How Maturity Changes the Two Estimates

Why Maturity Matters

A reserve estimate for an immature accident period must account for losses that have not yet emerged. As a period matures, a larger share of its ultimate losses is reported, and the estimated unreported share generally becomes smaller. Bornhuetter-Ferguson and Cape Cod both use this development information, but they use reported losses and prior expectations differently.

The distinction is important when interpreting results. Bornhuetter-Ferguson uses a selected expected loss ratio directly to estimate the unreported amount. Cape Cod estimates a common loss ratio from reported losses across periods, while allowing for how much of each period’s loss is expected to be reported. In both methods, the period ultimate on a reported-loss basis is the reported amount plus the reserve.

Definition: For accident period $i$, let $E_i$ be its exposure or premium on the selected basis, $R_i$ its cumulative reported losses, and $p_i$ its reported proportion. The reported proportion is the estimated fraction of ultimate losses reported by the valuation date, so the unreported proportion is $1-p_i$. A reported proportion may be obtained from the selected development pattern; for example, if the cumulative loss development factor to ultimate is $L_i$, then $p_i=1/L_i$. These quantities must be on consistent bases.

The formulas below assume that exposures, losses, and reported proportions are aligned: the same loss basis is used for each period, and $p_i$ represents the reported fraction for that basis. The selected development pattern must be appropriate for the periods being evaluated. These calculations do not adjust for changes in coverage, claim handling, or other differences unless those adjustments have already been incorporated into the inputs.

Bornhuetter-Ferguson as Maturity Changes

Let $\lambda_0$ be the selected expected loss ratio. The Bornhuetter-Ferguson reserve is the expected ultimate loss on the exposure basis, $E_i\lambda_0$, multiplied by the unreported proportion. The reported-loss-based ultimate adds that reserve to the losses already reported.

\widehat{U}_{i,\mathrm{BF}}=E_i\lambda_0(1-p_i), \qquad \widehat{T}_{i,\mathrm{BF}}=R_i+\widehat{U}_{i,\mathrm{BF}}.

Here $\widehat{U}_{i,\mathrm{BF}}$ denotes the reserve and $\widehat{T}_{i,\mathrm{BF}}$ the estimated ultimate. For fixed $E_i$ and $\lambda_0$, a larger reported proportion reduces the reserve in direct proportion to the remaining unreported share. Reported losses enter the ultimate estimate, but not the reserve formula itself.

Why the Bornhuetter-Ferguson reserve has this form: The expected ultimate on the exposure basis is $E_i\lambda_0$. If the reported proportion is $p_i$, the estimated unreported fraction is $1-p_i$. Thus the expected unreported amount is $E_i\lambda_0(1-p_i)$. Adding this amount to the observed reported losses gives the reported-loss-based ultimate. The argument depends on the selected expected loss ratio and reported proportion being appropriate; if they are not, the formula’s reserve estimate may be systematically misleading.

Worked Example: Comparing Three Maturity Levels

A hypothetical portfolio has three accident periods, each with exposure of $1,000 thousand. Their reported proportions are 0.25, 0.60, and 0.90, and their reported losses are $140 thousand, $450 thousand, and $640 thousand. Use an expected loss ratio of 70% for Bornhuetter-Ferguson. Calculate Bornhuetter-Ferguson reserves and ultimates, then calculate the Cape Cod estimates using the same periods.

Method and units. All exposure and loss amounts are in thousands of dollars. For Bornhuetter-Ferguson, multiply the expected ultimate $E_i\lambda_0$ by the unreported proportion. For Cape Cod, first estimate the common loss ratio from total reported losses divided by total exposure times reported proportion.

Bornhuetter-Ferguson. For each period, $E_i\lambda_0=1{,}000(0.70)=700$. The unreported proportions are $1-0.25=0.75$, $1-0.60=0.40$, and $1-0.90=0.10$. Therefore the reserves are:

700(0.75)=525,\qquad 700(0.40)=280,\qquad 700(0.10)=70.

The corresponding ultimates are reported losses plus reserves: $140+525=665$, $450+280=730$, and $640+70=710$, all in thousands of dollars.

