franchise deductible
Picture a deductible with an all-or-nothing twist. As long as your loss stays below a threshold, you get nothing. But the moment it crosses that threshold, the insurer pays the entire loss — not just the part above the line. That is a franchise deductible: a trip-wire, not a slice that is always removed. It is common in marine cargo and some health and travel policies.
Formally, with a franchise deductible d, the insurer pays nothing if the loss X is at or below d, and pays the full X if X exceeds d. So with d = 500: a loss of 400 pays 0, but a loss of 600 pays the whole 600 (under an ordinary deductible it would pay only 100). The contrast is sharp: an ordinary deductible always keeps the first d; a franchise deductible gives the first d back to you once you have crossed the line. This creates an unusual jump in the payment right at the threshold.
Franchise deductibles still cut out the nuisance of small claims, but they treat anyone over the line generously, which has a peculiar incentive: a claimant with a loss just below d has a strong reason to push it over the threshold (real or exaggerated) to flip from zero payout to full payout. That cliff-edge invites manipulation and a spike of claims clustered just above d, so actuaries pricing a franchise must watch for this behaviour. The common misconception is to treat 'franchise' and 'ordinary' deductibles as the same thing differing only in wording — they produce materially different expected payments and very different incentives.
A marine cargo policy has a 1,000 franchise deductible. A 900 loss pays nothing. A 1,001 loss pays the full 1,001 — a single extra dollar of loss flips the payout from 0 to over a thousand. Compare an ordinary 1,000 deductible, where the 1,001 loss would pay only 1.
Below the threshold: nothing. Above it: the whole loss — a cliff edge at d.
The jump at the threshold creates an incentive to push a near-threshold loss over the line, since doing so changes the payout from zero to the full amount. This invites claim inflation and clustering just above d that pure-loss data may hide.