Why Basis Alignment Matters
An expected loss estimate combines information from two sources: a selected expected loss relationship, such as an expected loss ratio, and a measure of the business to which that relationship applies. The arithmetic may be simple, but the estimate is only meaningful when both inputs describe the same coverage, period, premium or exposure basis, and loss scope. A mismatch can produce a plausible-looking number that is nevertheless too high or too low.
These checks are useful whether the expected loss is calculated from premium and an expected loss ratio or from exposure and an expected loss cost. They do not replace the earlier material on expected ultimate losses, loss and LAE scope, or segment-level estimates. Instead, they help ensure the inputs to those calculations belong together. This topic is aligned with the CAS Exam 5 (Basic Ratemaking and Estimated Claim Liabilities) syllabus.
The Expected Loss Calculation Requires Compatible Inputs
Let \(ELR\) denote an expected loss ratio and \(EP\) the earned premium for the coverage and period being estimated. Under the premium-based method, expected ultimate loss is \(ELR \times EP\). Here, \(ELR\) and \(EP\) must use a consistent coverage, time, and loss basis. For an exposure-based method, let \(e\) be the number of relevant exposure units and \(ELC\) the expected loss cost per unit. Expected ultimate loss is \(ELC \times e\); again, the unit in \(ELC\) must match the unit in \(e\).
These relationships follow directly from the meaning of each selected measure. An expected loss ratio is expected loss divided by the premium for the business represented by that loss. Multiplying by that same premium recovers the expected loss. Likewise, expected loss cost is expected loss divided by the relevant exposure count; multiplying by that count recovers expected loss. Changing the denominator’s basis without changing the selected ratio or cost breaks that relationship.
For example, an expected loss ratio for liability coverage should not be multiplied by premium that also includes physical damage unless the ratio was selected for the combined coverage. Similarly, a loss cost per vehicle-year should not be multiplied by a count of vehicles in force at a single date if the intended measure is earned vehicle-years.
| Basis to check | Question to ask | Typical mismatch |
|---|---|---|
| Coverage | Do the premium or exposures represent the same coverage as the selected expectation? | Applying a liability loss ratio to combined liability and physical damage premium |
| Period | Do the inputs represent the same accident, policy, or calendar period? | Using written premium for an accident-period estimate that requires earned premium |
| Rate level | Are premium and expected loss relationships expressed at compatible price levels? | Applying a current-level ratio to unadjusted historical premium |
| Exposure | Does the exposure count use the unit assumed by the expected loss cost? | Using policies in force instead of earned policy-years |
| Loss and expense scope | Do both the selected expectation and the target include the same components? | Adding an expense provision to a ratio that already includes it |
Worked Checks for Common Basis Mismatches
Worked Example: Applying a Coverage-Specific Loss Ratio
A hypothetical insurer is estimating ultimate liability loss and LAE for an accident period. The period has $1,500,000 of earned premium: $900,000 for liability coverage and $600,000 for physical damage coverage. The selected liability loss-and-LAE ratio is 65%. Estimate liability loss and LAE, and quantify the result of applying the ratio to the wrong premium.
Method. This is a premium-based expected loss calculation. The 65% ratio is specific to liability, so the compatible premium is the $900,000 liability premium.
Correct estimate. Multiply the selected ratio by liability earned premium:
The expected liability loss and LAE is $585,000.
Wrong calculation. Applying the same ratio to all earned premium gives:
The overstatement is \(\$975{,}000-\$585{,}000=\$390{,}000\). This is not a rounding issue: it occurs because the calculation implicitly treats physical damage premium as if it were liability premium. Common mistake avoided: using a total account premium when the selected ratio applies only to one coverage.
Check. The excess premium is $600,000, and \(0.65\times \$600,000=\$390,000\), which equals the overstatement.
Worked Example: Earned Premium Versus Written Premium
For a hypothetical accident period, earned premium is $1,200,000 and written premium is $1,500,000. A selected expected loss ratio of 60% is intended to estimate losses arising from earned exposure in that period. Calculate the estimate using the appropriate premium and compare it with the result from using written premium.
Method. Because the target is accident-period losses on earned exposure, use earned premium. Written premium can include amounts for coverage that has not yet been earned in the period, so it is not interchangeable with earned premium here.
Correct estimate.
The expected loss is $720,000.
Wrong calculation. Using written premium instead produces:
The overstatement is \(\$900{,}000-\$720{,}000=\$180{,}000\). Written premium exceeds earned premium by $300,000, and \(0.60\times \$300,000=\$180,000\), confirming the difference. The ratio’s intended period and the premium denominator must agree. Common mistake avoided: assuming that a larger premium measure is a more complete measure of exposure for any loss estimate.
Worked Example: Matching an Expected Loss Cost to Its Exposure Unit
A hypothetical commercial fleet program has an expected loss cost of $240 per earned vehicle-year. The accident period contains 3,600 earned vehicle-years. The administrative system also reports 4,000 distinct vehicles that appeared on at least one policy during the period. Estimate expected loss using the appropriate exposure measure and quantify the error from using the vehicle count.
