Why the Weighting Matters
A Cape Cod loss ratio uses observed losses together with an exposure measure adjusted for how much of each accident year’s loss experience has emerged. The adjustment matters because the same amount of earned premium can represent very different amounts of observable loss experience at different maturities. A recent accident year may have produced only a small share of its ultimate losses, while an older year may be nearly fully developed.
The calculation in this tutorial selects one loss ratio across multiple accident years. It does not treat each accident year as equally informative, nor does it average the accident-year loss ratios without regard to maturity. Instead, it weights each earned premium by that year’s proportion of losses emerged, then divides total observed losses by the resulting total weighted premium.
Model and Calculation
The method relies on a common expected ultimate loss ratio across the accident years being combined. In addition, the earned premiums must be comparable with one another, and the observed losses and development factors must use a consistent loss basis. For example, paid losses must be paired with appropriate paid development factors; incurred losses must be paired with appropriate incurred development factors. Premiums and losses should also reflect compatible coverage, limits, valuation, and any necessary on-level or trend adjustments.
The calculation is deterministic once the premiums, losses, and development factors have been selected. No interest discounting, mortality assumption, or probability distribution is required. The development factors summarize expected loss emergence; they are not themselves emergence proportions.
Here, \(P_i^{*}\) denotes emergence-weighted premium. If development is supplied directly as a proportion emerged rather than as a factor to ultimate, use that proportion as \(w_i\); do not take its reciprocal. All accident years included in the calculation must be on the same accounting and loss basis. The total loss amount and weighted-premium amount must also use the same monetary units.
Check that the premiums, observed losses, and development factors relate to comparable accident years, coverage, and loss definitions.
For a cumulative factor to ultimate \(f_i\), calculate \(w_i=1/f_i\).
Calculate \(P_i^{*}=P_iw_i\) separately for every accident year, then sum those amounts.
Sum the observed losses, divide by total emergence-weighted premium, and express the result as a percentage if requested.
Why the Formula Holds
The weighting follows from the model’s relationship between ultimate losses and losses observed at the valuation date. Suppose each accident year has the same underlying expected ultimate loss ratio \(\theta\). Under the selected development pattern, a proportion \(w_i\) of that year’s ultimate losses is expected to have emerged. The model therefore relates the expected observed losses to premium as follows:
Summing over the included accident years gives \(E[\sum_i C_i]=\theta\sum_i P_iw_i\). Equating aggregate observed losses to the model’s expected aggregate losses and solving for \(\theta\) gives the Cape Cod estimate:
This is an exposure-weighted estimate: the denominator reflects both the amount of premium and the share of losses expected to have emerged. The equation is a model-based justification, not a claim that the selected ratio is known without uncertainty. If accident years do not reasonably share a common expected loss ratio, the aggregate estimate may conceal meaningful differences among them.
Worked Example: Three Accident Years at Different Maturities
A hypothetical insurer is selecting a Cape Cod expected loss ratio for three accident years. Earned premiums, in thousands of dollars, are 800, 900, and 1,000. The cumulative factors to ultimate are 2.00, 1.25, and 1.00, respectively. Observed losses, also in thousands, are 320, 630, and 850. Calculate the weighted Cape Cod loss ratio. This is a deterministic calculation; no interest, mortality, or probability-distribution assumption is used.
Method. Convert each factor to a proportion emerged, weight the corresponding premium, and divide the total observed losses by the total weighted premium.
For accident year 1, \(w_1=1/2.00=0.50\), so its weighted premium is \(800(0.50)=400\). For accident year 2, \(w_2=1/1.25=0.80\), so its weighted premium is \(900(0.80)=720\). For accident year 3, \(w_3=1/1.00=1.00\), so its weighted premium is \(1{,}000(1.00)=1{,}000\).
The total weighted premium is \(400+720+1{,}000=2{,}120\) thousand dollars. Total observed losses are \(320+630+850=1{,}800\) thousand dollars. Thus,
Check. The individual observed-loss-to-weighted-premium ratios are \(320/400=80\%\), \(630/720=87.5\%\), and \(850/1{,}000=85\%\). The aggregate result, 84.91%, lies between them. The common mistake avoided here is treating the three years as equally mature: their factors imply emergence proportions of 50%, 80%, and 100%.
Worked Example: Converting Factors Before Aggregating
A hypothetical commercial property portfolio has earned premiums of 500, 750, 900, and 1,000 thousand dollars for four accident years. Its cumulative factors to ultimate are 5.00, 2.50, 1.25, and 1.00. The corresponding observed losses are 90, 270, 612, and 950 thousand dollars. Calculate the Cape Cod expected loss ratio to two decimal places. No interest discounting or mortality model is involved.
Method. The factors are cumulative factors to ultimate, so the emerged proportions are their reciprocals: \(0.20\), \(0.40\), \(0.80\), and \(1.00\).
The weighted premiums are \(500(0.20)=100\), \(750(0.40)=300\), \(900(0.80)=720\), and \(1{,}000(1.00)=1{,}000\), all in thousands of dollars. Their total is \(100+300+720+1{,}000=2{,}120\). The observed-loss total is \(90+270+612+950=1{,}922\).
