Fundamentals

Pure IBNR vs. Broad IBNR: Why the Distinction Matters More for Self-Insureds

The IBNR line on your actuarial report bundles two very different uncertainties into one number; understanding which piece is driving the estimate changes what you should ask and how you should fund.

By Sam · April 2026 · 11 min read

Your actuarial report shows a single line labeled IBNR. The number might be $4.2 million. What that label does not tell you is that the number bundles together two fundamentally different types of uncertainty, and the split between them determines whether your funding strategy and your claims-management response are pointed at the right target.

Two kinds of uncertainty in one line item

When actuaries talk about IBNR, they almost always mean the broad definition. Friedland states it directly: “Actuaries typically use the broad definition of IBNR” (Friedland, p. 388). That broad definition includes two components that behave differently, respond to different interventions, and carry different implications for your balance sheet.

Pure IBNR (sometimes called IBNYR, for incurred but not yet reported) is the estimated cost of accidents that have already occurred but that no one has reported to the claims administrator. There is no claim file. There is no case reserve. The only evidence these claims exist is the statistical pattern showing that, historically, claims from a given accident period continue to arrive for months or years after that period closes.

IBNER (incurred but not enough reported) is the expected future movement on claims that are already known and already carry a case reserve. The adjuster has set a number. IBNER is the actuarial expectation that the number will change, either upward or downward, as the claim matures, is investigated, litigated, and ultimately settled.

Broad IBNR equals pure IBNR plus IBNER. When the IBNR line on your financial schedule says $4.2 million, some portion of that is for claims no one has seen yet, and some portion is for claims the adjuster already has on file but has not fully reserved. Those are very different problems.

Why the distinction is not just academic

Consider what happens when you try to act on the number.

If IBNR is mostly pure IBNR (claims not yet reported), your primary risk-management response is to shorten reporting lags: tighten incident-reporting protocols, audit whether supervisors are delaying claim submissions, and check whether your TPA is batching reports. Funding the liability requires patience; the claims will arrive, and you need cash available when they do, but there is little you can do to influence their ultimate cost until they surface.

If IBNR is mostly IBNER (development on known claims), the response is different. The claims are already on file. The question is whether the case reserves are adequate. Your response is to review case reserve adequacy with your adjuster, examine whether large claims are reserved at realistic levels, and check whether settlement patterns have changed. Funding requires a different posture because the claims are already in the system and their trajectory is partially visible.

When the two components are mixed into a single number with no decomposition, you cannot tell which lever to pull.

A worked example

Suppose you run a workers compensation self-insured program. Your actuary estimates ultimate losses for the most recent accident year at $10 million. As of the evaluation date, paid losses are $2 million and case reserves (also called case outstanding) on known open claims total $3 million.

The total indicated IBNR is:

Ultimate ($10M) minus paid ($2M) minus case outstanding ($3M) = $5M in broad IBNR.

Now decompose that $5 million.

Your actuary estimates that claims from the most recent accident year are only 70% reported by count. Based on average severity for late-reported claims in your program, the pure IBNR component is approximately $1.5 million. These are claims from workplace injuries that occurred during the accident year but have not yet been filed, perhaps because the employee initially treated the injury as minor or because a supervisor delayed the report.

The remaining $3.5 million is IBNER: expected future development on the claims already on file. That number reflects the actuarial expectation that the $3 million in current case reserves will grow as claims mature, litigation develops, and medical treatment extends. Some of that development is normal claim maturation. Some may reflect systematic case reserve inadequacy.

The split ($1.5 million in pure IBNR versus $3.5 million in IBNER) tells you that most of the uncertainty in your reserve is about whether known claims are adequately reserved, not about whether unknown claims are lurking. That is an actionable insight. Without the split, all you have is a $5 million number and no direction.

The five-percent rule of thumb and its limits

Some actuaries estimate pure IBNR as approximately 5% of the most recent accident year’s ultimate losses. This shortcut has wide informal use, and Friedland references it (Friedland, p. 389). For a program with $10 million in estimated ultimate losses, the rule of thumb would put pure IBNR at $500,000 for that year.

The rule of thumb has two problems.

First, 5% is an average across many different lines and many different reporting patterns. For a short-tailed line like auto physical damage, where nearly all claims are reported within weeks of the accident, pure IBNR as a percentage of ultimate is much less than 5%. For a long-tailed line like general liability with latent-injury exposure, pure IBNR can far exceed 5% of ultimate because claims may not surface for years. Using a single percentage across all lines in a program flattens differences that matter.

Second, the rule of thumb is often applied without support from the program’s actual experience. The right way to estimate pure IBNR is to analyze the program’s own reporting patterns: how many claims from a given accident year are reported in the first quarter, the second quarter, the sixth quarter, and so on. That analysis produces a reporting pattern specific to your program, your line of business, and your claims-handling environment. When an actuary uses 5% without documenting why 5% is appropriate for your specific program, the number is a placeholder, not an estimate.

