IBNR Reserving Explained: Methods, Assumptions and Software

Written for actuaries, CFOs, and finance leads at insurers who need to pick the right IBNR method, defend its assumptions, and close each quarter on time.

TL;DR

IBNR reserving puts a number on claims that already happened but haven’t been reported. This guide shows how the main IBNR methods work and which assumptions matter most. You’ll see a worked chain ladder example beside a Bornhuetter-Ferguson check. We also cover when spreadsheets stop being safe and what IBNR software should do. Written for actuaries and finance leads at insurers across the GCC and beyond. Expect plain language and real arithmetic. Read the methods section first. Then test your own triangles against the assumption list before your next reserve review.

When a reserve turns out too small, nobody finds out in the quarter that matters. They find out two years later, in a rating note or a board meeting nobody wanted to attend.

Incurred but not reported claims are the reason. These are losses that already happened, sit on your books as a liability, and haven’t reached the claims system yet. IBNR reserving is how you put a number on them. Get it wrong and every figure built on it, from pricing to capital, inherits the error.

Prima’s actuaries work on reserving for insurers across the GCC, Europe, and Asia-Pacific. The pattern we see is simple. Teams argue about method, then lose the quarter to messy data and version control.

This guide covers both sides. You’ll get the methods, the assumptions behind them, a worked example, and a plain view on software.

What this article covers

• How the main IBNR methods work and where they fail
• Which assumptions move your reserve the most
• When spreadsheets stop being safe at quarter-end

What Is IBNR Reserving and Why Do Late Claims Wreck Quarter-End?

IBNR reserving is the actuarial estimate of claims that are incurred but not reported: they’ve happened, but the insurer hasn’t heard yet. It sits next to case reserves, which cover the claims you already know about.

Think of it as the part of your liability you can’t see in the claims system. Full stop.

What incurred but not reported actually includes

In practice, IBNR means two things. First, pure IBNR: claims nobody has told you about. Second, development on claims you already carry at a case reserve that turns out too low or too high. Most teams report both as one number.

A motor claim reported in week two and a liability claim reported in year four are the same problem on different clocks. That’s why reporting lag matters more than almost any other input.

Where IBNR sits in the reserve stack

In IBNR reserving, ultimate claims has three layers: what you’ve paid, what your claims team has reserved on open files, and IBNR. Subtract the first two from ultimate and you’ve got the IBNR figure.

It also feeds the wider work of reserving, pricing and insurance analytics, so a weak IBNR number leaks into your pricing too.

So how do you estimate something nobody has reported yet?

Which IBNR Methods Hold Up, and Where Does Each One Break?

You might think the most sophisticated IBNR reserving method wins. It doesn’t. The best method is the one your data can support.

Chain ladder and development factors

Chain ladder is still the workhorse of IBNR reserving. It takes your claims triangle, measures how claims grew from one period to the next, and projects that pattern forward to ultimate claims.

It assumes the past repeats. That’s the whole bet, and it’s a big one. AM Best flagged the weak spot in 2024: COVID-19 disrupted the payment patterns actuaries lean on, and the agency expected reserve strengthening on accident years 2015 to 2019 to continue.

Bornhuetter-Ferguson for thin or volatile data

Bornhuetter and Ferguson published their method in 1972. It keeps the claims you’ve actually seen and fills the unreported share with a prior expected loss ratio.

Use it on your newest accident year, or on a low-frequency, high-severity line, where one large claim would otherwise get multiplied by a huge development factor. The catch is the prior. It’s a judgment call, and a stale prior is a quiet way to be wrong.

Expected loss ratio, Cape Cod and average cost per claim

Expected loss ratio works when you’ve got premium but no usable history. Cape Cod refines Bornhuetter-Ferguson by deriving the loss ratio from your own data. Average cost per claim suits medical books, where inflation drives severity.

How to choose between IBNR methods

• Mature year, stable pattern: chain ladder.
• Immature year or volatile line: Bornhuetter-Ferguson.
• No claims history: expected loss ratio.
• Inflation-driven severity: average cost per claim.

We can’t tell you which one fits your book without seeing your triangles. Nobody honestly can. Run two IBNR methods side by side and study the gap.

What Assumptions Quietly Drive Your IBNR Reserving Result?

Method choice is the easy half of IBNR reserving. Our technical reserves valuation work usually starts with the harder half, the assumptions underneath it.

Two actuaries can run the same triangle through IBNR reserving and book very different reserves. The arithmetic matches. The judgment doesn’t.

Development factors and the tail

Which averages do you select? Do you exclude the outlier year? How long is the tail, and why? Tail factors deserve more argument than they usually get.

S&P Global Market Intelligence found $7.3 billion of adverse one-year development in US other liability (occurrence) lines for 2025. About 43.3% sat in the three most recent accident years, not in the old tail where reserves normally unwind. The report blamed legal and social inflation moving faster than insurers expected.

That’s an assumption failure. Not a method failure.

Reporting lags, claims handling and large events

A faster claims team breaks your historical pattern overnight. So does a new claims system, a policy wording change, or a flood. In the Actuaries Institute’s 2024 survey of Australian practice, 78% of respondents model catastrophe losses separately, for exactly this reason.