Cape Cod. The total reported losses are $140+450+640=1{,}230$. The exposure adjusted by reported proportion is $1{,}000(0.25)+1{,}000(0.60)+1{,}000(0.90)=250+600+900=1{,}750$. Thus:

\widehat{\lambda}_{\mathrm{CC}} =\frac{1{,}230}{1{,}750} =0.702857\ldots

The Cape Cod reserves are $1{,}000(0.702857\ldots)(0.75)=527.143$, $1{,}000(0.702857\ldots)(0.40)=281.143$, and $1{,}000(0.702857\ldots)(0.10)=70.286$. Adding the reported losses gives ultimates of $667.143$, $731.143$, and $710.286$, respectively. Values are rounded at the end to $0.001 thousand, or one dollar.

Check. The Cape Cod reserves sum to $878.572$ thousand after rounding; the unrounded sum is $878.571\ldots$. Adding the reported total gives an aggregate ultimate of $2{,}108.571\ldots$ thousand. The small displayed difference is solely rounding. In both methods, the least mature period has the largest reserve because its unreported proportion is largest.

How Cape Cod Uses Maturity and Experience

Cape Cod estimates a common loss ratio, $\widehat{\lambda}_{\mathrm{CC}}$, by dividing total reported losses by the total exposure adjusted for reported proportion. Its reserve and reported-loss-based ultimate for each period are:

\widehat{\lambda}_{\mathrm{CC}} =\frac{\sum_i R_i}{\sum_i E_i p_i}, \qquad \widehat{U}_{i,\mathrm{CC}} =E_i\widehat{\lambda}_{\mathrm{CC}}(1-p_i), \qquad \widehat{T}_{i,\mathrm{CC}} =R_i+\widehat{U}_{i,\mathrm{CC}}.

The numerator includes each period’s reported losses. The denominator counts each period’s exposure in proportion to how much loss is expected to have emerged by the valuation date. Therefore, for equal exposures, a more mature period contributes more exposure to the denominator than a less mature one. The estimated loss ratio is shared across periods, so a period’s reported losses can affect the reserves calculated for other periods as well as its own.

Why the Cape Cod loss-ratio formula works: If the common ultimate loss ratio is $\lambda$, then period $i$ has expected ultimate losses $E_i\lambda$, of which the reported fraction is $p_i$. Its expected reported losses are therefore $E_i\lambda p_i$. Summing over periods gives expected total reported losses $\lambda\sum_i E_i p_i$. Equating this to observed total reported losses and solving for $\lambda$ gives $\widehat{\lambda}_{\mathrm{CC}}=\sum_iR_i/\sum_iE_ip_i$. This relies on a common loss ratio and appropriate reported proportions across the periods.

The Cape Cod reserve is the expected unreported amount $E_i\widehat{\lambda}_{\mathrm{CC}}(1-p_i)$. Its period ultimate on the reported-loss basis is that reserve plus $R_i$, not automatically $E_i\widehat{\lambda}_{\mathrm{CC}}$. That latter quantity is an expected ultimate based on the portfolio loss ratio; subtracting $R_i$ from it gives a different period-level reserve. The portfolio totals can reconcile even when those period-level calculations differ, so the definitions must be kept consistent.

Worked Example: A Mature Period Has More Weight in the Cape Cod Denominator

A hypothetical two-period portfolio has exposures of $800 thousand and $1,200 thousand. The reported proportions are 0.25 and 0.75, and reported losses are $150 thousand and $700 thousand. Calculate the Cape Cod loss ratio, reserves, and reported-loss-based ultimates. Compare the reserves with Bornhuetter-Ferguson reserves using an expected loss ratio of 65%.

Cape Cod loss ratio. Total reported losses are $150+700=850$ thousand. Adjusted exposure is $800(0.25)+1{,}200(0.75)=200+900=1{,}100$ thousand. Thus $\widehat{\lambda}_{\mathrm{CC}}=850/1{,}100=0.772727\ldots$, or 77.27% to two decimal places.

Cape Cod reserves. The unreported proportions are 0.75 and 0.25. The reserves are $800(0.772727\ldots)(0.75)=463.636$ thousand and $1{,}200(0.772727\ldots)(0.25)=231.818$ thousand. The period ultimates are $150+463.636=613.636$ thousand and $700+231.818=931.818$ thousand.

Bornhuetter-Ferguson comparison. The expected ultimate amounts are $800(0.65)=520$ thousand and $1{,}200(0.65)=780$ thousand. Multiplying by the respective unreported proportions gives reserves of $520(0.75)=390$ thousand and $780(0.25)=195$ thousand. The reported-loss-based ultimates are $150+390=540$ thousand and $700+195=895$ thousand.

