In his column yesterday, David Brooks echoed Paul Ryan’s argument that the Medicare Prescription Drug Benefit’s “costs are 41 percent below expectations” and that the program’s success suggests that Ryan-like reforms to Medicare will work. I wrote about this at length in my recent back-and-forth with Ryan, but since that got pretty deep in the weeds, let’s do it again.

There are three problems with the Part-D theory of Ryan’s reforms: The first is that it’s misleading on what’s driving cost growth in Part D. The second is that it’s misleading on how much cost growth there’s been in Part D. The third is that it’s misleading on how Ryan’s plan is structured. Taken together, this talking point isn’t wrong so much as its deceptive in a fairly sophisticated way. But let’s start by taking the separate pieces in order.

It’s true that Part D’s costs have come in beneath expectations. The question is, “why?” Brooks suggests that that’s due to the structure of the program, which “uses a competition model.” But then, why is national drug spending 35 percent (pdf) below expectations? There was no major reform of the non-Medicare market for pharmaceuticals, after all.

The answer, basically, is that pharmaceutical spending is down because pharmaceutical innovation is down. Medicare’s trustees, whom you might expect to trumpet their success controlling costs in Part D, are very straightforward about this: “The reduced estimates reflect a higher market penetration of generic drugs and a decline in the number of new drug products that are expected to reach the market during this period.” In other words, old drugs are slipping out of patent and new ones aren’t being invented as quickly as we’d expected, or hoped. That’s not the program’s fault. But no one should cheer cost control that comes from a slowdown in innovation.

There’s another reason Part D has been cheaper than projected: Seniors aren’t signing up. The Congressional Budget Office estimated that 93 percent of Medicare enrollees would participate. Instead, 77 percent did. That’s obviously led to lower costs than expected, but not because the program is working better than expected.

So that explains the difference between Medicare Part D’s projected costs and its actual costs. But note how carefully that measure was chosen: Ryan didn’t says that the program’s costs are low. He said that they’re below expectations and implied that that’s the same thing. But it’s not. If only 50 percent of seniors sign up, costs will be very low against projections, but that doesn’t mean premiums aren’t growing at 20 percent a year. The system’s costs can be low even as cost growth is high.

And so too for Part D. Since 2006 — the first year of the benefit — Medicare Part D’s average premium has risen by 57 percent (pdf). Between 2010 and 2011, premiums rose by 10 percent. And going forward, the program’s actuaries expect (pdf) expenditures “to grow at an average annual rate of 9.7 percent for the 10-year period 2011 to 2020.” That may be an excellent performance when compared with the Congressional Budget Office’s initial projections. But it’s a lot faster than inflation, which is what Ryan needs for his plan to work.

Finally, if you look at them closely, Ryan’s plan and Part D don’t look all that similar. For one thing, Part D only covers drugs, while Ryan’s plan covers all health-care services. It’s not at all clear how applicable the Part D experience is to, say, hospital insurance, particularly given that drugs are experiencing a sector-wide slowdown right now.

But the bigger issue is that Ryan’s plan is capped while Medicare Part D isn’t. In Part D, the federal government pays, on average, 74 percent of program’s costs. And that support grows alongside the program’s costs. Ryan’s plan covers about a third of beneficiary costs, and that support grows at the rate of inflation — so much more slowly than the rest of the program, or than Medicare Part D. This has always been the main criticism of Ryan’s plan, and Medicare Part D’s structure shows what a radical decision he made to structure it like that.