Why Numerator and Denominator Must Share a Basis

Why a Shared Basis Matters

An indication compares a measure of the cost expected under a proposed set of conditions with premium representing those same conditions. The arithmetic may be correct while the conclusion is not: a numerator for one period divided by a denominator for another, or costs for one coverage divided by premium for another, does not measure the intended relationship. The key control is to make the numerator and denominator comparable before interpreting their ratio.

In this tutorial, “basis” means the conditions represented by the amounts being compared. Important dimensions include the time period, coverage, rate level, exposure base, and business mix. The earlier material on on-level premium and projecting losses addresses how amounts can be adjusted along these dimensions. Here the emphasis is on why a mismatch matters algebraically and how to diagnose its effect.

Definition: A matched basis is a common set of conditions for the numerator and denominator of an indication, including the same prospective time period, coverage definition, current rate level, and relevant exposure population or mix. A numerator and denominator on different bases are mismatched; their ratio combines effects that the indication is intended to separate.

For a simple illustration, let $C$ be projected costs on the intended basis, $P$ be premium on that same basis, and $q$ be the proportion of premium available for those costs under the selected provision. In this simplified setup, the indicated factor is the cost divided by the cost provision in the current-level premium base:

F=\frac{C}{qP}.

This formula is used here to isolate the effect of basis mismatches, not to derive a complete rate indication. Assume $q$ is appropriate and unchanged across the comparison, and that all relevant expense and profit provisions are already represented by it. If the numerator or denominator does not represent the same target conditions, the quotient no longer has that interpretation.

How a Mismatch Changes the Quotient

Suppose the matched cost and premium are $C$ and $P$. If a mismatched input is $a$ times its matched value in the numerator and $b$ times its matched value in the denominator, the calculated factor is:

F_{\text{mismatch}} =\frac{aC}{qbP} =\frac{a}{b}F_{\text{matched}}.

This gives a direct diagnostic: the relative distortion is $a/b$. If only the numerator is inflated by 8%, then $a=1.08$ and $b=1$, so the factor is 8% too high relative to the matched factor. If only the denominator is 8% too large, then $a=1$ and $b=1.08$, so the factor is too small by the reciprocal effect. The direction depends on which input is mismatched and how.

Key result: A ratio is meaningful only when its numerator and denominator describe the same underlying conditions. Adjustments must translate each input to a common basis; merely making the amounts look similar in size does not establish comparability.

Worked Examples

Worked Example: A Time-Period Mismatch

A hypothetical book of small-boat policies has projected costs of $792,000 for the prospective period. For this simplified illustration, 70% of premium is available for those costs. The current-rate-level premium for the same coverage and prospective exposure mix is $1,200,000. An analyst also has $1,080,000 of premium from the earlier experience period and proposes using it without adjustment. Calculate the matched factor, the factor from the mismatched denominator, and the relative distortion.

Method. Use the prospective-period projected cost and the premium representing that same prospective period, coverage, and mix. The historical premium is not a substitute unless it is brought to the required basis.

Matched calculation. The cost provision in the matched premium base is:

$$ 0.70(1{,}200{,}000)=840{,}000. $$

Therefore the matched indicated factor is:

F_{\text{matched}} =\frac{792{,}000}{840{,}000} =0.942857\ldots \approx 0.9429.

The factor is below 1.0000 in this simplified setup. Now use the earlier-period premium of $1,080,000 without adjustment. Its cost provision is:

0.70(1{,}080{,}000)=756{,}000, \qquad F_{\text{mismatch}} =\frac{792{,}000}{756{,}000} =1.047619\ldots \approx 1.0476.

The mismatched factor relative to the correct factor is:

\frac{1.047619\ldots}{0.942857\ldots} =\frac{1{,}200{,}000}{1{,}080{,}000} =1.111111\ldots.

Thus the mismatched result is 11.11% higher than the matched factor. Check: the denominator premium is 10% lower than the matched premium, so dividing by its smaller cost provision raises the factor by the reciprocal, $1/0.90=1.111111\ldots$. The common mistake is to compare prospective costs with unadjusted historical-period premium as though the periods were interchangeable.

Worked Example: A Coverage Mismatch

A hypothetical equipment policy is being indicated for a target coverage that excludes a specified machinery breakdown endorsement. For the same prospective exposure mix and current rate level, expected costs are $420,000 for the target coverage. The relevant premium base is $700,000, and 70% of premium is available for costs. A report instead supplies $500,000 of projected costs that includes $80,000 from the excluded endorsement. Calculate the matched factor and the factor using the broader-coverage numerator.

Method. Remove the cost for coverage not included in the target contract. Keep the premium denominator fixed because it already represents the target coverage.

The numerator that matches the target coverage is $420,000. The cost provision in the matched premium base is:

$$ 0.70(700{,}000)=490{,}000. $$

The matched factor is:

F_{\text{matched}} =\frac{420{,}000}{490{,}000} =0.857142\ldots \approx 0.8571.

If the broader-coverage cost is used without removing the excluded endorsement, the factor becomes:

F_{\text{mismatch}} =\frac{500{,}000}{490{,}000} =1.020408\ldots \approx 1.0204.

The numerator is $500{,}000/420{,}000=1.190476\ldots$ times the matched cost, and the denominator is unchanged. The mismatched factor is therefore 19.05% higher relative to the matched factor. Check: the $80,000 difference is exactly the cost assigned to the excluded endorsement. The common mistake is to use costs for broader coverage while leaving the target-coverage premium unchanged.

