As soon as DNA-based genetic testing became available, questions were asked about its use by insurers to discriminate against carriers of deleterious alleles. The statistical interest in these questions lies mainly in the application of genetic epidemiology to the actuarial models used to price life and health insurance. This chapter surveys the relevant research, including early work mostly focussed on single-gene disorders, and more recent attempts to predict the relevance of multifactorial genetics to insurance practice.Insurance is an unusual branch of commerce because the basis of its sound operation is mathematical in nature. Just as airlines (say) must take the mathematics of aerodynamics as they find it, insurance companies may be in peril if they ignore the actuarial mathematics and statistics that govern their businesses. But actuaries do not deal with precise and impersonal qualities like airflow over a wing; they deal with people and their personal attributes, that lead to an assessment of the risk of making a claim under an insurance contract. Some personal attributes (sex, race, disability) give rise to stronger sensitivities than others (age, weight, smoking habits) and the experience of the last 10 years suggests that any kind of personal genetic information falls in the first group.The mathematical nature of insurance, and life insurance in particular, is straightforward. The simplest contract, called whole-of-life insurance, pays an agreed sum (the sum assured) immediately on the death of the insured person. Suppose a person wishes to buy whole-of-life insurance with sum assured S. The problem is to determine the premium P to be paid at outset (in practice, premiums are usually payable monthly but we ignore this complication). Death being certain, so the insurer will pay out S at some future time with certainty. Intuitively the 'fair' premium before allowing for expenses, profit, etc. appears 1346 Handbook of Statistical Genetics, Third Edition . E dited by D . J. Balding, M . Bishop and C. Cannings.