Solving Backward for Converted Losses

Why Work Backward From a Retrospective Premium?

A retrospective premium is calculated from a fixed premium component and a loss-responsive component, then adjusted by a tax multiplier and subject to contractual minimum and maximum premiums. When a charged premium is known, it may be possible to work backward and recover the converted losses. The key question is whether the observed premium came from the unbounded formula or was determined by a binding limit.

This distinction matters because an interior premium determines one converted-loss amount, while a binding minimum or maximum generally corresponds to a range of amounts. The minimum observation has one important exception: under the nonnegativity condition on converted losses, that range can be a singleton.

Definition: Let $B$ be the fixed basic premium component, $C$ the converted losses, and $T$ the positive tax multiplier. The unbounded retrospective premium is $P^*=T(B+C)$. If the contractual minimum and maximum premiums are $P_{\min}$ and $P_{\max}$, respectively, the charged premium is $$ P=\min\bigl(P_{\max},\max(P_{\min},P^*)\bigr). $$ Assume $B\geq 0$, $C\geq 0$, $T>0$, and $P_{\min}\leq P_{\max}$. Here, $C$ is already converted; it is not necessarily the raw reported loss amount.

For the calculations below, no interest rate, mortality assumption, or probability distribution is involved. Each example is a deterministic application of the stated contract values. The tax multiplier applies to the entire sum $B+C$, not just to the converted losses.

Inverting an Interior Premium

An observed premium is interior when it is strictly above the minimum and strictly below the maximum. In that case neither bound changes the unbounded premium: $P=P^*$. Since $T>0$, the equation can be rearranged directly to recover one value of $C$.

P=T(B+C) \quad\Longrightarrow\quad \frac{P}{T}=B+C \quad\Longrightarrow\quad C=\frac{P}{T}-B.

The result is valid only if it satisfies $C\geq 0$ and the stated premium is genuinely interior. If the result is negative, the data and assumptions are inconsistent with a nonnegative converted loss amount; it is not appropriate to replace the result with zero and call that an inverse solution.

Worked Example: Recovering Converted Losses From an Interior Premium

A hypothetical account has $B=\$120{,}000$, $T=1.08$, $P_{\min}=\$110{,}000$, and $P_{\max}=\$210{,}000$. Its charged premium is $\$162{,}000$. Find the converted losses. No interest, mortality, or loss distribution is assumed.

Step 1: Check whether the premium is interior. Since $110{,}000<162{,}000<210{,}000$, neither bound is binding, so $P^*=\$162{,}000$.

Step 2: Remove the tax multiplier. $\$162{,}000/1.08=\$150{,}000$. Thus $B+C=\$150{,}000$.

Step 3: Subtract the fixed component. $C=\$150{,}000-\$120{,}000=\$30{,}000$.

Check. The unbounded premium is $1.08(\$120{,}000+\$30{,}000)=\$162{,}000$, which is between the bounds. Therefore the inferred converted losses are $\$30{,}000$.

What a Binding Minimum Tells You

Suppose $P_{\min}<P_{\max}$, and the observed charged premium equals $P_{\min}$. Under the contract formula, that observation means the unbounded premium is at or below the minimum. It does not generally mean that the unbounded premium equals the minimum. Solving the inequality gives the compatible converted-loss values:

T(B+C)\leq P_{\min} \quad\Longleftrightarrow\quad C\leq \frac{P_{\min}}{T}-B.

Combining this with $C\geq 0$, the compatible range is $0\leq C\leq P_{\min}/T-B$, provided the upper endpoint is nonnegative. If that endpoint is negative, no nonnegative $C$ can produce the observed minimum under the stated assumptions. If it is positive, there are multiple compatible values. If it is exactly zero, the range consists only of $C=0$, so the minimum observation does identify a unique converted-loss amount.

Worked Example: A Minimum Premium With Multiple Compatible Loss Amounts

A hypothetical contract has $B=\$80{,}000$, $T=1.10$, $P_{\min}=\$99{,}000$, and $P_{\max}=\$160{,}000$. The charged premium is $\$99{,}000$. Find all compatible converted losses. No interest, mortality, or loss distribution is assumed.

Step 1: Interpret the observation. The bounds are distinct, and the charged premium equals the minimum. Therefore $P^*\leq \$99{,}000$.

Step 2: Solve the inequality. Divide by $1.10$: $\$99{,}000/1.10=\$90{,}000$. Hence $\$80{,}000+C\leq \$90{,}000$, or $C\leq \$10{,}000$.

Step 3: Apply nonnegativity. The compatible range is $0\leq C\leq \$10{,}000$, not one uniquely determined amount.

Check. At $C=0$, the unbounded premium is $1.10(\$80{,}000)=\$88{,}000$, so the minimum is charged. At $C=\$10{,}000$, the unbounded premium is exactly $\$99{,}000$, also giving the minimum. Every value between these endpoints produces an unbounded premium no greater than $\$99{,}000$. The observed premium alone cannot distinguish them.

Worked Example: When a Minimum Observation Identifies One Value

A hypothetical plan has $B=\$20{,}000$, $T=1.10$, $P_{\min}=\$22{,}000$, and $P_{\max}=\$30{,}000$. The charged premium is $\$22{,}000$. Determine the compatible converted losses. No interest, mortality, or loss distribution is assumed.

Step 1: Use the minimum condition. Because $P_{\min}<P_{\max}$, the observation implies $1.10(\$20{,}000+C)\leq \$22{,}000$.

Step 2: Solve and impose nonnegativity. Dividing by $1.10$ gives $\$20{,}000+C\leq \$20{,}000$, so $C\leq 0$. Together with $C\geq 0$, this gives the singleton range $C=0$.

