What is IBNR (Incurred But Not Reported) in insurance?

IBNR (Incurred But Not Reported) is an actuarial reserve for insurance losses that have occurred but have not yet been reported to the insurer. For long-tail lines (liability, workers' comp, professional liability), claims can take years to be reported and years more to settle. The IBNR reserve covers this 'tail.' IBNR is estimated using development factors (also called link ratios) — ratios of cumulative losses at different development ages — applied to reported losses to project ultimate losses. The most common method is the chain ladder (development) method.

What is a loss development factor (LDF)?

A loss development factor (LDF or link ratio) measures how much cumulative claims payments or incurred losses grow from one evaluation period to the next. For example, a 12-to-24 month LDF of 1.45 means claims paid at 24 months are on average 45% higher than at 12 months. LDFs are computed from the claims triangle — a matrix of cumulative losses by accident year and development age. Actuaries select LDFs from historical patterns, often blending volume-weighted averages with simple averages, and adjusting for trend changes.

AI for Insurance Claims Analytics: Loss Development, IBNR, and Claims Triage with Claude (2026)

How actuaries and claims managers use Claude AI for insurance claims analysis: loss development factor selection from claims triangles, IBNR estimation (chain ladder and Bornhuetter-Ferguson), claims fraud indicators, medical severity trending, and reserve adequacy assessment.

Insurance Claims Analytics and AI

Claims analytics drives reserving accuracy, pricing adequacy, and loss management. An actuarial team computing IBNR for a 10-line-of-business portfolio manually selects development factors from loss triangles, applies methods, blends results, and documents conclusions — a process that takes weeks. Claude with ClaudeFinLab handles the computations, method comparisons, and documentation, compressing the cycle significantly.

Loss Development Factor Selection

"Select loss development factors for commercial auto liability from this loss triangle: [paste 10-year cumulative incurred loss triangle]. Compute the volume-weighted average LDF for each development period: 12→24 months 1.52, 24→36 months 1.28, 36→48 months 1.12, 48→60 months 1.05, 60→72 months 1.02, 72→84 months 1.01 (tail). Compute simple averages for comparison. For each period, assess: does the latest 3-year average differ significantly from the 10-year weighted average? Any trend in factors (increasing or decreasing pattern suggests reserve development is changing — flag for actuary review). Recommend selected LDFs with supporting rationale."
"Flag anomalies in the loss development pattern: accident year 2022 shows 12-month incurred losses of $8.4M vs peer years average $6.2M (36% above expected) — may reflect a large loss or adverse development in that accident year. LDF for 2022 at 12→24 months: 1.81 vs selected 1.52 — 19% above selected. Possible explanations: (1) large individual loss; (2) late reporting pattern change; (3) coverage expansion. Recommendation: isolate the 2022 anomaly in the analysis, consider capping the loss at expected level for LDF selection, and perform a large loss review."

IBNR Estimation Methods

"Estimate IBNR using the chain ladder method: apply selected LDFs to the latest diagonal of the loss triangle. Accident year 2023 (12 months development): reported losses $6.8M. Apply cumulative development factor (CDF) to ultimate: 12-ult CDF = 1.52 × 1.28 × 1.12 × 1.05 × 1.02 × 1.01 = 2.37. Ultimate loss estimate: $6.8M × 2.37 = $16.1M. IBNR = $16.1M - $6.8M = $9.3M. Repeat for all accident years. Sum IBNR across years = total reserve requirement. Compare to case reserves (already established by claims): total IBNR + case reserves = total indicated reserves."
"Cross-check with Bornhuetter-Ferguson method: BF blends the development method with an a priori loss ratio estimate. For accident year 2023: a priori loss ratio 68% (based on pricing), earned premium $22.5M → a priori ultimate = $15.3M. BF IBNR = a priori ultimate × (1 - 1/CDF) = $15.3M × (1 - 1/2.37) = $15.3M × 0.58 = $8.87M. Compare to chain ladder IBNR $9.3M. Difference of $0.43M (4.6%) — within normal range. BF is preferred when the latest diagonal has limited statistical credibility (young accident year with CDF > 2.0)."

Claims Fraud Indicators

"Analyze this commercial auto claim for fraud indicators: Accident: 3-car rear-end collision, 2 AM on a Saturday, suburban highway. Claimants: 4 passengers in a 5-seat sedan (crowded vehicle late at night). Injuries: all 4 claim soft-tissue (whiplash) injuries, all treated by the same clinic, all represented by the same attorney. Settlement demands: $45K, $48K, $52K, $38K (similar amounts, uncommon for 4 unrelated individuals). Prior claims history: 2 of 4 claimants have previous soft-tissue claims. Red flags for Special Investigations Unit: pre-arranged collision indicators, medical mill involvement, attorney concentration. Recommend: refer to SIU, request independent medical examination, conduct attorney/clinic background check."

Reserve Adequacy Analysis

"Assess reserve adequacy for this workers' compensation book: actuarial IBNR indication $42.5M. Carried reserves: $38.8M. Reserve deficiency: $3.7M (9.5% deficient). Medical severity trend: +8.2% per year (2022-2025) — elevated due to pharmaceutical cost inflation. If trend continues, next year's development will further increase IBNR indication. Required reserve strengthening: $3.7M immediate, plus $2.1M prospective reserve for trend impact. Present the reserve analysis memo for the audit committee: current carried reserves, actuarial indication, deficiency, trend context, and management's planned response (reserve increase schedule)."

Where to Start

Paste your 5-10 year loss triangle (cumulative incurred losses by accident year and development age) and ask Claude to compute the volume-weighted average development factors and the 3-year and 5-year simple averages for comparison. Once you identify any anomalous years or shifting development patterns, apply the chain ladder and BF methods to estimate IBNR. That analytical workflow — which typically takes an actuarial team 2-3 days — can be structured in an hour with Claude.

Get your free API key