Why Sequential Adjustments Must Be Compounded
A rate change or a trend rate describes a percentage change from one level to another. When several changes occur in sequence, each percentage applies to the level reached after the preceding change. As a result, the calculation is multiplicative: the first change affects the amount to which the second change applies.
This matters when carrying a historical amount through successive rate changes or trend intervals. Adding the percentages can be a tempting shortcut, but it generally treats each change as though it were calculated from the same starting amount. That is not what sequential percentage changes mean.
The same rule applies to successive trend intervals. If a cost level increases by 4% in one year and by 3% in the next, the two-year factor is \(1.04 \times 1.03\). When the annual trend rate differs from interval to interval, use the rate that applies to each interval rather than treating all years as having one average rate. The calculation assumes that each stated rate applies to the level at the start of its interval.
Rate Factors and Trend Factors Have Different Roles
The multiplication rule is the same for rate changes and trend rates, but the factors apply to different quantities in a ratemaking calculation. A rate-level factor adjusts premium from one rate level to another. A trend factor adjusts a loss or loss cost from one cost level to another. Keep track of which quantity each factor changes before combining them.
If a rate level changes by successive percentages, the cumulative factor from the initial level to the final level is the product of the rate-change factors. For instance, increases of 8% and then 4% produce \(1.08 \times 1.04=1.1232\): the final level is 12.32% above the initial level, not 12%.
For a sequence of trend intervals with rates \(t_1,t_2,\ldots,t_n\), the cumulative trend factor is:
For sequential rate changes \(c_1,c_2,\ldots,c_m\), where a decrease is entered as a negative percentage, the cumulative rate-level factor is:
In these formulas, each factor maps the level at the start of one step to the level at the end of that step. Therefore, the order in which the factors are multiplied does not change the numerical product, but the order of the underlying changes still matters for determining which factors apply. A listed change that occurred before the starting level is not part of the adjustment.
Worked Example: Three Successive Rate Changes
A hypothetical policy has a premium of $1,200 at a starting rate level. The applicable rate changes are an increase of 7%, then an increase of 5%, then a decrease of 3%. Calculate the premium at the final level, the cumulative factor, and the net percentage change. Use the sequential-change method.
Step 1: Convert each change to a factor. The increases give factors of \(1.07\) and \(1.05\). The 3% decrease gives a factor of \(1-0.03=0.97\).
Step 2: Multiply the factors.
The cumulative rate-level factor is 1.089795. The final level is therefore \(1.089795-1=0.089795\), or 8.9795%, above the starting level.
Step 3: Apply the factor to the starting premium.
Rounded to the nearest cent, the final premium is $1,307.75. As a check, the successive amounts are $1,200, then \(1{,}200 \times 1.07=\$1{,}284\), then \(1{,}284 \times 1.05=\$1{,}348.20\), and finally \(1{,}348.20 \times 0.97=\$1{,}307.754\). The direct calculation agrees.
Common mistake avoided: adding \(7\%+5\%-3\%=9\%\) gives a factor of 1.09 and a premium of $1,308.00. That is not the correct sequential result because the second and third changes apply to changing rate levels, not each to the original $1,200.
Combining Trend and On-Level Premium in a Loss Ratio
A combined calculation must respect the direction of each adjustment. Consider a historical loss ratio, defined as historical losses divided by historical premium. To adjust it for a loss-cost trend factor \(T\) and a premium on-level factor \(R\), losses are multiplied by \(T\), while premium is multiplied by \(R\). If the same historical losses and premium are being adjusted, the resulting ratio is:
This formula assumes the trend factor applies to the historical losses and the rate-level factor applies to the historical premium. It also assumes the starting loss ratio and both factors are defined on compatible bases. Do not multiply a loss ratio by both factors in the same direction: one adjusts the numerator, and the other adjusts the denominator.
Worked Example: Trend Losses and On-Level Premium Together
A hypothetical experience period has a historical loss ratio of 65%. The losses are to be trended over three successive annual intervals at 4%, 3%, and 2%. The premium rate level is to be adjusted through a rate increase of 8% followed by a decrease of 3%. Calculate the adjusted loss ratio using the combined-factor method.
Step 1: Calculate the cumulative trend factor.
Step 2: Calculate the cumulative premium rate-level factor.
The trend factor raises the loss level by a cumulative 9.2624%, while the premium rate-level factor raises the premium level by a cumulative 4.76%.
Step 3: Apply the factors in the correct directions. Convert 65% to 0.65 for the calculation:
As a percentage, the adjusted loss ratio is 67.79%, rounded to two decimal places. The numerator check is \(0.65 \times 1.092624=0.7102056\); dividing by the premium factor of 1.0476 gives the displayed result.
Common mistake avoided: multiplying the historical loss ratio by both \(T\) and \(R\) would give the premium increase the wrong effect. Since premium is the denominator of a loss ratio, its factor belongs in the denominator of the adjustment, as \(T/R\).
