This paper reviews the movement among multiple health plan options between 1994 and 1998 for Minnesota state employees whose work site was located in the Minneapolis/St. Paul metropolitan area. During this period the employer contribution was based on the lowest family premium bid from a qualified plan in the county of the employee’s work site. In 1995 the largest individual practice association model HMO in the state, Medica, reduced its state premium by 25%, becoming the lowest-priced option. This resulted in massive transfers of enrollees between plans. The point of this study was to estimate the risk changes that resulted from these movements between plan options. We obtained enrollment data by age and gender from Blue Cross Blue Shield of Minnesota (Blue Cross) and applied age/gender risk weight factors derived from actuarial rate tables to the Blue Cross cells. Annual changes in risk weights by 10-20% were common in a number of Blue Cross subpopulations, and in one case, by more than 50%. The Blue Cross POS plan experienced increases in risk and went into a death spiral, while a second Blue Cross plan with a more restrictive provider network started with low risk, but experienced increases in risk when the Medica plan was withdrawn. Similar demographic data were not available from other plans offered by the state and claim costs were confidential, so the results pertain only to Blue Cross risks. The question is raised as to whether managed competition can work without some means of adjusting premium rates to the expected cost level of the enrollees of a particular health plan. All carriers seemed reluctant to guarantee premium rates after the 1994-1998 experience, and the state soon became self-insured.