Calculating Expected Losses from Rates and Exposures

From an Expected Loss Rate to Expected Losses

An expected loss rate is useful only when it is applied to exposure measured on the same basis as the rate. A rate of dollars per $100 of payroll, for example, must be multiplied by the number of $100 units in the payroll—not by the payroll amount stated in dollars. A rate per policy-year must be multiplied by policy-years, not by payroll or by an unrelated count.

The calculation is simple once the units match, but unit mismatches can create errors by factors of 100, 1,000, or more. The objective is to preserve the meaning of the rate through every conversion and then check that the resulting expected loss is plausible in scale.

Definition: An expected loss rate is expected loss per stated unit of exposure. If $r$ is the rate per exposure unit and $e$ is the number of matching exposure units, then expected losses are $L_{\mathrm{exp}}=r e$. The rate and exposure must refer to the same coverage, classification, and exposure basis for this multiplication to be appropriate.

For a payroll-based rate, the exposure might be stated per $100 of payroll or per $1,000 of payroll. For a policy-based rate, it might be stated per policy-year. The denominator is part of the rate: $3.10 per $100 is not the same rate as $3.10 per $1,000.

The expected loss calculated here is a benchmark amount implied by the rate and exposure. It is not the observed losses for a particular account, nor is it automatically the account’s premium. Other plan features and premium components are separate matters.

Why the Product Gives Expected Losses

Suppose a fixed exposure consists of $e$ identical exposure units. Let $Z_j$ be the loss associated with unit $j$, for $j=1,\ldots,e$, and suppose each unit has expected loss $E[Z_j]=r$. Total loss is $L=\sum_{j=1}^{e}Z_j$. By linearity of expectation,

E[L] = E\left[\sum_{j=1}^{e}Z_j\right] = \sum_{j=1}^{e}E[Z_j] = \sum_{j=1}^{e}r = er.

This justification does not require the losses from different units to be independent; linearity of expectation is sufficient. It does require the expected loss per unit to be $r$ for the units being counted. If the rate is not appropriate to the exposure’s coverage or classification, the multiplication may be arithmetically correct but actuarially unsuitable.

In practice, the stated rate is a supplied expected loss rate, and the exposure is often a known aggregate amount. The calculation therefore uses the rate as the expected loss per unit rather than modeling each individual loss separately.

A Reliable Calculation Process

Before multiplying, write down the rate’s unit and the exposure’s unit. Convert the exposure so that its units cancel the denominator in the rate. Then multiply and label the result in dollars.

Read the rate denominator.
Record whether the rate is per $100 of payroll, per $1,000 of payroll, per policy-year, or another stated unit.

Convert the exposure to matching units.
For a rate per $100 of payroll, divide payroll dollars by 100. For a rate per $1,000, divide by 1,000. For a per-policy-year rate, count policy-years.

Multiply and check.
Multiply the rate by the matching exposure units. Confirm that the units reduce to dollars and that the answer has a plausible scale.

Worked Calculations

Worked Example: Payroll Rate per $100

A hypothetical classification has an expected loss rate of $2.40 per $100 of payroll. An account has $1,275,000 of payroll in that classification. Calculate expected losses and check the unit conversion.

Method. Convert payroll dollars into $100 exposure units, then multiply by the rate. The rate denominator is $100, so dividing payroll by $1,000 or multiplying the rate by the unconverted payroll would not be appropriate.

Step 1: Convert payroll. The number of $100 units is

\frac{\$1{,}275{,}000}{\$100}=12{,}750\text{ units}.

Step 2: Multiply.

12{,}750\text{ units}\times \frac{\$2.40}{\text{unit}} = \$30{,}600.

Check. The units cancel to dollars. As a scale check, $1,275,000 of payroll contains 12,750 groups of $100, and 12,750 multiplied by $2.40 is $30,600. The expected loss is therefore $30,600.

Common mistake avoided. Multiplying $2.40 by $1,275,000 treats the rate as dollars per dollar of payroll and overstates the result by a factor of 100.

Worked Example: Policy-Year Rate and Policy Count

A hypothetical program uses an expected loss rate of $184.60 per policy-year. An account has 137 policies, each in force for one full year on the exposure basis used by the rate. Calculate the expected losses.

Method. Because the rate is stated per policy-year and each policy contributes one policy-year, convert the count to policy-years before multiplying.

Step 1: Determine exposure.

137\text{ policies}\times 1\frac{\text{policy-year}}{\text{policy}} = 137\text{ policy-years}.

