Incentives in the Medicare Prescription Drug Benefit

The introduction of Medicare's new Part D prescription drug benefit is perhaps the most significant expansion of the program since its inception in 1965. The cost of the drug benefit for 2004-2013 is projected to be $410 Billion. As of mid-April 2006, 19.7 million Medicare beneficiaries were enrolled in a Part D plan, including 8.1 million beneficiaries who signed up for new stand alone prescription drug plans (PDPs).

While private insurance markets have generally failed to provide stand alone prescription drug insurance, there are several features that may allow Medicare to succeed. First, the Centers for Medicare and Medicaid Services (CMS) can impose mandatory minimum quality standards, as CMS must approve each PDP's formulary, or the specific drugs covered by the plan. Second, the program is heavily subsidized, with Medicare paying 74.5% of total plan premiums plus 80% of catastrophic costs (annual drug costs exceeding $5,100). Third, payments to providers are risk adjusted, meaning that they are adjusted up or down based on the expected cost of patients enrolled in the plan. This can help to avoid a "race to the bottom," whereby providers design their plans to make them unappealing to costly patients.

As the market for PDPs is brand new, it remains to be seen whether it will function as lawmakers envision. In "Perverse Incentives in the Medicare Prescription Drug Benefit," (NBER Working Paper 12008), David McAdams and Michael Schwarz examine issues in the design of the new drug benefit that may put upward pressure on drug prices and downward pressure on drug plan quality.

One issue the authors examine is risk adjustment. Although risk adjustment is commonly used for traditional health insurance plans, the authors suggest that it may be more problematic in the case of PDPs. If risk adjustment is very fine, for example based on a specific drug a patient is taking, then the manufacturer of that drug has an incentive to raise its price, knowing that the provider has little incentive to encourage the patient to switch to a lower cost substitute given that Medicare will reimburse most of the cost.

This might suggest that coarse risk adjustment is preferable, but this too has drawbacks. If patients with different preexisting conditions receive the same risk adjustment, PDP providers will have an incentive to discourage patients with the more costly condition from joining their plan. This is easy for providers to do, for example by moving a specific drug to a different tier or adding it to a preapproval list. CMS has said that it will monitor plans to identify those that are outliers in terms of what drugs they exclude; however, if many plans engage in this behavior, it will not be easy to identify discriminatory PDPs in this manner.

While risk adjustment is probably a necessary part of any reimbursement scheme, the authors argue that the drug benefit could be improved by having all PDPs charge the same premium, which would be determined based on the amount budgeted by Congress. This would have the obvious advantage of eliminating uncertainty about the total cost of the program. Moreover, the authors contend that there would be less upwards pressure on price and downwards pressure on quality in such a scheme.

With a fixed price, providers would try to assemble the most generous formulary possible within their budget. This would put pressure on drug manufacturers to keep prices low so that their drugs would be included. In terms of quality, the race to the bottom occurs when providers offer cheaper and less generous plans to appeal to healthier patients. With premiums fixed, less generous plans would be less attractive to all seniors.

The authors envision that CMS will need to actively regulate PDPs. They note that this is particularly likely in the case of "fuzzy line" rules, such as that PDPs must cover all drugs "presenting unique and important therapeutic advantages" or "most commonly used by the Medicare population." Providers are likely to interpret such requirements narrowly and offer a limited formulary, in order to extract greater price concessions for drugs that are included in the formulary. This may lead CMS to demand that more drugs be covered, which increases the upwards price pressure for drugs.

Another fuzzy line rule is that PDPs may not discriminate against specific groups of patients. Since patients with a certain condition may all rely on a particular drug, enforcing this rule may lead CMS to mandate that specific drugs be covered by all PDPs, which can be expected to raise the drugs' price. As a final example, since Medicare will pay up to 80% of the difference between a PDP's projected and actual costs, CMS will need to be very involved in setting methods for projecting plan costs.

The authors conclude that "CMS will have to continue to closely regulate the benefit, especially formulary design, for the foreseeable future. Any minimum standard that CMS imposes on formularies, however, will put additional upward pressure on drug prices. Ultimately, this could jeopardize the benefit's budgetary viability."

NBER Videos

National Bureau of Economic Research, 1050 Massachusetts Ave., Cambridge, MA 02138; 617-868-3900; email:

Contact Us