Why an Inflation Index Is Not a Claim
A claim amount recorded at one date may not be comparable with an amount at another date: prices, labor costs, and other relevant costs can change. An inflation index summarizes price movement over time and can be used to calculate a factor for relating index levels on two dates. But the index is not itself a claim, and its percentage change is not automatically the percentage change in claim severity.
This distinction matters when an actuary interprets historical amounts or makes a projection under an explicit assumption. First calculate what the index implies. Then state whether, and why, claim severity is being assumed to move in line with that index. A claim could change by a different percentage because its cost components, coverage, or mix differ from what the index measures.
Interpreting the Ratio
An index level depends on how the index is scaled. For example, the same index series might be reported with a reference level of 100 or rescaled to a reference level of 1. The levels differ, but the ratio between two dates does not, provided both values come from the same consistently scaled series. The ratio, rather than the difference between the index numbers, is the trend factor.
The percentage change in the index from \(s\) to \(t\) is the trend factor minus 1, expressed as a percentage:
A factor greater than 1 means the index increased; a factor below 1 means it decreased. A factor of 1 means no change in the index. The difference \(I_t-I_s\) is a number of index points, not a percentage change. It cannot be interpreted as a dollar increase in a claim.
To use the factor to project a claim amount, an additional modeling assumption is needed: the claim amount is assumed to move in proportion to the selected index over the specified dates. If a starting amount is \(C_s\), the corresponding projected amount under that assumption is \(C_sT_{s,t}\). This is a conditional calculation, not a general law about claims.
Why the Trend Factor Is a Ratio
The index measures relative price levels. To find how many times as large the target level is as the starting level, divide the target level by the starting level. This also shows why changing the index’s scale does not affect the trend factor.
Suppose every index value is multiplied by the same positive constant \(k\). The reported levels become \(kI_s\) and \(kI_t\).
The ratio using the rescaled levels is \(\frac{kI_t}{kI_s}\).
Since \(k\) is positive, \(\frac{kI_t}{kI_s}=\frac{I_t}{I_s}\). Thus the trend factor is unchanged by a consistent rebasing of the index.
The derivation establishes a property of the index ratio; it does not establish that claim severities follow the index. That requires a separate assumption or evidence.
Worked Calculations
Worked Example: Calculating a Factor and Projecting an Amount
A hypothetical repair-cost index is 125 at the start of a period and 140 at its end. A covered repair amount at the starting date is $6,250. Calculate the index trend factor and percentage change. Then find the projected amount if the repair amount is assumed to follow the index.
Method. Divide the target index by the starting index, convert the factor to a percentage change, and apply the factor to the amount only under the stated assumption. This avoids treating an index-point difference as either a percentage or a dollar claim increase.
Step 1: Calculate the factor. The starting value is \(I_s=125\), and the target value is \(I_t=140\).
Step 2: Calculate the index percentage change.
The index increased by 15 index points, from 125 to 140, but its percentage increase is 12%, not 15%.
Step 3: Project the repair amount under the assumption.
Check. The increase in the projected amount is $7,000.00 − $6,250.00 = $750.00, and \(\$750/\$6{,}250=0.12\). The amount therefore increased by 12%, matching the index only because the example explicitly assumes it follows the index.
Worked Example: Index Change Versus Observed Severity Change
A hypothetical claims portfolio has an average severity of $4,000 at a starting date and $4,520 at a later date. Over the same dates, a selected cost index increases from 200 to 230. Calculate the index trend factor and the observed percentage change in average severity. If the earlier average severity is projected using the index, compare that projection with the later observed average.
Method. Calculate the index ratio and the severity ratio separately, each relative to its own starting value. This avoids assuming that the index percentage change and the observed severity percentage change are identical.
Step 1: Calculate the index trend factor and index change.
The index increased by 30 points, which corresponds to a 15% increase from its starting level.
