Tail factor selection is the step in actuarial reserving where the actuary chooses a development factor to project claims from the oldest age observed in a claims triangle to their ultimate settled value. A triangle with ten years of history shows development only up to age ten; in long-tail lines such as liability or workers' compensation, claims may keep developing for decades, so the tail factor can be a material part of the reserve.
Common approaches are fitting a curve, such as exponential decay or an inverse power curve, to the observed age-to-age development factors and extrapolating it; comparing paid and incurred development, since incurred data often reach ultimate sooner; using longer company or industry data, such as published benchmark patterns; and examining open claims individually. The tail factor multiplies every origin year's projection in the chain ladder and affects Bornhuetter-Ferguson results, so it is a key assumption under assumption governance and is often tested with sensitivity analysis.