Risk Modelling and Survival Analysis · Loss distributions, with and without risk sharing

Effect of Inflation on Loss Distributions and Deductibles

Updated 11 October 2026 · Fact-checked

If every claim grows by a factor (1 + r), the new claim is Y = (1 + r)X. For scale-family distributions, this multiplies the scale parameter by (1 + r) and leaves the shape unchanged. A fixed deductible or retention does not inflate, so the insurer or reinsurer pays more than proportionally more (leverage). Solve by rescaling the limit, not the claims.

Understand Effect of Inflation on Loss Distributions

Claim inflation means claim amounts rise over time. The simplest model says every claim is multiplied by the same factor. If last year's claim is X and inflation is r, this year's claim is Y = (1 + r)X. Here r is the total inflation over the period, not the annual rate, unless the period is one year.

Some distributions are closed under this change. A scale family keeps its type when you multiply by a positive constant. Only the scale parameter changes. The shape stays the same. For the Pareto with parameters α and λ, Y is Pareto with α and (1 + r)λ. For the exponential with rate λ, Y is exponential with rate λ ÷ (1 + r). For the gamma with α and λ, Y is gamma with α and λ ÷ (1 + r). For the Weibull and the loglogistic, the scale parameter is multiplied by (1 + r) in the same way.

The lognormal is different in its parameters but works the same way. If X is lognormal(μ, σ²), then ln Y = ln X + ln(1 + r). So Y is lognormal with μ increased by ln(1 + r) and σ unchanged. Here e^μ acts as the scale parameter.

The key idea for deductibles and reinsurance: inflation acts on the claim, but the deductible or retention is a fixed rupee amount. Claims move up past a fixed point, so more claims exceed it and each excess is larger. The expected payment E[(Y − d)+] of the insurer or reinsurer therefore rises by more than r in percentage terms. This is called leverage. It applies to the expected payment, not to the exceedance probability P(Y > d), which can rise by less than r. Do not inflate the deductible unless the question says it is indexed.

To find the probability or expected payment after inflation, work with Y directly. P(Y > d) = P(X > d ÷ (1 + r)). So you deflate the limit and use the old distribution, or you use the new parameters and the old limit. Both give the same answer.

Key rules to remember

Inflated claim: Y = (1 + r)X; r is total inflation over the period. For n years at annual rate i, use (1 + i)ⁿ.
Probability after inflation: P(Y > d) = P(X > d ÷ (1 + r)); Equivalent to using the new distribution of Y with the same d.
Pareto: X ~ Pareto(α, λ) ⇒ Y ~ Pareto(α, (1 + r)λ); Shape α unchanged. Mean is λ ÷ (α − 1) for α > 1.
Exponential and gamma: X ~ Exp(λ) ⇒ Y ~ Exp(λ ÷ (1 + r)); X ~ Gamma(α, λ) ⇒ Y ~ Gamma(α, λ ÷ (1 + r)); Rate parameter is divided. Scale (1 ÷ λ) is multiplied.
Lognormal: X ~ LN(μ, σ²) ⇒ Y ~ LN(μ + ln(1 + r), σ²); σ is unchanged.
Mean and variance: E[Y] = (1 + r)E[X]; Var(Y) = (1 + r)² Var(X); Standard deviation scales by (1 + r). The coefficient of variation is unchanged.
Expected payment above a fixed deductible: E[(Y − d)+] = (1 + r) × E[(X − d ÷ (1 + r))+]; Use this to price excess of loss cover after inflation with a fixed d.

How to solve Effect of Inflation on Loss Distributions questions

Use this method for any question on inflation with a distribution, deductible or retention.

1Identify the inflation factor (1 + r). Check whether r is annual or total, and compound over several years if needed.
2Write the distribution of the old claim X and name its family.
3Find the distribution of Y = (1 + r)X by changing the scale parameter only. For the lognormal, add ln(1 + r) to μ.
4Decide whether the deductible, limit or retention is fixed or indexed. Unless stated, treat it as fixed.
5Convert the question to one distribution. Either use Y with the fixed d, or use X with d ÷ (1 + r).
6Calculate the probability, mean or expected payment using the stated formula or the given distribution function.
7Check the expected payment for sense: with a fixed deductible and positive inflation, E[(Y − d)+] should rise, and the rise should exceed r in percentage terms. Do not apply this check to P(Y > d), which can rise by less than r.

Quickest way: Deflate the limit and reuse the old distribution

When to use it: Use this when the old distribution is easy to work with, such as Pareto or exponential, and you need a probability or an excess payment.

Divide the fixed limit d by (1 + r) to get d′.
Compute the required quantity for X at d′.
For probabilities, stop: P(Y > d) = P(X > d′).
For expected excess payments, multiply the result by (1 + r).
For a lognormal, standardise with (ln d − μ − ln(1 + r)) ÷ σ.

