Cape Cod Ultimate Losses and Unpaid Reserves

From Expected Losses to Ultimate Losses

Cape Cod first supplies an expected ultimate loss amount for each accident year. The next step is to estimate the losses that have not yet emerged, add those losses to the amount already observed, and calculate the unpaid reserve. The central discipline is to keep the exposure, development measure, and observed losses on a consistent basis.

In this tutorial, observed losses in the calculations are cumulative paid losses at the valuation date. Therefore, ultimate loss less observed loss is an estimate of unpaid claim payments. The same arithmetic can be applied to an incurred-loss basis, but ultimate less incurred is additional incurred loss development, not the total unpaid cash liability; total unpaid cash is measured against paid losses.

Definition: For accident year \(i\), let \(P_i\) be the aligned premium or exposure base, \(\theta_{CC}\) the selected Cape Cod expected loss ratio, and \(w_i\) the expected proportion of ultimate losses reported by the valuation date. The expected ultimate loss is \(\widehat{E}_i=P_i\theta_{CC}\). If \(C_i\) is observed cumulative paid loss, the expected unreported loss is \((1-w_i)\widehat{E}_i\), the Cape Cod ultimate estimate is \(\widehat{U}_i=C_i+(1-w_i)\widehat{E}_i\), and the estimated unpaid reserve is \(\widehat{R}_i=\widehat{U}_i-C_i=(1-w_i)\widehat{E}_i\). Here \(0\leq w_i\leq1\).

The weight \(w_i\) is a cumulative proportion of ultimate losses expected to have emerged by the valuation date, not a development factor. If a selected cumulative development factor is \(f_i\), the corresponding reported proportion is \(w_i=1/f_i\), provided the factor and losses refer to the same cumulative measure. Thus the unreported proportion is \(1-w_i\).

The Cape Cod expected loss ratio is commonly obtained by dividing total observed losses by the total premium-weighted reported proportion. This is the exposure-weighted selection from the earlier material on calculating the weighted Cape Cod loss ratio. For the calculations that follow, it is enough to remember that the selected ratio produces an expected ultimate amount for each accident year; the reserve calculation then uses the unreported portion of that amount.

\theta_{CC}=\frac{\sum_i C_i}{\sum_i P_iw_i}, \qquad \widehat{E}_i=P_i\theta_{CC}, \qquad \widehat{R}_i=(1-w_i)\widehat{E}_i, \qquad \widehat{U}_i=C_i+\widehat{R}_i.

The model assumes that premiums are on the selected basis and aligned with the accident years, the development weights reasonably represent the emergence pattern for the losses being measured, and the selected expected loss ratio is appropriate across the included years. It also assumes that observed losses and development weights use a consistent loss basis. If these conditions fail, the arithmetic may still be correct while the resulting reserve is not appropriate.

Why the Ultimate Formula Adds Expected Unreported Losses

The Cape Cod estimate does not replace observed losses with expected losses. It preserves the observed amount and estimates only the portion not yet observed. To see why, let \(\widehat{E}_i\) denote the expected ultimate loss. Under the selected development pattern, the expected amount reported by valuation is \(w_i\widehat{E}_i\). The expected amount still unreported is therefore the difference between expected ultimate and expected reported loss:

\widehat{E}_i-w_i\widehat{E}_i=(1-w_i)\widehat{E}_i.

The current observed amount \(C_i\) takes the place of the expected reported portion. Adding the expected unreported portion gives the estimate:

\widehat{U}_i=C_i+(1-w_i)\widehat{E}_i.

Subtracting the observed paid amount from both sides gives \(\widehat{R}_i=(1-w_i)\widehat{E}_i\). This is an estimation procedure, not a claim that the ultimate must equal the expected ultimate. Actual observed losses can differ from the amount expected to have emerged; Cape Cod retains that observed experience and adds the expected remainder.

Establish the basis.
Confirm the premium, observed losses, and development proportions refer to compatible accident years, valuation dates, and loss measures.

Calculate expected ultimate losses.
Multiply each accident year's premium or exposure base by the selected Cape Cod loss ratio.

Estimate the unreported portion.
Multiply expected ultimate loss by one minus that accident year's reported proportion.

Complete and reconcile.
Add the expected unreported amount to observed loss for each year, then check that total ultimate equals total observed loss plus total unpaid reserve.

Worked Calculations

Worked Example: Calculating the Cape Cod Ultimate and Reserve

A hypothetical insurer reviews three accident years, with amounts measured in thousands of dollars. Premiums are \(1{,}000\), \(800\), and \(600\); reported proportions are \(0.80\), \(0.50\), and \(0.20\); and cumulative paid losses are \(528\), \(264\), and \(66\). Calculate the Cape Cod loss ratio, each year's ultimate loss and unpaid reserve, and the portfolio totals.

