Deriving Uncapped Premium Sensitivity to Losses

Why Derive Premium Sensitivity?

A retrospective premium changes when the losses used in the plan change. A useful question is not only what premium results at a particular loss amount, but also how much premium changes for each additional dollar of loss. In an uncapped plan, that marginal response follows directly from the premium formula. It separates the effect of converting losses from the effect of applying the tax multiplier.

This tutorial considers an unbounded retrospective premium: no minimum or maximum premium is applied. The basic premium, loss conversion factor, and tax multiplier are held fixed while losses vary. Those conditions matter. A premium bound or a parameter that changes with losses can alter the response, so the derivative derived here applies only to the stated unbounded formula with fixed plan parameters.

Definition: Let $B$ be the basic premium, $L$ the aggregate losses used in the retrospective calculation, $k$ the loss conversion factor, and $T$ the tax multiplier. Converted losses are $C=kL$. The unbounded retrospective premium is $$ P(L)=T(B+kL)=T(B+C). $$ Here $B$, $k$, and $T$ are treated as fixed. The formula does not include a minimum or maximum premium.

The loss conversion factor converts the loss amount into the loss component used in the premium calculation. The tax multiplier then multiplies the sum of the basic premium and converted losses. These operations are in a specified order: first calculate $C=kL$, then add $B$, then multiply by $T$. A candidate should not treat the tax multiplier as applying only to losses or treat the loss conversion factor as applying to the basic premium.

Deriving the Marginal Response

There are two related derivatives, depending on which loss quantity is changing. With respect to converted losses $C$, the formula is $P=T(B+C)$. With respect to aggregate losses $L$, use $C=kL$, so that $P=T(B+kL)$. Differentiating each expression gives the response on its corresponding scale.

\frac{dP}{dC}=T \qquad\text{and}\qquad \frac{dP}{dL}=Tk.

The first derivative says that, with converted losses increasing by one dollar, premium increases by $T$ dollars. The second says that, with aggregate losses increasing by one dollar, converted losses increase by $k$ dollars and premium increases by $Tk$ dollars. Both are marginal responses; they are not the premium itself.

Start with the unbounded formula.
Write $P=T(B+kL)$, with $B$, $T$, and $k$ fixed as $L$ varies.

Differentiate with respect to aggregate losses.
The derivative of the constant $B$ is zero, and the derivative of $kL$ is $k$. Therefore $dP/dL=Tk$.

Interpret the result in context.
$Tk$ is the premium change in dollars per dollar of aggregate loss change, provided the plan remains unbounded and its factors stay fixed.

The same result can be justified by considering a finite loss change. If aggregate losses increase by $\Delta L$, then converted losses increase by $k\Delta L$. The premium change is therefore

\Delta P =T\bigl(B+k(L+\Delta L)\bigr)-T(B+kL) =Tk\Delta L.

This identity holds for any finite $\Delta L$ under the stated formula, not just for a very small change. The derivative $Tk$ is constant, so the premium is linear in aggregate losses. This also provides a direct check on a calculation made by computing the initial and revised premiums separately.

Worked Applications

Worked Example: Premium Change From an Aggregate Loss Increase

A hypothetical unbounded retrospective plan has a basic premium of $18,000, a loss conversion factor of 0.80, and a tax multiplier of 1.10. Aggregate losses increase from $42,500 to $48,750. Calculate the marginal response, each premium, and the premium increase.

Method and assumptions. Use $P=T(B+kL)$. The basic premium and both factors remain fixed, and neither premium is subject to a bound.

Step 1: Find the response per dollar of aggregate loss. $$ \frac{dP}{dL}=Tk=(1.10)(0.80)=0.88. $$ Thus each additional dollar of aggregate losses adds $0.88 to premium under these assumptions.