Check and interpretation. The Cape Cod reserves total $695.454\ldots$ thousand; adding the $850 thousand reported total gives an aggregate ultimate of $1{,}545.454\ldots$ thousand. The more mature period contributes $900$ thousand to the adjusted-exposure denominator, versus $200$ thousand for the less mature period. Cape Cod uses the portfolio’s observed losses to estimate its common loss ratio, while Bornhuetter-Ferguson retains the selected 65% ratio. Results are rounded to $0.001 thousand for amounts and two decimal places for the loss ratio.

Following a Period as It Matures

A simple development path makes the role of the unreported proportion explicit. Consider a hypothetical period with exposure of $1,000 thousand and a Bornhuetter-Ferguson expected loss ratio of 80%. At an early valuation, suppose $p_i=0.30$ and $R_i=240$ thousand. At a later valuation, suppose $p_i=0.75$ and $R_i=600$ thousand. These reported amounts are consistent with the same $800 thousand ultimate emerging in the stated proportions.

At the early valuation, the reserve is $1{,}000(0.80)(1-0.30)=560$ thousand, so the ultimate is $240+560=800$ thousand. At the later valuation, the reserve is $1{,}000(0.80)(1-0.75)=200$ thousand, and the ultimate is $600+200=800$ thousand. As the reported proportion rises, the reserve falls by $360 thousand while reported losses rise by the same amount. Under these inputs, the ultimate remains unchanged.

For a single-period Cape Cod calculation, the estimated ratio is $R_i/(E_ip_i)$. It equals $240/(1{,}000(0.30))=0.80$ at the early valuation and $600/(1{,}000(0.75))=0.80$ at the later one. The resulting reserves are also $560 thousand and $200 thousand. This single-period identity illustrates the formula; it does not supply the cross-period experience that makes a multi-period Cape Cod estimate useful.

Common Trap: Mixing Cape Cod Reserve and Ultimate Definitions

A frequent error is to calculate the Cape Cod expected unreported amount $E_i\widehat{\lambda}_{\mathrm{CC}}(1-p_i)$, then describe $E_i\widehat{\lambda}_{\mathrm{CC}}$ as the period ultimate. Those statements do not form a consistent period-level reserve and ultimate pair. If the reserve is expected unreported losses, the matching ultimate is reported losses plus that reserve.

In the preceding two-period example, the first period’s expected ultimate on the portfolio-ratio basis is $800(0.772727\ldots)=618.182$ thousand. Subtracting its reported losses would give $468.182$ thousand, whereas the Cape Cod expected-unreported reserve is $463.636$ thousand. For the second period, the corresponding amounts are $1{,}200(0.772727\ldots)-700=227.273$ thousand and $231.818$ thousand. The differences offset in total, but the period-level reserves are not interchangeable.

Key takeaway: Bornhuetter-Ferguson reserve estimates respond directly to the selected expected loss ratio and the unreported proportion; reported losses are added to obtain its ultimate. Cape Cod estimates a portfolio loss ratio from reported experience adjusted by reported proportions, then calculates each period’s expected unreported amount. In either method, define the reserve first and calculate the matching ultimate as reported losses plus reserve.

Check Your Understanding

Use consistent exposure and loss units, and keep the reserve distinct from the ultimate.

A hypothetical period has exposure of $500 thousand, an expected loss ratio of 72%, and reported proportion 0.40. Find its Bornhuetter-Ferguson reserve. (Answer: $216 thousand.)
A period has exposure of $600 thousand, expected loss ratio 65%, reported proportion 0.50, and reported losses of $260 thousand. Find its Bornhuetter-Ferguson reserve and reported-loss-based ultimate. (Answer: Reserve = $195 thousand; ultimate = $455 thousand.)
Two periods have exposures of $400 thousand and $600 thousand, reported proportions 0.50 and 0.80, and reported losses of $140 thousand and $390 thousand. Find the Cape Cod loss ratio, both expected-unreported reserves, and both reported-loss-based ultimates. (Answer: Loss ratio = 77.94%; reserves = $155.882 thousand and $93.529 thousand; ultimates = $295.882 thousand and $483.529 thousand.)
Two periods each have exposure of $1,000 thousand. Their reported proportions are 0.20 and 0.80, and their reported losses are $130 thousand and $620 thousand. Calculate the Cape Cod loss ratio, reserves, and reported-loss-based ultimates. (Answer: Loss ratio = 75%; reserves = $600 thousand and $150 thousand; ultimates = $730 thousand and $770 thousand.)

What You'll Learn