Integrated Expected Loss and Weight Calculations

From Class Rates to Component Weights

A plan-weight calculation is only as reliable as the expected-loss amounts supplied to it. The complete sequence is to calculate expected losses for each class, combine the classes, allocate the total between primary and excess losses, and then apply the plan’s specified credibility rule separately to each component. Keeping these steps distinct helps prevent a correct formula from being applied to the wrong amount.

The calculations below assume that each class rate is on the same basis as its stated exposure, that the rates apply to the period being analyzed, and that the D-ratios match the applicable classes and split point. All amounts are hypothetical. The credibility thresholds are plan provisions, not universal constants.

Definition: For class $j$, let $r_j$ be the expected-loss rate per exposure unit and $e_j$ the number of exposure units. The class expected loss is $L_j=r_je_j$, and total expected loss is $L=\sum_j L_j$. If $D_j$ is the class D-ratio at the plan’s split point, the class primary and excess expected losses are $P_j=D_jL_j$ and $E_j=(1-D_j)L_j$. Aggregate component amounts are $P=\sum_jP_j$ and $E=\sum_jE_j$.

A D-ratio is the proportion of expected loss assigned to the primary component at the specified split point. It is between 0 and 1. The corresponding excess proportion is $1-D_j$. If a plan uses a single D-ratio for all classes, that same value is used for each class. When class-specific D-ratios are provided, calculate each class’s components before aggregating; applying one average D-ratio without a stated basis can change the allocation.

For a plan that specifies separate square-root weights and full-credibility expected-loss thresholds $K_P$ and $K_E$, calculate $Z_P=\min(1,\sqrt{P/K_P})$ and $Z_E=\min(1,\sqrt{E/K_E})$. The primary weight uses primary expected loss and its own threshold; the excess weight uses excess expected loss and its own threshold. A weight is capped at 1. Other plans may prescribe different formulas, so the written plan governs.

Why the Allocation and Weight Formulas Work

The D-ratio allocation follows from the definition of the ratio. For class $j$, $D_j$ is the share of expected loss in the primary layer, so multiplying total class expected loss by that share gives the expected primary amount. The remaining share is $1-D_j$. Therefore:

$$ P_j+E_j=D_jL_j+(1-D_j)L_j=L_j. $$

This identity is both the justification for the allocation and a useful check: allocating a class’s expected loss to its two components must not create or lose expected loss.

The square-root weight is a plan rule motivated by how relative statistical fluctuation can decline with volume. For independent, comparable exposure units with finite variance and a stable loss distribution, aggregate expected loss grows in proportion to exposure while the standard deviation grows in proportion to the square root of exposure. Thus relative fluctuation is proportional to the reciprocal of the square root of volume. A square-root credibility rule reflects this relationship, with $K$ setting the volume at which the plan assigns full weight. It is not a universal consequence of every loss model; use it here only when the plan specifies it.

Z=\min\left(1,\sqrt{\frac{L}{K}}\right)

Here $L$ is the expected-loss volume for the component being weighted, not necessarily the total expected loss. The cap means that a component above its threshold receives weight 1 rather than a weight greater than 1.

Worked Calculations

Worked Example: Two Classes and Separate Component Weights

A hypothetical plan has two classes. Class A has 1,200 exposure units and an expected-loss rate of $42 per unit; its D-ratio is 0.70. Class B has 800 exposure units and a rate of $65 per unit; its D-ratio is 0.55. The plan specifies $K_P=$$80,000 and $K_E=$$160,000 and uses square-root weights capped at 1. Calculate class and total expected losses, the aggregate primary and excess expected losses, and both weights.

Method and assumptions. Rates are dollars per exposure unit for the same period as the exposures. Use the stated class-specific D-ratios at the applicable split point, then use the plan’s separate component thresholds.

1. Calculate class expected losses. For Class A, $L_A=1{,}200\times\$42=\$50{,}400$. For Class B, $L_B=800\times\$65=\$52{,}000$. Thus $L=\$50{,}400+\$52{,}000=\$102{,}400$.

2. Allocate Class A. Its primary expected loss is $P_A=0.70\times\$50{,}400=\$35{,}280$. Its excess expected loss is $E_A=0.30\times\$50{,}400=\$15{,}120$. The check is $\$35{,}280+\$15{,}120=\$50{,}400$.

3. Allocate Class B. Its primary expected loss is $P_B=0.55\times\$52{,}000=\$28{,}600$. Its excess expected loss is $E_B=0.45\times\$52{,}000=\$23{,}400$. The check is $\$28{,}600+\$23{,}400=\$52{,}000$.

