The Association of American Medical Colleges surveyed the principal business officers of all 126 accredited US medical schools in late 1991 in order to learn about their retirement benefit programs for faculty and whether early-retirement incentive programs were being used. A total of 115 of the schools provided usable responses, which the author reports for all schools, public schools, and private schools. For all schools, defined-contribution plans (based on employer's and employee's contributions plus investment earnings) were the preferred type of retirement benefit program (77%, or 89, of the responding schools). Defined-benefit plans (based on a formula that uses variables of age at retirement, years of service, and salary) were available at 37%(43) of the schools. Early-retirement incentive programs had been used by 70% of the responding schools in the five-year period 1987–1991. Sixteen schools provided descriptions of their formal early-retirement programs, which are summarized. The author observes that, although nearly three-fourths of the responding schools are familiar with the use of the incentive programs, these programs have resulted in few actual early retirements. He discusses why this may be true and compares the pros and cons of formal and ad hoc programs. He concludes that no single program can be considered best; each institution must work with its faculty to design programs to meet institutional goals and faculty interests.