Why Overall Rate Changes Need Weights
A rate indication may differ by class: one class may need an increase while another needs a decrease. To summarize those indications for a portfolio, the relevant question is not how many classes have each result. It is how much current premium each class represents. A change affecting a class with a large current premium volume generally changes aggregate premium more than the same percentage change affecting a small class.
The calculation in this tutorial assumes that class indications are already expressed as percentage changes from current rates. It focuses on combining those indications, not deriving them. The exposures and mix represented by the current premiums are held fixed when calculating the aggregate result. No interest rate, mortality assumption, or loss distribution is needed for this aggregation.
A positive \(r_j\) is an increase; a negative \(r_j\) is a decrease. For example, an indicated increase of 6% means \(r_j=0.06\). If the class indication is instead stated as a new rate factor, such as 1.06, its percentage change is the factor minus 1, or 0.06.
Deriving the Aggregate Formula
The formula follows by comparing the aggregate premium before and after the indicated changes. For the derivation, assume the class premium amounts \(P_j\) are the current premiums for the same portfolio and that the indicated change \(r_j\) applies uniformly to the current premium in class \(j\). There are no changes in exposure mix or other adjustments between the two totals.
After applying each class’s indicated percentage change, its premium becomes \(P_j(1+r_j)\). Thus the new aggregate premium is the sum of the changed class premiums, and the overall percentage change is the increase in aggregate premium divided by current aggregate premium:
This also explains why the weights sum to 1: \(\sum_j w_j=\sum_j P_j/\sum_j P_j=1\). The weighted average is therefore a percentage change relative to the current total premium. The formula does not say that the rate changes themselves are necessarily correct; it only aggregates the supplied class indications on the stated premium basis.
Use current premiums for the same portfolio and period, with each class premium measured consistently.
Use decimals in calculations; subtract 1 from a rate factor to obtain its percentage change.
Multiply its current premium by its indicated percentage change.
Sum the class premium changes and divide by total current premium. This gives the overall percentage change.
Worked Applications
Worked Example: Combining Three Class Indications
A hypothetical insurer has current annual premiums of $1,200,000 in Class A, $600,000 in Class B, and $200,000 in Class C. The indicated rate changes are +8%, +2%, and −5%, respectively. Find the premium-weighted overall rate change and the new aggregate premium if the indications are applied to the same portfolio.
Method and assumptions. Weight each class change by current class premium. Treat the class premium amounts as fixed exposure weights for this calculation; no interest, mortality, or loss-distribution assumption is involved.
Step 1: Find total current premium.
Step 2: Calculate each class’s dollar premium change. Class A contributes \(1{,}200{,}000(0.08)=\$96{,}000\). Class B contributes \(600{,}000(0.02)=\$12{,}000\). Class C contributes \(200{,}000(-0.05)=-\$10{,}000\).
Step 3: Aggregate. The net premium change is \(96{,}000+12{,}000-10{,}000=\$98{,}000\). Therefore \(R=98{,}000/2{,}000{,}000=0.049\), or +4.90%.
Step 4: Check using the new aggregate premium. The new class premiums are $1,296,000, $612,000, and $190,000. Their sum is $2,098,000, which is $98,000 above the original $2,000,000. The implied change is again 4.90%.
Common mistake avoided: The unweighted average, \((8\%+2\%-5\%)/3=1.67\%\) (rounded to two decimals), gives each class equal influence despite their very different premium volumes. It does not reproduce the aggregate premium change.
Worked Example: Why Policy Counts Are Not the Weights
For a hypothetical portfolio, Class A has 1,200 policies with average current premium $600 and an indicated increase of 4%. Class B has 600 policies with average current premium $300 and an indicated increase of 12%. Class C has 200 policies with average current premium $500 and an indicated decrease of 3%. Calculate the overall rate change.
Method and assumptions. First convert policy counts and average premiums into current class premiums. The count alone is not the weight. Assume the indicated changes apply to these current premiums and that the policy mix is held fixed.
Step 1: Calculate current premium by class. Class A has \(1{,}200(600)=\$720{,}000\). Class B has \(600(300)=\$180{,}000\). Class C has \(200(500)=\$100{,}000\). Total current premium is \(720{,}000+180{,}000+100{,}000=\$1{,}000{,}000\).
