Advertise - Write For Us

Actuaries: this one is written for your CFO. Forward it this week, while there is still time to send the October email it describes.

Every February, somewhere, a CFO discovers a problem. The actuarial opinion is due March 1. That opinion is the actuary's signed conclusion that the company holds enough reserves, meaning the money set aside today to pay tomorrow's claims, to meet its obligations. Filing it needs data nobody extracted from a system nobody owns, reconciled to a ledger that closed late.

The reaction is predictable. Everyone works longer hours. Data requests become urgent. Finance calls actuarial. Actuarial calls finance. Somebody discovers that the policy administration system, the software of record for every policy's detail, changed 6 months ago. Another person is trying to explain why the reinsurance balance does not match. The auditors are waiting.

By then, most of the damage has already been done. A smooth March filing is usually determined months earlier. For many companies, the most important month in the actuarial year-end process is November.

For U.S. statutory reporting, March 1 is the filing date for the annual statement and the actuarial opinion for life and fraternal insurers. Exact requirements can vary by state, and related actuarial submissions, such as memoranda or summaries filed later in the spring, can carry different deadlines; check with your domiciliary regulator for specifics.

The useful way to manage the process is backwards. Ask a simple question: what must already be finished for March 1 to become an ordinary working day? Then keep moving backwards until you reach November.

By March, the analytical work should largely be finished: the actuarial opinion is finalized and filed with the annual statement, required approvals and final checks are complete, and any separate supporting submissions follow the applicable regulatory timetable.

March should be the end of the process, not the month someone first discovers a material data problem.

Newsletter continues after job posts…

By February, management should be reviewing results. Fundamental inputs should already be in hand. Reserve results are discussed with management. Material movements from the prior year are explained. Unusual experience is investigated. Opinion language is settled. Auditor questions are addressed. Supporting actuarial documentation continues toward completion.

There will always be some judgment here; a reserve analysis can raise questions nobody anticipated in November, and that is normal actuarial work. Discovering in February that a major block of policies was omitted from the year-end extract is a different matter: a process failure that could have been caught months earlier.

January is when the final December 31 information becomes available. The actuary receives the final year-end inforce extract, the file listing every active policy and the details that drive its liability, along with the claims information. Asset information arrives from finance or the investment manager. Reinsurance balances and settlements are reconciled. The actuarial team validates the files, compares them with the prior period, and begins the final reserve runs and adequacy analysis.

For the CFO, the distinction that matters is between finalizing data in January and designing the data request that late. Finalizing is expected. Redesigning the request in January is a warning sign that November was skipped.

December is the month that prevents unpleasant surprises. By now, extract specifications should already be frozen. A dry run using November data can test whether the company can actually produce what the actuary expects. This step matters most after a policy administration conversion.

Suppose the old system recorded a benefit under 1 field, while the new system splits it across 3. The total policy count may still look reasonable, and the extract may load without error. Yet the information feeding the reserve calculation may no longer mean what it meant last year. That problem is cheap to investigate in December. It is far more expensive once discovered during the final valuation.

The dry run should answer a short set of practical questions: can the file be produced, do totals reconcile, have field definitions changed, and are the asset and reinsurance files arriving in the expected format. Keep an open-items list, and give every item an owner and a date.

November is where an organized year-end begins. The valuation calendar is circulated. Data owners are named. Extract specifications are confirmed. Material methodology and assumption discussions start early enough to avoid surprises. The scenario set and testing requirements are agreed where applicable. Known changes to systems, products, investments and reinsurance arrangements are communicated to the actuary.

None of this requires weeks of meetings. In many cases, one disciplined 30-minute call between the CFO, the controller, the data owner and the actuary prevents days of February troubleshooting. The statutory deadline sits in March. The operational deadline begins in November.

Ask your actuary one question on that November call: if we ran this exact data file today, what would break? The answer says more about year-end readiness than any calendar review.

Actuarial data requests can look unnecessarily technical from the finance side. Translated into operational language, most become easier to manage.

The actuary needs an inforce file: the list of every policy actually on the books, carrying the characteristics that drive its liability, such as issue dates, ages, benefits, premiums, product codes, policy status and guarantees. The actuary also needs credible claims, lapse and other experience information, which shows how the portfolio has behaved and whether current assumptions still hold.

