An illustrative diagnostic for CFOs and risk management actuaries when a South African insurer’s SAM quarterly return cannot be reconciled to the management pack.
A South African insurer can submit its SAM quarterly return on time and still have a serious control problem: the numbers do not agree to the management pack used by the executive committee.
This is not only an actuarial issue. It is not only a finance issue. It is a governance issue around how policy, claims, actuarial, reinsurance, investment and ledger data are defined, moved, adjusted and approved before they appear in a QRT or board pack.
This article sets out an illustrative diagnostic for a CFO and risk management actuary when a SAM reporting reconciliation insurance South Africa problem emerges. It is not based on a client case. It reflects a common pattern in insurers where regulatory reporting, finance reporting and management information have developed at different speeds, often across legacy systems, spreadsheets and manual controls.
For wider context on insurance data foundations, see Insurance Data Strategy. For related illustrative scenarios, see Insurance Examples. For a deeper discussion of regulatory and actuarial traceability, see Regulatory Reporting and Actuarial Lineage.
Assume a short-term insurer has completed its quarterly SAM return. The regulatory reporting team has prepared the QRT pack. The risk management actuary has reviewed the solvency calculations. Finance has produced the monthly management pack for Exco and the board finance committee.
Then the CFO notices three uncomfortable differences.
First, gross written premium in the management pack does not align with the premium base used in the SAM reporting process. Second, outstanding claims used in the actuarial reserving file differ from the claims liability reported in finance. Third, reinsurance recoveries in the prudential return are materially different from the reinsurance view presented to Exco.
None of the numbers is obviously wrong. Each team can explain its own position. Finance refers to the general ledger close. Actuarial refers to a valuation extract. Underwriting says policy corrections were processed after the finance cut-off. Claims operations says several large claims were reopened late in the quarter. Reinsurance points to treaty adjustments not yet reflected in the ledger.
The problem is not the presence of differences. In insurance, timing differences and valuation adjustments are normal. The problem is that the insurer cannot produce a clear, agreed reconciliation showing which differences are expected, which are judgemental, which are errors, and which require correction before submission or board sign-off.
Executives often expect one official number. In practice, different reports serve different purposes.
A management pack may focus on commercial performance: premium growth by channel, loss ratios by product, expense trends, cash collection, broker performance and operational risks. A SAM return is designed for prudential supervision. It needs regulatory classifications, solvency measures, technical provisions, capital requirements and prescribed templates.
The two reporting views may legitimately use different bases. For example, the management pack may show gross written premium by broker segment, while the QRT requires a different allocation across lines of business. The finance pack may reflect ledger postings at month-end, while actuarial reserving may rely on claims development data extracted before final finance journals. An internal performance view may exclude once-off operational adjustments that still need to be treated correctly in prudential returns.
The CFO’s concern should therefore not be: “Why are the numbers different?” The better question is: “Can we explain the differences in a way that is complete, evidenced and repeatable?”
If the answer depends on one senior analyst’s spreadsheet or the memory of an actuarial manager, the insurer is relying on personal knowledge rather than an institutional control.
The first diagnostic step is to establish the data cut-off for each report.
In many insurers, reconciliation problems begin before any calculation is performed. Finance closes the ledger on one date. Claims data is extracted on another. Policy administration corrections continue after the actuarial extract has been taken. Reinsurance bordereaux may arrive late from brokers, coverholders or external administrators. Investment balances may be sourced from custodian reports that follow a different timetable.
A practical diagnostic should map the exact dates and times of the key extracts used for:
This is especially important in South Africa, where operating conditions can interfere with clean reporting cycles. Load-shedding, connectivity failures, delayed batch processing and system downtime can all affect whether a data file is complete when it is extracted. A late-running claims batch can create a real difference between two reports even when both teams followed their usual process.
The control question is whether the insurer knows this happened and can evidence it.
Once cut-offs are understood, the next issue is definition.
Terms such as “active policy”, “written premium”, “earned premium”, “large claim”, “recoverable”, “lapse”, “expense allocation” and “line of business” may appear simple. In practice, different teams may interpret them differently.
