When Paul Ryan introduced his proposal for converting Medicare into a system of “premium support,” he reminded everybody that it closely resembled another Medicare proposal—one he’d developed with Alice Rivlin, the Brooking scholar and former Clinton Administration advisor. “Alice Rivlin is a great, proud Democrat,” Ryan said. “This path to prosperity builds upon those Ryan-Rivlin plans that we put in here.”

It’s true that Rivlin is a Democrat, albeit a fairly conservative one on fiscal issues. And it’s true she worked with Ryan on a premium support proposal for Medicare. But Rivlin says she cannot support the new Ryan plan for two key reasons—and she has spent the last 24 hours telling anybody who would listen about them.

The concept of premium support for Medicare is fairly straightforward. Instead of giving seniors insurance, the federal government would give seniors a fixed some of money—what most of us would call a voucher—for buying private insurance. But Rivlin told Politico (among others) that she prefers maintaining traditional Medicare as an option for anybody that wanted to enroll. (I’m not sure this was explicit, or even implicit, in the original Ryan-Rivlin plan, at least based on this Congressional Budget Office evaluation. But Rivlin is saying it now.) In the new Ryan plan, Medicare would no longer be available to new retirees as of 2022.

Whether traditional Medicare could survive as an option, in the sort of system Rivlin says she envisions, is open to question and would depend, in part, on the regulations governing the whole system. But, at least in principle, she wouldn’t force people into private insurance. The new Ryan plan would.

The other key difference is the money. In a premium support system, the value of the voucher grows by some fixed formula. Ryan-Rivlin set that formula at GDP+1—in other words, the voucher would increase every year, enough to accommodate inflation plus productivity growth (which is what GDP represents) plus one additional percentage point.

That’s still not a lot. In fact, many critics argued—persuasively, in my view—that giving seniors vouchers that increased in value so slowly would force seniors to bear ever higher costs for their medical care, imposing real hardship. The Center on Budget and Policy Priorities was among those making that case.

But the voucher in the new Ryan plan would increase in value even more slowly than the voucher in Ryan-Rivlin. Instead of rising at GDP+1, it would rise at inflation—in other words, there'd be no adjustment for productivity growth and no extra percentage point. “In the Ryan version, he has lowered the rate of growth and I don’t think that’s defensible,” Rivlin said. “It pushed too much of the cost onto the beneficiaries.” As Ezra Klein observed, “it’s totally unrealistic. … either those savings aren’t real or we’re assuming America is going to abandon seniors and the disabled in a way that has no recent precedent.”

I agree. And there’s a broader lesson here, one that applies to any efforts at controlling health care costs, from the left or the right (or some combination thereof). We need to do something to make health care less expensive and we need to get started right away. But, particularly early on, we need to move slowly. The system can absorb only so much shock at once. If you clamp down on spending too quickly, it’s going to cause serious harm to somebody, whether it’s providers or insurers, employers or beneficiaries. The best strategy is to apply cost control gradually and broadly, as part of a system-wide reform like the Affordable Care Act, so that the sacrifice is tolerable as well as shared.

Update: I originally misstated the value of the voucher in the Ryan plan; I've since corrected that. Also, Ezra just posted a full interview with Rivlin, in which she not only reiterates her rejection of the new Ryan plan but explains why she supports, as an alternative on health care, the Affordable Care Act:

...there’s a great deal in the Affordable Care Act in terms of research, pilot programs, alternative payment structures, alternative delivery systems, research on the effectiveness of treatments, that is needed. That’s why we need to keep the Affordable Care Act and strengthen the parts that hopefully give us more cost-effective care in the future. If you just control the federal spending without changing the delivery system, you just get cost shifting.