Burning Cost Reinsurance Pricing: Method and Pitfalls

Burning Cost Analysis in Reinsurance Pricing: Method, Adjustments and Pitfalls

Last reviewed: November 2026

Burning cost reinsurance pricing estimates the expected cost of a layer from the cedent's own loss history, expressed as a percentage of subject premium or another exposure base. Done properly, every historical loss and every year of exposure is restated to the conditions of the treaty being priced before one is divided by the other.

The raw version, losses to layer divided by premium, is easy to calculate and nearly always wrong. On excess layers, the adjustments routinely change the answer by a factor of two or more.

This guide is for treaty underwriters, pricing actuaries, chief actuaries and heads of outwards reinsurance. It goes step by step through trending, development, IBNR, exposure, as-if adjustments, credibility and loadings, with a worked casualty XoL example.

Key Takeaways

• A burning cost is only as good as its adjustments; the raw ratio of past layer losses to past premium understates excess layers.
• Leveraged trend means a modest ground-up inflation rate can more than double losses to an excess layer.
• The large-loss reporting threshold must sit below the retention divided by the cumulative trend factor, or losses that trend into the layer go missing.
• Excess development factors are larger and longer than ground-up factors, and immature years need a credibility-weighted view of IBNR.
• As-if adjustments are where cedents and reinsurers disagree most, so each one needs evidence.
• Credibility should follow the number of layer losses, with the balance from exposure rating, before loadings convert the loss cost into a technical rate.

What Is Burning Cost in Reinsurance Pricing?

Burning cost is the experience-based estimate of how much a reinsurance layer will cost, measured as historical layer losses relative to an exposure base. It answers one question: if the past repeated under today's conditions, what would the layer pay?

The exposure base is usually subject premium income (SPI), the cedent's premium for the business protected by the treaty. Some lines use other bases, such as vehicle years, payroll or sums insured.

When Is Burning Cost Used?

Burning cost is the main pricing method for working layers and lower middle layers, where losses are frequent enough to be informative. It is also a key cross-check on exposure rating, even where it does not drive the final price.

For high layers that have rarely or never been hit, burning cost says little on its own. Those layers lean on exposure rating, market curves and scenario views, which we discuss in how reinsurers price risk they have never seen before.

How Does the Burning Cost Method Work?

The burning cost method works in eight steps. Each step restates the past to the conditions of the future treaty period.

1. Collect data. Gather individual large losses above a suitable threshold and subject premium by year, usually for at least five to ten years.
2. Trend losses. Inflate each loss from its date of occurrence to the average date of the treaty period.
3. Apply the layer. Calculate the part of each trended loss that falls in the layer.
4. Develop to ultimate. Adjust layer losses for future development on known claims and for claims not yet reported.
5. Adjust exposure. Restate subject premium to the future rate and exposure level.
6. Make as-if adjustments. Restate losses and exposure for changes in limits, portfolio, law or structure.
7. Weight for credibility. Blend the burning cost with an exposure-rated or market view.
8. Load. Add brokerage, expenses and a margin for profit and cost of capital, and allow for reinstatement terms.

The order matters. Trending must happen before the layer is applied, and development is best applied to layer losses rather than ground-up losses.

How Should Losses Be Trended?

Each loss should be trended individually from its occurrence date to the average loss date of the treaty period, using a severity trend that reflects the line's claims inflation. Trending must happen before the layer is applied, because the layer amplifies the effect.

Why Excess Layers Feel Inflation More

This effect is called leveraged trend. Inflation increases the part of every loss above the retention by more than the loss itself, and it pushes losses that were below the retention into the layer.

With a fixed retention, a 6% ground-up trend can produce layer trend well above 6% a year. The worked example later shows layer losses rising by about 150% from trending alone.

Setting the Large-Loss Threshold

The cedent must report every loss that could trend into the layer, not just those that already exceed the retention. The reporting threshold should be no higher than the retention divided by the cumulative trend factor for the oldest year.

For a USD 5 million retention, six years of 6% trend gives 5 ÷ 1.06⁶, or about USD 3.5 million. A loss list cut at USD 5 million would miss most of the losses that matter.

What Goes Into the Trend Rate?

Severity trend should reflect the claims inflation of the line, not general consumer prices. Casualty lines need to allow for wage growth, medical costs, court awards and social inflation, which we cover in social inflation in casualty reinsurance.