4. Aggregate the components. Primary expected loss is $P=\$35{,}280+\$28{,}600=\$63{,}880$. Excess expected loss is $E=\$15{,}120+\$23{,}400=\$38{,}520$. Together, $\$63{,}880+\$38{,}520=\$102{,}400=L$.

5. Calculate the weights. $Z_P=\sqrt{\$63{,}880/\$80{,}000}=\sqrt{0.7985}=0.8936$. $Z_E=\sqrt{\$38{,}520/\$160{,}000}=\sqrt{0.24075}=0.4907$. Neither exceeds 1, so neither is reduced by the cap. Rates are shown to four significant figures.

Check and common mistake avoided. The component totals reconcile to total expected loss, and the two weights use different component amounts and thresholds. Using the overall $L=\$102{,}400$ in both weight formulas would be wrong: the plan weights primary and excess separately.

Worked Example: A Component Reaches Full Credibility

A hypothetical equipment account has two classes. Class C has 500 exposure units, a rate of $90 per unit, and a D-ratio of 0.80. Class D has 300 units, a rate of $150 per unit, and a D-ratio of 0.40. The plan specifies $K_P=$$50,000 and $K_E=$$40,000 with separate square-root weights capped at 1. Find the total and component expected losses and weights.

Method and assumptions. Treat the rates as dollars per unit for the analyzed period. Apply each class’s own D-ratio before summing. The plan’s cap applies independently to each component weight.

1. Calculate class totals. Class C has $L_C=500\times\$90=\$45{,}000$. Class D has $L_D=300\times\$150=\$45{,}000$. Therefore total expected loss is $L=\$90{,}000$.

2. Split Class C. $P_C=0.80\times\$45{,}000=\$36{,}000$, and $E_C=0.20\times\$45{,}000=\$9{,}000$. These sum to $45,000.

3. Split Class D. $P_D=0.40\times\$45{,}000=\$18{,}000$, and $E_D=0.60\times\$45{,}000=\$27{,}000$. These also sum to $45,000.

4. Aggregate and reconcile. $P=\$36{,}000+\$18{,}000=\$54{,}000$, while $E=\$9{,}000+\$27{,}000=\$36{,}000$. Their sum, $90,000, equals the total expected loss.

5. Apply the plan rule. Before the cap, the primary value is $\sqrt{\$54{,}000/\$50{,}000}=\sqrt{1.08}=1.0392$. The cap makes $Z_P=1.0000$. The excess value is $\sqrt{\$36{,}000/\$40{,}000}=\sqrt{0.9}=0.9487$, so $Z_E=0.9487$. Weights are rounded to four decimals.

Check and common mistake avoided. The primary component exceeds its own threshold, but the excess component does not. Capping both weights at 1 just because one component is above threshold would be incorrect; each component is tested separately.

Worked Example: Class-Specific D-Ratios with Unequal Component Mix

A hypothetical portfolio has three classes: Urban, with 600 exposure units, a $30 rate, and D-ratio 0.75; Rural, with 400 units, a $50 rate, and D-ratio 0.60; and Coastal, with 200 units, an $80 rate, and D-ratio 0.25. A plan specifies $K_P=$$100,000 and $K_E=$$25,000, using square-root component weights capped at 1. Calculate all expected-loss amounts and weights.

Method and assumptions. The rates are dollars per exposure unit on a consistent basis. Use the supplied D-ratio for each class, rather than assuming the portfolio has one common ratio.

1. Calculate expected loss by class. Urban: $600\times\$30=\$18{,}000$. Rural: $400\times\$50=\$20{,}000$. Coastal: $200\times\$80=\$16{,}000$. Thus $L=\$18{,}000+\$20{,}000+\$16{,}000=\$54{,}000$.

2. Allocate Urban. Primary is $0.75\times\$18{,}000=\$13{,}500$; excess is $0.25\times\$18{,}000=\$4{,}500$.

3. Allocate Rural. Primary is $0.60\times\$20{,}000=\$12{,}000$; excess is $0.40\times\$20{,}000=\$8{,}000$.

4. Allocate Coastal. Primary is $0.25\times\$16{,}000=\$4{,}000$; excess is $0.75\times\$16{,}000=\$12{,}000$.

5. Aggregate and check. $P=\$13{,}500+\$12{,}000+\$4{,}000=\$29{,}500$. $E=\$4{,}500+\$8{,}000+\$12{,}000=\$24{,}500$. The component sum is $\$29{,}500+\$24{,}500=\$54{,}000$, matching $L$.

