Spending on Medicare and Medicaid is pushing the federal budget to the breaking point, but, in the aftermath of the cuts enacted in the 2025 reconciliation bill, Congress might have difficulty producing another round of necessary savings in the near term. To make continued progress on health care spending restraint while waiting for the congressional window to reopen, the executive branch should make maximum use of a powerful existing tool—the Center for Medicare and Medicaid Innovation, or CMMI—to cut health entitlement expenditures.

In 1980, combined federal spending on Medicare and Medicaid was equal to 1.7 percent of gross domestic product (GDP). The Congressional Budget Office (CBO) projects it will be 6.3 percent of GDP in 2026 and 7.3 percent in 2036. The added cost burden of these programs relative to four decades ago is a major reason the government continues to run large annual budget deficits.

With CMMI, the congressional authors of the Affordable Care Act (ACA) delegated substantial discretion to the Centers for Medicare and Medicaid (CMS) to test novel payment models under Medicare, Medicaid, and the Children’s Health Insurance Program (CHIP). As stated in CMMI’s authorizing legislation, the office’s mission is to “reduce program expenditures” on health entitlements, “while preserving or enhancing” the quality of care provided to patients. The law outlines a process for evaluating CMMI-tested policies against a standard of “budget neutrality” and no demonstrable harm to the quality of care provided to program beneficiaries. Beyond those two considerations, CMMI is free to consider and test many different types of payment changes, including those that might produce substantial savings. Congress wanted to make it easier for CMS to implement promising payment changes without having to continuously work through multi-year legislative processes.

In practice, CMMI has been overly cautious and heavily provider-focused. CMMI’s flagship programs are often called “alternative payment models” and zero in on getting more patients into provider-led managed care programs of various types that can weed out unnecessary expenses through better care coordination. Some of these models have reduced costs, but the overall record has been unimpressive.

Further, CMMI has shied away from beneficiary incentives. CMS officials sometimes point to a law that bars “inducements” as an excuse for their reticence, but that is not what would be tested by well-constructed CMMI policies. Rather, the government, not outside parties, should be in charge of rewarding beneficiaries for appropriately economizing choices. It should be remembered that Medicare law requires 20 percent co-insurance for outpatient care to incentivize price-conscious choices.

While there is nothing wrong with continuing to refine APMs, after 15 years of trial and error, CMMI should make deep cost reduction its central objective. The Trump administration’s early announcement of a new orientation for CMMI was a first step in this direction, but it has not yet led to the announcement of models with dramatic cost-cutting potential (its drug reimbursement test is a possible exception, but it is too early to know how it will play out). With nearly three years left in its term, the administration still has time to make a concerted push for cost-cutting experiments.

There are numerous early targets to explore:

• Bryan Dowd and Anthony T. LoSasso argue for allowing those accountable care organizations (ACOs) with proven cost-cutting records to offer to Medicare beneficiaries discounted cost-sharing, premium reductions, and possibly rebates in return for enrolling in their care networks. (CMMI announced similar possibilities in a separate model revamp recently, but the policy is delayed and not yet defined well enough to assess.) With these incentives, provider-led managed care networks could compete more effectively with the big Medicare Advantage plans.
  
• Reference-based pricing for high-volume surgical procedures has worked in other settings. Under this model, surgeons would be required to submit their all-in fees for standardized bundles of services attached to common procedures, such as joint replacements. Current CMMI “bundled payment” models use CMS regulations to build all-in fees rather than submitted bids from the providers. Beneficiaries choosing surgeons with prices that are below what Medicare would have paid under its standard regulations should get to keep a portion of the savings.
  
• Many Medicare-covered services and products could be procured through competitive pricing models. For instance, Medicare could test competitive bidding for diagnostic imaging services and then reward beneficiaries who use the program’s preferred network. Similar tests could be implemented for home health care and lab tests (an existing authority allows competitive bidding for durable medical equipment).

Although these policies are directed toward Medicare, CMMI could work with states to implement similar experiments in Medicaid.

Congress might have been too eager to offload its responsibilities on CMMI. However, as long as CMMI exists, it might as well be used for a good purpose, like implementing sensible Medicare and Medicaid cost reductions that Congress would have trouble approving.