Worked Example: An Unadjusted Rate-Level Mismatch

A hypothetical annual travel policy has $900,000 of premium from an experience period at the old rate level. A uniform rate change of 8% has since been implemented. The selected prospective costs are $680,400, and 70% of premium is available for those costs. Assume the premium is otherwise appropriate for the same coverage, period, and exposure mix. Calculate the indicated factor using premium on the current rate level, then show the error from using the unadjusted premium.

Method. Convert the historical premium to the current rate level before using it as the denominator. Under the stated assumption of a uniform 8% change with no other differences:

P_{\text{current level}} =900{,}000(1.08) =972{,}000.

The matched cost provision is:

$$ 0.70(972{,}000)=680{,}400. $$

The matched factor is:

F_{\text{matched}} =\frac{680{,}400}{680{,}400} =1.0000.

Using the old-rate-level premium without adjustment gives:

0.70(900{,}000)=630{,}000, \qquad F_{\text{mismatch}} =\frac{680{,}400}{630{,}000} =1.0800.

The factor is 8% too high because the denominator is only $1/1.08$ of the matched premium. Check: the current-level premium is $72,000 greater than the old-level premium, and its 70% cost provision is $50,400 greater; adding $50,400 to the unadjusted $630,000 gives the matched $680,400. The common mistake is to update the cost estimate but leave the premium at an obsolete rate level.

A Practical Basis Check

A disciplined comparison starts by naming the basis intended for the indication and then checking each input against it. If one input differs, identify the adjustment needed and its direction. Do not apply an adjustment merely because a number is historical or because a factor is available: it must correspond to the specific mismatch, and it should not duplicate an adjustment already reflected in the data.

State the target basis.
Specify the prospective period, coverage, rate level, exposure population, and mix that the comparison is intended to represent.

Label both inputs.
Record what period, coverage, rate level, and exposure basis the selected cost and premium actually represent. Do not infer that two amounts match simply because they come from the same report.

Translate each input as needed.
Bring costs and premium to the target basis using adjustments appropriate to each. Avoid using an adjustment that changes a different dimension from the one that is mismatched.

Reconcile and interpret.
Verify the adjusted inputs are comparable, calculate the quotient, and explain what the result means for that basis. If the basis is still inconsistent, the factor is not interpretable as the intended indication.

Time period, coverage, and rate level are distinct dimensions. A trend adjustment may project costs to a prospective period, but it does not by itself change the coverage represented by those costs. An on-level adjustment may translate premium to a current rate level, but it does not automatically correct a different exposure mix. Likewise, changing a numerator without checking whether the denominator requires a corresponding translation can leave the comparison inconsistent.

Mismatch	Typical symptom	Basis check
Time period	Prospective costs compared with historical premium	Do both amounts represent the same target period?
Coverage	Costs include an endorsement or benefit absent from the premium base, or vice versa	Are covered events, limits, and other relevant contract terms consistent?
Rate level	Costs are projected for a current or prospective setting while premium remains at an older rate level	Has premium been brought to the rate level required for the comparison?
Exposure or mix	Costs reflect one set of risks while premium reflects another	Do the inputs refer to a comparable exposure base and distribution of risks?

Common Exam Trap

A frequent error is to assume that an adjustment to one input repairs every possible mismatch. It does not. If losses have been projected to a prospective period but the premium is still on an old rate level, the loss projection does not make the premium current-level. Conversely, on-leveling premium does not project historical costs to the prospective period. Each transformation must be tied to a stated basis difference.

A useful algebraic check is to compare the matched and mismatched factors. If the numerator alone is multiplied by $a$, the factor changes by $a$. If the denominator alone is multiplied by $b$, the factor changes by $1/b$. If both are adjusted, the factor changes by $a/b$. This check can catch a reversed adjustment or an unexplained change in the result.

Key takeaway: Before accepting an indicated factor, ask whether its costs and premium refer to the same time period, coverage, rate level, exposure base, and mix. If they do not, quantify the mismatch and bring the inputs to a common basis before interpreting the ratio.

Check Your Understanding

For each problem, identify the basis issue and calculate the requested factor. Round factors to four decimal places where needed.

A hypothetical policy group has $630,000 of prospective costs, a matched prospective premium base of $1,000,000, and a 70% cost provision. An analyst instead uses $900,000 of earlier-period premium without adjustment. Find the matched factor and the mismatched factor. (Answer: matched $0.9000$; mismatched $1.0000$.)
For a target contract, projected costs are $360,000 and the matching current-level premium is $600,000. The cost provision is 65%. A separate $45,000 of costs relates to an excluded benefit. Find the matched factor and the factor if the excluded-benefit cost is mistakenly included. (Answer: matched $0.9231$; mismatched $1.0385$.)
A hypothetical portfolio has $800,000 of premium at an old rate level. A uniform 5% rate change applies, and the same-period, same-coverage premium is otherwise appropriate. Projected costs are $588,000 and the cost provision is 70%. Find the factor using on-level premium and the factor using old-level premium. (Answer: on-level premium is $840,000; matched factor $1.0000$; old-level factor $1.0500$.)
Projected costs on the target basis are $686,400. A $980,000 premium base is matched to the target period, coverage, rate level, and mix; the cost provision is 70%. An adjusted historical premium of $945,000 is also available, but it reflects a different exposure mix. Calculate the factor using the matched base and state whether the adjusted historical premium should replace it. (Answer: $0.70(980{,}000)=686{,}000$, so the matched factor is $686{,}400/686{,}000=1.0006$; do not substitute the $945,000 premium because its mix is different.)
A matched factor is $0.9600$. A proposed numerator is 4% higher than the matched cost, while a proposed denominator is 2% higher than the matched premium; the cost provision is unchanged. Find the resulting factor and its relative change from the matched factor. (Answer: $0.9600(1.04/1.02)=0.9788$, approximately 1.96% higher than the matched factor.)

What You'll Learn