Calculating Expected Unpaid Loss From Paid-To-Date Data

From an A Priori Estimate to Expected Unpaid Loss

An a priori loss assumption provides an expected ultimate loss estimate before the period’s full claim experience is known. As claims are paid, the unpaid portion changes. To estimate that remaining amount, first calculate expected ultimate losses on the appropriate premium or exposure basis, then account for cumulative paid losses as of the valuation date.

The arithmetic is simple, but the inputs must describe the same business. A mismatch in accident period, premium basis, coverage, or loss and adjustment expense (LAE) scope can make a correct subtraction produce an unsuitable estimate. The method also estimates total unpaid losses, not just amounts associated with claims that have not been reported.

Definition: Let $\widehat U_{\mathrm{exp}}$ denote an a priori estimate of ultimate losses for a defined group of policies or claims, and let $L_{\mathrm{paid}}$ be cumulative losses paid for that same group through a stated valuation date. The estimated total unpaid losses are $\widehat R_{\mathrm{unpaid}}=\widehat U_{\mathrm{exp}}-L_{\mathrm{paid}}$. The result represents the estimated ultimate amount remaining to be paid, including any amounts associated with claims that have already been reported but not fully paid.

There are two common ways to construct $\widehat U_{\mathrm{exp}}$. With an expected loss ratio, multiply the premium on the selected basis by the expected loss ratio. With an expected loss cost, multiply the matching exposure count by the expected loss cost per exposure. The expected loss ratio or loss cost is the a priori assumption; it should be selected for the relevant coverage and period, not inferred from the paid amount being subtracted.

\widehat U_{\mathrm{exp}}= \begin{cases} \text{premium}\times\text{expected loss ratio},\\ \text{exposures}\times\text{expected loss cost per exposure}, \end{cases} \qquad \widehat R_{\mathrm{unpaid}}=\widehat U_{\mathrm{exp}}-L_{\mathrm{paid}}.

Why Subtracting Paid Losses Gives Unpaid Loss

For a fixed group and valuation date, ultimate loss can be divided into the amount already paid and the amount still unpaid. This is an accounting identity, not a claim-development pattern: no assumption about how quickly claims are reported or settled is needed to state it.

Partition ultimate loss.
For the defined group, write ultimate loss $U$ as paid-to-date loss $L_{\mathrm{paid}}$ plus the remaining unpaid loss $R_{\mathrm{unpaid}}$: $U=L_{\mathrm{paid}}+R_{\mathrm{unpaid}}$.

Rearrange the identity.
Subtract the known paid amount from both sides: $R_{\mathrm{unpaid}}=U-L_{\mathrm{paid}}$.

Use the selected expected ultimate.
When ultimate loss is represented by an a priori estimate, substitute $\widehat U_{\mathrm{exp}}$ for $U$. The resulting expected-loss estimate is $\widehat R_{\mathrm{unpaid}}=\widehat U_{\mathrm{exp}}-L_{\mathrm{paid}}$.

In probabilistic terms, if cumulative paid loss is known at the valuation date, the conditional expectation identity is $E[U-L_{\mathrm{paid}}\mid\text{available information}]=E[U\mid\text{available information}]-L_{\mathrm{paid}}$. In practice, an a priori expected ultimate may be a selected estimate rather than a fully updated conditional expectation. The subtraction still defines the expected-loss estimate being calculated; it does not guarantee that the actual remaining payments will equal that estimate.

Conditions for a meaningful estimate: The expected ultimate and paid amount must refer to the same accident or policy period, coverage, segment, gross-or-net basis, and loss-and-LAE scope. Paid losses must be cumulative through the stated valuation date. The method assumes no additional discounting in the subtraction: both amounts are compared on the same nominal dollar basis. If the selected ultimate is for losses excluding LAE, for example, paid losses including LAE are not a consistent deduction.

