Integrated Inflation and LER Calculation Workshop

Why Inflation and Deductibles Must Be Considered Together

When claim severities trend upward, a deductible that stays fixed in dollars becomes smaller relative to typical losses. More of each loss can then exceed the deductible, changing both the insurer's expected payment and the proportion of ground-up losses eliminated by the deductible. Indexing the deductible along with losses can produce a different result.

The key is to calculate each period consistently: first trend the ground-up losses, then apply that period's deductible, and finally calculate the loss elimination ratio using that period's expected ground-up loss as the denominator. This tutorial uses hypothetical claim models and assumes no policy limit, coinsurance, expenses, or other coverage modifications.

Definition: For a nonnegative ground-up loss $X$ with finite mean, an ordinary deductible $d$ gives the insurer payment per loss $Y^L=(X-d)_+=\max(X-d,0)$. The loss elimination ratio (LER) is the proportion of expected ground-up loss not paid by the insurer: $$ \operatorname{LER}(d)=\frac{E[X]-E[Y^L]}{E[X]}=1-\frac{E[Y^L]}{E[X]}, \qquad E[X]>0. $$ The payment formula assumes an ordinary deductible with no limit or other coverage modification. The LER definition does not require a particular severity distribution.

For an ordinary deductible, the amount eliminated on a particular loss is $\min(X,d)$. Thus the numerator of the LER can also be calculated as $E[\min(X,d)]$. This gives two equivalent routes: calculate expected insurer payments and subtract from expected ground-up loss, or calculate expected eliminated loss directly.

Trend Factors and Indexed Deductibles

Let $X$ denote the ground-up loss at a base period. If losses increase proportionally by a deterministic cumulative factor $c_t>0$ by period $t$, the trended loss is $X_t=c_tX$. For annual inflation rates $r_1,\ldots,r_t$, the cumulative factor is $c_t=\prod_{j=1}^{t}(1+r_j)$. The expected ground-up loss in that period is therefore $E[X_t]=c_tE[X]$.

A fixed deductible remains at its stated dollar amount in every period. An indexed deductible begins at a base amount $d_0$ and is multiplied by the same cumulative trend factor as the loss, so $d_t=c_td_0$. These definitions assume the index tracks the same proportional trend as the modeled losses. If trend changes the distribution in some other way, the proportional scaling results below need not hold.

Theorem: If $X_t=c_tX$ and the deductible is indexed proportionally as $d_t=c_td_0$, then the insurer payment is $Y_t^L=c_t(X-d_0)_+$. Consequently, the expected payment and expected ground-up loss both scale by $c_t$, and the LER is unchanged from its base-period value.

The result follows directly from the positive-part payment formula. Since $c_t>0$, multiplying a positive part by $c_t$ preserves whether it is zero and scales its amount. Taking expectations then scales both numerator and denominator of the LER by the same factor.

(X_t-d_t)_+=(c_tX-c_td_0)_+=c_t(X-d_0)_+, \qquad \operatorname{LER}_t =1-\frac{c_tE[(X-d_0)_+]}{c_tE[X]} =\operatorname{LER}_0.

For a fixed deductible $d$, the period-$t$ payment is instead $(c_tX-d)_+$. It cannot generally be obtained by multiplying the base-period payment by $c_t$. Equivalently, factoring out $c_t$ gives $c_t(X-d/c_t)_+$: the effective deductible measured in base-period loss units falls as $c_t$ rises. That is why the LER can change under inflation when the deductible is fixed.

A Reliable Calculation Sequence

For a problem comparing multiple periods or deductible designs, use the same sequence for each period. In particular, do not apply a base-period deductible to trended losses unless the contract specifies that the deductible is fixed in dollars.

Set the trend factor.
Calculate the cumulative factor from the stated period-by-period inflation rates. Apply it to each ground-up loss outcome.

Set the period's deductible.
Use the unchanged dollar amount for fixed coverage, or multiply the base deductible by the cumulative factor for indexed coverage.

Calculate insurer payments and their expectation.
For every possible loss, calculate the positive part of loss minus deductible, then weight those payments by their probabilities.

Calculate the LER and check the result.
Use the period's expected ground-up loss as the denominator. Confirm that expected payment plus expected eliminated loss equals expected ground-up loss.

Worked Calculations

Worked Example: A Discrete Severity and an Ordinary Deductible

A hypothetical insurer models one claim with ground-up loss $X$ equal to $1,000, $3,000, or $7,000 with probabilities 0.25, 0.45, and 0.30, respectively. For a $2,500 ordinary deductible, calculate the expected ground-up loss, expected insurer payment, and LER. State the model and identify a common calculation error.