Why segmentation matters

Averages lie. Assured Research’s year-end 2025 analysis put industry carried reserves $20.7 billion redundant across all lines, yet $12.5 billion deficient in other liability (occurrence) alone.

Which segment in your book is hiding behind a healthy total?

Where IFRS 17 changes the conversation

Under IFRS 17, the unpaid claims estimate becomes the liability for incurred claims. It carries a discount and a risk adjustment for non-financial risk. That’s where IFRS 17 advisory and reserving meet, and where sloppy IBNR gets expensive.

How Do You Calculate IBNR? A Worked Chain Ladder Example

Here’s one simplified accident year of IBNR reserving. The numbers are illustrative, not from a client.

1. Reported incurred claims, paid plus case reserves, stand at 8.4 million.
2. Your selected development factor to ultimate is 1.25.
3. Chain ladder ultimate is 8.4 million times 1.25, or 10.5 million.
4. IBNR is ultimate minus reported incurred: 2.1 million.

That’s it for chain ladder. Four lines of arithmetic, and the rest is judgment.

Now run Bornhuetter-Ferguson. Earned premium is 14 million and your prior loss ratio is 70%, so expected ultimate is 9.8 million. A factor of 1.25 means 80% of claims are reported and 20% aren’t, so BF IBNR is 9.8 million times 20%, or 1.96 million.

BF ultimate comes to 10.36 million. Chain ladder says 10.5 million. The gap is small here. On a volatile line it won’t be.

That gap is the conversation. Have it before you book the number.

Prima’s team of actuaries, CPAs, and CFAs serves insurers across the GCC, Europe, and APAC, including appointed actuary services for signed reserve opinions.

IBNR Software: When Do Spreadsheets Stop Being Safe?

Spreadsheets aren’t the villain. They’re how most IBNR reserving teams learned the trade. But they carry a risk nobody writes into the reserving memo.

A February 2026 FinTech Global piece describes the pattern. As reserving cycles repeat, broken links and undocumented assumptions pile up, especially when the person who built the workbook leaves. WTW’s 2023 IFRS 17 survey listed data, skilled people, and systems as the primary challenges for production.

What IBNR software should do at quarter-end

Good IBNR software doesn’t replace the actuary. It removes the busywork around the actuary. Look for four things in any actuarial reserving software:

• Chain ladder, Bornhuetter-Ferguson, and expected loss ratio side by side
• A reserve movement waterfall against the last valuation
• A log of every override and excluded data point
• A clean hand-off to IFRS 17 measurement, with no re-keying

The IBNR software names you’ll meet on most shortlists are WTW’s ResQ, Arius, and Addactis. We’ve written a buyer’s guide to insurance reserving software for GCC insurers, including what to ask vendors about SAMA templates.

If you’d rather see a finished process, our actuarial reserving solution page walks from triangle to report. The Aegis reserving solution is the platform Prima’s own actuaries run on live engagements.

Controls that catch a bad IBNR before the auditor does

In IBNR reserving, controls matter more than features. Lock the data cut-off. Reconcile triangles to the ledger. Compare this quarter’s selected factors to last quarter’s. Make every override carry a name and a reason.

The appointed actuary then reviews a documented trail instead of a pile of tabs.

Getting IBNR Reserving Right Starts Before Quarter-End

IBNR reserving isn’t a math problem you solve once. It’s a judgment you defend every quarter. Pick the method your data supports. Write the assumptions down. Compare at least two IBNR methods.

If your IBNR software is still a chain of workbooks, start with one question. Could a new team member rerun last quarter’s IBNR and land on the same number? If not, that’s your first fix.

IBNR Reserving FAQs

What is IBNR in insurance?

IBNR stands for incurred but not reported. It’s the reserve an insurer holds for claims that have already happened but haven’t been reported yet. IBNR reserving estimates it from claims triangles and reporting lags, then adds it to case reserves to reach total outstanding claims.

How do you calculate IBNR?

Estimate ultimate claims first, usually by applying development factors to reported claims with the chain ladder method. Then subtract paid claims and case reserves. What’s left is IBNR. Many teams cross-check the result with Bornhuetter-Ferguson, especially for recent accident years.

What is the difference between IBNR and case reserves?

Case reserves are your claims team’s estimates on claims already reported and still open. IBNR covers incurred but not reported claims, plus expected development on known claims. Both sit inside outstanding claims, and actuarial reserves evaluation reviews them together at each valuation date.

Which IBNR method is best?

None wins everywhere. Chain ladder suits mature years with stable development. Bornhuetter-Ferguson suits recent or volatile years. Expected loss ratio works when history is thin. Run at least two IBNR methods and investigate the gap before you select a final number.

When should an insurer move from spreadsheets to IBNR software?

Move when reviews run late, formulas change without a record, or one person holds the model in their head. Also move when you need IFRS 17 outputs your workbook can’t produce. Dedicated IBNR software adds audit trails, version control, and repeatable quarter-end runs.