Why Reference Levels Need an Explicit Label
A multiplicative rating plan expresses a risk’s indicated premium as a base rate multiplied by the applicable rating factors. A factor table assigns a value to each category of a rating variable, such as territory or vehicle class. Within each factor, one category is the reference level: the category against which the other levels are expressed. A table should identify that level clearly, rather than requiring a reader to guess from the numbers.
The reference level is a role assigned to a category, not simply a category whose displayed factor equals 1.00. A non-reference category may also have a factor of 1.00, and a reference category may be missing from the table or mislabeled. Both the label and the factor values must be checked.
For a risk with base rate \(B\), suppose the applicable factors are \(r_1,\ldots,r_m\), one from each of \(m\) rating variables. The indicated premium is
This calculation assumes the stated rating plan is multiplicative, that the risk can be assigned to one category in each factor, and that the table supplies an unambiguous factor for every applicable category. It does not account for additional interactions or adjustments that are not included in the plan.
What a Correct Reference Entry Looks Like
A well-formed factor table makes three things easy to verify: the factor name, the full set of applicable category labels, and exactly which category is the reference. It is useful to put “Reference” in a separate column. The number 1.00 in the factor column is not a substitute for that label.
| Territory | Factor | Reference? |
|---|---|---|
| Central | 1.00 | Yes |
| Coastal | 1.12 | No |
| Inland | 0.94 | No |
Here Central is explicitly identified as the reference. Its factor is 1.00, while Coastal and Inland factors compare their indicated rating levels with Central, all else equal. The table has one row for each listed territory and no duplicate territory labels.
Worked Example: Checking a Complete Factor Table
A hypothetical insurer uses a base rate of $540. Its rating plan has three factors: territory, vehicle class, and annual mileage. The tables identify Suburban, Standard, and Typical, respectively, as the reference levels. A risk is Urban, Premium, and High mileage. Check the reference entries and calculate the indicated premium.
Method. First verify each factor’s reference label and its factor of 1.00. Then select one factor for the risk from each table and multiply by the base rate.
Territory has Suburban, the designated reference, at 1.00; Urban is 1.18 and Rural is 0.82. Vehicle class has Standard, the reference, at 1.00; Compact is 0.90 and Premium is 1.25. Annual mileage has Typical, the reference, at 1.00; Low is 0.85 and High is 1.20. Each factor has one explicitly designated reference, and every listed category has one entry.
The risk’s selected factors are Urban \(1.18\), Premium \(1.25\), and High mileage \(1.20\). The intermediate products are
and then
The indicated premium is $955.80. As a check, a risk in all three reference categories would have premium \(540 \times 1.00 \times 1.00 \times 1.00 = \$540.00\), exactly the base rate. The common mistake to avoid is treating a factor’s reference label as optional because its displayed value is 1.00; here the labels, not just the values, establish the intended comparisons.
Why the Reference Factor Is Set to One
The value 1.00 gives the reference level a neutral role in the product: it leaves the base rate unchanged when the risk is in that category and all other factors are also at their reference levels. It also fixes the scale used to express the other factors. The following short argument shows why an anchor is needed to interpret a factor table.
Consider one factor \(r_k\) for a category \(k\), with the other selected factors held fixed. Let \(C\) be the product of the base rate and those other factors. Then the indicated premium is \(Cr_k\). If every factor in this table is multiplied by a positive constant \(c\), and \(C\) is divided by \(c\), the predicted premiums do not change:
Thus premiums alone do not determine a unique scale for the factor table and the base rate. Designating a reference category and setting its factor to 1.00 fixes that scale. This does not mean a second category cannot also have a factor of 1.00; it means only one category is designated as the anchor for that factor.
This distinction matters when reviewing an unfamiliar or exported table. A numerical value of 1.00 may arise for a non-reference category. Conversely, if the table does not identify a reference category, a reader should not infer the intended anchor solely from whichever entry happens to equal 1.00.
Diagnosing Missing, Duplicated, and Mismatched Entries
A reference-level check is part of a broader table audit. For each factor, compare the category labels and their meanings with the rating plan’s allowed categories. Then check the reference designation and look for duplicate or unmatched labels. A table that fails one of these checks may not support a reliable premium calculation.
List the categories the rating plan says can occur for this factor. Check that the table covers them using consistent labels.
Confirm that exactly one category per factor is explicitly identified as the reference. Do not count numerical values of 1.00 as labels.
A repeated category label may have two different factors or appear twice in an export. Establish which entry is authoritative before using the table.
Use a documented correction only when a reliable source establishes the intended category, factor, or reference. If the intended entry is ambiguous, stop and request clarification.