Worked Calculations

The examples below use hypothetical insurance portfolios. No interest, mortality model, or probability distribution is assumed; the calculations are nominal-dollar estimates using the stated a priori assumptions. Amounts are rounded to the nearest dollar only at the final step.

Worked Example: Premium-Based Estimate for a Specialty Coverage

A hypothetical specialty insurer has $2,400,000 of earned premium on the selected basis for one accident year. Its a priori expected loss ratio is 62.5%. Cumulative paid losses for that same accident year and coverage are $420,000 at the valuation date. Estimate expected ultimate losses and total unpaid losses.

Method and assumptions. Use the premium-based expected loss method. The selected premium and loss ratio apply to the same coverage, accident year, and basis as the paid amount; all three values are in dollars. No discounting or separate distributional assumption is used.

Step 1: Calculate expected ultimate losses. Convert 62.5% to 0.625, then multiply:

\widehat U_{\mathrm{exp}}=\$2{,}400{,}000(0.625)=\$1{,}500{,}000.

Step 2: Subtract cumulative paid losses.

\widehat R_{\mathrm{unpaid}}=\$1{,}500{,}000-\$420{,}000=\$1{,}080{,}000.

Check. Expected ultimate equals paid plus estimated unpaid: $\$420{,}000+\$1{,}080{,}000=\$1{,}500{,}000$. The method avoids treating the paid amount as a new estimate of ultimate; it is deducted from the a priori ultimate.

Worked Example: Exposure-Based Estimate for Equipment Policies

A hypothetical insurer writes 18,000 equipment-policy exposure units for a defined period. The selected expected loss cost is $84 per exposure unit. Losses paid through the valuation date total $610,000 for those units. Find expected ultimate losses and estimated total unpaid losses.

Method and assumptions. Use the exposure-based expected loss cost method. Each of the 18,000 units is measured on the basis to which the $84 assumption applies, and paid losses have the same period and loss scope. No discounting or separate distributional assumption is used.

Step 1: Calculate expected ultimate losses.

\widehat U_{\mathrm{exp}}=18{,}000\times \$84=\$1{,}512{,}000.

Step 2: Calculate unpaid losses.

\widehat R_{\mathrm{unpaid}}=\$1{,}512{,}000-\$610{,}000=\$902{,}000.

Check. Paid plus unpaid is $\$610{,}000+\$902{,}000=\$1{,}512{,}000$, which matches the expected ultimate. The key unit check is that exposure units multiplied by dollars per exposure unit produce dollars.

Worked Example: Combining Two Segments

A hypothetical program has two segments with different a priori loss ratios. Segment A has $1,200,000 of premium, an expected loss ratio of 55%, and $400,000 paid. Segment B has $800,000 of premium, an expected loss ratio of 75%, and $500,000 paid. Estimate total expected ultimate and total unpaid losses. Also calculate the unpaid estimate for each segment.

Method and assumptions. Apply the premium-based method separately by segment, then aggregate. Each segment’s premium, loss ratio, and paid losses refer to the same period and loss scope. No discounting or separate distributional assumption is used.

Step 1: Calculate Segment A’s expected ultimate and unpaid estimate.

\widehat U_A=\$1{,}200{,}000(0.55)=\$660{,}000, \qquad \widehat R_A=\$660{,}000-\$400{,}000=\$260{,}000.

Step 2: Calculate Segment B’s expected ultimate and unpaid estimate.

\widehat U_B=\$800{,}000(0.75)=\$600{,}000, \qquad \widehat R_B=\$600{,}000-\$500{,}000=\$100{,}000.

Step 3: Aggregate the results.

\widehat U_{\mathrm{total}}=\$660{,}000+\$600{,}000=\$1{,}260{,}000, $$ $$ L_{\mathrm{paid,total}}=\$400{,}000+\$500{,}000=\$900{,}000, \qquad \widehat R_{\mathrm{total}}=\$260{,}000+\$100{,}000=\$360{,}000.

