Cape Cod Exposure Weights and Development

Why Cape Cod Weights Exposure by Development

Cape Cod estimates a common expected loss ratio from the experience observed across accident periods. But the reported losses for periods at different maturities do not represent the same fraction of their expected ultimate losses. An immature period has had less time for losses to emerge. Its exposure therefore must contribute less to the denominator used to estimate the common loss ratio.

The adjustment is the period’s reported proportion. Multiply each period’s exposure by that proportion, then add the weighted exposures. This denominator represents the exposure on a reported-loss basis. It is not simply total exposure, and it is not a credibility weight chosen independently of development.

Definition: For accident period $i$, let $E_i$ be its exposure, $R_i$ its cumulative reported losses at the valuation date, and $p_i$ the proportion of ultimate losses expected to be reported at its maturity. The Cape Cod exposure weight is $E_i p_i$. The Cape Cod loss ratio is $$ \widehat{\lambda}_{CC} = \frac{\sum_i R_i}{\sum_i E_i p_i}. $$ The estimated ultimate loss for period $i$ is $E_i\widehat{\lambda}_{CC}$, and its estimated reserve on the same reported-loss basis is $E_i\widehat{\lambda}_{CC}-R_i$.

When a selected cumulative development factor from period $i$’s maturity to ultimate is $C_i$, the reported proportion is $p_i=1/C_i$. For example, a cumulative factor of $2$ means expected ultimate losses are twice reported losses, so the reported proportion is $0.5$. The factor and proportion describe the same development assumption in reciprocal forms.

These formulas presume that the exposure amounts and reported losses use a consistent basis, that the selected reported proportions are appropriate for the periods being estimated, and that a common expected ultimate loss ratio is reasonable across those periods. The exposure measure itself is taken as given here. If a period is fully developed, then $p_i=1$, so its full exposure enters the denominator. If $p_i=0$, it contributes no observed reported emergence and no weight.

What the Exposure Weights Mean

Suppose a period has exposure $E_i$ and a common expected ultimate loss ratio $\lambda$. Its expected ultimate losses are $E_i\lambda$. If the expected proportion reported by the valuation date is $p_i$, then its expected reported losses are $E_i\lambda p_i$. Thus, per unit of exposure, a less mature period is expected to produce fewer reported losses at the valuation date.

Across periods, the Cape Cod estimate equates total observed reported losses with their expected value under a common loss ratio. Dividing the observed total by the sum of exposure multiplied by reported proportion solves for that loss ratio. The weights can also be normalized: $w_i=E_i p_i/\sum_j E_jp_j$. These weights sum to one and show each period’s share of the reported-basis exposure denominator.

E\!\left[\sum_i R_i\right] = \sum_i E_i\lambda p_i = \lambda\sum_i E_i p_i \quad\Longrightarrow\quad \widehat{\lambda}_{CC} = \frac{\sum_i R_i}{\sum_i E_i p_i}.

This is a moment-based justification: the estimate sets observed aggregate reported losses equal to expected aggregate reported losses under the assumed common ratio. It does not assert that every period’s actual reported losses equal its expected reported losses. Rather, it uses the aggregate experience to estimate a shared ratio.

Worked Cape Cod Calculations

Worked Example: Different Maturities and Weighted Exposure

A hypothetical portfolio has three accident periods. Exposures are in thousands of dollars. The reported proportions are $0.80$, $0.50$, and $0.25$, and reported losses are $640$, $510$, and $180$, respectively, in thousands. Calculate the Cape Cod loss ratio, each period’s exposure weight, ultimate loss, and reserve.

Method. Weight each exposure by its reported proportion, sum the weighted exposures, and divide total reported losses by that denominator.

Step 1: Calculate the exposure weights. The weighted exposures are $1{,}000(0.80)=800$, $1{,}200(0.50)=600$, and $800(0.25)=200$, all in thousands. The denominator is $800+600+200=1{,}600$ thousand.