6. Calculate the weights. $Z_P=\sqrt{\$29{,}500/\$100{,}000}=\sqrt{0.295}=0.5431$. $Z_E=\sqrt{\$24{,}500/\$25{,}000}=\sqrt{0.98}=0.9899$. Neither raw weight is over 1. Results are rounded to four decimals.

Check and common mistake avoided. The excess weight is close to full credibility because excess expected loss is close to its $25,000 threshold, even though total expected loss is only $54,000. Weighting from total loss alone would ignore the component-specific volume the plan requires.

Process and Practical Checks

A disciplined order of operations makes a multi-class calculation auditable. In particular, retain intermediate class amounts so that a questionable aggregate can be traced to its rate, exposure, or D-ratio.

Confirm the basis.
Check that each rate and exposure use matching units, periods, and class definitions. A rate per $1,000 of payroll cannot be multiplied directly by payroll stated in dollars without conversion.

Calculate class expected losses.
For every class, multiply its rate by its exposure. Sum these amounts to obtain total expected loss.

Allocate each class.
Multiply its expected loss by its D-ratio for primary loss and by one minus the D-ratio for excess loss.

Aggregate and reconcile.
Sum the primary amounts and excess amounts separately. Confirm each class allocation and the overall component sum.

Apply the specified weight rule.
Use each component’s expected loss and its matching threshold, then apply any stated cap.

The most consequential checks are simple. First, verify $0\leq D_j\leq1$, so neither component is negative. Second, confirm $P_j+E_j=L_j$ for every class, and $P+E=L$ overall. Third, verify that a weight is dimensionless: dollars divided by dollars inside the square root. Finally, confirm that any cap is applied after calculating the raw weight.

Common Exam Trap: Averaging Before Allocating

A tempting shortcut is to average class D-ratios and multiply that average by total expected loss. Unless the average is weighted by each class’s expected loss, it generally will not preserve the correct component amounts. Even an expected-loss-weighted average is only a shortcut for the aggregate allocation; it does not replace class-level calculations when class component amounts are needed.

For example, in the three-class calculation above, the unweighted average D-ratio is $(0.75+0.60+0.25)/3=0.5333$. Multiplying $54,000 by this value gives about $28,800 of primary loss, not the correct $29,500. The classes have different expected losses, so equal weighting of their D-ratios is inappropriate. The correct aggregate D-ratio, if one is needed, is $P/L=\$29{,}500/\$54{,}000=0.5463$; calculating each class’s primary amount and summing produces it directly.

Another error is to treat the component weights as portions that must sum to 1. They are separate measures of credibility under the specified plan rule, not an allocation of a single weight. A primary weight of 0.89 and an excess weight of 0.49 can both be correct; each is determined by its own component volume and threshold.

Key takeaway: Calculate class expected losses first, allocate each class using its applicable D-ratio, reconcile primary plus excess to total, and only then calculate separate weights from the plan’s component thresholds. Do not substitute total expected loss for either component amount.

Check Your Understanding

For each problem, use the stated plan rule and retain the class-level allocation until the component totals are checked.

A hypothetical account has 250 exposure units at $48 per unit and a D-ratio of 0.65. Find total, primary, and excess expected loss. (Answer: Total $12,000; primary $7,800; excess $4,200.)
Class A has 300 units at $40 per unit and D-ratio 0.50. Class B has 200 units at $70 per unit and D-ratio 0.75. Find total, primary, and excess expected loss. (Answer: Total $26,000; primary $16,500; excess $9,500.)
A plan has two classes: 400 units at $25 per unit with D-ratio 0.80, and 300 units at $40 per unit with D-ratio 0.50. Its thresholds are $K_P=$$20,000 and $K_E=$$32,000. Calculate the component amounts and square-root weights, capped at 1. (Answer: Total $22,000; primary $14,000; excess $8,000; $Z_P=0.8367$, $Z_E=0.5000$.)
Class A has 1,000 units at $60 per unit and D-ratio 0.40; Class B has 500 units at $80 per unit and D-ratio 0.70. With $K_P=$$64,000 and $K_E=$$36,000, calculate the component weights under the capped square-root rule. (Answer: Total $100,000; primary $52,000; excess $48,000; $Z_P=0.9014$, $Z_E=1.0000$.)
Three classes have expected losses of $10,000, $15,000, and $25,000, with D-ratios 0.90, 0.60, and 0.20, respectively. The plan thresholds are $K_P=$$50,000 and $K_E=$$30,000. Find total and component expected losses and the two capped weights. (Answer: Total $50,000; primary $23,000; excess $27,000; $Z_P=0.6782$, $Z_E=0.9487$.)

What You'll Learn