Diagnosing Implausible Loss Ratio Indications

Why an Unusual Indication Needs a Diagnosis

A very high or low loss ratio indication is a signal to investigate, not proof that the selected rate change is wrong. The result depends on a chain of inputs: losses must be developed to the intended maturity, trended to the intended period, and compared with premium adjusted to a consistent rate level. The resulting loss ratio must then be compared with a permissible loss ratio built on a consistent expense and profit basis.

The practical task is to find which part of that chain explains the surprising result. A large indication may be justified by the experience, but it may also result from a missing premium on-level adjustment, a trend period that is too short, or a loss ratio divided by the wrong permissible ratio. Conversely, an unexpectedly small or negative indication may reflect excessive adjustments rather than favorable underlying experience.

Definition: For a loss ratio indication, let $L^{\text{adj}}$ denote losses adjusted to the selected development and trend basis, $P^{\text{OL}}$ denote premium adjusted to the current rate level, and $r_{\text{perm}}$ denote the permissible loss ratio. The adjusted loss ratio is $L^{\text{adj}}/P^{\text{OL}}$, and the indicated percentage rate change is $$ I=\frac{L^{\text{adj}}/P^{\text{OL}}}{r_{\text{perm}}}-1. $$ This comparison is meaningful only when losses, premium, and the permissible ratio are on consistent coverage, expense, and time bases.

The equation helps separate the components of the result. The adjusted loss ratio measures the losses per dollar of on-level premium. Dividing that ratio by the permissible loss ratio compares actual adjusted loss experience with the portion of premium available for losses and loss adjustment expenses. Subtracting one expresses the difference as a percentage change. A positive result indicates an increase; a negative result indicates a decrease.

A Structured Diagnostic

Begin by reproducing the indication from the displayed inputs. Then test the components in a fixed order. This avoids changing several assumptions at once and makes it possible to explain the effect of each correction.

Confirm the comparison basis.
Check that the losses and premium refer to compatible coverage, experience periods, and business segments. Confirm that the permissible loss ratio applies to the same expense and loss definition.

Reconcile premium on-leveling.
Verify that historical premium has been adjusted to the selected current rate level. Inspect the rate changes and factors used, and confirm whether each factor should increase or decrease the premium.

Reconcile loss development and trend.
Check that the development adjustment brings losses to the intended maturity and that the trend adjustment moves them from the correct historical point to the selected target point.

Audit period alignment.
State the period represented by the losses, the period represented by the premium, and the target period explicitly. A factor can be mathematically correct and still be wrong for the selected dates.

Rebuild the permissible loss ratio.
Confirm that the expense, profit, and contingency provisions used in the permissible ratio match the assumptions intended for the indication. Recalculate the final indication only after the inputs reconcile.

There is no universal threshold at which an indication becomes implausible. A large change can be reasonable for a particular hypothetical book of business, while a modest change can still contain a basis error. The diagnostic question is whether the inputs and transformations explain the result, not whether the answer falls inside a generic range.

Check the Calculation by Rebuilding It

For a single experience period, a useful calculation bridge is

\text{historical losses} \longrightarrow L^{\text{adj}} \qquad \text{historical premium} \longrightarrow P^{\text{OL}} \longrightarrow \frac{L^{\text{adj}}}{P^{\text{OL}}} \longrightarrow \frac{L^{\text{adj}}/P^{\text{OL}}}{r_{\text{perm}}}-1.

This bridge is also a short justification of the indication formula. By definition, the adjusted loss ratio is the adjusted losses divided by premium at the current rate level. The permissible loss ratio is the share of premium available for losses under the assumed expense and profit structure. Thus the ratio of actual adjusted loss ratio to permissible loss ratio is the factor by which the current rate level would need to change, if the selected assumptions and experience are appropriate. Subtracting one converts that factor into a percentage change.

When several experience periods are combined, use the premium-weighted aggregate ratio rather than an unweighted average of the individual loss ratios. In symbols, with period $j$ having adjusted losses $L_j^{\text{adj}}$ and on-level premium $P_j^{\text{OL}}$, the combined ratio is $\sum_j L_j^{\text{adj}}/\sum_j P_j^{\text{OL}}$. Each period’s contribution should be traceable through the same diagnostic steps.

Worked Diagnostics

Worked Example: A High Indication That Falls After Premium On-Leveling

A hypothetical commercial equipment portfolio has $1,260,000 of losses after the selected development and trend adjustments. Historical earned premium is $1,200,000. The applicable premium on-level factor is 1.10, and the permissible loss ratio is 0.65. An initial worksheet reports a large indication using historical premium without the on-level adjustment. Diagnose the result.

Method: rebuild the adjusted loss ratio and then compare it with the permissible ratio. The on-level premium is

P^{\text{OL}}=\$1{,}200{,}000(1.10)=\$1{,}320{,}000.

The correct adjusted loss ratio is

\frac{\$1{,}260{,}000}{\$1{,}320{,}000}=0.954545\ldots

The indicated change is therefore

I=\frac{0.954545\ldots}{0.65}-1 =0.468531\ldots =46.9\%.

To see the source of the initial result, omit on-leveling temporarily. The unadjusted ratio would be $\$1{,}260{,}000/\$1{,}200{,}000=1.05$, giving $1.05/0.65-1=61.5\%$. The missing premium adjustment overstated the indicated change by about 14.7 percentage points. The corrected result remains high, so on-leveling explains part, but not all, of the indication.

Common mistake avoided: Treating historical premium as though it were already at the current rate level. Since the factor is greater than one, applying it increases the premium denominator and lowers the loss ratio. Check: the computed on-level premium exceeds historical premium, as the 1.10 factor requires.

