Distinguish Tail-to-Ultimate From Development-Age Factors

Why the Factor’s Age Anchor Matters

A development factor has meaning only in relation to the ages it connects. A factor of 1.10 might describe development from 36 months to 48 months, or it might describe all remaining development from 48 months to ultimate. Those factors are not interchangeable: they start at different ages and cover different periods of development.

In a chain-ladder projection, finite-age link ratios describe movement between specified development ages. A tail-to-ultimate factor describes the remaining movement from a specified age to ultimate. To project from an earlier age, the cumulative development factor combines every finite-age factor after that starting age with the tail factor anchored at the last finite age. This distinction is aligned with the CAS Exam 5 (Basic Ratemaking and Estimated Claim Liabilities) syllabus.

Definition: A selected finite-age development factor $f_{a,b}$ connects development age $a$ to a later finite development age $b$. A selected tail-to-ultimate factor $T_b$ connects age $b$ to ultimate. The cumulative development factor from age $a$ to ultimate is the product of all selected finite-age factors from $a$ through the last explicit age and the tail factor anchored at that last age.

Keep the Start and End Ages Attached to Each Factor

Suppose the valuation ages are 12, 24, 36, and 48 months. The finite-age factors are $f_{12,24}$, $f_{24,36}$, and $f_{36,48}$. If the tail factor is $T_{48}$, it applies from 48 months to ultimate; it does not apply from 36 months.

The cumulative development factor from age 36 to ultimate therefore includes the factor from 36 to 48 and the tail from 48 onward. By contrast, the cumulative factor from age 48 to ultimate is just the tail factor. The age printed on a tail factor is its starting age, not a label that can be dropped when multiplying.

\text{CDF}_{36}=f_{36,48}T_{48}, \qquad \text{CDF}_{48}=T_{48}.

More generally, if the last explicit finite development age is $t$, then for an earlier age $a$,

\text{CDF}_{a} = f_{a,a_1}f_{a_1,a_2}\cdots f_{a_{n-1},t}T_t, \qquad \text{CDF}_{t}=T_t,

where $a_1,a_2,\ldots,t$ are the subsequent selected development ages. The notation emphasizes the factor’s age anchor. A tail factor $T_t$ is already a factor from age $t$ to ultimate; multiplying it by the finite-age factors before $t$ produces a cumulative factor from an earlier age.

Why the Factors Multiply

The multiplication follows from the way ratios compose. For a single development history, let $C_a$, $C_b$, and $C_t$ denote cumulative values at successive ages. If $C_a$, $C_b$, and $C_t$ are positive, the ratios between ages satisfy

\frac{C_b}{C_a}\times\frac{C_t}{C_b} = \frac{C_t}{C_a}.

The intermediate amount $C_b$ cancels. If the remaining development from $t$ to ultimate is represented by $U/C_t$, where $U$ is ultimate, then

\frac{C_b}{C_a}\times\frac{C_t}{C_b}\times\frac{U}{C_t} = \frac{U}{C_a}.

Thus, successive development ratios multiply to give a ratio from the first age to ultimate. In practice, selected link ratios and a selected tail factor are estimates, not necessarily exact ratios from one observed claim history. The product is the chosen projection convention: each factor represents the next segment of development, and the age anchors must line up so that no segment is omitted or counted twice.

Worked Example: Building Cumulative Factors From Different Starting Ages

A hypothetical liability portfolio has selected factors $f_{12,24}=1.80$, $f_{24,36}=1.35$, and $f_{36,48}=1.15$, with a tail factor $T_{48}=1.08$. Calculate the cumulative factors from ages 36, 24, and 12 to ultimate. Identify a tempting but incorrect multiplication.

Method. Start with the factor whose starting age matches the requested age, continue through each later finite age, and then apply the tail anchored at 48 months.

From age 36, the remaining development consists of the 36-to-48 factor and the 48-to-ultimate tail:

\text{CDF}_{36}=1.15\times1.08=1.242.

From age 24, include the 24-to-36 factor as well:

\text{CDF}_{24} =1.35\times1.15\times1.08 =1.5525\times1.08 =1.6767.

From age 12, include the 12-to-24 factor:

\text{CDF}_{12} =1.80\times1.35\times1.15\times1.08 =1.80\times1.6767 =3.01806.

Check. The age-12 factor equals the age-12-to-24 factor times the age-24 cumulative factor: $1.80\times1.6767=3.01806$. The age-24 factor similarly equals $1.35\times1.242=1.6767$.

Common mistake avoided. Multiplying $f_{24,36}$ directly by $T_{48}$ gives $1.35\times1.08=1.458$, which incorrectly skips development from 36 to 48 months. The correction is to include $f_{36,48}=1.15$, giving $1.6767$.

Tail-to-Ultimate From an Earlier Age

A tail factor is often selected from the latest age at which finite link ratios are explicitly modeled. If a question supplies a tail-to-ultimate factor at an earlier age instead, that factor already includes all development after its own starting age. Do not append the same later factors again.

For example, if a factor is stated as $F_{36,\text{ult}}$, it means the cumulative factor from 36 months to ultimate. If the selected factors from 36 to 48 and from 48 onward are $f_{36,48}$ and $T_{48}$, respectively, then the two descriptions should be consistent:

F_{36,\text{ult}}=f_{36,48}T_{48}.

The notation $F_{a,\text{ult}}$ is used here to make the age anchor explicit for a direct age-to-ultimate factor. It is not an additional factor to multiply by $f_{a,b}$ when $F_{a,\text{ult}}$ already starts at age $a$.