Method and loss ratio. Use the exposure-weighted Cape Cod loss ratio. The premium-weighted reported amounts are \(1{,}000(0.80)=800\), \(800(0.50)=400\), and \(600(0.20)=120\), for a denominator of \(800+400+120=1{,}320\). Total observed loss is \(528+264+66=858\). Therefore:

\theta_{CC}=\frac{858}{1{,}320}=0.65.

Expected ultimate losses. Multiplying each premium by \(0.65\) gives \(1{,}000(0.65)=650\), \(800(0.65)=520\), and \(600(0.65)=390\).

Unpaid reserves and ultimates. The unreported proportions are \(0.20\), \(0.50\), and \(0.80\). The reserves are \(650(0.20)=130\), \(520(0.50)=260\), and \(390(0.80)=312\). Adding each reserve to observed paid loss gives ultimate estimates of \(528+130=658\), \(264+260=524\), and \(66+312=378\).

Reconciliation. Total unpaid reserve is \(130+260+312=702\). Total ultimate loss is \(658+524+378=1{,}560\), and total observed plus reserve is \(858+702=1{,}560\), as required. The calculation avoids adding the entire expected ultimate loss to observed loss, which would double count the expected reported portion.

Worked Example: A Year with Observed Losses Above Expectation

A hypothetical portfolio has three accident years, with premiums of \(500\), \(750\), and \(1{,}000\) thousand dollars. Their reported proportions are \(0.90\), \(0.60\), and \(0.25\), and paid losses are \(360\), \(324\), and \(144\) thousand dollars. Calculate the selected Cape Cod ratio, the year-by-year reserve and ultimate estimates, and the totals. Do not force the individual ultimate estimates to equal expected ultimate losses.

Selected ratio. The weighted exposure denominator is \(500(0.90)+750(0.60)+1{,}000(0.25)=450+450+250=1{,}150\). Total paid loss is \(360+324+144=828\). Thus \(\theta_{CC}=828/1{,}150=0.72\).

Expected ultimate losses. The amounts are \(500(0.72)=360\), \(750(0.72)=540\), and \(1{,}000(0.72)=720\), all in thousands.

Expected unreported losses. The unreported proportions are \(0.10\), \(0.40\), and \(0.75\). Reserves are \(360(0.10)=36\), \(540(0.40)=216\), and \(720(0.75)=540\).

Ultimate estimates. Add observed paid losses to those reserves: \(360+36=396\), \(324+216=540\), and \(144+540=684\). The first year's observed paid loss equals its expected ultimate loss, but its estimated ultimate is \(396\), not \(360\): the method retains the observed amount and adds the expected unreported amount.

Check. Total observed loss is \(828\), total reserve is \(36+216+540=792\), and total ultimate is \(396+540+684=1{,}620\). Since \(828+792=1{,}620\), the totals reconcile. The common mistake avoided here is substituting expected ultimate loss for the ultimate estimate after already observing actual losses.

Worked Example: Converting Development Factors Before Reserving

A hypothetical insurer has premiums of \(400\), \(600\), and \(800\) thousand dollars for three accident years, and a selected Cape Cod loss ratio of \(0.75\). The selected cumulative development factors to ultimate are \(1.25\), \(2\), and \(4\). Paid losses to date are \(270\), \(200\), and \(90\) thousand dollars. Calculate ultimate losses and unpaid reserves.

Convert factors to reported proportions. A cumulative factor is not itself the reported proportion. The corresponding proportions are \(1/1.25=0.80\), \(1/2=0.50\), and \(1/4=0.25\). The unreported proportions are \(0.20\), \(0.50\), and \(0.75\).

Expected ultimate losses. Premium times loss ratio gives \(400(0.75)=300\), \(600(0.75)=450\), and \(800(0.75)=600\), in thousands.

Reserves and ultimates. Expected unreported amounts are \(300(0.20)=60\), \(450(0.50)=225\), and \(600(0.75)=450\). The ultimate estimates are \(270+60=330\), \(200+225=425\), and \(90+450=540\).

Reconciliation. Total paid loss is \(270+200+90=560\); total reserve is \(60+225+450=735\); and total ultimate is \(330+425+540=1{,}295\). The identity \(560+735=1{,}295\) checks. Using the factors \(1.25\), \(2\), and \(4\) as though they were reported proportions would reverse the development meaning and produce incorrect unreported amounts.

