rating algorithm

All the ingredients of pricing — base rates, relativities, deductible factors, fees — are useless until something tells you the exact order to combine them in. The rating algorithm is that precise, step-by-step recipe: the formula a quoting system follows to turn one customer's answers into one final premium. It is the operational heart of a rate filing, the part the computer actually runs.

A typical algorithm reads like an ordered set of instructions: start with the base rate for the coverage; multiply by each applicable classification relativity (driver, vehicle, territory); apply the deductible relativity and any increased-limit factor; apply credits and surcharges (multi-policy discount, prior-claims surcharge); add fixed expense fees; then cap the result if a maximum change rule applies. Order matters because the plan can be multiplicative (factors multiply, the common case), additive, or a mix, and a discount applied before versus after another step gives different answers. For example: base 500 × territory 1.30 × age 0.90 × deductible 0.92, then add a 25 policy fee, gives 500 × 1.30 × 0.90 × 0.92 + 25 ≈ 563.

The rating algorithm is what makes a rate filing reproducible and auditable: a regulator, an agent's quoting screen, and the actuary must all arrive at the identical premium for the same risk. It is also where pricing theory meets messy reality — rounding rules, minimum premiums, fees, and rate-change caps all live in the algorithm and can quietly move the real price away from the pure indicated rate. A subtle but important point: because of capping and minimums, the average premium the algorithm actually produces may not equal the average the actuary intended, which is why an off-balance check is run after the algorithm is set.