A retail insurer might classify add-on products differently in underwriting reports and SAM templates. A life insurer might treat policy status changes differently across administration, actuarial and finance systems. A commercial insurer might allocate multi-risk policies to management segments that do not align neatly with prudential reporting categories.
These are not IT details. They affect solvency calculations, performance reporting and executive decisions.
For the CFO and risk management actuary, the diagnostic should highlight the definitions that materially affect the reconciliation. It is not necessary to document every field in the enterprise at once. Start with the line items that are causing the difference. For each one, confirm:
Where definitions differ by design, that should be explicit. Where they differ by accident, the insurer has found a control weakness.
A common mistake is to treat every reconciliation item as a data problem.
Some differences arise because actuarial and finance views answer different questions. For example, a claims reserve may include actuarial estimates for incurred but not reported claims, while the ledger reflects case estimates and journals. A risk margin calculation may introduce assumptions that are not present in the management pack. SAM may require look-through or classification treatments that do not appear in normal executive reporting.
These are valuation differences. They need explanation, approval and evidence, but they are not necessarily data defects.
Other differences are more concerning. Duplicate policies, missing claims statuses, inconsistent broker codes, outdated reinsurance treaty mappings, unexplained manual overrides and incomplete system interfaces point to data quality and lineage problems. These can weaken both prudential returns and management decisions.
The diagnostic should therefore classify reconciliation items into clear categories:
This prevents the reconciliation meeting from becoming a general debate. It also helps the CFO decide which items affect submission confidence, which affect board reporting, and which should become remediation actions after the quarter-end cycle.
Manual adjustments are often where the real risk sits.
In an insurer with legacy platforms, outsourced administration or multiple product systems, manual adjustments are sometimes unavoidable. A finance team may post a top-side journal to correct commission. Actuarial may adjust a claims triangle for a known system issue. Reinsurance may maintain an offline calculation for a treaty condition that the core system cannot handle.
The issue is not whether manual adjustments exist. The issue is whether they are controlled.
For each material adjustment affecting the SAM return or management pack, the insurer should be able to show who prepared it, what source evidence was used, who reviewed it, why it was necessary, and whether it is recurring. If the same adjustment appears every quarter, it should not remain an undocumented workaround. It should be built into a controlled reconciliation process, with ownership and a plan to resolve the root cause where practical.
This matters for POPIA as well. Reconciliation exercises often lead teams to share detailed policyholder or claims files by email or spreadsheet. If sensitive personal information is copied widely to solve a reporting problem, the insurer may be creating privacy risk while trying to fix a prudential reporting issue. Access should be limited to what is necessary, and evidence should be retained in a controlled environment.
The CFO and risk management actuary do not need to inspect every cell in every workbook. They do need a diagnostic pack that makes the control position visible.
A useful pack should include a one-page movement summary from management pack figures to SAM return figures for the material line items in dispute. It should show the major reconciling items, their value, their category, their owner and their status. It should also identify unresolved items separately from approved differences.
The pack should include lineage for the key numbers: where the data originated, how it was transformed, which models or calculations were applied, and where approvals were recorded. This does not need to be a complex architecture diagram. A clear flow from source system to reporting output is often enough to expose where the breaks occur.
Finally, the pack should distinguish between quarter-end actions and structural fixes. The quarter-end action may be to approve a specific reconciling item with evidence. The structural fix may be to align extract dates, standardise a product mapping, reduce manual handling, or formalise ownership of a key definition.
This is where executive discipline matters. If every reconciliation issue is treated as urgent only during submission week, the insurer will repeat the same failure next quarter.
A SAM return that cannot be reconciled to the management pack is not automatically a sign of incorrect reporting. It is a sign that the insurer must prove the difference rather than explain it informally.
The next step for the CFO and risk management actuary is to select the most material unreconciled QRT line item and trace it back to source, through definitions, extracts, adjustments, model outputs and approvals. If that path is clear, the issue may be manageable. If it is not, the insurer has found a data lineage control gap that should be addressed before it becomes a regulatory, audit or board confidence problem.