial fleets and delivery operators through four wholesale brokers.
The situation
Corrin Ridge planned the year around a premium target of $64,200,000 for its commercial fleet book, with a loss ratio assumption of 61 percent. By the end of the second quarter, written premium was running about 7 percent under plan because two wholesale brokers had shifted volume to competitors, and the reported loss ratio had climbed to 68 percent. Cash flow looked fine on the surface since claim payments were still lagging behind incurred losses, but Pilar, the director of financial planning, could see that the faster premium decline and the rising loss ratio would compress surplus within two quarters. The board wanted a revised forecast and a recommendation before the next capital call discussion.
The challenge
The forecast had three moving parts that interacted: the premium volume, the loss ratio, and the timing of paid claims. Each one could be right on its own and still produce a wrong surplus figure. The team also had to decide which assumptions to stress, because the regulatory capital floor, the reinsurance program's retention, and the expense ratio all responded differently to the same shock. A single downside case would hide the threshold at which the company needed to act.
Analysis
The team built three scenarios from the same general ledger and the same policy count, so every difference could be traced to an assumption rather than to a different starting point. The base case kept the planned premium trajectory but used the reported loss ratio. The broker case removed the lost volume and assumed the remaining book held its rate. The stressed case combined the broker loss with a second accident year of adverse development on commercial auto liability, which the claims department had flagged after reviewing open files. Each scenario was tied to a source: the premium register, the claim reserve listing and the reinsurance treaty schedule.
The decision rule was set before the stressed numbers were reviewed. If the stressed scenario pushed the capital ratio below 1.35 times the regulatory minimum, the company would pause new fleet binding and cut the expense budget. If it stayed above that line, the company would hold appetite and monitor monthly. The stressed case did breach the line, though only by a narrow margin, and the margin depended on a judgment about how much of the broker volume would return. Pilar documented that assumption as judgment, asked the treaty team to confirm the retention on the largest accounts and kept audited figures apart from estimates.
Actions taken
- Rebuild the forecast from the premium register and the claim reserve listing, then tag every assumption as sourced or judgment before the board sees it.
- Pause new fleet binding for accounts whose loss ratio runs above the stressed threshold, and review each renewal individually until the adverse development is settled.
- Cut the expense budget by $1,850,000 for the year, weighted toward discretionary marketing, and report the effect on the expense ratio at each monthly close.
- Ask the treaty team to confirm retention and reinstatement terms on the five largest accounts, and test whether the stop loss would respond under the stressed case.
Results
The board accepted the stressed case as the planning basis and approved the fleet binding pause for accounts above the loss ratio line, which covered about 22 percent of expiring premium. The expense cut of $1,850,000 was delivered in the third quarter, and the monthly close showed the expense ratio falling by roughly half a point. By year end, premium finished about 4 percent under plan and the loss ratio settled near 66 percent, leaving the capital ratio above the minimum but with little room.
Questions for discussion
- Which assumption in the stressed case would you defend to the board first, and what evidence would you bring for it?
- If the broker volume returned in the third quarter, how should the binding pause be unwound without repeating the same mistake?
- Where should a judgment about adverse development sit in a forecast that is meant to be audited?
Case Study 3
Early cancellations and refunds at one producing agency
Westmarch Indemnity Company, a mid-size writer of inland marine and small commercial property coverage placed through independent agencies.
The situation
Westmarch's compliance analyst, Owen, flagged a cluster during a routine review of premium refunds. Over the course of a quarter, one independent agency had produced 41 short-term inland marine policies that were canceled within the first 30 days, with refunds sent to 17 different payee accounts. Premium had been paid by wire from third parties rather than the named insureds. Three of the policies had already produced claims totaling $386,500, all paid within the first 60 days of inception. The agency's written commission statements did not match the premium register by $52,700, and two of the insured addresses matched a single rental property.
The challenge
The indicators pointed in two directions at once. Short-term cancellations with refunds to varied accounts can reflect legitimate customer behavior or an agency that churns policies to earn commission, while early claims on new inland marine policies can reflect fraud or simply poor underwriting. The compliance team had to decide whether this was a claims problem, a premium problem, a producer conduct problem or a financial crime question, because each one carried different reporting duties and different timelines for acting.
Analysis
The first step was to put a number on exposure without assuming the worst. The premium register showed $1,094,000 written through the agency, of which about $612,000 had been refunded and $386,500 had gone out as claim payments on three policies. That gross exposure was not the same as loss. The team separated amounts paid to the named insured from amounts paid to third parties, and it tied each refund to a bank record. Some files held signed applications, others only a typed name. Where the record was missing, the team logged the gap as a control failure rather than treating the absence of evidence as proof of anything.
