Calculate Interim Adjustments As Loss Estimates Mature

Why Interim Retrospective Premiums Change

A retrospective premium depends on the losses estimated for a policy period and on the plan parameters used to convert those losses into premium. At successive valuation dates, the loss estimate may change as claim information develops. The plan parameters may also be revised, for example when an updated calculation uses a different basic premium or loss conversion factor. To explain an interim adjustment, an actuary must identify which inputs changed and apply the plan consistently at each date.

An interim calculation is a valuation-date calculation, not necessarily a statement of the amount that must immediately be billed or paid. Any billing or settlement action depends on the applicable plan provisions. The calculations below concern the indicated premium under stated hypothetical plan terms.

Definition: For valuation date $t$, let $\hat L_t$ be the estimated ultimate loss amount used by the plan after any required loss limitations. Let $B_t$ be the basic premium, $c_t$ the loss conversion factor, and $T_t$ the tax multiplier. Before applying any premium floor or cap, the indicated interim premium is $$ R_t=(B_t+c_t\hat L_t)T_t. $$ If the plan has minimum premium $R_{\min}$ and maximum premium $R_{\max}$, the premium after those provisions is $$ P_t=\min\!\left(R_{\max},\max(R_{\min},R_t)\right). $$ All amounts and parameters must be defined on a consistent basis at each valuation date. In particular, $\hat L_t$ must reflect the loss basis specified by the plan; it should not be substituted with a different gross or limited-loss measure.

This formula is a stated plan form, not a universal formula for every retrospective rating plan. A contract may define expenses, taxes, loss limitations, or adjustments differently. Use the plan’s actual provisions; the calculations here assume the displayed form and apply only when its inputs and provisions match the hypothetical plan.

Recalculation and Attribution

First calculate the premium at each valuation date from the loss estimate and parameters effective for that calculation. Then compare the resulting premiums. When parameters are fixed and neither the minimum nor maximum premium applies, the loss-estimate effect has a particularly direct form.

R_1-R_0 =(B+c\hat L_1)T-(B+c\hat L_0)T =cT(\hat L_1-\hat L_0).

The basic premium cancels because it is unchanged. The conversion factor and tax multiplier scale the change in estimated losses. This relationship does not give the change in the premium after a floor or cap binds; in that case, calculate the premium at both dates using the floor or cap.

If both the loss estimate and plan parameters change, a useful exact attribution is to change the estimate first while holding the old parameters fixed, then change the parameters while holding the new estimate fixed. For any premium function $P(\hat L,\theta)$, where $\theta$ denotes the plan parameters,

P(\hat L_1,\theta_1)-P(\hat L_0,\theta_0) = \bigl[P(\hat L_1,\theta_0)-P(\hat L_0,\theta_0)\bigr] + \bigl[P(\hat L_1,\theta_1)-P(\hat L_1,\theta_0)\bigr].

The first bracket is the loss-estimate effect under the old parameters. The second is the parameter effect at the updated estimate. The sum is the total premium change because the intermediate value $P(\hat L_1,\theta_0)$ is added once and subtracted once. This is an exact accounting identity, but the separate attributions can depend on the order chosen when parameter effects interact with loss estimates or with a floor or cap. State the order used.

Identify each valuation basis.
Record the applicable loss estimate and all plan parameters for each date. Confirm that loss estimates use the same plan-defined basis before comparing them.

Calculate the indicated premium.
Apply the basic premium, loss conversion factor, and tax multiplier as specified. Then apply any minimum or maximum premium provision.

Attribute the difference.
For a loss-only change, hold parameters fixed. If parameters also change, use a stated sequence and verify that the attributed effects sum to the total change.

Worked Calculations

Worked Example: Loss Estimates Mature Under Fixed Parameters

A hypothetical plan uses a basic premium of $18,000, a loss conversion factor of 1.12, and a tax multiplier of 1.08. There is no minimum or maximum premium. At the first valuation date, the plan-defined estimated ultimate losses are $60,000; at the next date, they are $72,000. Calculate both indicated premiums and isolate the effect of the estimate change.

Method. Hold all plan parameters fixed and apply $R=(B+c\hat L)T$ at each date. This avoids treating the loss change as a dollar-for-dollar premium change.

At the first valuation date, converted losses are $1.12\times \$60{,}000=\$67{,}200$. The subtotal is $\$18{,}000+\$67{,}200=\$85{,}200$, so the indicated premium is $\$85{,}200\times1.08=\$92{,}016.00$.

At the second valuation date, converted losses are $1.12\times \$72{,}000=\$80{,}640$. The subtotal is $\$18{,}000+\$80{,}640=\$98{,}640$, so the indicated premium is $\$98{,}640\times1.08=\$106{,}531.20$.

The premium increase is $\$106{,}531.20-\$92{,}016.00=\$14{,}515.20$. As a check, the fixed-parameter formula gives $1.12\times1.08\times(\$72{,}000-\$60{,}000)=\$1.2096\times\$12{,}000=\$14{,}515.20$. The two calculations agree.

Worked Example: Separate Estimate and Parameter Changes

A hypothetical plan initially has $B_0=\$18{,}000$, $c_0=1.12$, and $T_0=1.08$, with estimated losses $\hat L_0=\$60{,}000$. At the next valuation date, the estimate is $\hat L_1=\$72{,}000$, and the parameters are revised to $B_1=\$19{,}000$, $c_1=1.10$, and $T_1=1.07$. Assume no minimum or maximum premium applies. Calculate the total change and attribute it by changing the estimate first.

Initial premium. Converted losses are $1.12\times \$60{,}000=\$67{,}200$. Thus, $P(\hat L_0,\theta_0)=(\$18{,}000+\$67{,}200)\times1.08=\$92{,}016.00$.

