AC599 Unit 4 Capstone Methodology Memo Example

What Counts as Proof a Reserve Is Right? A Unit 4 Methodology Memo for a Health Claims Estimate

[Student Name]

Purdue University Global

AC599: Graduate Capstone in Accounting

Unit 4 Methodology Memo

[Instructor Name]

[Date]

Hollis Creek Supply and every figure are composite; the actuarial guidance, standards and texts cited are real.

Memorandum

To: Chief Financial Officer From: Capstone Analyst Date: September 28, 2026 Re: Method for estimating the December 31 health claims liability

Population

The estimate uses the third-party administrator's paid claims for medical services incurred from January 2024 through December 2026, arranged by month incurred and month paid, thirty-six incurred months in all, and the pharmacy manager's equivalent report. Enrollment counts by month come from payroll. Claims for the one member whose costs passed the stop-loss deductible are tracked separately so they do not distort the pattern for everyone else.

Step 1: Build the Lag Triangle

For each month of service, the triangle shows how much had been paid by the end of the month of service, one month later, two months later, and so on. Older months are fully paid; recent months are only partly paid. The triangle is reconciled to the administrator's total paid claims for each calendar month before it is used.

Step 2: Completion Factors for Older Months

From the fully developed months, the method computes what share of a month's eventual cost is typically paid by each lag: for example, 52 percent in the month of service, 86 percent by one month later, 95 percent by two. Dividing what has been paid so far for each recent month by its completion factor gives the expected total, and the difference is the unpaid amount (Friedland, 2010). This works well for months that are at least two lags old, where enough has been paid for the factor to be reliable.

Step 3: Per-Member Projection for the Newest Months

For November and December 2026, little has been paid, and a completion factor applied to a small number magnifies any noise. For those months, the method instead projects cost per member per month from the prior twelve months, adjusted for medical inflation, times enrollment, and subtracts what has been paid (Bluhm, 2012).

Step 4: Pharmacy and Stop-Loss

Pharmacy claims are paid within weeks, so a one-lag completion factor is enough. The large claim above the stop-loss deductible is estimated separately from the administrator's case report, and the expected reimbursement is computed under the contract.

Step 5: The Hindsight Test

The method is applied as if it were December 31, 2025, using only data available then, and the result is compared with what the 2025 claims actually cost once they were paid through September 2026. If the method's 2025 estimate falls within 5 percent of the actual run-out, it is considered reliable for 2026. The same test is applied to the one-month rule of thumb.

The hindsight test is the proof the memo promises: a method that would have estimated last year's unpaid claims closely, using only what was known then, has earned trust for this year, while a rule that would have missed has not.

Why Two Methods Rather Than One

Using completion factors for every month would be simpler, but it fails where it matters most. For December, only about 1 percent of the month's cost has been paid by December 31, and dividing a small paid amount by a small completion factor turns any noise in that amount into a large error in the estimate. A per-member projection does not depend on what has been paid, so it is steadier for the newest months. For older months, the reverse is true: most of the cost has been paid, and the completion factor method uses that real information rather than an average. Combining the two methods at the point where each is more reliable is standard practice for health claims (Actuarial Standards Board, 2011).

Range and Best Estimate

The method produces a best estimate and a range, using the variation in completion factors across the thirty-six months. The recorded liability will be the best estimate, consistent with the accounting standard, with the range disclosed in the report (Financial Accounting Standards Board, 1975).

What Would Undermine the Method

A change in the administrator's processing speed, a backlog, or a benefit change would make past patterns a poor guide. The January deductible change affects only 2027 claims, but any processing change in late 2026 will be checked with the administrator before the estimate is final (Actuarial Standards Board, 2011).

References

Actuarial Standards Board. (2011). Incurred health and disability claims (Actuarial Standard of Practice No. 5). Actuarial Standards Board.

Bluhm, W. F. (Ed.). (2012). Group insurance (6th ed.). ACTEX Publications.

Financial Accounting Standards Board. (1975). Accounting for contingencies (Statement of Financial Accounting Standards No. 5). FASB.

Friedland, J. (2010). Estimating unpaid claims using basic techniques. Casualty Actuarial Society.