The escalation rule was set in advance. Any payee account receiving refunds from more than three policies, any third-party premium payment above $5,000, or any claim paid within 60 days of inception would move the file to the special investigations unit. Counsel confirmed that the federal suspicious activity reporting rule did not reach these covered products, so the reporting clock came from the state insurance fraud requirement. The $52,700 commission gap cleared the producer audit threshold, so the agency's appointment was suspended. Open claims were paused for review with written notice to each insured, because the unfair claims settlement rules limit how long a carrier can delay a decision.
- Detect Threshold rules flag refunds, third-party payments and early claims weekly.
- Document Pull signed applications, bank records and payer identity for each file.
- Escalate Route to investigations, pause open payments and start reporting clocks.
- Reassess thresholds Feed closed files back into the rules and retune the thresholds quarterly.
Actions taken
- Hold refunds to any payee account tied to more than three policies until bank records are matched to the named insureds and the premium source is documented.
- Send the agency's commission statements and remittance details to an outside accountant for a producer audit, with results due before any appointment is restored.
- Route each claim on the flagged policies to a senior adjuster who reviews coverage, loss timing and photographs, and records the decision in the claim file.
- Make signed applications, payer identity and premium source mandatory fields before any short-term binder can be issued, and audit a sample each quarter.
Results
The bank match showed that 29 of the 41 policies had premiums paid by the named insured, while 12 were paid from accounts with no documented link to the insured. Of the twelve, the team found nine traced to the agency principal's relatives, which supported the suspension and a referral to the state fraud unit. Claim review found that two of the three early claims were supported by photographs and repair invoices, and the third was denied after a site visit showed the property was vacant. The commission gap resolved to $48,900 in duplicated lines, leaving $3,800 unexplained.
Questions for discussion
- Which threshold would you retune first if the agency argued that its refund pattern reflected a legitimate customer base, and what evidence would settle the matter?
- How should claims handling change while a producer is under audit, and where does the duty to decide promptly limit a pause?
- What would a reviewer need to see before accepting that a signed application proves the named insured actually paid the premium?
Case Study 4
Funding a managing general agent's book acquisition
Solenne Crest Underwriting, a managing general agent placing commercial general liability for contractors in three regions, seeking capital to buy a renewal book.
The situation
Solenne Crest asked for a $12,000,000 credit facility and a $3,500,000 equity contribution to acquire a renewal book from a retiring agency. The target book carried $21,400,000 in annual written premium and had produced a loss ratio of 58 percent over three accident years, but the reserves on the two most recent years had been set by the seller rather than by an independent actuary. Solenne Crest's own financials showed commission income up sharply, but its cash balances were thin, and most of the earnings came from a single carrier's fronting arrangement. Ilse, the head of corporate lending, had to decide whether to recommend approval to the credit committee before the purchase agreement's exclusivity window closed.
The challenge
The question was not whether the book was profitable on paper. It was whether the reserves, the fronting dependency and the thin liquidity could support a $12,000,000 facility through a downside year. Solenne Crest's facility would be repaid from commission and fee income, so the lender's exposure rose steeply if the carrier withdrew its fronting paper or if the seller's reserves proved light. The committee also needed clear conditions, because a yes without covenants would be hard to unwind later.
Analysis
Start with the reserves, since nearly everything else depends on them. The seller's two most recent accident years carried case reserves that had been set by claims staff at the seller and then accepted without an independent review. Solenne Crest's analysts rebuilt the triangles from claim-level loss runs, which pushed the implied loss ratio on the latest year from 58 percent to roughly 67 percent. That still leaves a workable book at the current premium rate, but it narrows the margin of safety. Commission income was also concentrated: about 61 percent of the last twelve months came from one carrier's fronting arrangement, which could be terminated on notice.
The downside case was built around two shocks that could arrive together. The first was reserve development: the implied loss ratio moves to 72 percent, which cuts annual underwriting margin enough that debt service coverage falls below 1.25 times. The second was loss of the fronting carrier on ninety days' notice, which removes most of the commission base in one quarter. Under both shocks combined, cash covers debt service for about five months, which is the minimum runway the committee set. The facility was sized to existing cash flow, not the seller's pro forma, and the equity had to come from owners, not the facility.
Actions taken
- Require an independent actuarial review of the two most recent accident years before any funds move, with reserve changes flowing into the purchase price adjustment.