Method. The loss cost is stated per earned vehicle-year, so multiply it by earned vehicle-years, not by a count of distinct vehicles. A vehicle insured for only part of the year contributes less than one earned vehicle-year.
Correct estimate.
Wrong calculation. Substituting the 4,000 distinct vehicles gives:
The overstatement is \(\$960{,}000-\$864{,}000=\$96{,}000\). Equivalently, the count used exceeds the valid exposure by \(4{,}000-3{,}600=400\) units, and \(400\times \$240=\$96{,}000\). Common mistake avoided: matching a rate that is stated per unit to a count with a different time basis.
Worked Example: Loss and LAE Scope
A hypothetical line has $2,000,000 of earned premium. The selected expected indemnity loss ratio is 55%, excluding allocated loss adjustment expense (ALAE). Expected ALAE is estimated separately as 5% of the same earned premium. Estimate expected loss and ALAE, and show the effect of omitting or double-counting ALAE.
Method. The target is indemnity loss plus ALAE. Since the selected 55% ratio covers indemnity only, add the separately estimated ALAE once.
Indemnity estimate.
ALAE estimate.
Therefore expected loss and ALAE is \(\$1{,}100{,}000+\$100{,}000=\)$1,200,000. Omitting ALAE would produce $1,100,000, an understatement of $100,000. If the selected ratio instead already included ALAE, adding the same $100,000 again would overstate the result by $100,000. Common mistake avoided: relying on the ratio’s label without checking which loss components it includes.
Check. The combined ratio for the specified target is \(55\%+5\%=60\%\), and \(0.60\times \$2{,}000{,}000=\$1{,}200{,}000\), the same result.
A Practical Reconciliation Sequence
Before accepting an expected loss calculation, trace each input back to its definition. The following sequence makes the checks explicit and helps distinguish a true assumption difference from a basis error.
Name the coverage, period, and loss components to be estimated. For example, specify whether the target is liability loss only or liability loss and LAE.
Record the coverage, period, rate level, and loss scope represented by the expected loss ratio or expected loss cost.
For a premium-based estimate, determine whether the premium is earned or written and which coverage and period it represents. For an exposure-based estimate, confirm the unit and time basis.
Adjust or select inputs so they represent the intended basis. Do not use multiplication to conceal an unresolved difference in scope.
Recalculate using the mismatched input and compare the result with the aligned estimate. The difference indicates the direct impact of that mismatch, holding other assumptions fixed.
The same discipline applies when expected ultimate loss is later used to estimate unpaid loss. If a reserve is calculated as a fixed proportion \(q\) of expected ultimate loss, then changing the ultimate estimate by \(\Delta U\) changes that reserve by \(q\Delta U\), provided \(q\) is unchanged. A mismatch in the expected loss basis can therefore carry through to the reserve; it does not disappear merely because a later calculation is performed correctly.
Common Exam Traps and Interpretation
- Matching labels rather than definitions. “Loss ratio” does not by itself specify coverage, period, or whether LAE is included. Read the selection’s scope.
- Treating premium measures as interchangeable. Earned and written premium answer different questions. Use the measure consistent with the loss period and expected relationship.
- Ignoring units. A loss cost per vehicle-year requires vehicle-years, not policies, vehicles, or another count unless those units are explicitly equivalent.
- Mixing rate levels. Historical premium may need to be adjusted to the level represented by a selected ratio. Applying a current-level ratio to an unadjusted historical premium can distort the estimate.
- Adding expenses twice or not at all. Establish which loss adjustment expense components are in the selected ratio and which are in the target before combining amounts.
A useful final check is dimensional: the premium or exposure measure should have the same unit and scope as the denominator implicit in the selected expected loss relationship. Then check the direction and size of any difference. If the estimate changes substantially when the basis is corrected, the mismatch is material and should be resolved rather than treated as harmless rounding.
Check Your Understanding
For each problem, identify the compatible basis, calculate the estimate, and compare it with the stated incorrect calculation where requested.
- A hypothetical coverage has $800,000 of earned premium and an expected loss ratio of 58% for that same coverage. What is the expected loss? (Answer: $464,000)
- A total premium of $1,100,000 contains $700,000 for the target coverage and $400,000 for another coverage. The selected target-coverage loss ratio is 62%. What is the correct estimate, and how much would using total premium overstate it? (Answer: Correct estimate $434,000; overstatement $248,000)
- A selected expected loss cost is $250 per earned unit-year. The period has 2,800 earned unit-years, but a report lists 3,000 distinct units. Calculate the estimate using the correct basis and the overstatement from using the distinct-unit count. (Answer: Correct estimate $700,000; overstatement $50,000)
- A hypothetical portfolio has $3,000,000 earned premium and a 57% expected loss ratio that already includes ALAE. An analyst adds a separate expense provision equal to 9% of premium. What is the correct expected loss and LAE, and what is the analyst’s overstatement? (Answer: Correct estimate $1,710,000; overstatement $270,000)
- Historical earned premium is $1,800,000 at 96% of the current rate level. A 64% expected loss ratio is selected for the current rate level. Estimate expected loss on the current-level basis and quantify the understatement from using historical premium without adjustment. (Answer: Current-level premium $1,875,000; expected loss $1,200,000; understatement $48,000)