Estimate effect at old parameters. At $\hat L_1=\$72{,}000$, but still using $B_0,c_0,T_0$, the premium is $(\$18{,}000+\$80{,}640)\times1.08=\$106{,}531.20$. The estimate effect is $\$106{,}531.20-\$92{,}016.00=+\$14{,}515.20$.

Parameter effect at the new estimate. At the new parameters, converted losses are $1.10\times\$72{,}000=\$79{,}200$. The new premium is $(\$19{,}000+\$79{,}200)\times1.07=\$98{,}200\times1.07=\$105{,}074.00$. The parameter effect is $\$105{,}074.00-\$106{,}531.20=-\$1{,}457.20$.

Check. The attributed effects sum to $\$14{,}515.20-\$1{,}457.20=\$13{,}058.00$. Directly, the new premium less the initial premium is $\$105{,}074.00-\$92{,}016.00=\$13{,}058.00$, as required.

The parameter effect can also be split in a stated sequence at the new loss estimate. Raising the basic premium by $1,000 adds $\$1{,}000\times1.08=\$1{,}080.00$. Reducing the conversion factor from 1.12 to 1.10 changes the premium by $(1.10-1.12)\times\$72{,}000\times1.08=-\$1{,}555.20$. Reducing the tax multiplier from 1.08 to 1.07, after those changes, changes it by $(1.07-1.08)\times\$98{,}200=-\$982.00$. These parameter effects total $-\$1{,}457.20$.

Worked Example: Minimum and Maximum Premiums Clip the Adjustment

A hypothetical plan has a basic premium of $20,000, a loss conversion factor of 1.10, and a tax multiplier of 1.05. Its minimum premium is $75,000 and its maximum premium is $95,000. Estimated losses rise from $40,000 to $55,000, then to $70,000. Parameters are unchanged. Calculate the premium at each valuation date and each successive adjustment.

First date. Before applying the minimum, the indicated premium is $(\$20{,}000+1.10\times\$40{,}000)\times1.05=(\$20{,}000+\$44{,}000)\times1.05=\$67{,}200$. This is below the $75,000 minimum, so the premium is $75,000.

Second date. The indicated premium is $(\$20{,}000+\$60{,}500)\times1.05=\$80{,}500\times1.05=\$84{,}525$. It is between the minimum and maximum, so the premium is $84,525. The adjustment from the first date is $\$84{,}525-\$75{,}000=+\$9{,}525$.

Third date. The indicated premium is $(\$20{,}000+\$77{,}000)\times1.05=\$97{,}000\times1.05=\$101{,}850$. This exceeds the $95,000 maximum, so the premium is $95,000. The adjustment from the second date is $\$95{,}000-\$84{,}525=+\$10{,}475$.

Check. With no floor or cap, each $15,000 increase in estimated losses would change premium by $1.10\times1.05\times\$15{,}000=\$17{,}325$. The actual changes are smaller because the first calculation is lifted by the minimum and the third is limited by the maximum. The three premiums, after applying the plan provisions, are $75,000, $84,525, and $95,000.

Common Mistakes and Practical Checks

A frequent error is to compare loss estimates while silently changing other inputs, then label the entire premium difference a loss-development effect. Another is to calculate the uncapped premium difference and report it as the actual adjustment even though a floor or cap applies. Both errors obscure what the plan calculation says.

Common exam trap: Do not calculate an estimate effect using the new conversion factor or tax multiplier unless the attribution explicitly calls for those parameters. For the estimate-first attribution, apply the old parameters to both estimates; then calculate the parameter effect at the new estimate. Apply any floor or cap at every intermediate premium if those provisions are part of the plan.

Before reporting an interim change, check that the valuation dates use comparable loss bases, the correct parameters are paired with each date, and all plan limits have been applied. If a premium changes, state whether the amount is an uncapped indicated-premium change or the change after plan limits. When attributing several changes, show the intermediate premium and confirm that the components sum to the total.

This ability to distinguish emerging loss information from parameter revisions is aligned with the CAS Exam 8 (Advanced Ratemaking) syllabus. It is also useful in practice: a clear reconciliation makes an interim estimate easier to review, explain, and reproduce.

Check Your Understanding

For each hypothetical plan, calculate the requested premiums using the stated formula and provisions.

A plan has $B=\$12{,}500$, $c=1.08$, and $T=1.04$, with no floor or cap. Estimated losses rise from $48,000 to $53,000. Calculate both premiums and the increase. (Answer: $66,913.60 and $72,529.60; increase $5,616.00.)
A plan starts with $B_0=\$10{,}000$, $c_0=1.15$, $T_0=1.06$, and $\hat L_0=\$40{,}000$. At the next date, $\hat L_1=\$50{,}000$, $B_1=\$10{,}800$, $c_1=1.10$, and $T_1=1.05$. No floor or cap applies. Using estimate-first attribution, calculate the total change and the separate estimate and parameter effects. (Answer: estimate effect +$12,190.00; parameter effect -$2,460.00; total increase $9,730.00.)
A plan has $B=\$8{,}000$, $c=1.20$, $T=1.10$, a $50,000 minimum, and an $80,000 maximum. Estimated losses rise from $20,000 to $40,000, with no parameter changes. What are the premiums and the adjustment? (Answer: $50,000.00 and $61,600.00; adjustment +$11,600.00.)
At a fixed estimated loss of $60,000, a plan changes from $B_0=\$15{,}000$, $c_0=1.10$, $T_0=1.04$ to $B_1=\$16{,}000$, $c_1=1.08$, $T_1=1.03$. No floor or cap applies. Calculate the old and new premiums and identify the parameter-driven change. (Answer: old premium $84,240.00; new premium $83,224.00; parameter-driven change -$1,016.00.)

What You'll Learn