Step 2: Calculate class premium changes. Class A: \(720{,}000(0.04)=\$28{,}800\). Class B: \(180{,}000(0.12)=\$21{,}600\). Class C: \(100{,}000(-0.03)=-\$3{,}000\).
Step 3: Aggregate and divide. The net change is \(28{,}800+21{,}600-3{,}000=\$47{,}400\). Hence \(R=47{,}400/1{,}000{,}000=0.0474\), or +4.74%.
Check. The changed class premiums are $748,800, $201,600, and $97,000, for a total of $1,047,400. The increase of $47,400 on $1,000,000 is 4.74%.
Common mistake avoided: A policy-count-weighted average would give each policy equal influence, but policies in these classes have different average premiums. The correct weights are based on class premium: 72%, 18%, and 10%.
Worked Example: Converting Rate Factors Before Weighting
A hypothetical two-class portfolio has current premiums of $400,000 and $600,000. The proposed rate factors are 1.08 for the first class and 0.97 for the second. Find the overall rate change and revised aggregate premium.
Method and assumptions. Convert each rate factor to a percentage change by subtracting 1, then apply that change to current class premium. Assume the same portfolio is being repriced; no other adjustments are made.
Step 1: Convert factors to changes. The first class change is \(1.08-1=0.08\), or +8%. The second is \(0.97-1=-0.03\), or −3%.
Step 2: Find the current total. \(400{,}000+600{,}000=\$1{,}000{,}000\).
Step 3: Find dollar changes. The first class change is \(400{,}000(0.08)=\$32{,}000\). The second is \(600{,}000(-0.03)=-\$18{,}000\). The net increase is $14,000.
Step 4: Calculate and check. The overall change is \(14{,}000/1{,}000{,}000=0.014\), or +1.40%. The revised class premiums are $432,000 and $582,000, totaling $1,014,000, which confirms the result.
Common mistake avoided: Weighting the factors 1.08 and 0.97 and reporting the resulting average as a percentage change confuses a rate factor with a rate change. The weighted factor here is 1.014; the corresponding percentage change is \(1.014-1=1.40\%\).
Interpretation, Limits, and Exam Traps
The weights describe the current premium distribution, not the number of classes and not necessarily the distribution of policies. A high-premium class can carry substantial weight even if it has relatively few policies. Conversely, a class with many low-premium policies may contribute less to the overall change than its policy count suggests.
The calculation is exact for the stated fixed portfolio when each class’s change applies proportionally to its current premium. It is not automatically a forecast of the premium change for a different future portfolio. If exposure mix changes, the class weights may change too; the future aggregate result must then use the relevant future premium basis or otherwise account for that change. Likewise, do not use current-premium weights from one basis with class indications measured on a different or inconsistent basis.
| Calculation choice | Use when | Why the alternative can fail |
|---|---|---|
| Current-premium weights | Aggregating class percentage changes for the same current portfolio | They reflect each class’s contribution to current aggregate premium |
| Equal weights by class | Only when class premiums are equal, or when an equal-class summary is specifically requested | It gives a small class the same influence as a large class |
| Policy-count weights | Only when the requested measure is explicitly count-weighted | Policy counts ignore differences in premium per policy |
| New-premium weights | When a question explicitly asks for that different weighting basis | They generally do not yield the change relative to current aggregate premium |
Check Your Understanding
For each problem, treat the stated class indications as applicable to the stated current premium basis. Unless a problem says otherwise, assume the portfolio mix is fixed.
- A hypothetical portfolio has $300,000 of current premium in Class A at +5%, $500,000 in Class B at +1%, and $200,000 in Class C at −4%. What is the overall rate change? (Answer: +1.20%)
- Two classes have current premiums of $250,000 and $750,000, with indicated changes of +12% and −2%. Calculate the net dollar premium change and the overall rate change. (Answer: $15,000 increase; +1.50%)
- Class A has 100 policies averaging $900 in current premium and an indication of +4%. Class B has 300 policies averaging $500 and an indication of +2%. Find the premium-weighted overall change. (Answer: +2.75%)
- Current premiums are $400,000 and $600,000, and the proposed rate factors are 1.08 and 0.97. What is the overall percentage change? (Answer: +1.40%)
- Current premiums are $500,000, $300,000, and $200,000, with class changes of −4%, +10%, and +5%. Calculate the fixed-portfolio overall change. Would that necessarily be the actual change for a future portfolio with a different class mix? (Answer: +2.00%; no, the changed mix may require different premium weights.)