Buyers should require documentation showing how the pure IBNR percentage was derived. If the actuary used a percentage, was it based on the program’s own reporting lag data? If reporting patterns have changed (because of a TPA transition, a new incident-reporting system, or a change in employee demographics), has the percentage been updated?

Why the split matters more for self-insureds

For a large commercial insurer writing thousands of policies across many states, the law of large numbers smooths the distinction between pure IBNR and IBNER. Individual claim severity variance is small relative to total reserve volume. The insurer’s aggregate IBNR estimate can be reliable even without a precise decomposition.

For a self-insured program, the math is less forgiving. A self-insured workers compensation program might have 200 open claims. A single large claim (a severe traumatic injury, an extended-leave case that converts to permanent disability) can represent 10% or more of total case reserves. When per-claim severity variance is high relative to total reserve volume, the distinction between pure IBNR and IBNER becomes material.

If most of your IBNR is IBNER and a single large claim is driving it, that is a concentrated risk that can be investigated and managed. If most of your IBNR is pure IBNR, you have a diffuse risk that depends on how many claims are still unreported, and a single late-reported large claim can move the entire estimate.

Self-insured programs also face a practical problem that carriers do not: the person reading the actuarial report is often the same person responsible for the claims program. A CFO or risk manager who knows that $3.5 million of a $5 million IBNR is IBNER can walk into a claims review with a specific question: which open claims are most likely to develop adversely, and are the case reserves on those claims realistic? A CFO who knows only that IBNR is $5 million has no starting point.

When the decomposition breaks down

The pure IBNR versus IBNER split is conceptually clean, but estimating each component separately introduces its own uncertainties.

Pure IBNR depends on the reporting pattern, which is itself an estimate. If reporting patterns have shifted (because of a pandemic that delayed medical treatment and therefore delayed claim filings, or because a new incident-reporting system shortened the lag), the historical pattern may not reflect the current environment.

IBNER depends on case reserve adequacy assumptions. If case reserves are systematically inadequate, IBNER will be large. If the actuary does not detect the inadequacy (or detects it but underestimates its magnitude), the IBNER component will be understated. The diagnostic signals described in What’s Actually Driving Your IBNR Higher? are directly relevant here: shifts in case reserve adequacy show up as changes in reported development factors, and the actuary should be looking for them.

For very small programs with thin data, the decomposition may not be statistically credible. If your program has 50 open claims and three years of history, the reporting pattern is too volatile to support a reliable pure IBNR estimate, and the actuary may reasonably present only a broad IBNR number with a note explaining why the decomposition was not provided.

What a buyer should ask their actuary

Does your report decompose IBNR into pure IBNR and IBNER? If not, ask why. For programs with enough data to support the split, the decomposition is valuable. If the actuary does not provide it, you are missing information that would help you manage the reserve.
How was the pure IBNR percentage derived? If the actuary used a rule of thumb (such as 5% of ultimate), ask whether the percentage is supported by the program’s actual reporting lag data. If it is not, ask for an analysis based on your data.
Has the reporting pattern changed? TPA transitions, new incident-reporting systems, and changes in employee demographics can all shift reporting lags. If the pattern has changed, the pure IBNR estimate should reflect the new pattern, not the historical average.
What is driving the IBNER component? Is it broad-based development across many claims, or is it concentrated in a small number of large claims? If concentrated, which claims, and have you reviewed the case reserves on those claims?
How sensitive is the split to the case reserve adequacy assumption? If the actuary assumes current case reserves are adequate, IBNER will be smaller. If the actuary assumes case reserves are 15% inadequate, IBNER will be larger. Understanding that sensitivity tells you how much of your IBNR estimate depends on the adjuster getting it right.

What to require in documentation

Whether or not your actuary provides a formal decomposition, the reserve report should include the following.

Definition of IBNR as used in the report. The report should state explicitly whether the IBNR estimate reflects the broad definition (pure IBNR plus IBNER) or only one component. If the report uses the term IBNR without defining it, ask for clarification.
Reporting pattern analysis (if the split is provided). The report should show the reporting pattern (percentage of claims reported by development period) and identify any changes in the pattern over the experience period.
Case reserve adequacy assessment. Whether provided as part of an IBNER estimate or as a standalone diagnostic, the report should address whether case reserve adequacy has been stable across the experience period. If it has not, the report should describe the adjustment applied to the development factors.
Support for any rule-of-thumb percentages. If the actuary uses a percentage-based estimate for pure IBNR (such as the 5% rule), the report should document why that percentage is appropriate for this specific program, line of business, and reporting environment.
Sensitivity analysis on the decomposition. At minimum, the report should show how the total IBNR estimate changes under alternative assumptions about reporting patterns and case reserve adequacy. A range estimate is even better, but sensitivity analysis is the floor.