Worked Example: Repairing a Duplicated and Mismatched Table
A hypothetical rating plan has base rate $700. The policy system uses the service-use categories Pleasure, Commute, and Business. A table export reads: Pleasure 0.88; Commute 1.00, marked Reference; Business 1.22; Commute 1.00, marked Reference; Personal use 0.88. A policy is Urban, Pleasure use, and Mid-size. Its territory factor is 1.10 and its vehicle-size factor is 1.05. An authoritative table record confirms that Commute is the sole reference, that the second Commute row is an accidental duplicate, and that Personal use is a retired label for Pleasure. Audit the service-use table and calculate the premium after applying those documented corrections.
Method. Compare the exported labels to the policy system’s categories, count the reference designations, and use the authoritative record to resolve the duplicate and retired label. Then select one factor per rating variable.
Before correction, the export has two rows marked as the Commute reference, so it does not show exactly one reference entry. It also has a repeated Commute label and a Personal use label that does not match the policy system’s current categories. The table is not safe to use as exported. The authoritative record resolves the issues: retain one Commute row at 1.00 as the reference, remove the duplicate row, and remove the retired Personal use row. The resulting service-use entries are Pleasure 0.88, Commute 1.00 (Reference), and Business 1.22.
For the risk, select Urban at 1.10, Pleasure at 0.88, and Mid-size at 1.05. The premium calculation is
and
The indicated premium is $711.48. The wrong calculation would treat both repeated Commute rows as separate factors and, for a Commute risk, calculate \(700 \times 1.10 \times 1.00 \times 1.00 \times 1.05 = \$808.50\). That incorrectly applies the service-use variable twice. The correction is to use one factor for each rating variable, and to remove a duplicate only because the authoritative record confirms it is an accidental duplicate. Without that record, a reader should not choose between conflicting entries by guesswork.
Do Not Confuse a Reference Label with a Factor Value
A table can have exactly one designated reference even when another category also has a factor of 1.00. The designation tells the reader which category anchors comparisons; the number tells the reader the category’s indicated multiplicative value. Both pieces of information matter.
Worked Example: A Non-Reference Category Also Equals One
A hypothetical insurer has a base rate of $400. Its occupancy factor table is Owner 1.00 (Reference), Tenant 1.00, and Seasonal 1.30. Its roof-condition table is Sound 1.00 (Reference), Worn 1.15, and New 0.95. A policy is in Central territory, whose territory factor is 1.00, and is Tenant with a Worn roof. Check the reference labels and calculate the indicated premium.
Method. Use the explicit designations to identify the anchors; do not assume that every entry valued at 1.00 is a reference.
The occupancy table has one designated reference, Owner. Tenant also has factor 1.00, but it is not marked as the reference. The roof table has one designated reference, Sound. Central is given as the territory reference at 1.00. The selected factors are therefore Central 1.00, Tenant 1.00, and Worn 1.15.
The intermediate calculations are
and
The indicated premium is $460.00. The table remains interpretable because Owner, not Tenant, is explicitly named as the occupancy reference. The common mistake is to infer that Tenant must also be a reference because its factor equals 1.00. It is a non-reference category with the same numerical value; the displayed table provides no reason to change its label.
Common Exam Traps
- Assuming 1.00 identifies the reference. It does not. Look for an explicit designation and check that it occurs exactly once for each factor.
- Counting factors instead of variables. A duplicated row is not a second rating variable. Do not apply it twice in a premium product.
- Silently treating labels as synonyms. An old or alternate label may refer to a current category, but confirm that mapping rather than assuming it.
- Forcing a calculation from ambiguous data. If the reference designation or applicable factor cannot be determined, report the issue and request clarification instead of selecting a convenient value.
Check Your Understanding
For each problem, check both the reference designation and the arithmetic where a premium is requested.
- A hypothetical plan has a base rate of $300. Its region factor is North 1.00 (Reference), South 1.10, and West 0.95. Its construction factor is Masonry 1.00 (Reference) and Frame 1.20. A North, Frame risk has what indicated premium? (Answer: $360.00)
- A table lists Plan A 0.90, Plan B 1.00 (Reference), and Plan C 1.15, with Plan B also repeated as 1.00 (Reference). How many reference rows are shown, and can the table be accepted as written? (Answer: Two reference rows are shown; no, the duplicate must be resolved before the table is accepted.)
- A hypothetical factor table lists Low 0.80, Medium 1.00, and High 1.20, but does not identify a reference. Is Medium confirmed as the reference, and what should the reviewer do? (Answer: No; request confirmation of the intended reference rather than inferring it from the value 1.00.)
- A base rate is $500. A table identifies Standard use as the reference at 1.00 and lists Personal use at 0.90 and Commercial use at 1.25. The policy system uses Personal and Commercial, not the table’s full labels. Is the factor mapping sufficient without clarification? (Answer: No; confirm that the labels map to the policy-system categories before selecting a factor.)
- A hypothetical plan has a $250 base rate. The explicitly designated references are Zone 1 at 1.00 and Basic class at 1.00. Zone 2 is 1.08, and Advanced class is 1.30. A Zone 2, Basic risk has what indicated premium, and how many references are designated across these two factors? (Answer: $270.00, with two designated references total—one for each factor.)