Worked Example: A Low Indication Caused by a Short Trend Period

A hypothetical specialty property segment has $800,000 of losses at the selected historical maturity and $1,400,000 of on-level premium. Apply a loss development factor of 1.05. The selected annual loss trend is 4%, and the correct interval from the loss-period midpoint to the target midpoint is two years. The permissible loss ratio is 0.65. Diagnose an analyst’s result, which used only one year of trend.

Method: verify development and trend separately, then multiply their factors. The developed losses are

\$800{,}000(1.05)=\$840{,}000.

The correct two-year trend factor is $1.04^2=1.0816$. Thus the adjusted losses are

L^{\text{adj}}=\$840{,}000(1.0816)=\$908{,}544.

The adjusted loss ratio and indication are

\frac{\$908{,}544}{\$1{,}400{,}000}=0.64896, \qquad I=\frac{0.64896}{0.65}-1=-0.0016=-0.16\%.

Using only one year of trend gives losses of $\$840{,}000(1.04)=\$873{,}600$. The resulting loss ratio is $0.624$, and the indicated change is $0.624/0.65-1=-4.0\%$. The trend-period error therefore turns an indication close to zero into a materially larger decrease.

Common mistake avoided: Counting calendar years without identifying the actual trend interval. The correct interval here is explicitly two years between midpoints, so the factor is $1.04^2$, not $1.04$. Check: positive trend raises adjusted losses from $840,000 to $908,544.

Worked Example: A Sign Reversal From an Inconsistent Permissible Ratio

A hypothetical liability portfolio has $1,260,000 of adjusted losses and $2,000,000 of on-level premium. The intended expense provisions are 20% variable acquisition expense, 10% other variable expense, 5% fixed expense provision expressed as a share of premium for this indication, and 5% profit and contingency provision. An analyst uses a permissible loss ratio of 0.65, apparently omitting the fixed expense provision. Diagnose the result.

Method: rebuild the permissible ratio from the stated provisions. Under the given assumptions, the share of premium remaining for losses is

r_{\text{perm}}=1-0.20-0.10-0.05-0.05=0.60.

The adjusted loss ratio is

\frac{\$1{,}260{,}000}{\$2{,}000{,}000}=0.63.

Using the consistent permissible ratio, the indicated change is

I=\frac{0.63}{0.60}-1=0.05=5.0\%.

If the analyst instead uses 0.65, the result is $0.63/0.65-1=-0.030769\ldots=-3.1\%$. That incorrect denominator reverses the direction of the indication. The stated expense provisions sum to 40%, leaving 60%, not 65%, for losses.

Common mistake avoided: Accepting a permissible ratio without reconciling it to the expense and profit assumptions. Check: the adjusted loss ratio exceeds the correctly calculated permissible ratio, so the indicated change must be positive.

Common Traps and Interpretation

Reading a high ratio as a high rate need without checking its denominator. Historical premium that has not been brought to the current rate level can make the adjusted loss ratio appear too high.
Applying a correct trend rate over the wrong duration. The annual trend assumption and the number of years are separate inputs. Confirm the starting and ending points before calculating the factor.
Using a development factor as though it were a trend factor. Development brings losses to a selected maturity; trend adjusts the loss level between time periods. The factors have different purposes even when both increase losses.
Comparing different bases. If losses include a component that the permissible ratio or premium basis does not allow for, the resulting ratio is not a like-for-like comparison. Confirm the loss definition and coverage basis.
Changing several assumptions at once. If premium on-leveling, trend, and the permissible ratio are all changed together, it becomes difficult to determine which change drove the result. Recalculate one component at a time and preserve a bridge from the original to the corrected indication.

A sound diagnosis does not force the indication toward a preferred answer. It shows how the selected data and assumptions produce the result, identifies any inconsistent or misapplied input, and states what remains unexplained after correction. If the result is still unusual after the arithmetic and bases reconcile, the next step is to review whether the selected experience and assumptions are appropriate for the risk being priced.

Key takeaway: Rebuild an implausible indication in sequence: confirm the basis, on-level premium, development, trend duration and period alignment, then the permissible loss ratio. A corrected calculation should be both arithmetically reproducible and conceptually consistent.

Check Your Understanding

For each problem, identify the relevant diagnostic and calculate the requested result.

A hypothetical book has $540,000 of adjusted losses, $900,000 of historical premium, a premium on-level factor of 1.08, and a permissible loss ratio of 0.60. Calculate the indication using on-level premium. (Answer: $-7.4\%$)
A hypothetical segment has $1,000,000 of losses before adjustment, a loss development factor of 1.10, a 5% annual trend for one year, $1,400,000 of on-level premium, and a permissible loss ratio of 0.70. Calculate adjusted losses and the indicated change. (Answer: adjusted losses are $1,155,000; indication is $17.9\%$)
A portfolio’s intended provisions are 18% variable expense, 12% other expense, 4% fixed expense provision, and 6% profit and contingency provision. Its adjusted loss ratio is 0.63. Calculate the permissible loss ratio and indication. (Answer: permissible loss ratio $0.60$; indication $5.0\%$)
Losses are to be trended for three years at 3% annually. An analyst applies only two years of trend. Calculate both trend factors and the percentage by which the analyst’s factor is below the correct factor, using the correct factor as the denominator. (Answer: correct factor $1.0927$, analyst factor $1.0609$; analyst factor is $2.9\%$ below the correct factor)
A hypothetical portfolio has $720,000 of losses before adjustments, a development factor of 1.05, a 4% annual trend for two years, $1,000,000 of historical premium, a premium on-level factor of 1.12, and a permissible loss ratio of 0.65. Calculate the indication using consistent adjustments, then compare it with the result from incorrectly using historical premium. (Answer: adjusted losses are $817,689.60; correct indication is $12.3\%$; using historical premium gives $25.8\%$)

What You'll Learn