Legal changes need separate as-if treatment rather than a smoothed trend. In England and Wales, for example, the personal injury discount rate moved to 0.5% from 11 January 2025, which changes the value of large bodily injury claims settled before and after that date.

What About Indexation Clauses?

Some casualty treaties, particularly motor, carry an indexation or stability clause that moves the retention and limit with an agreed index. Where one applies, trend the losses and index the layer consistently, or the leveraged trend will be double-counted or missed.

How Do You Develop Layer Losses and Allow for IBNR?

Layer losses should be developed to ultimate using factors built for excess losses, not ground-up factors. Excess losses report later and develop further, because large claims take longer to emerge and to settle.

Two Kinds of IBNR

1. IBNER (incurred but not enough reported). Known claims whose case reserves will increase. On excess layers, this often pushes claims from below the retention into the layer.
2. Pure IBNR. Claims that have happened but have not been reported to the cedent, or not reported above the large-loss threshold.

Both matter more for excess layers. A claim reserved at USD 3 million today, below a USD 5 million retention, may settle at USD 7 million in five years and put USD 2 million into the layer.

Which Development Method Fits?

A chain ladder on the layer's own triangle is rarely stable because there are too few claims. Common alternatives include factors derived from ground-up triangles adjusted for the excess layer, market benchmark factors, and Bornhuetter-Ferguson style estimates that blend an a priori loss cost with reported experience.

For the most recent years, the a priori view should carry most of the weight. A chain ladder factor of 2 or more applied to one or two reported claims can swing the result either way.

The CAS study note Basics of Reinsurance Pricing covers experience rating, including trend and development of excess losses, in more technical depth.

How Do You Adjust Exposure and Make As-If Adjustments?

The exposure base must be restated to the rate level and exposure profile of the treaty period, and losses restated for any structural change. Otherwise the burning cost mixes future-value losses with past-value premium and past conditions.

On-Levelling Premium

Restate each year's subject premium to the future rate level using the cedent's rate changes. The definition of SPI, gross or net of commission, must be identical in the historical data and the treaty wording.

Exposure Change and Mix

Growth by itself is handled by dividing by premium, but a change in the mix of business is not. If the cedent has moved into higher limits, heavier industries or new territories, the historical burning cost will understate the future.

As-If Adjustments

As-if adjustments restate historical losses and exposure as if today's conditions had applied throughout. They are legitimate and necessary, and they are also where most pricing disagreements arise.

Removing losses without removing the matching premium is the most common error. The second is accepting that a large loss is unrepeatable without asking what replaced the business that produced it.

Worked Example: Burning Cost for a 5 xs 5 Casualty Layer

The illustration below is simplified and uses USD millions. It prices a 5 xs 5 casualty XoL on a losses-occurring basis for the 2027 treaty year.

Assumptions:

• Six large losses from accident years 2021 to 2025, as reported at 30 September 2026
• Severity trend of 6% a year, trended to 2027
• On-levelled subject premium for 2021 to 2025 of 300, 330, 360, 390 and 420, a total of 1,800
• Excess development factors to ultimate of 1.05, 1.10, 1.25, 1.50 and 2.00 for 2021 to 2025

Step 1: Trend Each Loss and Apply the Layer

Before trending, only two losses reach the layer, contributing 2.9 and 1.1, a total of 4.0. After trending, all six do.

Step 2: Develop and Divide by Exposure

Step 3: Test an As-If Request

The cedent argues that the 2023 loss came from a scheme it exited in 2024, which wrote 40 of that year's premium. Removing the loss (3.38 developed) and the premium gives 8.30 ÷ 1,760, or 0.47%.

That is a large move from one adjustment. Without evidence of the exit and of the premium removed, the reinsurer should not accept it in full.

Step 4: Credibility, Loading and Price

Six layer losses over five years is limited experience, so the example gives the burning cost a credibility of 40%. The remaining 60% goes to an exposure-rated loss cost of 0.55% of SPI.

The figures use unrounded intermediate values: 0.5894% for the loss cost, 0.8421% for the rate and 3.958 for the premium. On the same credibility and loadings, the raw 0.22% burning cost would have produced a technical premium of about 2.81.

If the as-if exclusion were accepted in full, the premium would fall to about 3.48. The single adjustment is worth roughly 0.48 of premium, more than 10% of the price.

The lesson for reinsurers: trending and development moved the burning cost from 0.22% to 0.65%, and that adjustment matters far more than any negotiation over the final loading. The lesson for cedents: an as-if request backed by clear data on exited premium and claims is far more likely to be accepted.

