A 2004 paper by Thomas Rice and Katherine Desmond in JHPPL includes the history of the premium support concept as it has been considered in the academic literature. (The paper also includes the history of premium support legislation.) An ungated version of the paper is available. Because such a version exists, I will not include full references at the end of this post. You’ll have to click through for those.
Proposals in the Academic Literature
The genesis of not only premium support, but most procompetitive proposals in health care, is the work of Alain Enthoven (Enthoven 1978; Enthoven and Kronick 1989). Under his original proposal first presented in the late 1970s, Enthoven envisioned that an array of health plans, including both capitated systems such as HMOs and those based on fee for service, would compete for patients on the basis of price and quality. Employers would be required to give their employees a choice of health plans and contribute equal premium amounts to each plan. The income tax system would be modified so that the choice of a more expensive plan would not be subsidized. Health plans would be precluded from charging lower premiums to the healthiest individuals and groups to attract favorable risks. Low-income individuals and families would be given subsidies to purchase coverage. […]
In the early 1980s, Walter McClure (1982) proposed enactment of a competitive-based health care system that included Medicare. He developed two alternatives for the program: either a series of competitive-based demonstration projects or, alternatively, direct enactment of a program in which beneficiaries would receive 95 percent of the adjusted area per capita costs (AAPCC, a formula in which payments to health plans are tied to what it would have cost to treat enrollees in the fee-for-service system) and use that to purchase traditional Medicare or any other qualified health plan. Later in the decade, Randall Bovbjerg (1988) provided a detailed analysis of the concept of a Medicare voucher system. He noted that the success and popularity of such a concept would depend on a number of design features, including whether Medicare would compete against private plans, the generosity of the voucher (or here, premium support amount), methods of risk adjustment employed, the degree of subsidization for those with low incomes, and the amount of government regulation of health plans that would remain.
Bryan Dowd and colleagues (1992) proposed a system similar to McClure’s as well as an updated version a few years later (Feldman and Dowd 1998). Under these proposals, the Medicare contribution would be set “as a function of the lowest price for basic Part A and Part B Medicare benefits . . . submitted by a qualified plan in a predefined market area” (Dowd et al. 1992: 444). One of the competing health plans would be the traditional Medicare fee-for-service system. Beneficiaries choosing health plans that were more expensive than the cheapest plan available would pay the difference out of pocket.
The academic proposal that probably led to the most policy interest was developed by Henry Aaron and Robert Reischauer (1995). Rather than remaining as what they call a service reimbursement system, Medicare would be converted into one based on premium support in which the federal contribution would be a “defined sum towards the purchase of an insurance policy that provided a defined set of services” (ibid.: 20). Beneficiaries would have to pay, in full, any difference in premiums between the plan they choose and the amount the federal government contributes. The proposal is largely similar to those developed by Bryan Dowd, Roger Feldman, and colleagues but contains two key differences. First, although health plans would submit bids on how much they would require to enroll Medicare beneficiaries, the amount paid by Medicare would be based not on these bids, but rather on a risk-adjusted administrative formula. Second, health plans would provide the standard Medicare benefits package for all beneficiaries; there would not be a competing Medicare-operated fee-for-service system.
Gail Wilensky and Joseph Newhouse (1999) also proposed a premium support system. Although perhaps more food for thought than a formal proposal, it did contain some notable modifications from previous iterations. The authors proposed, for example, that the level of premium support “would vary . . . according to the person’s age, sex, geographic area, health risk status, wealth, and use of services” (ibid.: 101). Of particular note is that wealthier beneficiaries would receive less in premium support. Furthermore, because of the tremendous impediments raised by selection bias and our present inability to successfully risk-adjust premiums, they proposed a “partial capitation” system of paying health plans, in which “a portion of the government payment would reflect the capitation payment otherwise calculated, and the remaining portion would vary with actual services used” (ibid.: 104).