The IRA's "Dessert First, Spinach Later" Budgeting Has Failed

The 2022 Inflation Reduction Act (IRA) failed - spectacularly - to reduce health care costs.  Michael F. Cannon, the Director of Health Policy Studies at the Cato Institute, explains why and details the Center for American Progress' (CAP) latest plan to limit drug prices and increase drug subsidies for Medicare enrollees.  CAP and its lovely and gracious Director Neera Tanden were the most active promoters of both ObamaCare and the IRA:

https://mfcannon.substack.com/p/cap-health-care-proposal-regulation

https://www.americanprogress.org/article/a-patients-bill-of-rights-to-lower-health-care-costs/

CAP Health Care Proposal: Regulation Hasn’t Delivered Affordability, So Let’s Try Regulation
To make health care more affordable and universal, we need a different approach.
Michael F. Cannon - April 10, 2026

In 2022, under the banner, “How the Inflation Reduction Act Reduces Health Care Costs,” the left-leaning Center for American Progress predicted that that law’s provisions to limit drug prices and increase drug subsidies for Medicare enrollees “will improve health care affordability for Americans” and “translate into lower premiums for Part D plans.”

Fast forward to 2024, the year before most of the IRA’s drug provisions took effect. The Congressional Budget Office (CBO) announced that insurer bids for Part D plans rose by 42 percent—16 percentage points more than the CBO expected. In 2026, the CBO announced that Part D plan bids increased by another 35 percent, leading to a whopping increase in the agency’s spending projections:

Part D spending per beneficiary in 2035 is now projected to be more than $4,000, up from less than $3,000 in the January 2025 baseline.

The agency correspondingly increased its 10-year spending projection for Part D by $600 billion. Repealing the enhanced matching rate for Obamacare’s Medicaid expansion (10-year savings: $561 billion), at which Republicans balked during last year’s budget debate, would not offset that much additional spending. The only non-health care, CBO-scored budget option that would involves cutting Social Security for 75 percent of new recipients (10-year savings: $607 billion).

According to the CBO, much of the cause of this increase in Part D spending is that the IRA’s drug provisions turned out to be a lot more expensive than the agency previously projected. Bundling long-term spending restraints with near-term subsidies—what I call “dessert first, spinach later” budgeting—hasn’t had the effect CAP predicted.

The IRA experience came to mind when I read CAP’s new health care affordability proposal, which would pair broader and tighter health care price controls with a prohibition on certain spending restraints. Specifically, the CAP authors propose:

• Tightening price controls on health insurance premiums by lowering the threshold for individual-market premium increases that regulators may reject, and imposing “rate review” regulation on employer plans as well.
• Indirect, Medicare-based price controls on hospitals in highly concentrated markets, that limit prices to three times what Medicare sets.
• Additional indirect price controls on hospitals whose prices exceed the statewide median, that prohibit price increases greater than the rate of general inflation plus one percentage point.
• Tightening indirect regulation of insurer profits, by limiting administrative expenses to a percentage of industry-average premiums (rather a figure that individual insurers can manipulate) for purposes of “medical loss ratio” (MLR) regulation.
• Imposing indirect regulation of insurer profits on self-funded employer plans. (MLR again.)
• Prohibiting integration, specifically banning health insurance companies from owning “providers, pharmacies, and PBMs,” to prevent insurers from maximizing government subsidies by acquiring downstream providers and then overcharging themselves.
• Additional indirect regulation of insurer profits, where regulators determine whether provider-subsidiaries are overcharging parent insurers, and count the excess against administrative expenses rather than claims.
• Prohibiting prior authorization for “routine, emergency, and essential care” and requiring insurers to obtain permission from a government agency before denying any other claim.

These proposals double down on the very ideas that are currently making health care so unaffordable. Contrary to conventional wisdom, the US health sector already suffers from extensive government price-setting, which more often than not increases health care prices, in both government programs and private markets, including by encouraging inefficient market concentration.

To their credit, the authors admit, implicitly and explicitly, that many such regulations have failed. Obamacare already regulates premiums via both “community rating” (which links price floors to price ceilings) and “rate review” (which imposes subjective price ceilings). If those regulations were making health insurance affordable, there would be no need for additional proposals. Yet Obamacare premiums have grown at three times the rate of inflation and increased 26 percent in 2026, so here we are. Despite Obamacare’s “protections” for patients with preexisting conditions, the authors write, “health insurance companies can still discriminate against sick people by requiring prior authorization of claims.” It’s worse than that: research shows those Obamacare provisions are increasing prior authorization hurdles.

The authors explicitly admit that MLR regulation is increasing prices and premiums. Thanks to those regulations, “insurers are insensitive to increases in provider prices” because “higher prices translate to higher premiums and higher profits,” which counterproductively encourages insurers to “boost profits by increasing spending and premiums.” The authors acknowledge that MLR regulation encourages insurers to acquire downstream providers, who then overcharge their parent-insurers, leading to still-higher prices, claims, and premiums. We’ve come a long way since the Obama administration boasted that MLR regulation “has saved consumers a lot of money” because insurance companies “are charging lower premiums and operating more efficiently.”

Where the authors claim that regulation has succeeded, they set the bar very low. They cite one study finding that after Rhode Island introduced price caps, regulated (fully insured) and unregulated (self-insured) hospital prices both fell by 9 percent relative to prices in comparator states. (See Exhibit 1.) Curiously, the authors interpret this to mean that regulation reduced even the unregulated prices. Obviously, that’s not the only available interpretation. The authors claim that rate-review regulation “clearly” reduced Obamacare premiums. Yet the data they cite can’t rule out that insurer gaming defeated this regulation too.

The authors’ proposals are not likely to perform any better than existing government interventions. Take the proposal to limit hospital prices to three times what Medicare pays. Medicare is not a good price negotiator. In the words of former administrator Tom Scully, Medicare is just “a big, dumb price fixer.” It overpays hospitals for cataract removal by 100 percent. It overpays long-term care hospitals by more than 200 percent. Under the CAP proposal, private insurers could (use government subsidies to) overpay long-term care hospitals by 800 percent—i.e., three times the Medicare-set price—and supporters would still call this idea a success on which Congress should build.

The fact that the authors also propose to prohibit prior authorization—a spending restraint—will win them favorable attention from doctors, hospitals, and pharmaceutical companies. The proposal would effectively require insurers to pay more invalid claims and then seek repayment from providers later. In Medicare, we call it “pay and chase.” It doesn’t work very well. And, as with the IRA, it all but guarantees that this package would increase rather than reduce health care spending and health insurance premiums.

If we’re going to make health care more affordable and universal, we need a different approach.