Method and assumptions. This is a discrete severity model with the three stated outcomes and probabilities, which sum to $0.25+0.45+0.30=1.00$. There is one ordinary deductible, no limit, and no other coverage modification. The payment for each loss is $(X-2500)_+$.

Expected ground-up loss.

E[X]=(1000)(0.25)+(3000)(0.45)+(7000)(0.30) =250+1350+2100=\$3{,}700.

Payments and expected payment. The three insurer payments are $(1000-2500)_+=0$, $(3000-2500)_+=500$, and $(7000-2500)_+=4500$. Therefore,

E[Y^L]=(0)(0.25)+(500)(0.45)+(4500)(0.30) =0+225+1350=\$1{,}575.

LER and check. The expected eliminated loss is $3700-1575=\$2{,}125$, so

\operatorname{LER} =\frac{2125}{3700} =0.574324\ldots \approx 0.5743.

The calculation checks because the expected payment and expected eliminated loss sum to $1575+2125=\$3{,}700$. The common error is to subtract the deductible from every loss, including the $1,000 loss, and allow a negative payment. An ordinary deductible floors the payment at zero; the correct payment is $(X-d)_+$.

Worked Example: Comparing Fixed and Indexed Deductibles Over Two Years

A hypothetical portfolio has base-period claim severity $X$ equal to $1,000, $3,000, or $6,000, with probabilities 0.20, 0.50, and 0.30. Losses trend by 10% per year, compounded. Compare expected insurer payments and LERs under a fixed $2,000 deductible and a deductible indexed from $2,000 at the same rate. Calculate results for the base period and the next two years.

Method and assumptions. The severity probabilities remain unchanged, and every loss is multiplied by the same deterministic cumulative trend factor. The indexed deductible is multiplied by that factor too. There are no limits or other modifications.

First calculate the base expected loss:

E[X]=(1000)(0.20)+(3000)(0.50)+(6000)(0.30) =200+1500+1800=\$3{,}500.

The cumulative factors are $c_0=1$, $c_1=1.10$, and $c_2=(1.10)^2=1.21$. Thus expected ground-up losses are $3,500, $3,850, and $4,235. In the base period, payments under either design are $0, $1,000, and $4,000, giving expected payment $0+500+1200=\$1{,}700$ and LER $(3500-1700)/3500=0.5143$.

In year 1, the trended losses are $1,100, $3,300, and $6,600. Under the fixed $2,000 deductible, payments are $0, $1,300, and $4,600. The expected payment is

(0)(0.20)+(1300)(0.50)+(4600)(0.30)=650+1380=\$2{,}030.

The fixed-deductible LER is $(3850-2030)/3850=0.4727$. The year-1 indexed deductible is $2000(1.10)=\$2{,}200$, so payments are $0, $1,100, and $4,400. Its expected payment is $0+550+1320=\$1{,}870$, and its LER is $(3850-1870)/3850=0.5143$.

In year 2, the trended losses are $1,210, $3,630, and $7,260. With the fixed deductible, payments are $0, $1,630, and $5,260. The expected payment is

(0)(0.20)+(1630)(0.50)+(5260)(0.30)=815+1578=\$2{,}393.

The fixed-deductible LER is $(4235-2393)/4235=0.4349$. The year-2 indexed deductible is $2000(1.21)=\$2{,}420$, producing payments of $0, $1,210, and $4,840. Its expected payment is $0+605+1452=\$2{,}057$; its LER is $(4235-2057)/4235=0.5143$.

Period	Expected ground-up loss	Fixed payment	Fixed LER	Indexed payment	Indexed LER
Base	$3,500	$1,700	0.5143	$1,700	0.5143
Year 1	$3,850	$2,030	0.4727	$1,870	0.5143
Year 2	$4,235	$2,393	0.4349	$2,057	0.5143

The indexed results agree with the proportional-scaling derivation: both loss and deductible rise by the same factor, preserving the LER. The fixed deductible does not keep pace with losses, so its LER falls in this example. A common trap is to use $2,000 as the year-2 indexed deductible; the correct amount is $2,420 because the factor 1.21 is cumulative, not merely the 10% rate for one year.

Worked Example: Fixed Versus Indexed Coverage With Uniform Losses

Suppose a hypothetical base-period ground-up loss $X$ is uniformly distributed from $0 to $8,000. Losses trend by a factor of 1.25. Compare expected insurer payments and LERs for a fixed $2,000 deductible and a deductible indexed from $2,000. Give dollar amounts to the cent and LERs to four decimal places.