Check. Direct subtraction gives $\$1{,}260{,}000-\$900{,}000=\$360{,}000$, the same total unpaid estimate. The segment calculation preserves the distinct expected loss ratios rather than applying one unweighted average to both segments.

Interpretation and Common Mistakes

The result is an estimate of all loss dollars still unpaid at the valuation date, not a separate estimate for each claim status. A claim already reported may have an unpaid balance; a claim not yet reported may also contribute to ultimate loss. Subtracting cumulative paid losses leaves both types of unpaid amount within the total.

A common exam trap is to deduct an incurred amount instead of paid losses. Incurred losses generally include paid losses plus case outstanding amounts, so subtracting incurred rather than paid changes the question and understates total unpaid under this calculation. For example, in the first worked example, suppose incurred losses are $590,000 while paid losses are $420,000. The incorrect calculation $\$1{,}500{,}000-\$590{,}000=\$910{,}000$ deducts $170,000 of case outstanding in addition to paid losses. The correct total-unpaid calculation is $\$1{,}500{,}000-\$420{,}000=\$1{,}080{,}000$. The $910,000 figure is not the total unpaid amount under the stated definition.

Do not mix bases. Premium, exposure, expected loss assumptions, and paid losses must cover the same business and period.
Do not subtract a percentage. Convert an expected loss ratio to a dollar expected ultimate first; subtract the paid dollar amount from that dollar estimate.
Do not deduct the same amount twice. If an amount has already been removed from the selected ultimate or paid total, do not subtract it again.
Do not confuse paid with incurred. The required deduction here is cumulative paid loss, not paid plus case outstanding.
Check the reconciliation. Estimated unpaid plus paid should equal the selected expected ultimate, subject only to rounding.

These checks matter because the formula is intentionally direct: it does not adjust the a priori ultimate for how much loss is expected to emerge by the valuation date. Instead, it starts with the selected expected ultimate and subtracts the amount actually paid. A different estimate can result if the underlying a priori assumption, the paid data, or their bases differ.

Key takeaway: First convert the a priori assumption into an expected ultimate dollar amount. Then subtract cumulative paid losses for the same business and valuation date. Verify that paid plus estimated unpaid reconciles to expected ultimate.

Check Your Understanding

For each hypothetical problem, calculate the requested estimate and check that expected ultimate equals paid plus unpaid.

A small-business portfolio has $960,000 of earned premium and an expected loss ratio of 58%. Paid losses for the same period are $310,000. What are expected ultimate and expected unpaid losses? (Answer: Expected ultimate = $556,800; expected unpaid = $246,800.)
A hypothetical marine policy group has 7,500 exposure units and a selected expected loss cost of $126 per unit. Paid losses are $525,000. Calculate expected ultimate and total unpaid losses. (Answer: Expected ultimate = $945,000; expected unpaid = $420,000.)
A program has two segments. Segment X has $500,000 premium, a 64% expected loss ratio, and $210,000 paid. Segment Y has $900,000 premium, a 52% expected loss ratio, and $350,000 paid. Find total expected ultimate and total unpaid losses. (Answer: Total expected ultimate = $788,000; total paid = $560,000; total unpaid = $228,000.)
A hypothetical book has $1,750,000 premium and an expected loss ratio of 48%. At the valuation date, $690,000 of losses have been paid and $140,000 of case outstanding is reported. Under the total-unpaid definition in this tutorial, calculate expected unpaid losses. (Answer: Expected ultimate = $840,000; expected unpaid = $150,000. Deduct paid losses, not paid plus case outstanding.)
A crop-insurance portfolio has 12,400 exposure units and a selected expected loss cost of $95 per unit. Cumulative paid losses are $780,000. Calculate the expected ultimate and total unpaid amount, and state the reconciliation. (Answer: Expected ultimate = $1,178,000; expected unpaid = $398,000; $780,000 + $398,000 = $1,178,000.)

What You'll Learn