Check. With $C=0$, the unbounded premium is $1.10(\$20{,}000)=\$22{,}000$, exactly the minimum. No other nonnegative converted-loss value can satisfy the inequality. Thus the minimum observation identifies $C$ uniquely in this special case.

What a Binding Maximum Tells You

When $P_{\min}<P_{\max}$ and the charged premium equals $P_{\max}$, the unbounded premium must be at least the maximum. Rearranging this inequality gives a lower bound on converted losses rather than a single value:

T(B+C)\geq P_{\max} \quad\Longleftrightarrow\quad C\geq \frac{P_{\max}}{T}-B.

After incorporating $C\geq 0$, the compatible range is $C\geq\max(0,P_{\max}/T-B)$. It extends without an upper limit under this formula. If $P_{\max}/T-B<0$, every nonnegative $C$ is compatible with the maximum observation; if the expression is nonnegative, it is the smallest compatible value.

Worked Example: A Maximum Premium Gives a Lower Bound

A hypothetical account has $B=\$50{,}000$, $T=1.05$, $P_{\min}=\$70{,}000$, and $P_{\max}=\$115{,}500$. Its charged premium is $\$115{,}500$. Find the compatible converted losses. No interest, mortality, or loss distribution is assumed.

Step 1: Translate the observation. The bounds are distinct and the maximum is charged, so $P^*\geq \$115{,}500$.

Step 2: Solve for the threshold. Dividing by $1.05$ gives $\$115{,}500/1.05=\$110{,}000$. Therefore $\$50{,}000+C\geq \$110{,}000$, which gives $C\geq \$60{,}000$.

Step 3: State the full range. Every $C\geq \$60{,}000$ is compatible; the premium observation supplies no upper bound.

Check. At $C=\$60{,}000$, the unbounded premium is $1.05(\$110{,}000)=\$115{,}500$. At $C=\$75{,}000$, it is $1.05(\$125{,}000)=\$131{,}250$, which is reduced to the charged maximum of $\$115{,}500$. Both loss amounts produce the observed premium.

Equal Bounds and the Inverse-Calculation Trap

The implications for binding minimums and maximums require distinct bounds. If $P_{\min}=P_{\max}$, the charged premium is always their common value, regardless of $P^*$. The charged premium then gives no information about $C$ beyond constraints imposed separately by the contract or other data.

Worked Example: Equal Bounds Provide No Loss Information

A hypothetical plan has $B=\$60{,}000$, $T=1.10$, and equal bounds $P_{\min}=P_{\max}=\$90{,}000$. The observed premium is $\$90{,}000$. Determine what can be inferred about $C$. No interest, mortality, or loss distribution is assumed.

Step 1: Apply the contractual bounds. When both bounds equal $\$90{,}000$, the formula returns $\$90{,}000$ for every unbounded premium.

Step 2: Check distinct possible loss amounts. For $C=0$, $P^*=1.10(\$60{,}000)=\$66{,}000$, and the charged premium is $\$90{,}000$. For $C=\$30{,}000$, $P^*=1.10(\$90{,}000)=\$99{,}000$, and the charged premium is still $\$90{,}000$.

Conclusion and check. Both different converted-loss amounts, and indeed every $C\geq0$, produce the same charged premium because the bounds coincide. The observation provides no information about $C$.

A frequent exam trap is to treat any observed bound as though it were an unbounded premium. For example, from the $\$99{,}000$ minimum observation in the earlier example, the incorrect calculation $C=\$99{,}000/1.10-\$80{,}000=\$10{,}000$ reports the upper endpoint as if it were the only answer. The correction is to retain the inequality $P^*\leq P_{\min}$, giving the full range $0\leq C\leq\$10{,}000$. Also avoid reversing the tax adjustment: when an interior premium is given, divide the premium by $T$ first, then subtract $B$.

Key takeaway: An interior charged premium gives $C=P/T-B$, provided the result is nonnegative. With distinct bounds, a minimum observation gives an upper range for $C$, which is a singleton only when its upper endpoint is zero; a maximum observation gives a lower range. Equal bounds make the charged premium independent of $C$.

Check Your Understanding

For each problem, use the stated contract formula and nonnegativity condition. Treat each scenario as hypothetical and deterministic.

A plan has $B=\$72{,}000$, $T=1.05$, $P_{\min}=\$75{,}000$, and $P_{\max}=\$130{,}000$. The charged premium is $\$105{,}000$. Find $C$. (Answer: $\$28{,}000$)
A contract has $B=\$40{,}000$, $T=1.10$, $P_{\min}=\$49{,}500$, and $P_{\max}=\$90{,}000$. If the minimum is charged, give the full compatible range for $C$. (Answer: $0\leq C\leq \$5{,}000$)
A plan has $B=\$25{,}000$, $T=1.20$, $P_{\min}=\$30{,}000$, and $P_{\max}=\$50{,}000$. The minimum is charged. Is $C$ uniquely determined? (Answer: Yes; $C=\$0$)
A contract has $B=\$90{,}000$, $T=1.04$, $P_{\min}=\$100{,}000$, and $P_{\max}=\$124{,}800$. If the maximum is charged, give the compatible range for $C$. (Answer: $C\geq \$30{,}000$)
A plan has $B=\$50{,}000$, $T=1.10$, and $P_{\min}=P_{\max}=\$88{,}000$. The charged premium is $\$88{,}000$. What does this observation imply about nonnegative $C$? (Answer: It imposes no further restriction; every $C\geq0$ is compatible.)

What You'll Learn