Step 2: Multiply by the rate.

137\text{ policy-years}\times \frac{\$184.60}{\text{policy-year}} = \$184.60(137) = \$25{,}290.20.

The multiplication can be checked in parts: $184.60(100)=18{,}460.00$, $184.60(30)=5{,}538.00$, and $184.60(7)=1{,}292.20$. Their sum is $25{,}290.20$. Thus expected losses are $25,290.20.

Common mistake avoided. A policy count is not always a policy-year count. Here they are numerically equal only because every policy contributes one full policy-year, as the problem states. Do not assume that equality when the exposure basis or policy duration differs.

Worked Example: Payroll Rate per $1,000

A hypothetical rate is $7.85 per $1,000 of payroll. The payroll exposure is $2,460,000. Calculate expected losses, then verify the calculation using an equivalent per-$100 rate.

Method. First match the payroll to the stated $1,000 denominator. For the verification, express the same rate per $100 and use the corresponding number of $100 units.

Step 1: Convert payroll to $1,000 units.

\frac{\$2{,}460{,}000}{\$1{,}000} = 2{,}460\text{ units}.

Step 2: Calculate expected losses.

2{,}460\times \$7.85 = \$19{,}311.00.

For the multiplication, $7.85(2{,}000)=15{,}700$, $7.85(400)=3{,}140$, and $7.85(60)=471$. These sum to $19{,}311$.

Step 3: Check with a per-$100 rate. A $1,000 payroll unit contains ten $100 units, so the equivalent rate per $100 is $7.85/10=\$0.785$. The payroll contains $2{,}460{,}000/100=24{,}600$ $100 units. Therefore,

24{,}600\times \$0.785 = \$19{,}311.00.

Both methods give $19,311.00. No intermediate rounding was needed.

Common mistake avoided. Using 24,600 units with the original $7.85 per $1,000 rate mixes two denominators. That would count each $1,000 of payroll ten times without reducing the rate accordingly.

Unit and Reasonableness Checks

A sound answer should pass both a dimensional check and a scale check. In a payroll calculation, the exposure units should be expressed as payroll dollars divided by payroll dollars per unit; multiplying those units by dollars per unit leaves dollars. In a policy-year calculation, policy-years multiplied by dollars per policy-year also leaves dollars.

The scale check asks whether the conversion factor makes sense. A rate per $100 applied to payroll stated in dollars requires division by 100. A rate per $1,000 requires division by 1,000. If the rate is per policy-year and the policies each represent one policy-year, the expected losses should be the policy count multiplied by the per-policy-year rate.

Keep full precision through the multiplication unless the problem specifies a rounding convention. Report money to the cent when appropriate. If the rate itself is supplied to a limited number of decimal places, do not invent additional precision; the result is based on the given rate.

Key takeaway: Expected losses equal expected loss rate multiplied by exposure only after the exposure has been expressed in the rate’s stated units. Write the units beside the numbers; they often reveal a conversion error before it reaches the final answer.

Common Exam Traps

Using payroll dollars as exposure units. With a rate per $100 of payroll, convert payroll to $100 units before multiplying.
Missing a factor of ten or one hundred. Rates per $100 and per $1,000 differ in denominator, even when their displayed dollar amounts look similar.
Equating policy count with policy-years without support. This is valid in the example above because each policy is explicitly one full year on the stated basis; the count alone does not establish that condition.
Reporting the wrong quantity. The result is expected losses implied by the selected rate and exposure. It should not be relabeled as observed losses or total premium without additional information.

Check Your Understanding

For each problem, show the exposure conversion before multiplying by the rate.

A hypothetical rate is $4.20 per $100 of payroll. Payroll is $860,000. What are the expected losses? (Answer: $36,120)
A hypothetical program has an expected loss rate of $315 per policy-year and 124 policies, each contributing one policy-year. Calculate expected losses. (Answer: $39,060)
A rate is $1.26 per $100 of payroll, and payroll is $2,375,000. Calculate expected losses. (Answer: $29,925)
A hypothetical policy-count rate is $22,400 per 100 policies. The exposure is 375 policies on the same basis. Convert to rate units and calculate expected losses. (Answer: $84,000)
A hypothetical rate is $9.60 per $1,000 of payroll. Payroll is $1,875,000. Calculate expected losses and verify that the exposure is expressed in $1,000 units. (Answer: $18,000)

What You'll Learn