Variant of Bornhuetter-Ferguson that does not take the a priori loss ratio from expert judgment but computes it from the triangle itself, weighted by already developed exposure.
Bornhuetter-Ferguson is worth exactly what its a priori loss ratio is worth, and that a priori usually comes from pricing, that is, from the very source reserving is supposed to check. Cape Cod removes the dependency by estimating a single loss ratio from the triangle itself: divide the sum of reported cumulatives across all accident years by the sum of premiums weighted by each year's percent developed. A year developed to 15% contributes 15% of its premium to the denominator, which stops recent immature years from dragging the loss ratio down. The Bornhuetter-Ferguson mechanics are then applied with that loss ratio. The method assumes every accident year shares the same expected loss ratio, an assumption that is false across an underwriting cycle: in a market that has moved from soft to hard, Cape Cod blends years priced at 95% and at 62% and returns an average that describes neither. The decay-weighted variant, which also weights by recency, softens the flaw.
Professional liability book as of December 31, 2025, six accident years. Total reported cumulatives 52.4M EUR, total premium 96M EUR, weighted developed exposure 71.8M EUR. Cape Cod loss ratio = 52.4 / 71.8 = 73%, against a pricing a priori of 66% that had been carried for three closings. The seven-point gap, applied to the undeveloped share, adds 4.9M EUR of reserve and triggered a rate review for the line.
Stanard-Buhlmann, méthode de Stanard-Bühlmann, Cape Cod reserving, loss ratio pondéré par l'exposition développée