- Condition the facility on a fronting agreement with at least twelve months of notice, or on a second carrier signing a letter of intent covering at least half of commission income.
- Set a minimum liquidity covenant that keeps at least five months of debt service in cash, tested monthly, with a breach triggering a cash sweep and a fresh review.
- Decline the request if the independent review moves the implied loss ratio above 72 percent, since debt service coverage would then fall below the 1.25 times floor.
Results
The independent review placed the implied loss ratio at 67 percent, which was above the seller's figure but below the rejection line. The purchase price fell by $1,150,000 to reflect the reserve gap, and the committee approved the $12,000,000 facility with the fronting and liquidity conditions attached. The owners put in the $3,500,000 equity from their own funds. The fronting carrier agreed to twelve months of notice in the final documents. In the first quarter after closing, the liquidity covenant was tested twice and passed both times, with cash covering about six months of debt service.
Questions for discussion
- Which of the two downside shocks would you weight more heavily for a fronting-dependent agency, and why?
- If the seller insists its reserves are sound, what evidence would move you away from an independent review as the basis for price?
- How tight should a liquidity covenant be for a business whose cash arrives in lumps around renewal dates?
Case Study 5
Ranking control gaps in a reinsurance recoverables process
Halvard Ridge Treaty Services, an intermediary that places treaty cessions for three regional carriers and administers their recoverable accounts.
The situation
Halvard Ridge's internal audit team reviewed the recoverable accounts after a carrier client disputed a $740,000 balance in the third week of the review period. The audit covered the full cession cycle: treaty setup, claim notification, cession recording, reinsurer confirmation, and cash application. Over the course of four weeks, auditors sampled 64 cessions across three treaties and found that 11 had been recorded in the wrong layer, 6 lacked a signed reinsurer confirmation, and cash receipts were applied to the oldest open items without a documented rule. Javier, the operations director, asked the audit team to rank the findings so the remediation budget could go to the most material ones first.
The challenge
Several findings looked small on their own. A layer error on one cession moved a few thousand dollars, and a missing signature could be fixed with a phone call. Taken together, though, they meant the recoverable balance could not be relied on as a financial asset in the statement, and the cash application rule could misstate which carrier client was owed money. The audit team had to rank by materiality using the balance at risk, the likelihood of a loss, and whether the gap touched a reported figure.
Analysis
Materiality here was a dollar question first and a control question second. The recoverable balance at issue was $740,000, and the sample suggested that layer errors across the population could move reported recoverables by roughly $1,380,000 in either direction. That exceeded the threshold the firm used for statement-level findings, so the layer error went to the top of the list. Missing reinsurer confirmations ranked second because the balances they supported were large enough to matter even though no error had yet been proven. Cash application without a documented rule ranked third because it was pervasive, touching every receipt against an open item.
Scores combined balance at risk, likelihood and reporting exposure. A finding scored 8 or higher needed an accountable owner, a dated remediation plan and a closure test approved by internal audit before the item could be marked closed. For the layer error, the test was a re-performance of the full population of cessions for the affected treaties, with each correction traced to a signed confirmation. For the missing confirmations, the test was a request to each reinsurer and a count of responses received. For cash application, the test required a written rule approved by the controller and a sample of receipts showing the rule was applied.
Actions taken
- Assign the layer correction to the treaty accountant, Marisol, with a deadline for re-performing the cession population on the three affected treaties.
- Obtain signed confirmations for the six open cessions from each reinsurer within fifteen working days, and hold any recoverable on those items from the carrier statement until received.
- Write a cash application rule that matches each receipt to the named cession first and to the oldest item only when no reference exists, approved by the controller.
- Give the operations director, Javier, a quarterly report on the three closure tests, and do not mark any finding closed until internal audit has re-performed the test.
Results
The re-performance found three more layer errors, bringing the total to 14, and the net recoverable misstatement came to $1,020,000, which was corrected before the period closed. Reinsurer confirmations returned for five of the six open cessions within the window. The sixth was reserved in full after the reinsurer disputed its terms. The disputed client balance of $740,000 was settled once the layer correction was posted. Javier now receives the quarterly report, and the closure tests for all four findings have been approved.
Questions for discussion
- How would you weigh a small balance that recurs across many treaties against one large but isolated error?
- What should count as acceptable closure evidence when a reinsurer never answers a confirmation request?
- Who should own a cash application rule when operations and finance disagree about which receipt belongs to which cession?
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.