I've gotten some confused e-mails over David Rogers' report on Paul Ryan's "0.5 percent solution" -- his decision to cap the growth of Medicare spending at GDP+0.5 percentage points (a rate that is slower than health-care costs typically grow) to make the rest of his budget numbers add up.

The first thing to know here is that over the past few years the GOP vice presidential nominee has proposed capping Medicare spending at all sorts of different levels, depending on whom he was working with and what else he needed to pay for.

J.D. Pooley/AP

In Romney's 2010 "Roadmap for America's Future," Medicare was capped at "a blended rate of the CPI and the medical care component of the CPI." In his bipartisan proposal with Alice Rivlin, it was capped at GDP+1 percentage point. In his 2011 budget, it was capped at inflation, which was an incredibly implausible assumption that would, if followed, have led to massive cost-shifting onto seniors. In the Ryan-Wyden proposal, it was back up to GDP+1 percentage point. And in his most recent budget, it went down, as Rogers explains, to GDP+0.5 percentage points.

Ryan needs these caps because there's no good evidence that any of his premium support plans will actually save much money in Medicare.* He thinks they will, and he hopes they will, but that's not enough for the Congressional Budget Office to actually say they will. So he includes these caps and tells the CBO that if his premium support plan fails, he'll save the money in a different way. In his 2011 budget proposal, that's through shifting costs to seniors. In Ryan's most recent budget, it's left blank. But the CBO has no choice. They have to assume Congress will abide by the cap.

This isn't a trick unique to Paul Ryan. President Obama's budget proposal also includes a GDP+0.5 percentage point cap on Medicare's spending growth. Obama proposes to hit that cap by moving Medicare away from fee-for-service and toward pay-for-quality. But the CBO doesn't think there's enough evidence to certify that plan, and so Obama, like Ryan, included a promise that if his plan fails, Congress will have to find another way to save the money. Congress could, of course, choose to ignore that promise, as they've done with other such promises in the past.

The Simpson-Bowles plan is also thick with spending caps. Health costs are capped at GDP+1 percentage point. Discretionary spending sees some cuts and then is capped at inflation. And contrary to the report's reputation for specificity and bold choices, it's not even clear how they plan to hit these caps. There are a number of options and possible recommendations, but because the Simpson-Bowles plan didn't want to simply tell Congress what to do, they've left a lot of the actual decisions up to lawmakers. 

One of our worst tendencies in Washington is to debate budget plans by referencing their bottom-line numbers, many of which are dependent on these kinds of caps. The question with a budget plan isn't whether it says it will save money, but whether the policies in it will actually save money. If Ryan's fond hopes for premium support don't pan out, then his budget has no chance of meeting its targets. Similarly, if Medicare proves incapable of saving money by paying for quality, Obama's numbers aren't going to happen. And for all that you hear about how Congress should "just pass Simpson-Bowles," if Congress can't figure out the right mix of policies to hit those caps, then Simpson-Bowles won't solve our deficit problems. 

These questions, in the end, require a judgment about what's actually the best way to save money in the health-care system, how we can raise more tax revenue, or what will move the needle on growth. But too often, when we say we're debating budget policy, we're actually just debating budget caps. And budget caps don't work if the underlying policies fail.

* Ryan also caps other forms of spending, and his most recent budget implausibly assumes that all non-entitlement spending will fall from 12.5 percent of GDP today to 3.75 percent of GDP in 2050. This is more important to his numbers than his Medicare plan, and it's also a complete joke. See page 13 of the CBO report for more.