Check and common mistake. The four accident-year ratios against weighted premium are 90%, 90%, 85%, and 95%; 90.66% is plausible as their weighted aggregate. Using the factors themselves as weights would produce a denominator of \(500(5)+750(2.5)+900(1.25)+1{,}000=6{,}500\), and a ratio of \(1{,}922/6{,}500=29.57\%\). That is wrong because a factor to ultimate is not a proportion emerged; the correct conversion is \(w_i=1/f_i\).
Worked Example: Combining a Direct Emergence Proportion With a Factor
A hypothetical liability portfolio has three accident years. Earned premiums are 600, 1,000, and 1,400 thousand dollars, and observed losses are 150, 680, and 1,330 thousand dollars. For the first year, the proportion emerged is supplied directly as 0.25. The second year has a cumulative factor to ultimate of 2.00, and the third has a factor of 1.00. Find the weighted Cape Cod loss ratio, rounding only the final percentage to two decimal places.
Method. Use a supplied emergence proportion directly. Convert only the factors to ultimate by taking their reciprocals. Thus, the three emergence proportions are \(0.25\), \(1/2.00=0.50\), and \(1/1.00=1.00\).
The weighted premiums are \(600(0.25)=150\), \(1{,}000(0.50)=500\), and \(1{,}400(1.00)=1{,}400\) thousand dollars. Their total is \(150+500+1{,}400=2{,}050\). The observed losses total \(150+680+1{,}330=2{,}160\) thousand dollars.
Check. The accident-year ratios using weighted premium are \(150/150=100\%\), \(680/500=136\%\), and \(1{,}330/1{,}400=95\%\). The aggregate is within their range. A common mistake would be to invert the first year’s supplied 0.25, producing a weight of 4.00. That is incorrect: it is already the proportion emerged, so its weighted premium is \(600(0.25)=150\), not \(600/0.25\).
Interpretation, Limits, and Exam Traps
The denominator is not total earned premium and is not total premium multiplied by one arbitrarily selected emergence proportion. Each accident year contributes its own \(P_iw_i\). In particular, a mature year contributes more weighted premium per dollar of earned premium than an immature year. This reflects that a larger share of the mature year’s expected ultimate loss is already represented in its observed losses.
The selected ratio is sensitive to both the premiums and the emergence proportions. A factor that is too large implies too small a proportion emerged and therefore too little weighted premium; all else equal, that raises the indicated ratio. A factor that is too small has the opposite effect. Applying factors inconsistently, such as using paid factors with incurred losses, can make the denominator and numerator represent incompatible emergence patterns.
The method is not appropriate without qualification when accident years have materially different expected loss ratios, premiums have not been brought to a consistent basis, or the selected development factors do not describe the observed loss measure. A year with no observed emergence has \(w_i=0\) and contributes no weighted premium; it provides no direct observed-loss information for estimating this ratio. A very small positive emergence proportion can make the result particularly sensitive to the development assumption. These are model limitations, not arithmetic exceptions.
- Using the LDF instead of its reciprocal. If the input is a factor to ultimate, use \(1/f_i\) as the emerged proportion.
- Weighting the loss total. In this calculation, sum the observed losses as given. Apply the emergence weight to premium, not to observed losses.
- Averaging accident-year ratios equally. The Cape Cod ratio is the ratio of aggregate losses to aggregate weighted premium, not the simple mean of the individual ratios.
- Mixing bases. Confirm that loss development, observed losses, and earned premiums are consistent before interpreting the result.
- Rounding too early. Retain adequate precision in reciprocals and weighted premiums; round the final ratio as requested.
Check Your Understanding
For each hypothetical portfolio, calculate the weighted Cape Cod expected loss ratio. Use the stated monetary units consistently and round the final percentage to two decimal places.
- Earned premiums are 400, 600, and 800 thousand dollars; cumulative factors to ultimate are 2.00, 1.50, and 1.00; observed losses are 120, 270, and 680 thousand dollars. (Answer: 76.43%)
- Earned premiums are 1,000, 1,200, and 1,500 thousand dollars; the factors to ultimate are 4.00, 2.00, and 1.25; observed losses are 200, 510, and 1,020 thousand dollars. (Answer: 84.39%)
- Earned premiums are 800, 1,000, and 1,200 thousand dollars. The proportions emerged are supplied directly as 0.25, 0.60, and 1.00; observed losses are 180, 570, and 1,140 thousand dollars. (Answer: 94.50%)
- Earned premiums are 700, 900, and 1,300 thousand dollars; factors to ultimate are 3.50, 1.50, and 1.00; observed losses are 190, 540, and 1,170 thousand dollars. (Answer: 90.48%)
- Four accident years have premiums of 500, 800, 1,000, and 1,200 thousand dollars, proportions emerged of 0.20, 0.50, 0.75, and 1.00, and observed losses of 95, 360, 675, and 1,140 thousand dollars. Calculate the ratio using the supplied proportions directly. (Answer: 92.65%)