Buhlmann Credibility
Buhlmann credibility estimates a risk by blending its own observed experience with the collective mean. The credibility weight increases when the risk has more exposure or less process noise.
Quick Answer
Buhlmann credibility is the cleanest credibility formula: estimate equals Z times the risk's observed mean plus 1-Z times the collective mean. The model decides Z from a signal-to-noise ratio, not from judgment alone.
The Spring 2026 ASTAM syllabus names Buhlmann and Buhlmann-Straub models inside the credibility block, and the FAM syllabus introduces credibility as a short-term insurance topic. That makes Buhlmann the bridge between basic credibility intuition and advanced ASTAM calculation.
Core Formula
For one risk with n observations, let m be the collective mean, EPV be expected process variance, and VHM be variance of hypothetical means. The Buhlmann credibility premium is a weighted average of the risk's sample mean and m.
K is the ratio EPV / VHM. A smaller K means the collective contains more between-risk signal relative to within-risk noise, so credibility rises faster as n increases.
Worked Example
A class has five years of experience with average claim cost 1,200. The collective mean is 900. Suppose K = 15. Then Z = 5 / (5 + 15) = 0.25.
The credibility estimate is 0.25(1,200) + 0.75(900) = 975. The class moves upward from the collective mean, but it does not receive full credibility because five observations are not enough to ignore process variance.
Interpretation
Buhlmann credibility is actuarial shrinkage. A high individual average is pulled down toward the collective mean unless the risk has enough exposure to earn a high Z. A low individual average is pulled up for the same reason.
The model is not saying the collective mean is always right. It is saying the observed individual mean contains both signal and random noise, and the balance depends on EPV, VHM, and exposure.
Common Traps
Trap 1: treating K as a fixed magic constant. K comes from EPV and VHM, so it changes when the portfolio structure or variance assumptions change.
Trap 2: confusing Buhlmann with Buhlmann-Straub. Buhlmann gives equal weight to each observation; Buhlmann-Straub lets exposures vary across observations.
Trap 3: reporting only the estimate. ASTAM-style answers usually need the reason: how much individual experience was trusted and why.