Interpretation, Reconciliation, and Exam Traps

The portfolio calculation is a direct sum of the accident-year calculations. If all values use the same units and basis, then:

\sum_i\widehat{U}_i = \sum_i C_i+\sum_i\widehat{R}_i.

This equality is a useful arithmetic check, but it does not establish that the selected loss ratio or development weights are appropriate. Reconciliation verifies the calculation structure; actuarial judgment is still needed to assess the assumptions and inputs.

A more mature year generally has a larger \(w_i\), so a smaller fraction of expected ultimate remains unreported. In the limiting cases, if \(w_i=1\), the estimated reserve is zero and ultimate equals observed loss; if \(w_i=0\), the entire expected ultimate amount remains unreported and the ultimate estimate is observed loss plus expected ultimate. These checks are useful for detecting an inverted proportion.

A frequent exam trap is applying the reported proportion to expected ultimate when the question asks for the unreported amount. For example, if expected ultimate is \(520\) and \(w=0.50\), the wrong calculation \(520(0.50)=260\) happens to give the same answer as the unreported amount only because the proportions are equal in this special case. If \(w=0.80\), however, the unreported amount is \(520(1-0.80)=104\), not \(520(0.80)=416\). Always identify whether the requested proportion is reported or unreported before multiplying.

Do not double count. The ultimate estimate is observed loss plus expected unreported loss, not observed loss plus expected ultimate loss.
Do not confuse a factor with a proportion. Convert a cumulative factor to the reported proportion before finding the unreported share.
Keep the basis consistent. For an unpaid cash reserve, subtract paid-to-date losses from ultimate losses. If observed losses are incurred, the difference has a different interpretation.
Check units and totals. A reserve in thousands must not be added to losses stated in dollars without conversion; aggregate ultimate must equal aggregate observed losses plus aggregate reserve.

Key takeaway: For each accident year, Cape Cod ultimate loss equals observed loss plus expected unreported loss. On a paid basis, unpaid reserve is that expected unreported amount, and the portfolio totals must reconcile as ultimate equals paid-to-date loss plus unpaid reserve.

Check Your Understanding

For each problem, calculate on the stated loss basis and check the relationship between observed loss, reserve, and ultimate.

A hypothetical accident year has premium of \(500\) thousand dollars, a selected Cape Cod loss ratio of \(0.70\), a reported proportion of \(0.80\), and paid loss of \(290\) thousand dollars. Find its expected ultimate loss, reserve, and ultimate. (Answer: Expected ultimate \(350\), reserve \(70\), ultimate \(360\), all in thousands of dollars.)
Two hypothetical accident years have premiums of \(250\) and \(750\) thousand dollars, a selected loss ratio of \(0.64\), reported proportions of \(0.60\) and \(0.20\), and paid losses of \(100\) and \(200\) thousand dollars. Find each reserve and ultimate, then the portfolio reserve and ultimate. (Answer: Reserves \(64\) and \(384\); ultimates \(164\) and \(584\); total reserve \(448\) and total ultimate \(748\), all in thousands of dollars.)
For two hypothetical years, premiums are \(200\) and \(400\) thousand dollars, reported proportions are \(0.75\) and \(0.25\), and paid losses are \(90\) and \(90\) thousand dollars. Calculate the weighted Cape Cod loss ratio, each reserve and ultimate, and the aggregate reconciliation. (Answer: Loss ratio \(0.72\); reserves \(36\) and \(216\); ultimates \(126\) and \(306\); aggregate paid loss \(180\), reserve \(252\), and ultimate \(432\), all in thousands of dollars.)
A hypothetical portfolio has premiums of \(300\), \(500\), and \(400\) thousand dollars, selected loss ratio \(0.80\), cumulative factors of \(1.25\), \(2.5\), and \(4\), and paid losses of \(210\), \(190\), and \(80\) thousand dollars. Convert the factors and calculate the reserve and ultimate for each year and in total. (Answer: Reported proportions \(0.80\), \(0.40\), and \(0.25\); reserves \(48\), \(240\), and \(240\); ultimates \(258\), \(430\), and \(320\); total reserve \(528\) and total ultimate \(1{,}008\), all in thousands of dollars.)
A hypothetical year has expected ultimate loss of \(300\) thousand dollars, a reported proportion of \(0.50\), and paid loss of \(110\) thousand dollars. Find the unpaid reserve and ultimate on a paid basis. Explain why the reserve is not simply expected ultimate less expected reported loss. (Answer: Reserve \(150\) thousand dollars and ultimate \(260\) thousand dollars; the observed paid amount is retained, so ultimate is observed paid loss plus the expected unreported amount.)

What You'll Learn