If you run risk or finance at a self-insured company or a captive, you have seen the letters IBNR on a schedule and you have probably been asked to sign off on the number next to them. This article explains what that number represents, where it comes from, and what questions to ask before you accept it.
The claim lifecycle in one paragraph
An injury or loss occurs on a specific date. That is the accident date. At some point after that, someone tells the insurer or the claims administrator the claim exists. That is the report date. Case reserves are then set and revised as the claim is investigated, negotiated, and eventually paid or closed. The gap between the accident date and the final payment can be days for a small medical bill, or years for a disputed liability claim.
Reserving is the process of estimating, at a specific evaluation date, how much money is still owed on claims that have already occurred, whether or not anyone has told you about them yet.
The three pieces of unpaid losses
At any evaluation date, unpaid losses break into three conceptual buckets.
- Case reserves. The adjuster’s estimate of remaining payments on each known open claim. These sit on the claim file.
- Case development. An expected movement, up or down, on those known open claims as investigation continues. Sometimes called IBNER, for incurred but not enough reported.
- Pure IBNR. Losses from accidents that have already happened but have not yet been reported to anyone. There is no claim file. There is no adjuster. There is only the statistical expectation that they exist.
In common usage, the label IBNR on a financial schedule almost always means the sum of case development plus pure IBNR, not just the pure piece. When someone asks whether your IBNR is adequate, they are asking whether the number on the schedule is big enough to cover both the adjuster optimism on open claims and the claims no one has yet seen.
Why IBNR is the biggest unknown
Case reserves are set by people who can look at the file. Pure IBNR is set by statistics applied to aggregate history. The second is harder, and for long-tailed lines like workers compensation and general liability, it is also larger. For a mature workers compensation program, IBNR can exceed case reserves in magnitude by the time you look at accident years three or four years old. For medical stop-loss it is smaller in dollars but proportionally huge against case reserves because the tail is short.
The reason it matters is simple. If the IBNR estimate is wrong, every downstream number is wrong. Accrued liability on the balance sheet is wrong. The current year expense is wrong. The funding rate for next year is wrong. The premium your captive is charging its parent is wrong. And when the estimate corrects, it corrects all at once, in a single quarter, with everyone watching.
Where the number actually comes from
An actuarial reserve estimate for IBNR is not a single calculation. It is usually the result of running several methods and then picking a point or a range. The most common methods a reader will see referenced are:
- Chain ladder (also called development factor method), which projects future development by applying historical development factors to the current diagonal of a loss triangle.
- Bornhuetter-Ferguson, which blends an a priori expected loss with actual reported losses to date, leaning more on the expectation for green accident years and more on the actuals for mature years.
- Expected loss ratio, which projects ultimate losses from exposure and an assumed loss ratio, useful when loss history is too thin to trust.
- Cape Cod, which estimates the a priori loss ratio from the triangle itself and then applies Bornhuetter-Ferguson.
These names will show up in any reserve report you receive. The methods are not interchangeable, they make different assumptions about what the recent diagonal tells you, and a reserving actuary will usually present several as a cross-check.
What to ask before signing off
You do not need to run the numbers yourself. You do need to ask enough questions to understand what the actuary believed and what they assumed.
- Which methods did you give most weight to, and why? A reasonable answer names the methods, explains what behavior in the data drove the selection, and does not hide behind phrases like professional judgment.
- How did the selected ultimates move from the prior review? Every accident year should have a story. If a year moved materially, the story should be in the report.
- What is the effect of changing the tail assumption by one period? For long-tailed lines, the tail factor is where most of the subjective judgment lives. A small change in tail can move the answer a lot.
- Is this a point estimate or a range, and what would push us out of the range? Point estimates give false precision. Ranges are more honest, but only if the limits mean something specific.
- What did you change in methodology since last year? Methodology changes are fine, but they should be documented and their effect isolated from changes in the data.
Bottom line
IBNR is not a rounding item. For most self-insureds and captives, it is the single largest reserving uncertainty on the balance sheet. You do not need to become an actuary to read an IBNR estimate responsibly, but you do need to ask the questions above and expect substantive answers. When those answers do not come, that is itself useful information.