Step 2: Calculate the initial converted losses and premium. $$ C_0=(0.80)(42{,}500)=34{,}000, \qquad P_0=(1.10)(18{,}000+34{,}000) =(1.10)(52{,}000)=57{,}200. $$

Step 3: Calculate the revised converted losses and premium. $$ C_1=(0.80)(48{,}750)=39{,}000, \qquad P_1=(1.10)(18{,}000+39{,}000) =(1.10)(57{,}000)=62{,}700. $$

Step 4: Check the premium change using the marginal response. The aggregate loss increase is $48{,}750-42{,}500=6{,}250$. Therefore, $$ \Delta P=(0.88)(6{,}250)=5{,}500. $$ The separately calculated premiums also give $62{,}700-57{,}200=5{,}500$, so the result checks. This method avoids the common mistake of applying the tax multiplier but omitting the loss conversion factor from the response to aggregate losses.

Worked Example: Starting From a Change in Converted Losses

A hypothetical plan has a basic premium of $24,000, a loss conversion factor of 1.15, and a tax multiplier of 1.06. Aggregate losses rise from $31,200 to $36,000. Calculate the premium response both per dollar of converted losses and per dollar of aggregate losses, then verify the total premium change.

Method and assumptions. Converted losses equal $kL$; the tax multiplier applies to the basic premium plus converted losses. All plan parameters remain fixed, and the formula is unbounded.

Step 1: Calculate both marginal responses. With respect to converted losses, the response is $T=1.06$, or $1.06 of premium per additional dollar of converted losses. With respect to aggregate losses, it is $$ Tk=(1.06)(1.15)=1.219. $$ So each additional dollar of aggregate losses increases premium by $1.219.

Step 2: Compute the initial premium. $$ C_0=(1.15)(31{,}200)=35{,}880, \qquad P_0=(1.06)(24{,}000+35{,}880) =(1.06)(59{,}880)=63{,}472.80. $$

Step 3: Compute the revised premium. $$ C_1=(1.15)(36{,}000)=41{,}400, \qquad P_1=(1.06)(24{,}000+41{,}400) =(1.06)(65{,}400)=69{,}324.00. $$

Step 4: Check the change by either route. Aggregate losses increase by $4,800, so the response formula gives $$ \Delta P=(1.219)(4{,}800)=5{,}851.20. $$ Alternatively, converted losses increase by $41{,}400-35{,}880=5{,}520$, and $(1.06)(5{,}520)=5{,}851.20$. Finally, $69{,}324.00-63{,}472.80=5{,}851.20$. The three results agree. The key trap is using $T$ alone for a change in aggregate losses: $T$ is the response per dollar of converted losses, not per dollar of unconverted losses.

Worked Example: Comparing Two Plans’ Sensitivities

Two hypothetical unbounded plans cover the same account. Plan A has $B_A=$$20,000, $k_A=0.70$, and $T_A=1.12$. Plan B has $B_B=$$17,500, $k_B=0.82$, and $T_B=1.05$. Aggregate losses rise from $50,000 to $59,000. Find each plan’s marginal response and premium increase, and explain why the plan with the lower initial premium need not have the smaller response.

Method and assumptions. For each plan, calculate $Tk$ separately. Hold its basic premium and factors fixed. No premium bound applies.

Step 1: Calculate the marginal responses. $$ T_Ak_A=(1.12)(0.70)=0.784, \qquad T_Bk_B=(1.05)(0.82)=0.861. $$ Plan A adds $0.784 of premium per dollar of additional aggregate loss; Plan B adds $0.861.

Step 2: Calculate Plan A’s initial and revised premiums. $$ C_{A,0}=(0.70)(50{,}000)=35{,}000, \qquad P_{A,0}=(1.12)(20{,}000+35{,}000)=61{,}600. $$ $$ C_{A,1}=(0.70)(59{,}000)=41{,}300, \qquad P_{A,1}=(1.12)(20{,}000+41{,}300)=68{,}656. $$ Thus Plan A’s increase is $68{,}656-61{,}600=$$7,056.