Step 2: Estimate the common loss ratio. Total reported losses are $640+510+180=1{,}330$ thousand. Therefore,

\widehat{\lambda}_{CC} = \frac{1{,}330}{1{,}600} = 0.83125.

Step 3: Estimate ultimates and reserves. For the first period, the ultimate is $1{,}000(0.83125)=831.25$, and the reserve is $831.25-640=191.25$. For the second, the ultimate is $1{,}200(0.83125)=997.50$, and the reserve is $997.50-510=487.50$. For the third, the ultimate is $800(0.83125)=665.00$, and the reserve is $665.00-180=485.00$.

All amounts are in thousands of dollars and are reported to two decimal places. The estimated ultimates sum to $831.25+997.50+665.00=2{,}493.75$. The reserves sum to $191.25+487.50+485.00=1{,}163.75$, and $2{,}493.75-1{,}330=1{,}163.75$, as required. The common mistake this calculation avoids is putting all $3{,}000$ of exposure in the denominator without adjusting for maturity.

A period’s weight is determined by both its exposure and its reported proportion. A large exposure can still have a modest denominator contribution if the period is very immature. Conversely, a smaller but nearly fully developed period can contribute substantially. Normalizing the weights is useful for seeing these relative contributions, but it does not change the loss-ratio estimate.

Worked Example: Development Factors and Normalized Weights

A hypothetical book has exposures of $500$, $1{,}000$, and $1{,}500$ thousand dollars. The cumulative development factors to ultimate are $1.25$, $2$, and $4$. Reported losses are $360$, $420$, and $180$ thousand dollars. Find the reported proportions, normalized exposure weights, and Cape Cod loss ratio.

Method. Convert each cumulative development factor to a reported proportion using its reciprocal, then use the resulting exposure weights.

Step 1: Convert factors to reported proportions. The proportions are $1/1.25=0.80$, $1/2=0.50$, and $1/4=0.25$. A factor of $4$ does not mean that $4$ of every $1$ dollars is reported; it means ultimate losses are four times reported losses, so one quarter is reported.

Step 2: Calculate and normalize the exposure weights. The weights are $500(0.80)=400$, $1{,}000(0.50)=500$, and $1{,}500(0.25)=375$ thousand. Their sum is $1{,}275$ thousand. The normalized weights are $400/1{,}275=0.3137$, $500/1{,}275=0.3922$, and $375/1{,}275=0.2941$, to four decimal places. They sum to $1.0000$.

Step 3: Estimate the loss ratio. Total reported losses are $360+420+180=960$ thousand. Thus $\widehat{\lambda}_{CC}=960/1{,}275=0.752941\ldots$, or $0.7529$ to four decimal places. The normalized weights describe the relative contributions to the denominator; they are not the reported proportions themselves.

As a check, the weighted average of the individual ratios $R_i/(E_i p_i)$ gives the same answer: $0.313725(360/400)+0.392157(420/500)+0.294118(180/375)=0.282353+0.329412+0.141176=0.752941$. This confirms how the weights combine each period’s experience.

Development, Reserves, and a Common Trap

The exposure weight $E_i p_i$ is the amount of exposure translated to the reported basis. If two periods have equal exposure but one has a smaller reported proportion, the less mature period contributes less expected reported emergence under the same loss ratio. Its actual reported losses still belong in the numerator; it is the denominator contribution that reflects how much of its expected ultimate has emerged.

Once the common ratio is estimated, the ultimate estimate for a period is $E_i\widehat{\lambda}_{CC}$. The reported proportion is not multiplied into this ultimate a second time. It was used to estimate the ratio from reported experience. The estimated reported amount under the fitted common ratio would be $E_i\widehat{\lambda}_{CC}p_i$, while the estimated ultimate remains $E_i\widehat{\lambda}_{CC}$.

Common exam trap: Do not use $\sum_i E_i$ as the Cape Cod denominator when periods have different maturities. The relevant denominator is $\sum_i E_i p_i$. Using unweighted exposure treats every period as if its full ultimate losses had already been reported.

