Exam STAM: Short-Term Actuarial Mathematics Study Guide
Frequency and severity models, aggregate loss models, credibility theory, ratemaking, and reserving.
Overview
Exam STAM focuses on short-term actuarial mathematics used in property and casualty insurance. It covers frequency and severity modeling, aggregate loss distributions, credibility theory, ratemaking, and reserving. This exam is central to the CAS track and requires comfort with both statistical modeling and insurance operations.
Exam Format
- Duration
- 210 minutes (3.5 hours)
- Questions
- 35 multiple-choice questions
- Pathway
- CAS
- Format
- Computer-based testing (CBT)
Topic Breakdown
Severity Models
15-20%Parametric distributions, coverage modifications (deductibles, limits, coinsurance), limited expected values
Frequency Models
10-15%Poisson, Negative Binomial, Binomial, (a,b,0) and (a,b,1) classes, zero-modified and zero-truncated models
Aggregate Loss Models
15-20%Compound distributions, Panjer recursion, normal and lognormal approximations, stop-loss premiums
Credibility
15-20%Limited fluctuation, Bayesian credibility, Buhlmann and Buhlmann-Straub models, empirical Bayes
Ratemaking
15-20%Loss development, trending, on-level factors, indicated rate changes, expense loading
Reserving
10-15%Chain-ladder, Bornhuetter-Ferguson, expected loss ratio, IBNR estimation
Recommended Study Approach
- 1
Master the core severity distributions (Exponential, Pareto, Lognormal, Weibull) and know how to modify them for deductibles and policy limits.
- 2
Practice computing aggregate loss distributions using both the recursive (Panjer) method and the normal approximation.
- 3
Study credibility formulas (Buhlmann, Buhlmann-Straub) until you can set up the credibility-weighted estimate automatically.
- 4
Work through ratemaking problems involving loss development triangles, trend factors, and on-level premium adjustments.
- 5
For reserving, practice the chain-ladder method, Bornhuetter-Ferguson method, and the expected loss ratio method on sample triangles.
- 6
Understand the relationship between limited expected values and excess loss distributions, as this concept appears in many problem types.