Schnieper method

In 1991 Schnieper published "Separating true IBNR and IBNER claims" in the ASTIN Bulletin, a method that splits the cost of newly reported claims from the deterioration of claims already known. He started from a reinsurer's problem: on an excess of loss treaty, late emerging cost comes from two sources that chain ladder adds together without telling them apart. There are the claims nobody knew about, incurred and not yet reported, the true IBNR. There are known files getting worse, the IBNER, often claims open at the cedant below the retention that pierce it in year three or year six, when the medical report or the judgment comes in. So Schnieper builds two triangles from the same history. The first holds the cost of new claims in each development year, made proportional to the accident year's exposure (subject premium, vehicle count) through a rate specific to each age. The second tracks the change in incurred on claims already known, proportional to the known cumulative of the previous column, a chain ladder restricted to those files. You can then see what moves the reserve, files arriving or files being revalued, and young accident years are projected on exposure rather than on a nearly empty cumulative. The price is claim-level data, with the first notification date of every claim, which many reinsurance bordereaux do not supply cleanly. And the reporting threshold to the reinsurer, often a fraction of the retention, blurs the line between new and known. Liu and Verrall derived predictive distributions from the model in 2009. Not to be confused with IBNER itself, which is a component of the reserve, nor with Bornhuetter-Ferguson, which also leans on exposure but does not separate the two sources.