(Catherine Rampell/NY Times)

“We’ve seen how good government is at restraining costs,” writes Andrew Sullivan in praise of Ryan’s budget. “I suspect consumers will be better.”

Well, Medicare controls costs better than private insurance in this country, and every other developed country controls costs using government-managed systems and most spend around half what we do, so I’d actually say the government comes out pretty well on this measure. But that’s not really the point I want to make in this post.

Rather, I think a lot of very smart people — Tyler Cowen, Andrew Sullivan, Megan McArdle, etc. — are making a category error. They’re looking at Ryan’s Medicare reforms and seeing an ideological “bureaucrats vs. consumers” contrast from Democratic plans. There is some of that, though it’s mostly on Ryan’s side: Democrats try out the consumer-driven exchange model for most non-seniors and a mostly bureaucratic model for controlling costs in Medicare and Medicaid, while Ryan only uses a consumer-driven exchange model in his quest for savings.

But from the perspective of health-care cost control, what matters is the sectoral contrast. The basic theory of Ryan’s plan is that you can control costs by focusing on the insurance system. Seniors become consumers and their decision-making holds down costs. The Affordable Care Act has a lot of the same insurance-system reforms that Ryan does, but the basic theory of that plan is you control costs through the care delivery system. It’s about knowing what treatments work and what treatments don’t, paying for value rather than quantity, cutting down on unnecessary readmissions and errors, doing more to manage chronic diseases, etc.

It’s only once you can do all that that exchanges really become useful, because then and only then will insurers — be they private or public — really be able to control costs. The reason none of them really do it right now is that none of them can force hospitals and doctors to hold down costs. It’s not like there’d be no competitive advantage in being the cheapest, best insurer in the country. But you can’t find one example of the model Ryan proposes slowing costs to even half what Ryan predicts. That’s why his plan really isn’t a plan — it’s just a way to make the numbers work so he can “balance the budget” without tax increases and make some conservative changes to entitlements.

I’m a big fan of exchanges. I’d use them systemwide if I could, as the Swiss and the Dutch do. But health-care costs are not all about, or even mostly about, insurance. Indeed, so much as I like exchanges, it’s very possible that you need the government putting pressure on the delivery system to control costs. What we tend to see with private insurers is that they just don’t have enough leverage over hospitals or doctors to get major changes done. Medicare does have that leverage, which is why it makes sense to use Medicare as a tool to reform the health-care system, as opposed to just the health-insurance system.

People in D.C. — and I include myself here — know a lot more about insurance than about care delivery, and are much more comfortable with arguments that pit consumers against bureaucrats than with arguments about various ways to cut down on hospital readmissions. So there’s a tendency to try to map health-care policy onto those debates. But it’s not the right way to think about it. At the top, I quote Andrew saying government has failed at cost control, which, as I note, ignores some inconvenient facts. But what’s undeniably true is that American politics has failed at cost control, and I think partly for this reason.