Ross Douthat argues that looking deep into the numbers that drive Rep. Paul Ryan’s budget proposal towards balance misses the point. The specific projections in Ryan’s budget should be seen as “largely irrelevant to its real policy purpose, which is to make structural changes to existing programs in the hopes of changing the underlying incentives that are driving spending growth and hampering economic expansion.”

Douthat calls this “a slightly different way to look at it.” I’d call it a very different way to look at it. If the numbers in Ryan’s budget are absurdly implausible -- and Douthat admits that they are -- that’s a real problem. Ryan has built a reputation off the idea that his numbers work, and the literal justification for the policies he’s proposing is that they make his numbers work. And the reason the numbers don’t work is because the underlying policies don’t work.

And that’s why Ryan’s numbers matter. He had to direct the Congressional Budget Office to use his assumptions because they don’t think his policies will save anywhere near the amount of money he thinks they will. That is to say, the evidence is that they won’t work. But Ryan is publicly selling his plan using numbers that assumes otherwise. That seems relevant to the discussion over whether his plan is actually a good idea.

So where Douthat writes that “the projected dollar figures for Medicaid spending in 2021 or 2031 matter less than the budget’s proposal to devolve control over Medicaid to state governments and allow more room for policy experimentation at that level,” I’d put it the opposite way: the projected dollar figures for Medicaid spending in 2021 and 2031 matter much more than the budget’s proposal to devolve control over Medicaid to state governments because there’s no reason to believe that the budget’s proposal will save very much money. And remember, Ryan’s plan includes deep, automatic cuts to Medicaid in the event that the states can’t magically control costs.

Similarly, the reason it matters that there’s a $6.2 trillion gap between the tax plan he’s got in his budget and the tax plan he told the CBO to score is because we have no idea how he closes that gap. It’s possible he intends to close it through assuming a gusher of supply-side revenues -- in which case his plan could dramatically increase the deficit if it was implemented. That would make it a very, very bad idea.

Which is to say, that waving away the plausibility or implausibility of Ryan’s numbers sort of defeats the whole purpose of a long-term budget, which is exactly what Ryan offered, and how he’s framed the goal of the program. Insofar as Ryan’s plan is going to be one side of the national debate over the budget, it’s important that people accurately understand how it works.

And that requires looking closely at the numbers. Because when politicians aren’t held to the numbers, they get to say things like Mitt Romney did today, when he assured voters that Ryan’s proposal, which he supports, “does not balance the budget on the backs of the poor and the elderly.” He said that because balancing the budget on the backs of the poor and elderly is something most Americans don’t want to do. And yet, balancing the budget on the backs of the poor is exactly what Ryan’s budget actually does. But you have to go past the rhetoric, and into the numbers, to see that.

Related: Ryan’s budget isn’t about Medicare cuts.