Step 3: Calculate Plan B’s initial and revised premiums. $$ C_{B,0}=(0.82)(50{,}000)=41{,}000, \qquad P_{B,0}=(1.05)(17{,}500+41{,}000)=61{,}425. $$ $$ C_{B,1}=(0.82)(59{,}000)=48{,}380, \qquad P_{B,1}=(1.05)(17{,}500+48{,}380)=69{,}174. $$ Thus Plan B’s increase is $69{,}174-61{,}425=$$7,749.

Step 4: Verify the increases from the responses. The aggregate loss increase is $9,000. Hence Plan A’s change is $(0.784)(9{,}000)=$$7,056, while Plan B’s is $(0.861)(9{,}000)=$$7,749. The checks agree. Plan B starts at a slightly lower premium in this scenario, but its marginal response is greater. Starting premium depends on $B$, $L$, $k$, and $T$; sensitivity is determined by $Tk$. Confusing these two comparisons is a common exam error.

Interpreting the Factors and Avoiding Traps

The conversion factor and tax multiplier act multiplicatively on the marginal response to aggregate losses. A larger $k$, with $T$ fixed, increases the converted-loss increase associated with a given aggregate loss increase. A larger $T$, with $k$ fixed, scales the resulting premium response upward. The basic premium $B$ affects the premium level but not the derivative with respect to losses, because it is constant as losses vary.

It is important to name the loss quantity in a sensitivity statement. The premium response is $T$ per dollar of converted losses and $Tk$ per dollar of aggregate losses. Those are not interchangeable descriptions. They coincide in value when $T=Tk$, which means $T(1-k)=0$: this occurs if $T=0$ or if $k=1$. In a usual plan with a positive tax multiplier, they coincide when $k=1$. Separately, if there is no loss change, then the total premium change is zero regardless of either factor.

The derivative does not imply that the premium is always responsive at that rate in every contract situation. It relies on a fixed basic premium, fixed factors, the stated multiplication order, and an unbounded formula. If a premium floor or ceiling applies, the response can differ when that bound controls the premium. The sensitivity derived here should therefore be used only after confirming that the assumptions match the calculation.

Key takeaway: For the unbounded formula $P=T(B+kL)$, the premium response per dollar of converted losses is $T$, while the response per dollar of aggregate losses is $Tk$. For any aggregate loss change $\Delta L$, the exact premium change is $\Delta P=Tk\Delta L$, provided the plan parameters remain fixed and no premium bound applies.

Check Your Understanding

For each problem, identify whether the change is stated in aggregate or converted losses, then use the corresponding response. All plans are hypothetical, unbounded, and have fixed parameters.

A plan has $k=0.78$ and $T=1.08$. Find the marginal premium response per dollar of aggregate loss and the premium increase for a $5,000 increase in aggregate losses. (Answer: $0.8424 per dollar; $4,212 increase.)
A plan has $k=0.90$, $T=1.04$, and converted losses increase by $7,200. Find the premium increase and the corresponding aggregate loss increase. (Answer: $7,488 premium increase; $8,000 aggregate loss increase.)
A plan has $B=$$12,000, $k=1.10$, and $T=1.05$. Aggregate losses increase from $20,000 to $23,500. Calculate the initial premium, revised premium, and increase, and verify the increase using the marginal response. (Answer: $35,700.00 initial; $39,742.50 revised; $4,042.50 increase.)
Compare two plans with $(k,T)=(0.65,1.20)$ and $(0.80,1.05)$. For a $10,000 increase in aggregate losses, calculate each marginal response and premium increase. (Answer: responses $0.780 and $0.840 per dollar; increases $7,800 and $8,400, respectively.)
A plan has $B=$$14,500, $k=0.72$, and $T=1.09$. Losses rise from $28,000 to $33,000. Calculate each premium and check the increase using the marginal response. (Answer: initial premium $37,779.40; revised premium $41,703.40; increase $3,924.00.)

What You'll Learn