Worked Example: Diagnosing an Unweighted-Denominator Error

Use the first example’s exposures, reported proportions, and reported losses. A reviewer calculates the loss ratio as $1{,}330/3{,}000=0.4433$, using total exposure without development adjustment. Identify the error and calculate the correct ratio and total reserve.

Wrong calculation. The calculation $1{,}330/3{,}000=0.4433$ treats all $3{,}000$ thousand of exposure as fully reported. But the expected reported-basis exposure is $1{,}000(0.80)+1{,}200(0.50)+800(0.25)=800+600+200=1{,}600$ thousand, not $3{,}000$.

Correction. The correct loss ratio is $1{,}330/1{,}600=0.83125$. Total estimated ultimate losses are total exposure times that ratio: $3{,}000(0.83125)=2{,}493.75$ thousand. Subtracting the $1{,}330$ thousand already reported gives a total reserve of $2{,}493.75-1{,}330=1{,}163.75$ thousand.

The wrong ratio would imply total ultimate losses of $3{,}000(0.4433)=1{,}330$ thousand, approximately equal to the amount already reported, and therefore almost no reserve. That result is not a valid Cape Cod estimate: it arises because the denominator incorrectly assumes full reporting for immature periods. The corrected calculation reconciles to the accident-period reserves in the first example.

Applying the Weighting Logic Reliably

A disciplined calculation keeps the exposure, development, and loss steps separate. First identify the reported proportion appropriate to each period. Next calculate exposure times reported proportion. Only then sum the denominator and divide total reported losses by it. Finally apply the common loss ratio to each period’s exposure to obtain ultimate losses.

Establish the maturity measure.
Use the reported proportion $p_i$ for each period. If the input is a cumulative factor $C_i$ to ultimate, calculate $p_i=1/C_i$.

Weight the exposure.
Calculate $E_i p_i$ period by period, retaining units consistently. Add these quantities to form the Cape Cod denominator.

Estimate and apply the ratio.
Divide total reported losses by the weighted exposure denominator. Multiply the result by each unweighted exposure $E_i$ to estimate that period’s ultimate.

Check the result.
Subtract reported losses from estimated ultimates for reserves. The sum of period reserves must equal total estimated ultimate losses less total reported losses.

The weighting step matters because Cape Cod uses reported experience from periods that have not necessarily reached the same maturity. Treating their exposures as equally mature can bias the estimated common ratio downward when the denominator is too large. Weighting by reported proportion puts the observed losses and exposure denominator on a consistent reported basis, while leaving the ultimate estimate on the ultimate basis.

Check Your Understanding

For each problem, use the reported proportions or derive them from the stated cumulative development factors. Amounts are in thousands of dollars where indicated.

A period has exposure $400$, reported proportion $0.75$, and reported losses $210$. Calculate its Cape Cod loss ratio. (Answer: $0.7000$)
Two periods have exposures $600$ and $900$, reported proportions $0.80$ and $0.40$, and reported losses $300$ and $270$. Find the total weighted exposure, the normalized weights, and the common ratio. (Answer: weighted exposure $840$; normalized weights $0.5714$ and $0.4286$; loss ratio $0.6786$)
Two periods have exposures $800$ and $1{,}000$, cumulative development factors $1.5$ and $3$, and reported losses $400$ and $250$. Calculate the reported proportions and Cape Cod loss ratio. (Answer: reported proportions $0.6667$ and $0.3333$; weighted exposure $866.67$; loss ratio $0.7500$)
Two periods have exposures $500$ and $1{,}000$, reported proportions $0.90$ and $0.50$, and reported losses $315$ and $400$. Find the loss ratio and each period’s reserve. (Answer: loss ratio $0.7526$; reserves $61.32$ and $352.63$)
Two periods have exposures $1{,}000$ and $1{,}500$, cumulative development factors $2$ and $5$, and reported losses $320$ and $180$. Calculate the normalized exposure weights, common ratio, and total reserve. (Answer: weights $0.6250$ and $0.3750$; loss ratio $0.6250$; total reserve $1{,}062.50$)

What You'll Learn