Worked Example: Projecting From an Observed Age With a Tail Anchored Later

A hypothetical claims group has a cumulative value of $2,400,000 at 36 months. The selected factors are $f_{36,48}=1.12$, $f_{48,60}=1.06$, and $T_{60}=1.03$. Calculate the projected ultimate value from age 36, first using the segmented factors and then identifying the equivalent direct factor from 36 to ultimate.

Method. Multiply the factors in age order. The tail begins at 60 months, so both finite factors are needed to reach its starting age.

First combine the finite-age factors:

1.12\times1.06=1.1872.

Then include the tail:

\text{CDF}_{36}=1.1872\times1.03=1.222816.

Apply this cumulative factor to the value at 36 months:

\$2{,}400{,}000\times1.222816=\$2{,}934{,}758.40.

The equivalent direct factor from 36 months to ultimate is $F_{36,\text{ult}}=1.222816$. It covers exactly the same development as the three segmented factors.

Check. The projected increase is $534,758.40, and $\$2{,}400{,}000+\$534{,}758.40=\$2{,}934{,}758.40$. The calculation is consistent with applying the full age-36-to-ultimate factor once.

Common mistake avoided. Treating $T_{60}=1.03$ as the entire age-36-to-ultimate factor would project only $\$2{,}400{,}000\times1.03=\$2{,}472{,}000$, omitting the two finite-age intervals. Conversely, if the direct factor 1.222816 were supplied, multiplying it by 1.12 and 1.06 again would count those intervals twice.

Reconcile Factors Stated at Different Ages

A direct age-to-ultimate factor can be checked against its component factors. This is useful when a factor table contains values stated at several starting ages. Consistency requires that the earlier age-to-ultimate factor equal the first finite link ratio times the age-to-ultimate factor beginning at the next age.

F_{a,\text{ult}}=f_{a,b}F_{b,\text{ult}}.

This is an algebraic check on age alignment, not a separate method for estimating factors. If the equality fails materially, first check that the factors use the same development basis and that their age labels have not been misread. Differences can also arise when factors are rounded; in that case, small discrepancies may be expected.

Worked Example: Checking a Direct Factor Against Its Components

A hypothetical factor selection gives $f_{24,36}=1.25$, $f_{36,48}=1.10$, and $T_{48}=1.04$. A separate summary lists a direct factor $F_{24,\text{ult}}=1.43$. Check whether the direct factor agrees with the components, calculate $F_{36,\text{ult}}$, and find the projected ultimate from a cumulative value of $500,000 at 24 months.

Method. First form the tail-to-ultimate factor from age 36, then combine it with the 24-to-36 link ratio and compare with the stated direct factor.

From age 36, the remaining factors are the 36-to-48 link and the tail at 48:

F_{36,\text{ult}}=1.10\times1.04=1.144.

The component-based factor from age 24 is therefore

f_{24,36}F_{36,\text{ult}} =1.25\times1.144 =1.430.

This agrees with the listed direct factor of 1.43. Applying it to the 24-month value gives

\$500{,}000\times1.43=\$715{,}000.

Check. The projected increase is $\$715{,}000-\$500{,}000=\$215{,}000$. The direct and component-based factors agree before applying the value.

Common mistake avoided. Multiplying the stated direct factor 1.43 by $f_{24,36}=1.25$ would produce 1.7875 and overstate development. The 1.43 factor already starts at 24 months and includes all later development.

Common Traps in Factor Interpretation

A calculation can be arithmetically correct and still be wrong because the factors refer to different age intervals. Before multiplying, write the age path in order and check that each factor starts where the preceding factor ends.

Calling every factor a tail factor. A factor between two finite ages is a link ratio; it does not cover all remaining development.
Applying a tail at the wrong age. A tail $T_{60}$ begins at 60 months. It cannot replace factors needed to reach 60 months from an earlier age.
Double-counting a segment. A direct factor from age $a$ to ultimate already includes the development after $a$. Do not multiply it by later factors that it already contains.
Ignoring the factor’s starting age. A number such as 1.04 is not enough information by itself. Confirm the stated age range or tail anchor before using it.

When factors are reported with limited precision, products may differ slightly from a separately reported direct factor. Carry the given values at their stated precision through the calculation, and distinguish small rounding differences from a genuine age-anchor mismatch.

Key takeaway: A finite-age factor connects two specified development ages; a tail-to-ultimate factor connects its stated starting age to ultimate. To form a cumulative factor, multiply each needed segment once, in age order, and stop when ultimate is reached.

Check Your Understanding

For each problem, identify the relevant age anchors before multiplying.

A hypothetical portfolio has $f_{24,36}=1.30$, $f_{36,48}=1.10$, and $T_{48}=1.05$. Find the cumulative factor from age 24 to ultimate. (Answer: 1.5015)
A direct factor from age 36 to ultimate is 1.20, and $f_{24,36}=1.25$. Find the factor from age 24 to ultimate. (Answer: 1.50)
A hypothetical cumulative value at 48 months is $800,000. The selected factors are $f_{48,60}=1.04$ and $T_{60}=1.02$. Find the projected ultimate value. (Answer: $848,640)
The selected factors are $f_{12,24}=1.50$, $f_{24,36}=1.20$, $f_{36,48}=1.10$, and $T_{48}=1.05$. Find the cumulative factor from age 12 to ultimate. (Answer: 2.079)
A direct factor from age 24 to ultimate is reported as 1.43. Its components are $f_{24,36}=1.25$, $f_{36,48}=1.10$, and $T_{48}=1.04$. Check whether the components agree with the direct factor, and find the direct factor from age 36 to ultimate. (Answer: Yes; the component product from age 24 is 1.43, and the age-36-to-ultimate factor is 1.144.)

What You'll Learn