How Do Credibility and Loadings Turn Burning Cost Into a Price?

Credibility decides how much the burning cost counts against exposure rating, and loadings convert the blended loss cost into a technical rate. Both should be recorded with reasons so the next renewal can test them.

How Much Credibility Should Burning Cost Get?

Credibility should rise with the number of losses in the layer and the stability of the portfolio. The table below shows when to lean on burning cost and when to lean on exposure rating.

Formal credibility methods tie the weight to the expected number of claims in the layer. Whatever method is used, the weight is a judgement that should be visible in the pricing file.

How Are Loadings Applied?

Loadings convert an expected loss cost into a price that covers the reinsurer's costs and capital. A common form is the loss cost divided by one minus the total loading percentage.

Typical components are brokerage, the reinsurer's internal expenses, and a margin for profit and the cost of the capital the layer consumes. Capital-intensive layers, such as those with high volatility or long tails, carry a larger margin.

Reinstatement terms also matter. Where paid reinstatements are expected, the reinsurer receives extra premium after losses, so the upfront rate can be lower for the same expected result.

Aggregate deductibles, annual aggregate limits and profit commissions change the expected loss to the reinsurer. They should be modelled on simulated or scenario losses rather than on the average burning cost alone.

Which Burning Cost Pitfalls Distort the Price?

Seven pitfalls account for most mispriced layers. Each one is easy to check if the data is available.

What Data Should Reinsurers Request?

Reinsurers should request data that makes every adjustment testable. The table below lists the core items.

Which Questions Should Reinsurers Ask?

• What threshold was used for the loss list, and does it allow for the full trend period?
• Which development factors were used, and how were they derived for excess losses?
• Which as-if adjustments have been made, and where is the matching premium adjustment?
• Has the limit or industry profile shifted towards larger risks?
• How do recent claims reports compare with case reserves on the same claims last year?

Late or incomplete claims data undermines every one of these steps. We explain the reserving and cash effects in claims bordereaux lag.

How Does Market Context Affect Burning Cost?

Burning cost looks backwards, so it reacts slowly when conditions change. That matters most when claims inflation shifts or when long-tail reserves start to develop.

Casualty lines have seen sustained concern about social inflation and reserve development on recent accident years. Those trends affect both the trend rate and the development factors, and they compound on excess layers.

In property, the relevance of burning cost depends on the layer. Large per risk losses suit it, while catastrophe exposure is better priced with models and scenarios.

A softer market adds pressure to accept favourable as-if adjustments and lighter trend assumptions. Discipline on those inputs protects results long after rate changes have been forgotten.

Checklist for Treaty Underwriters and Pricing Actuaries

Use this checklist before you sign off a burning cost.

• The loss threshold is at or below the retention divided by the cumulative trend factor
• Each loss is trended individually before the layer is applied
• Layer losses are developed with excess-specific factors, with a blended view for recent years
• Subject premium is on-levelled and the SPI definition matches the treaty wording
• Every as-if adjustment removes both losses and matching premium and is evidenced
• The credibility weight and the exposure-rated alternative are recorded with reasons
• Loadings, reinstatement terms and aggregate features are applied consistently
• The loss list is held in a controlled system and reconciled year on year

For the last point, keeping treaty terms, losses and settlements together makes the burning cost reproducible. A centralised reinsurance treaty management platform holds the structure that each year's losses must be measured against.

Key Terms

Burning cost: Historical layer losses divided by an exposure base, after restating both to the conditions of the treaty period.

Subject premium income (SPI): The cedent's premium for the business protected by the treaty, used as the exposure base.

Leveraged trend: The effect by which inflation increases losses to an excess layer by more than it increases ground-up losses.

On-levelling: Restating historical premium to the rate level expected in the treaty period.

Excess development factor: A factor that develops excess layer losses to ultimate, usually larger and longer than ground-up factors.

IBNER: Incurred but not enough reported. The future increase on claims already known.

As-if adjustment: A restatement of historical losses or exposure as if current conditions had applied.

Credibility: The weight given to the cedent's own experience when blending it with another estimate, such as exposure rating.

Technical rate: The rate that covers expected losses plus brokerage, expenses and a margin for profit and capital.

Sources

• David R. Clark, Basics of Reinsurance Pricing, Casualty Actuarial Society study note
• Casualty Actuarial Society: Basics of Reinsurance Pricing abstract
• GOV.UK: Personal injury discount rate