Method and assumptions. Under the continuous uniform model, the trended loss $X_t=1.25X$ is uniform from $0 to $10,000, with expected value $5,000. For $Z$ uniform on $[0,b]$, an ordinary deductible $d$ between 0 and $b$ has expected payment

E[(Z-d)_+]=\frac{1}{b}\int_d^b(z-d)\,dz =\frac{(b-d)^2}{2b}.

This follows by integrating the payment $z-d$ over the range where it is positive, $d<z\leq b$; losses below $d$ contribute zero.

Fixed deductible. With $b=\$10{,}000$ and $d=\$2{,}000$,

E[Y^L]=\frac{(10000-2000)^2}{2(10000)} =\frac{64{,}000{,}000}{20{,}000} =\$3{,}200.

The LER is $(5000-3200)/5000=0.3600$. The expected eliminated loss is $1,800, and $3200+1800=5000$, as required.

Indexed deductible. The indexed amount is $1.25(2000)=\$2{,}500$. Therefore,

E[Y^L]=\frac{(10000-2500)^2}{2(10000)} =\frac{56{,}250{,}000}{20{,}000} =\$2{,}812.50.

The LER is $(5000-2812.50)/5000=0.4375$. For a check, the base-period expected payment is $(8000-2000)^2/[2(8000)]=\$2{,}250$; multiplying it by 1.25 gives $2,812.50. The indexed LER is therefore the base ratio $(4000-2250)/4000=0.4375$. The fixed deductible does not scale with the loss, so that invariance check does not apply to the fixed design.

Common Traps and Interpretation

A sound comparison keeps the payment calculation and the LER denominator tied to the same period. The denominator is expected ground-up loss, not expected insurer payment, the base-period expected loss, or the deductible amount. The ratio of expected eliminated loss to expected ground-up loss is not the average of individual claim-level ratios; claims and their probability weights matter.

Trend applied in the wrong place. Trend the ground-up outcomes before applying a fixed deductible. For indexed coverage, trend both losses and the deductible.
Annual rate mistaken for cumulative trend. With repeated annual inflation, multiply the annual factors. Two years at 10% gives $1.10^2=1.21$, not 1.10.
Wrong LER denominator. Calculate the expected loss for that same period and use it as the denominator.
Simple averaging across periods. A combined multi-period LER is weighted by expected ground-up losses, not generally the arithmetic mean of the period LERs. The combined ratio is total expected eliminated loss divided by total expected ground-up loss.
Assuming indexed coverage always preserves LER. Preservation requires proportional trend in every loss and the deductible, with no other changes to the model or coverage. A change in severity shape, probabilities, or contract terms can change the ratio.

For example, if two periods have expected ground-up losses of $2,000 and $8,000 and LERs of 0.40 and 0.60, the combined LER is $[2000(0.40)+8000(0.60)]/(2000+8000)=0.56$, not the unweighted average 0.50. This calculation reflects how much expected loss each period contributes.

Key takeaway: For each period, trend the losses, apply the deductible specified for that period, calculate expected insurer payment, and divide expected eliminated loss by expected ground-up loss. Proportional indexing preserves the LER only when both loss and deductible scale by the same factor.

Check Your Understanding

Use the ordinary-deductible payment formula and calculate each period's LER from its expected ground-up loss.

A hypothetical claim loss is $500, $1,500, or $3,500 with probabilities 0.20, 0.50, and 0.30. With a $1,000 deductible, calculate expected ground-up loss, expected insurer payment, and LER. (Answer: expected loss $1,900; expected payment $1,000; LER 0.4737.)
Base losses are $1,000 and $4,000 with probabilities 0.40 and 0.60. After one year at 10% trend, a fixed $2,500 deductible applies. Calculate expected ground-up loss, expected payment, and LER. (Answer: expected loss $3,080; expected payment $1,140; LER 0.6299.)
Use the same base severity and one-year trend as question 2, but index the $2,500 base deductible by 10%. Calculate the indexed deductible, expected payment, and LER. (Answer: indexed deductible $2,750; expected payment $990; LER 0.6786.)
A base-period loss is uniform on $0 to $8,000 and trends by a factor of 1.25. A deductible is indexed from $2,000. Calculate expected ground-up loss, expected payment, and LER. (Answer: expected loss $5,000; expected payment $2,812.50; LER 0.4375.)
Two periods have expected ground-up losses of $2,000 and $8,000, and their respective LERs are 0.40 and 0.60. Calculate the combined multi-period LER and compare it with the arithmetic mean of the two period LERs. (Answer: combined LER 0.5600; arithmetic mean 0.5000.)

What You'll Learn