Common mistakes in Effect of Inflation on Loss Distributions

Inflating the deductible or retention along with the claims
Students think everything in the question grows with inflation.
Fix: Read the wording. A fixed rupee deductible stays fixed. Only change it if the question says it is indexed.
Changing the shape parameter of the Pareto or gamma
Students are unsure which parameter is the scale.
Fix: The shape (α) never changes. Multiply the scale λ for the Pareto. Divide the rate λ for the gamma and exponential.
Multiplying the exponential rate by (1 + r)
Rate and scale are mixed up. The mean is 1 ÷ λ, so a larger mean needs a smaller rate.
Fix: Check by the mean. The new mean must be (1 + r) times the old, so the new rate is λ ÷ (1 + r).
Assuming the insurer's cost rises by exactly r
Students apply inflation to the total cost instead of the excess part.
Fix: With a fixed deductible, the expected payment E[(Y − d)+] rises by more than r. Compute it directly. Do not expect the same leverage in the exceedance probability P(Y > d).
Adding r to μ for the lognormal, or changing σ
Students copy the rule for a normal shift without the log.
Fix: Add ln(1 + r) to μ. Keep σ the same.
Using the annual rate when the gap is several years
The time period is missed in the question.
Fix: Compound: the factor is (1 + i)ⁿ. Check the stated years.

Worked examples

Example 1

Claims X follow a Pareto distribution with α = 3 and λ = ₹60,000. Claim inflation is 10% over the year. An excess of loss reinsurer has a fixed retention of ₹30,000. Find the percentage increase in the probability that a claim exceeds the retention, the new mean claim, and the percentage increase in the reinsurer's expected payment per claim.

Show the solution

The old distribution is Pareto(3, 60,000). P(X > x) = (λ ÷ (λ + x))^α.
Old probability: (60,000 ÷ 90,000)³ = (2/3)³ = 8/27 = 0.2963.
After inflation, Y is Pareto(3, 66,000).
New probability: (66,000 ÷ 96,000)³ = (0.6875)³ = 0.3250.
Check: 0.32495 ÷ 0.29630 = 1.0967, so the probability rises by about 9.7%. This is less than the 10% inflation, so the exceedance probability does not show leverage.
Old mean = 60,000 ÷ 2 = ₹30,000. New mean = 66,000 ÷ 2 = ₹33,000, which is 1.1 × 30,000.
For a Pareto, E[(X − d)+] = λ ÷ (α − 1) × (λ ÷ (λ + d))^(α − 1).
Old expected payment: 30,000 × (2/3)² = 30,000 × 4/9 = ₹13,333.
New expected payment: 33,000 × (66,000 ÷ 96,000)² = 33,000 × 0.47266 = ₹15,598.
Increase: 15,598 ÷ 13,333 = 1.1698, so about 17.0%. This is more than 10%, which shows leverage in the expected payment.

Answer: The probability of exceeding the retention rises from 0.2963 to 0.3250, an increase of about 9.7%, which is less than 10%. The new mean claim is ₹33,000. The reinsurer's expected payment per claim rises from about ₹13,333 to about ₹15,598, an increase of about 17.0%, which is more than 10%.

Example 2

Claims X follow an exponential distribution with mean ₹20,000. Claims inflate by 25% and a fixed deductible of ₹10,000 applies. Find the expected payment per claim by the insurer before and after inflation.

Show the solution

For an exponential with mean m and deductible d, E[(X − d)+] = m × e^(−d ÷ m).
Before: 20,000 × e^(−0.5) = 20,000 × 0.60653 = ₹12,130.66.
After inflation, Y is exponential with mean 25,000.
After: 25,000 × e^(−10,000 ÷ 25,000) = 25,000 × e^(−0.4).
e^(−0.4) = 0.67032, so the value is ₹16,758.
Increase: 16,758 ÷ 12,131 = 1.381, so about 38.1%.
This is more than the 25% claim inflation, which shows leverage.

Answer: Expected payment per claim rises from about ₹12,131 to about ₹16,758, an increase of about 38%.

Exam tips

Always state the new distribution and its parameters before calculating. Marks are usually given for this step.
Say clearly whether the deductible or retention is fixed, and give the reason in one line.
Mention leverage in written answers: a fixed deductible makes the insurer's or reinsurer's cost grow faster than claim inflation.
In MCQs, check which of the two parameters changes. Options often include the wrong shape change or the wrong direction for a rate.
For the lognormal, quote μ + ln(1 + r) and say σ is unchanged. Show the standardisation step.

Practice questions from Loss distributions, with and without risk sharing

Effect of Inflation on Loss Distributions in other exams

The same ground in other exams, if you are preparing for more than one or want another angle on it.

ACCA Applied SkillsEffects of Inflation and Historical Cost Accounting

Effect of Inflation on Loss Distributions: frequently asked questions

How does inflation change a Pareto claim distribution?

If claims grow by a factor (1 + r), the Pareto scale parameter λ becomes (1 + r)λ. The shape α stays the same. The mean rises by the same factor when α > 1.

Why does inflation hurt an excess of loss reinsurer more than it hurts the claims?

The retention is fixed in rupees. As claims rise, more of them pass the retention and each excess is larger. The reinsurer's expected payment therefore rises by a bigger percentage than r.

How do I handle inflation for a lognormal distribution?

Add ln(1 + r) to μ and keep σ unchanged. This is because ln Y = ln X + ln(1 + r). Then use the usual normal tables with the new μ.

What if the deductible is indexed to inflation?

Then both the claim and the deductible grow by (1 + r). The ratio between them is unchanged. The expected payment then rises by exactly the factor (1 + r).