A Step-by-Step Method for Factor Chains
For a calculation with several changes, a short sequence of checks prevents most compounding errors. Identify the starting and ending levels first. Then make a list of the changes that actually move the amount between those levels. This avoids including a rate change or trend interval that is outside the specified path.
State whether the factor changes losses, loss costs, premium, or a ratio, and identify the level being moved from and to.
Use \(1+r\) for an increase of \(r\) and \(1-r\) for a decrease of \(r\). Do not turn a percentage decrease into a negative factor.
Use one factor for each successive change or interval included in the adjustment.
Multiply the amount being adjusted by its factor. For a loss ratio, place a premium factor in the denominator.
Compare the direction and approximate size with the changes. A sequence of mostly increases should usually produce a final amount above its starting level.
Why Adding Percentages Fails
Adding percentages works exactly only in special cases, such as when all changes are stated as amounts based on the same fixed starting value rather than as sequential percentage changes. In ordinary sequential adjustments, each percentage is applied to the current level. For example, a 10% increase followed by a 10% decrease has factor \(1.10 \times 0.90=0.99\), so the final amount is 1% below the start. Adding the changes suggests no net change, which is wrong.
The discrepancy can be small when the percentages are small, but a small discrepancy can still matter in a rate indication. It can also grow across more intervals. The reliable method is not to estimate the effect by summing percentages: convert each change to a factor, multiply, and calculate the net percentage only at the end by subtracting 1 from the cumulative factor.
Worked Example: Different Trend Rates Across Three Years
A hypothetical loss cost is $240 at the beginning of a three-year projection. The selected trend rates are 2% in the first year, 6% in the second year, and 4% in the third year. Calculate the projected loss cost and compare it with the result from incorrectly adding the rates.
Step 1: Form the annual factors. The factors are 1.02, 1.06, and 1.04.
Step 2: Compound the intervals.
The cumulative trend factor is 1.124448, a net increase of 12.4448%.
Step 3: Project the loss cost.
Rounded to the nearest cent, the projected loss cost is $269.87. The interval-by-interval check is \(240 \times 1.02=\$244.80\), then \(244.80 \times 1.06=\$259.488\), then \(259.488 \times 1.04=\$269.86752\).
Adding the rates gives 12%, which would project \(240 \times 1.12=\$268.80\). That shortcut understates the correctly compounded projection by $1.07 when rounded to the nearest cent. The difference arises because the second and third annual changes apply to the cost level after the earlier trends.
Exam Traps and Practical Checks
Several recurring errors are really failures to track the base to which a percentage applies. A rate decrease after an increase is applied to the increased level; it does not simply erase the earlier increase by subtracting the same number of percentage points. Likewise, if losses are trended and premium is on-leveled, the two factors do not both affect the loss ratio in the same direction.
- Adding sequential percentages. Add only when the problem explicitly defines changes against a common fixed base. Otherwise, multiply the factors.
- Using the wrong decrease factor. A decrease of 4% gives \(0.96\), not \(1.04\) and not \(-0.04\).
- Applying a factor to the wrong amount. A trend factor changes the loss or cost level; a rate-level factor changes premium.
- Reversing the loss-ratio adjustment. If losses are trended by \(T\) and premium is adjusted by \(R\), the adjusted ratio uses \(T/R\).
- Rounding too early. Keep the full product of the factors through the calculation, then round the final dollar amount or percentage as requested.
The central discipline is to label each factor before multiplying it. A factor chain is meaningful only if each factor applies to the amount produced by the preceding step. Used this way, compounding gives a consistent calculation for successive rate changes, successive trend intervals, and a loss ratio adjusted for both trend and premium rate level. These techniques are aligned with the CAS Exam 5 (Basic Ratemaking and Estimated Claim Liabilities) syllabus.
Check Your Understanding
For each problem, calculate with unrounded factors and round dollar amounts to the nearest cent and percentages to two decimal places.
- A hypothetical premium of $900 is subject to a 6% increase, a 2% decrease, and then a 4% increase. What is the cumulative factor and final premium? (Answer: factor \(1.080352\); premium $972.32)
- A hypothetical loss cost of $1,200 is trended through three successive annual intervals at 2%, 5%, and a 1% decrease. What is the cumulative trend factor and projected loss cost? (Answer: factor \(1.06029\); loss cost $1,272.35)
- A historical loss ratio is 72%. Losses are trended by 6% and then 4%; premium is on-leveled by a 5% increase followed by a 2% decrease. What is the adjusted loss ratio? (Answer: 77.14%)
- A hypothetical loss cost of $500 is subject to trend rates of 8%, 8%, and a 4% decrease in successive years. Calculate the compounded projection, then state why simply adding the rates gives a different answer. (Answer: factor \(1.119744\); projected cost $559.87; adding the rates gives 12% and $560.00, which is not the sequential result.)
- A historical loss ratio is 60%. The loss trend factors are 1.03 and 1.05, and successive premium changes are increases of 4% and 6%. Calculate the adjusted loss ratio. (Answer: 58.86%)