Step 2: Calculate the observed severity change. The average severity increased by \(\$4{,}520-\$4{,}000=\$520\). Relative to the starting severity, this is
Step 3: Calculate the index-based projection. If the $4,000 starting severity were assumed to follow the index, the projected severity would be
The index-based projection is $80.00 above the observed $4,520 average. The index increased by 15%, whereas the observed average severity increased by 13%; these are distinct calculations. The difference alone does not explain why they differ. The index may not track every cost component in the claims, and the two severity measurements must also be comparable in claim definition and mix for their change to be meaningfully interpreted.
Check. The observed increase is $520, which is 13% of $4,000. The projected increase is $600, which is 15% of $4,000. Their difference is $600 − $520 = $80, consistent with $4,600 − $4,520 = $80.
Worked Example: A Falling Index
A hypothetical index is 150 at the starting date and 144 at the target date. A claim amount at the starting date is $8,750. Calculate the trend factor and percentage index change, and find the amount implied if the claim is assumed to move in proportion to the index.
Method. Use the target-to-starting ratio even when the index falls, and apply the resulting factor only under the stated assumption. This avoids reporting a positive “increase” merely because both index values are positive.
Step 1: Find the trend factor.
Step 2: Interpret its percentage change.
The index fell by 6 points, corresponding to a 4% decrease from its starting level.
Step 3: Apply the factor conditionally.
Check. The projected reduction is $8,750 − $8,400 = $350, and \(\$350/\$8{,}750=0.04\). The projected amount is therefore 4% below the starting amount, as implied by the index under the assumption.
Conditions and Common Traps
A trend factor is meaningful only when the two index values refer to the same index series and a clearly specified starting and target date. If an index is reported monthly, for example, the chosen monthly observations should correspond to the dates or periods in the comparison. Using a mismatched observation can produce a correct ratio for the wrong period.
The index also needs to be relevant to the intended comparison. A general price index may not reflect the costs underlying a particular type of claim as closely as a specialized index. Even with a relevant index, applying the factor to claim severity remains an assumption unless supported by the modeling context. In particular, a change in average severity can reflect more than price movement; the claims being averaged may differ in other ways.
- Using the index-point difference as the percentage. An increase from 200 to 230 is 30 points but a 15% increase, because the starting level is 200.
- Reversing the ratio. For a factor from starting date \(s\) to target date \(t\), use \(I_t/I_s\). Reversing it gives the factor in the opposite direction.
- Calling the index change the observed severity change. The index ratio describes the index. A separate comparison of severity amounts is needed to calculate observed severity change.
- Applying the factor without stating an assumption. Multiplying a claim amount by the factor is a projection under an assumption that the claim follows the index, not a conclusion supplied by the index alone.
Check Your Understanding
For each problem, identify whether the requested percentage refers to the index or to observed claim severity. Apply a trend factor to a claim amount only when the problem supplies that assumption.
- A hypothetical index rises from 80 to 92. A $3,200 amount is assumed to follow the index. Find the factor, index percentage change, and projected amount. (Answer: Factor \(1.15\); index increase \(15\%\); projected amount $3,680.00.)
- A hypothetical index rises from 250 to 275. A $12,000 claim amount is assumed to follow it. State the index-point change and percentage change, then find the projected amount. (Answer: Increase of 25 index points; percentage increase \(10\%\); projected amount $13,200.00.)
- A hypothetical index falls from 160 to 152. A $7,600 amount is assumed to move in proportion to it. Find the factor, percentage change, and projected amount. (Answer: Factor \(0.95\); index decrease \(5\%\); projected amount $7,220.00.)
- A portfolio’s average severity rises from $5,100 to $5,355 while a selected index rises from 300 to 318 over the same dates. Find the observed severity percentage change and the index percentage change. (Answer: Observed severity increase \(5\%\); index increase \(6\%\).)
- A hypothetical index rises from 110 to 121. A $9,450 severity is assumed to follow the index, but the later observed severity is $10,300. Find the index-based projection, the observed severity percentage increase, and the difference between the projected and observed amounts. (Answer: Factor \(1.10\); projected amount $10,395.00; observed increase approximately \(8.99\%\); projection is $95.00 higher than observed.)