Asset information matters because a liability's adequacy, meaning whether the assets backing it are enough to pay the future claims under a range of scenarios, cannot always be judged in isolation from the assets supporting it. Depending on the framework, the actuary may need duration, yields, cash flows or credit characteristics.

Reinsurance requires its own discipline: treaty terms, ceded balances and settlements need to agree closely enough across actuarial, accounting and reinsurance records. And finally, the actuary needs something no database provides: a person who knows what the data means. When a field suddenly shows zeros, the gap between an answer tomorrow and an answer 3 weeks later can matter more than the sophistication of the valuation software.

A handful of causes explain most year-end fire drills, and they repeat from company to company.

Ownership is the most common. Everyone assumes somebody else produces the actuarial extract: finance assumes IT owns it, IT assumes the administrator owns it, and the administrator remembers producing something similar last year but cannot find the specification.

System change is close behind. A policy administration conversion can alter codes, definitions, dates and record structures without anyone intending to change the actuarial data, and the system can work exactly as designed while producing an extract no longer comparable with the prior year.

Investment data causes its own problems, usually because nobody agreed in advance exactly what the actuary needed and when, so the asset file arrives late or in a different structure than expected.

Reinsurance is a recurring source of friction too. Open settlements, disputed balances and timing differences can sit quietly all year and become material the moment year-end numbers have to reconcile.

The last cause is organizational rather than technical: the actuary is treated as a February vendor rather than part of the November close process. That one is the easiest to fix. If your actuary only appears in February, that is usually a process choice, not a necessity. Invite the actuary earlier.

A late actuarial input rarely stays an actuarial problem. Take a common scenario: the final usable data arrives 2 weeks late. The reserve analysis moves 2 weeks. Management review moves with it. Questions that should have been answered calmly now compete with the financial close. Audit support moves too, and if a filing deadline comes under threat, regulatory communication may become necessary as well.

As an illustrative order of magnitude, not a figure drawn from any single engagement, a 2-week slip at a small carrier commonly adds $15,000 to $40,000 in incremental actuarial overtime, audit extension fees and diverted staff time; on a $150,000 combined actuarial and audit budget, that is a 10% to 25% cost increase for no additional insight. The number will vary by carrier; the point is that the cost is real and avoidable.

The larger cost is not the extra hours themselves. Senior finance staff are pulled off other close activities, auditors spend time chasing open items, and executives review numbers on a tighter deadline. Errors become more likely because every reviewer has less time.

That is why the cheapest year-end improvement is rarely another model. It is a better calendar.

There is a common perception that a spreadsheet goes to the actuary in January and a reserve number returns in February. The work between those 2 points is substantial: the team checks the data's completeness and internal consistency, compares it with prior periods, and investigates unexpected movements in exposures, claims, premiums, policy counts or reserves before relying on it.

Only then does the actuary run the reserve calculations, evaluate material assumptions, and carry out the applicable adequacy or scenario testing. Every liability movement has to be explained: new business, claims experience, lapses, interest rates, an assumption change, a data correction, or something else. An apparently favorable movement can turn out to be a data issue. The February workload depends heavily on the quality of what happened before January.

A system conversion deserves special treatment. A successful migration does not by itself establish actuarial continuity.

Run the old-system and new-system inventories against each other: compare policy counts, face amounts, premiums, reserves where available, status codes and other important valuation fields, and map every changed definition explicitly. Do this before year-end, not after. December is the last comfortable month to discover that Issue Date in one system does not represent precisely the same event as Issue Date in the other.

None of the most effective controls require a new system or a consulting project.

Naming one person who owns delivery of the actuarial data matters most. That person does not have to create every file, only have the authority to know where each one comes from and chase it when it does not arrive. Scheduling a 30-minute year-end readiness call with the actuary in the first week of November is nearly as valuable, and costs nothing beyond a calendar invite sent now, before the year gets busy. The cheapest move of all is the third: send one email in October, and keep sending it every year.

That email takes 5 minutes to send. It can save weeks.

The objective is simple: March should be boring. A well-run year-end actuarial process does not eliminate difficult questions. It creates enough time to answer them properly.

If your last year-end felt like a fire drill, I will walk your team through this reverse timeline against your actual close calendar. 30 minutes, no charge.

Syed Raza, FSA, is an appointed actuary working with life insurers, fraternal benefit societies, and preneed and final expense writers. Requirements vary by state; check with your domiciliary regulator on specific filing dates and submissions.