Ask what your program owes for claims and you will get a number that is really two numbers added together. The first is the case reserves: the amount your adjusters have estimated, claim by claim, for the claims they know about. The second is IBNR: the actuary’s provision for everything that estimate cannot see. The two are produced by different people, using different methods, at different levels of granularity, and they behave differently as your program’s year unfolds.
Most reserve misunderstandings between a CFO, a TPA, and an actuary trace back to blurring this split. This article defines each piece, shows how they interact, and walks through the specific mistakes that follow from treating one as a substitute for the other.
What case reserves are
A case reserve is the adjuster’s estimate of the remaining cost of a specific, reported claim. When a claim is opened, the adjuster (usually at your TPA) sets an initial reserve, then revises it as facts develop: medical records arrive, an attorney appears, surgery is scheduled, a demand letter lands. Add the case reserve to what has already been paid and you have the claim’s incurred value, the number that shows up on your loss run.
Three properties matter for everything that follows:
- Case reserves are claim-level. Each one is an estimate about one claim’s future, informed by that claim’s facts.
- Case reserves only cover known claims. A claim that has not been reported has no file, no adjuster, and no case reserve.
- Case reserves reflect current knowledge, not ultimate outcomes. Adjusters reserve to what is known and probable today. On long-tailed lines, what is known today systematically understates where a claim cohort ends up.
Case reserving philosophy also varies by TPA and over time: some shops reserve to the probable outcome early, others stair-step upward as facts force the issue. That variation is invisible in any single claim file and very visible in aggregate, which becomes important below.
What IBNR is
IBNR is an aggregate provision, estimated by an actuary, for the claim costs the case reserves cannot capture. Despite the name (incurred but not reported), the provision in most reserve studies is “broad” IBNR, which covers two distinct things:
- Unreported claims: injuries and incidents that have happened but have no file yet. This is pure IBNR.
- Development on known claims: the tendency of the existing claims, in aggregate, to close at values above their current case reserves.
The split between the two, and why it matters for self-insureds, is covered in Pure IBNR vs. Broad IBNR. The key structural point here is that IBNR is not claim-level. The actuary is not second-guessing any adjuster’s estimate on any file. The estimate comes from patterns in the aggregate data, typically through development factors applied to the program’s history.
Under the actuarial standard that governs this work (ASOP No. 43), the actuary’s real assignment is the total unpaid claim estimate. IBNR is what remains of that total after subtracting the case reserves:
IBNR = estimated ultimate losses - paid to date - case reserves
That subtraction is the hinge of the whole relationship.
The seesaw: why the pieces move opposite each other
Because IBNR is a remainder, the two components sit on a seesaw. Suppose the actuary estimates an accident year’s ultimate at $2.0 million, with $600,000 paid and $650,000 in case reserves; IBNR is $750,000. Total unpaid: $1.4 million.
Now your TPA runs a reserve adequacy review and raises case reserves by $200,000, touching nothing else. Incurred losses jump on the loss run. If the ultimate estimate is right, the liability did not change: IBNR should drop to $550,000, the seesaw tips, and the total holds at $1.4 million.
In practice it is rarely that clean, because a case reserving change is also information. Higher case reserves might mean the adjusters got more conservative (the ultimate is unchanged and IBNR should absorb the shift), or that claims are genuinely deteriorating (the ultimate itself is too low). Worse, the change distorts the reported development pattern the actuary projects from, in ways that can double-count or hide the shift. Untangling which one happened is a diagnostic exercise of its own, covered in Case Reserve Strengthening.
What a buyer needs to retain is simpler: case reserves and IBNR are not independent numbers. Any conversation about one that ignores the other is incomplete, and any change in TPA reserving practice must reach your actuary before the next study.
The mistakes this split causes
“Our TPA reserves conservatively, so we do not need much IBNR.” Conservative case reserves reduce the development component of IBNR; they do nothing for unreported claims. On a workers compensation program, a claim from late in the year may not be reported until after year end, and on public entity or abuse-exposed liability lines, claims can surface years later. No case reserving philosophy covers a claim with no file.
Booking the loss run as the liability. Paid plus case is what is known, not what is owed. For an immature accident year of a long-tailed line, the majority of the eventual cost is often in neither number yet. An organization that accrues only case incurred is recognizing losses on a lag, and the catch-up arrives in the worst possible form: several accident years correcting at once. How the full liability should land on the financial statements is covered in When Self-Insured Reserves Hit the Financial Statements.
Netting the seesaw manually. A finance team that trims booked IBNR because “case reserves went up” (without an actuarial update) is doing the actuary’s subtraction with none of the actuary’s analysis. The offset is only valid if the ultimate is unchanged, which is exactly the question that needs the analysis.
Assuming IBNR is always positive. On short-tailed lines, or mature accident years where case reserves have proven redundant, the indicated IBNR can be negative: the actuary expects known claims to close below their case estimates by more than any late-reported claims will add. Negative IBNR on an old accident year is unremarkable. Negative IBNR on a young one deserves a hard question.
Who owns what
The clean division of labor:
- Adjusters own claim outcomes. Their case reserves should reflect each claim’s probable cost, consistently over time. Their job is not to provision for the aggregate.
- The actuary owns the total. The unpaid claim estimate, and therefore IBNR, is an aggregate statistical judgment. It should never be produced by marking up individual files.
- Finance owns the booked number. What lands on the balance sheet is a management decision informed by the actuary’s estimate, including where in a range to book.
The system fails when a role is skipped: adjusters pressured to hold case reserves down to manage the incurred number, IBNR booked without a study, or an actuary’s total quietly overridden. Each looks like a small internal shortcut, and each is how a program ends up explaining a reserve charge to its board. The oversight questions that keep the roles honest are in Audit Committee Reserve Governance.
What to ask when you review the split
1. What share of the total unpaid estimate is IBNR, by accident year? Young years should be IBNR-heavy, old years case-heavy. A young year that is mostly case reserves suggests the estimate is leaning on adjuster judgment exactly where adjuster judgment sees the least.
2. Has case reserving practice changed during the experience period? New TPA, new claims leadership, adequacy reviews, system conversions. If yes, ask the actuary specifically how the analysis adjusted for it.
3. Where is IBNR negative, and why? Legitimate on mature or short-tailed segments; worth a documented explanation everywhere else.
4. Do the adjusters know the actuary’s view of their case adequacy? The development analysis quantifies whether case reserves tend to run high or low. Feeding that back to the claims operation improves both sides of the split.
5. When case reserves moved sharply, did the booked total move with them, against them, or not at all? The answer reveals whether the seesaw is being managed with analysis or by reflex.
Further reading
For the foundation, see Loss Reserves Explained and IBNR, Explained Without the Jargon. For the finer split inside IBNR, see Pure IBNR vs. Broad IBNR. For how case reserve changes distort the actuarial data that estimates IBNR, see Case Reserve Strengthening. For what happens when the diagnosis is missed, see What’s Actually Driving Your IBNR Higher?.